From e88ad8055584e6fc51c0d10db5ed94f5ca3c6fd0 Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Mon, 3 Mar 2025 16:37:25 -0800 Subject: [PATCH 01/25] biolog templates --- src/ontogpt/templates/biolog.yaml | 403 +++++++++++++++++++++++ src/ontogpt/templates/biolog_simple.yaml | 98 ++++++ 2 files changed, 501 insertions(+) create mode 100644 src/ontogpt/templates/biolog.yaml create mode 100644 src/ontogpt/templates/biolog_simple.yaml diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml new file mode 100644 index 000000000..06a10475c --- /dev/null +++ b/src/ontogpt/templates/biolog.yaml @@ -0,0 +1,403 @@ +id: https://example.org/PaperExtractionSchema +name: PaperExtractionSchema +title: Paper Extraction Schema +description: "Schema for extracting structured data from papers, including Biolog Phenotype MicroArray experiments." +license: https://creativecommons.org/publicdomain/zero/1.0/ +version: 0.1.0 + +prefixes: + rdf: "http://www.w3.org/1999/02/22-rdf-syntax-ns#" + linkml: "https://w3id.org/linkml/" + paperex: "https://example.org/PaperExtractionSchema/" + +default_prefix: paperex +default_range: string + +imports: + - linkml:types + +######################## +# ENUMS +######################## + +enums: + FlexibleBooleanEnum: + description: > + A minimal enumeration for capturing yes/no/true/false in lowercase, + plus 'not provided' if data are missing or ambiguous. + permissible_values: + 'yes': {} + 'no': {} + 'true': {} + 'false': {} + 'not provided': {} + +######################## +# CLASSES +######################## + +classes: + NamedEntity: + description: "A generic named entity for all non-root classes." + + Paper: + tree_root: true + description: "A single paper or study." + attributes: + study_title: + description: "The paper's title." + examples: + - value: "Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas." + description: "From Mavrodi et al. 2021" + annotations: + prompt: "Extract the paper's title. If not found, return 'Not provided'." + range: string + required: true + + authors: + description: "Authors of the paper." + examples: + - value: "Olga V. Mavrodi" + - value: "Janiece R. McWilliams" + annotations: + prompt: | + Extract a list of authors (each as a separate entry). + If multiple authors appear in one string, please split them carefully. + If not found, use 'Not provided'. + range: Author + multivalued: true + required: true + + doi: + description: "DOI of the publication." + examples: + - value: "10.3389/fmicb.2021.651282" + annotations: + prompt: "Extract the DOI. If not found, use 'Not provided'." + range: string + required: false + + date: + description: "Publication date." + examples: + - value: "14 April 2021" + annotations: + prompt: "Extract the publication date. If not found, use 'Not provided'." + range: string + required: false + + experiments: + description: "Experiments (e.g., Biolog) included in the paper." + examples: + - value: "Biolog Phenotype MicroArray; RNA-seq for gene expression" + annotations: + prompt: | + Extract a list of experiments described in the paper. + If nothing is mentioned, use 'Not provided'. + range: Experiment + multivalued: true + required: true + + Author: + is_a: NamedEntity + description: "An author of the paper." + attributes: + name: + description: "Name of the author." + examples: + - value: "Olga V. Mavrodi" + annotations: + prompt: "Extract the full name of this author. If none, return 'Not provided'." + range: string + required: true + + Experiment: + is_a: NamedEntity + description: "A single experiment from the paper." + attributes: + experiment_motivation: + description: "Rationale for the experiment." + examples: + - value: "Assess the metabolic impact of root exudates on Pseudomonas strains." + annotations: + prompt: | + Explain the motivation or purpose of the experiment. + If missing, use 'Not provided'. + range: string + required: false + + experiment_design: + description: "Summary of how the experiment was designed." + examples: + - value: "Comparative analysis of Pseudomonas gene expression with/without root exudates." + annotations: + prompt: "Summarize the overall design. If missing, use 'Not provided'." + range: string + required: false + + environment: + description: "Location or conditions in which the experiment occurred." + examples: + - value: "Lab-based assay under 25°C" + annotations: + prompt: "Extract the environment or setting. If none, 'Not provided'." + range: string + required: false + + host: + description: "One or more hosts in the experiment." + examples: + - value: "Brachypodium distachyon" + annotations: + prompt: | + Identify the host organisms or systems used. + If not mentioned, use 'Not provided'. + range: Host + multivalued: true + required: false + + host_type: + description: "Type of host (e.g., plant, animal)." + examples: + - value: "Plant" + annotations: + prompt: | + Describe the host type if applicable (e.g. 'plant'). If nothing specified, 'Not provided'. + range: string + required: false + + target_microbes: + description: "Microbes targeted in the experiment." + examples: + - value: "Pseudomonas fluorescens" + annotations: + prompt: | + Extract a list of target microbes studied. If none are mentioned, 'Not provided'. + range: Microbe + multivalued: true + required: false + + biological_system: + description: "Biological system under study." + examples: + - value: "Root exudates system" + annotations: + prompt: "Extract the focal biological system (e.g., rhizosphere). If none, 'Not provided'." + range: string + required: false + + conditions: + description: "Experimental conditions (temperature, pH, etc.)." + examples: + - value: "pH 7, 25°C" + annotations: + prompt: | + Identify relevant experimental conditions (e.g., 25°C, pH 7). + If not stated, 'Not provided'. + range: string + required: false + + experimental_factors: + description: "Factors tested or measured." + examples: + - value: "Root exudates presence vs. absence" + annotations: + prompt: | + List experimental factors or treatments. + If no info, 'Not provided'. + range: ExperimentalFactor + multivalued: true + required: false + + is_biolog_experiment: + description: "Whether the experiment is a Biolog experiment." + examples: + - value: "yes" + - value: "no" + annotations: + prompt: | + Indicate if it's a Biolog experiment ('yes' or 'no'). + If not stated, 'not provided'. + range: FlexibleBooleanEnum + required: false + + type_of_biolog_experiment: + description: "Type or subtype of the Biolog experiment." + examples: + - value: "Phenotype MicroArray" + annotations: + prompt: | + If it is a Biolog experiment, specify the subtype (e.g., Phenotype MicroArray). + Otherwise, 'Not provided'. + range: string + required: false + + plates: + description: "Biolog plates used (e.g., EcoPlate, GEN III, PM01, PM02A)." + examples: + - value: "PM01" + - value: "PM02A" + annotations: + prompt: | + Extract the name(s) of any Biolog plates (e.g., PM01, PM02A). + If not mentioned, 'Not provided'. + range: string + multivalued: true + required: false + + replicates: + description: "Number of replicates." + examples: + - value: "3" + annotations: + prompt: | + Extract the number of replicates. + If not mentioned, 'Not provided'. + range: string + required: false + + protocol_steps: + description: "Key steps in the experimental protocol." + examples: + - value: > + 1) Grow cells overnight on LB agar + 2) Harvest and suspend in IF-0 + 3) Adjust transmittance to 42% using a turbidimeter + 4) Mix suspension to final transmittance of 85% + 5) Inoculate PM01 and PM02A plates + 6) Incubate in OmniLog at 25°C for 48 h + 7) Record formazan formation every 15 min + 8) Data analysis with OmniLog Parametric Analysis software + annotations: + prompt: | + List the full protocol steps or key experimental procedures, including + inoculation details, incubation times/temperatures, data recording intervals, etc. + If not provided, 'Not provided'. + range: string + multivalued: true + required: false + + plate_reader_types_of_data_collected: + description: "Data types collected (OD, respiration, formazan formation, etc.)." + examples: + - value: "Formazan production, recorded by OmniLog system" + annotations: + prompt: | + Identify data types measured by the plate reader (e.g. formazan, OD). + If none, 'Not provided'. + range: string + multivalued: true + required: false + + od_protocol: + description: "Protocol for measuring optical density." + examples: + - value: "OD600 measured every 2 hours" + annotations: + prompt: | + Describe OD measurement protocol (if relevant). + If not described, 'Not provided'. + range: string + required: false + + respiration_protocol: + description: "Protocol for measuring respiration or related metabolic activity." + examples: + - value: "Colorimetric change at 590 nm for formazan detection" + annotations: + prompt: | + Describe the respiration or metabolic activity measurement protocol (e.g. formazan production). + If none, 'Not provided'. + range: string + required: false + + instrument_used: + description: "Instrument or equipment used for the experiment." + examples: + - value: "OmniLog Phenotype MicroArray System (Biolog)" + annotations: + prompt: | + Extract the name(s) of the instrument/equipment used (e.g., OmniLog system). + If none, 'Not provided'. + range: string + multivalued: true + required: false + + analysis_software: + description: "Software used to analyze or interpret data." + examples: + - value: "OmniLog Parametric Analysis software v1.20.02 (Biolog)" + annotations: + prompt: | + Extract any software or platform used for data analysis. + If not mentioned, 'Not provided'. + range: string + multivalued: true + required: false + + incubation_temperature: + description: "Temperature used for incubating the plates or cultures." + examples: + - value: "25°C" + annotations: + prompt: | + Extract the incubation temperature if provided (e.g. 25°C). + If not mentioned, 'Not provided'. + range: string + required: false + + incubation_duration: + description: "Total duration of incubation (e.g., 48 h)." + examples: + - value: "48 h" + annotations: + prompt: | + Extract how long the incubation lasted (e.g. 48 hours). + If not stated, 'Not provided'. + range: string + required: false + + Host: + is_a: NamedEntity + description: "A host organism or system." + attributes: + placeholder_attribute: + description: "Placeholder attribute for Host." + examples: + - value: "Host name: B. distachyon" + annotations: + prompt: | + Extract details about this host (placeholder). + If not found, 'Not provided'. + range: string + required: false + + Microbe: + is_a: NamedEntity + description: "Information about a microbial species or strain." + attributes: + placeholder_attribute: + description: "Placeholder attribute for Microbe." + examples: + - value: "Pseudomonas fluorescens" + annotations: + prompt: | + Extract details about this microbe. + If none, 'Not provided'. + range: string + required: false + + ExperimentalFactor: + is_a: NamedEntity + description: "A factor manipulated or measured in an experiment." + attributes: + placeholder_attribute: + description: "Placeholder attribute for ExperimentalFactor." + examples: + - value: "Root exudates concentration" + annotations: + prompt: | + Extract details about this experimental factor (e.g. concentration). + If absent, 'Not provided'. + range: string + required: false diff --git a/src/ontogpt/templates/biolog_simple.yaml b/src/ontogpt/templates/biolog_simple.yaml new file mode 100644 index 000000000..e5e7df749 --- /dev/null +++ b/src/ontogpt/templates/biolog_simple.yaml @@ -0,0 +1,98 @@ +id: https://example.org/PaperExtractionSchema +name: PaperExtractionSchema +title: Paper Extraction Schema +description: > + A simplified schema describing the fields to extract from a paper that includes + study metadata and a minimal notion of experiments. +version: 0.1.0 + +prefixes: + linkml: "https://w3id.org/linkml/" + xsd: "http://www.w3.org/2001/XMLSchema#" + +imports: + - linkml:types + +types: + string: + base: str + boolean: + base: bool + +######################## +# CLASSES +######################## + +classes: + # The top-level class (root) for the paper + Paper: + description: "Top-level class representing a single paper/study." + tree_root: true + slots: + - study_title + - authors + - doi + - date + - experiments + + Author: + description: "Represents an author of the paper." + slots: + - name + + Experiment: + description: "A minimal experiment class (could be Biolog or otherwise)." + slots: + - experiment_title + - is_biolog_experiment + +######################## +# SLOTS +######################## + +slots: + # Paper-level slots + study_title: + description: "Title of the study." + range: string + required: true + + authors: + description: "List of authors of the paper." + range: Author + multivalued: true + required: true + + doi: + description: "DOI of the publication." + range: string + required: false + + date: + description: "Date of the publication or the study." + range: string + required: false + + experiments: + description: "List of experiments mentioned in the paper." + range: Experiment + multivalued: true + required: false + + # Author-level slot + name: + description: "Name of the author." + range: string + required: true + + # Experiment-level slots + experiment_title: + description: "Title or short description of the experiment." + range: string + required: false + + is_biolog_experiment: + description: "Flag indicating if this experiment is a Biolog experiment." + range: boolean + required: false + From c20019892457b743ece25283abf96bbab251b1cf Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 10:50:18 -0500 Subject: [PATCH 02/25] Add py classes for biolog templates --- src/ontogpt/templates/biolog.py | 389 +++++++++++++++++++++++++ src/ontogpt/templates/biolog_simple.py | 124 ++++++++ 2 files changed, 513 insertions(+) create mode 100644 src/ontogpt/templates/biolog.py create mode 100644 src/ontogpt/templates/biolog_simple.py diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py new file mode 100644 index 000000000..2964af4da --- /dev/null +++ b/src/ontogpt/templates/biolog.py @@ -0,0 +1,389 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "0.1.0" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'paperex', + 'default_range': 'string', + 'description': 'Schema for extracting structured data from papers, including ' + 'Biolog Phenotype MicroArray experiments.', + 'id': 'https://example.org/PaperExtractionSchema', + 'imports': ['linkml:types'], + 'license': 'https://creativecommons.org/publicdomain/zero/1.0/', + 'name': 'PaperExtractionSchema', + 'prefixes': {'linkml': {'prefix_prefix': 'linkml', + 'prefix_reference': 'https://w3id.org/linkml/'}, + 'paperex': {'prefix_prefix': 'paperex', + 'prefix_reference': 'https://example.org/PaperExtractionSchema/'}, + 'rdf': {'prefix_prefix': 'rdf', + 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, + 'source_file': 'src/ontogpt/templates/biolog.yaml', + 'title': 'Paper Extraction Schema'} ) + +class FlexibleBooleanEnum(str, Enum): + """ + A minimal enumeration for capturing yes/no/true/false in lowercase, plus 'not provided' if data are missing or ambiguous. + + """ + yes = "yes" + no = "no" + true = "true" + false = "false" + not_provided = "not provided" + + + +class NamedEntity(ConfiguredBaseModel): + """ + A generic named entity for all non-root classes. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + pass + + +class Paper(ConfiguredBaseModel): + """ + A single paper or study. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema', 'tree_root': True}) + + study_title: str = Field(default=..., description="""The paper's title.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Extract the paper's title. If not found, " + "return 'Not provided'."}}, + 'domain_of': ['Paper'], + 'examples': [{'description': 'From Mavrodi et al. 2021', + 'value': 'Root Exudates Alter the Expression of Diverse ' + 'Metabolic, Transport, Regulatory, and Stress Response ' + 'Genes in Rhizosphere Pseudomonas.'}]} }) + authors: List[Author] = Field(default=..., description="""Authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract a list of authors (each as a ' + 'separate entry).\n' + 'If multiple authors appear in one ' + 'string, please split them carefully.\n' + "If not found, use 'Not provided'.\n"}}, + 'domain_of': ['Paper'], + 'examples': [{'value': 'Olga V. Mavrodi'}, {'value': 'Janiece R. McWilliams'}]} }) + doi: Optional[str] = Field(default=None, description="""DOI of the publication.""", json_schema_extra = { "linkml_meta": {'alias': 'doi', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Extract the DOI. If not found, use 'Not " + "provided'."}}, + 'domain_of': ['Paper'], + 'examples': [{'value': '10.3389/fmicb.2021.651282'}]} }) + date: Optional[str] = Field(default=None, description="""Publication date.""", json_schema_extra = { "linkml_meta": {'alias': 'date', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the publication date. If not ' + "found, use 'Not provided'."}}, + 'domain_of': ['Paper'], + 'examples': [{'value': '14 April 2021'}]} }) + experiments: List[Experiment] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract a list of experiments described ' + 'in the paper.\n' + "If nothing is mentioned, use 'Not " + "provided'.\n"}}, + 'domain_of': ['Paper'], + 'examples': [{'value': 'Biolog Phenotype MicroArray; RNA-seq for gene ' + 'expression'}]} }) + + +class Author(NamedEntity): + """ + An author of the paper. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the full name of this author. If ' + "none, return 'Not provided'."}}, + 'domain_of': ['Author'], + 'examples': [{'value': 'Olga V. Mavrodi'}]} }) + + +class Experiment(NamedEntity): + """ + A single experiment from the paper. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + experiment_motivation: Optional[str] = Field(default=None, description="""Rationale for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_motivation', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Explain the motivation or purpose of the ' + 'experiment.\n' + "If missing, use 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Assess the metabolic impact of root exudates on ' + 'Pseudomonas strains.'}]} }) + experiment_design: Optional[str] = Field(default=None, description="""Summary of how the experiment was designed.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_design', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Summarize the overall design. If ' + "missing, use 'Not provided'."}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Comparative analysis of Pseudomonas gene expression ' + 'with/without root exudates.'}]} }) + environment: Optional[str] = Field(default=None, description="""Location or conditions in which the experiment occurred.""", json_schema_extra = { "linkml_meta": {'alias': 'environment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the environment or setting. If ' + "none, 'Not provided'."}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Lab-based assay under 25°C'}]} }) + host: Optional[List[Host]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Identify the host organisms or systems ' + 'used.\n' + "If not mentioned, use 'Not " + "provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Brachypodium distachyon'}]} }) + host_type: Optional[str] = Field(default=None, description="""Type of host (e.g., plant, animal).""", json_schema_extra = { "linkml_meta": {'alias': 'host_type', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe the host type if applicable ' + "(e.g. 'plant'). If nothing specified, " + "'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Plant'}]} }) + target_microbes: Optional[List[Microbe]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract a list of target microbes ' + "studied. If none are mentioned, 'Not " + "provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) + biological_system: Optional[str] = Field(default=None, description="""Biological system under study.""", json_schema_extra = { "linkml_meta": {'alias': 'biological_system', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the focal biological system ' + "(e.g., rhizosphere). If none, 'Not " + "provided'."}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Root exudates system'}]} }) + conditions: Optional[str] = Field(default=None, description="""Experimental conditions (temperature, pH, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'conditions', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Identify relevant experimental ' + 'conditions (e.g., 25°C, pH 7).\n' + "If not stated, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'pH 7, 25°C'}]} }) + experimental_factors: Optional[List[ExperimentalFactor]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'List experimental factors or ' + 'treatments.\n' + "If no info, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Root exudates presence vs. absence'}]} }) + is_biolog_experiment: Optional[FlexibleBooleanEnum] = Field(default=None, description="""Whether the experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Indicate if it's a Biolog experiment " + "('yes' or 'no').\n" + "If not stated, 'not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'yes'}, {'value': 'no'}]} }) + type_of_biolog_experiment: Optional[str] = Field(default=None, description="""Type or subtype of the Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'type_of_biolog_experiment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'If it is a Biolog experiment, specify ' + 'the subtype (e.g., Phenotype ' + 'MicroArray).\n' + "Otherwise, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Phenotype MicroArray'}]} }) + plates: Optional[List[str]] = Field(default=None, description="""Biolog plates used (e.g., EcoPlate, GEN III, PM01, PM02A).""", json_schema_extra = { "linkml_meta": {'alias': 'plates', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the name(s) of any Biolog plates ' + '(e.g., PM01, PM02A).\n' + "If not mentioned, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'PM01'}, {'value': 'PM02A'}]} }) + replicates: Optional[str] = Field(default=None, description="""Number of replicates.""", json_schema_extra = { "linkml_meta": {'alias': 'replicates', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the number of replicates.\n' + "If not mentioned, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': '3'}]} }) + protocol_steps: Optional[List[str]] = Field(default=None, description="""Key steps in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'protocol_steps', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'List the full protocol steps or key ' + 'experimental procedures, including\n' + 'inoculation details, incubation ' + 'times/temperatures, data recording ' + 'intervals, etc.\n' + "If not provided, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': '1) Grow cells overnight on LB agar 2) Harvest and ' + 'suspend in IF-0 3) Adjust transmittance to 42% ' + 'using a turbidimeter 4) Mix suspension to final ' + 'transmittance of 85% 5) Inoculate PM01 and PM02A ' + 'plates 6) Incubate in OmniLog at 25°C for 48 h 7) ' + 'Record formazan formation every 15 min 8) Data ' + 'analysis with OmniLog Parametric Analysis ' + 'software\n'}]} }) + plate_reader_types_of_data_collected: Optional[List[str]] = Field(default=None, description="""Data types collected (OD, respiration, formazan formation, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'plate_reader_types_of_data_collected', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Identify data types measured by the ' + 'plate reader (e.g. formazan, OD).\n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Formazan production, recorded by OmniLog system'}]} }) + od_protocol: Optional[str] = Field(default=None, description="""Protocol for measuring optical density.""", json_schema_extra = { "linkml_meta": {'alias': 'od_protocol', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe OD measurement protocol (if ' + 'relevant).\n' + "If not described, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'OD600 measured every 2 hours'}]} }) + respiration_protocol: Optional[str] = Field(default=None, description="""Protocol for measuring respiration or related metabolic activity.""", json_schema_extra = { "linkml_meta": {'alias': 'respiration_protocol', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe the respiration or metabolic ' + 'activity measurement protocol (e.g. ' + 'formazan production).\n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Colorimetric change at 590 nm for formazan ' + 'detection'}]} }) + instrument_used: Optional[List[str]] = Field(default=None, description="""Instrument or equipment used for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'instrument_used', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the name(s) of the ' + 'instrument/equipment used (e.g., OmniLog ' + 'system).\n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'OmniLog Phenotype MicroArray System (Biolog)'}]} }) + analysis_software: Optional[List[str]] = Field(default=None, description="""Software used to analyze or interpret data.""", json_schema_extra = { "linkml_meta": {'alias': 'analysis_software', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract any software or platform used ' + 'for data analysis.\n' + "If not mentioned, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'OmniLog Parametric Analysis software v1.20.02 ' + '(Biolog)'}]} }) + incubation_temperature: Optional[str] = Field(default=None, description="""Temperature used for incubating the plates or cultures.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_temperature', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the incubation temperature if ' + 'provided (e.g. 25°C).\n' + "If not mentioned, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': '25°C'}]} }) + incubation_duration: Optional[str] = Field(default=None, description="""Total duration of incubation (e.g., 48 h).""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_duration', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract how long the incubation lasted ' + '(e.g. 48 hours).\n' + "If not stated, 'Not provided'.\n"}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': '48 h'}]} }) + + +class Host(NamedEntity): + """ + A host organism or system. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Host.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract details about this host ' + '(placeholder).\n' + "If not found, 'Not provided'.\n"}}, + 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], + 'examples': [{'value': 'Host name: B. distachyon'}]} }) + + +class Microbe(NamedEntity): + """ + Information about a microbial species or strain. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Microbe.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract details about this microbe.\n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], + 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) + + +class ExperimentalFactor(NamedEntity): + """ + A factor manipulated or measured in an experiment. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for ExperimentalFactor.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract details about this experimental ' + 'factor (e.g. concentration).\n' + "If absent, 'Not provided'.\n"}}, + 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], + 'examples': [{'value': 'Root exudates concentration'}]} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +NamedEntity.model_rebuild() +Paper.model_rebuild() +Author.model_rebuild() +Experiment.model_rebuild() +Host.model_rebuild() +Microbe.model_rebuild() +ExperimentalFactor.model_rebuild() + diff --git a/src/ontogpt/templates/biolog_simple.py b/src/ontogpt/templates/biolog_simple.py new file mode 100644 index 000000000..ee20a9b27 --- /dev/null +++ b/src/ontogpt/templates/biolog_simple.py @@ -0,0 +1,124 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "0.1.0" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'https://example.org/PaperExtractionSchema/', + 'description': 'A simplified schema describing the fields to extract from a ' + 'paper that includes study metadata and a minimal notion of ' + 'experiments.\n', + 'id': 'https://example.org/PaperExtractionSchema', + 'imports': ['linkml:types'], + 'name': 'PaperExtractionSchema', + 'prefixes': {'linkml': {'prefix_prefix': 'linkml', + 'prefix_reference': 'https://w3id.org/linkml/'}, + 'xsd': {'prefix_prefix': 'xsd', + 'prefix_reference': 'http://www.w3.org/2001/XMLSchema#'}}, + 'source_file': 'src/ontogpt/templates/biolog_simple.yaml', + 'title': 'Paper Extraction Schema', + 'types': {'boolean': {'base': 'bool', + 'from_schema': 'https://example.org/PaperExtractionSchema', + 'name': 'boolean'}, + 'string': {'base': 'str', + 'from_schema': 'https://example.org/PaperExtractionSchema', + 'name': 'string'}}} ) + + +class Paper(ConfiguredBaseModel): + """ + Top-level class representing a single paper/study. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema', 'tree_root': True}) + + study_title: str = Field(default=..., description="""Title of the study.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', 'domain_of': ['Paper']} }) + authors: List[Author] = Field(default=..., description="""List of authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', 'domain_of': ['Paper']} }) + doi: Optional[str] = Field(default=None, description="""DOI of the publication.""", json_schema_extra = { "linkml_meta": {'alias': 'doi', 'domain_of': ['Paper']} }) + date: Optional[str] = Field(default=None, description="""Date of the publication or the study.""", json_schema_extra = { "linkml_meta": {'alias': 'date', 'domain_of': ['Paper']} }) + experiments: Optional[List[Experiment]] = Field(default=None, description="""List of experiments mentioned in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'domain_of': ['Paper']} }) + + +class Author(ConfiguredBaseModel): + """ + Represents an author of the paper. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', 'domain_of': ['Author']} }) + + +class Experiment(ConfiguredBaseModel): + """ + A minimal experiment class (could be Biolog or otherwise). + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + experiment_title: Optional[str] = Field(default=None, description="""Title or short description of the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_title', 'domain_of': ['Experiment']} }) + is_biolog_experiment: Optional[bool] = Field(default=None, description="""Flag indicating if this experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', 'domain_of': ['Experiment']} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +Paper.model_rebuild() +Author.model_rebuild() +Experiment.model_rebuild() + From ac0db1c8b4c6ab8d8911571b4c0d576c242f542c Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 11:07:04 -0500 Subject: [PATCH 03/25] Import core NamedEntity and use example for paper title --- src/ontogpt/templates/biolog.py | 300 +++++++++++++++++++++++++++++- src/ontogpt/templates/biolog.yaml | 9 +- 2 files changed, 296 insertions(+), 13 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 2964af4da..0d4c2c663 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -68,7 +68,7 @@ def __contains__(self, key:str) -> bool: 'description': 'Schema for extracting structured data from papers, including ' 'Biolog Phenotype MicroArray experiments.', 'id': 'https://example.org/PaperExtractionSchema', - 'imports': ['linkml:types'], + 'imports': ['linkml:types', 'core'], 'license': 'https://creativecommons.org/publicdomain/zero/1.0/', 'name': 'PaperExtractionSchema', 'prefixes': {'linkml': {'prefix_prefix': 'linkml', @@ -80,6 +80,12 @@ def __contains__(self, key:str) -> bool: 'source_file': 'src/ontogpt/templates/biolog.yaml', 'title': 'Paper Extraction Schema'} ) +class NullDataOptions(str, Enum): + UNSPECIFIED_METHOD_OF_ADMINISTRATION = "UNSPECIFIED_METHOD_OF_ADMINISTRATION" + NOT_APPLICABLE = "NOT_APPLICABLE" + NOT_MENTIONED = "NOT_MENTIONED" + + class FlexibleBooleanEnum(str, Enum): """ A minimal enumeration for capturing yes/no/true/false in lowercase, plus 'not provided' if data are missing or ambiguous. @@ -93,15 +99,148 @@ class FlexibleBooleanEnum(str, Enum): -class NamedEntity(ConfiguredBaseModel): +class ExtractionResult(ConfiguredBaseModel): """ - A generic named entity for all non-root classes. + A result of extracting knowledge on text """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + input_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_id', 'domain_of': ['ExtractionResult']} }) + input_title: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_title', 'domain_of': ['ExtractionResult']} }) + input_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_text', 'domain_of': ['ExtractionResult']} }) + raw_completion_output: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'raw_completion_output', 'domain_of': ['ExtractionResult']} }) + prompt: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'prompt', 'domain_of': ['ExtractionResult']} }) + extracted_object: Optional[Any] = Field(default=None, description="""The complex objects extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'extracted_object', 'domain_of': ['ExtractionResult']} }) + named_entities: Optional[List[Any]] = Field(default=None, description="""Named entities extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'named_entities', 'domain_of': ['ExtractionResult']} }) + + +class NamedEntity(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class CompoundExpression(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) pass +class Triple(CompoundExpression): + """ + Abstract parent for Relation Extraction tasks + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject', 'domain_of': ['Triple']} }) + predicate: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'predicate', 'domain_of': ['Triple']} }) + object: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object', 'domain_of': ['Triple']} }) + qualifier: Optional[str] = Field(default=None, description="""A qualifier for the statements, e.g. \"NOT\" for negation""", json_schema_extra = { "linkml_meta": {'alias': 'qualifier', 'domain_of': ['Triple']} }) + subject_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the subject of the statement, e.g. \"high dose\" or \"intravenously administered\"""", json_schema_extra = { "linkml_meta": {'alias': 'subject_qualifier', 'domain_of': ['Triple']} }) + object_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the object of the statement, e.g. \"severe\" or \"with additional complications\"""", json_schema_extra = { "linkml_meta": {'alias': 'object_qualifier', 'domain_of': ['Triple']} }) + + +class TextWithTriples(ConfiguredBaseModel): + """ + A text containing one or more relations of the Triple type. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + triples: Optional[List[Triple]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'triples', 'domain_of': ['TextWithTriples']} }) + + +class TextWithEntity(ConfiguredBaseModel): + """ + A text containing one or more instances of a single type of entity. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + entities: Optional[List[str]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'entities', 'domain_of': ['TextWithEntity']} }) + + +class RelationshipType(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core', + 'id_prefixes': ['RO', 'biolink']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Publication(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: Optional[str] = Field(default=None, description="""The publication identifier""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'domain_of': ['NamedEntity', 'Publication']} }) + title: Optional[str] = Field(default=None, description="""The title of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'title', 'domain_of': ['Publication']} }) + abstract: Optional[str] = Field(default=None, description="""The abstract of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'abstract', 'domain_of': ['Publication']} }) + combined_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'combined_text', 'domain_of': ['Publication']} }) + full_text: Optional[str] = Field(default=None, description="""The full text of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'full_text', 'domain_of': ['Publication']} }) + + +class AnnotatorResult(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject_text', 'domain_of': ['AnnotatorResult']} }) + object_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_id', 'domain_of': ['AnnotatorResult']} }) + object_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_text', 'domain_of': ['AnnotatorResult']} }) + + class Paper(ConfiguredBaseModel): """ A single paper or study. @@ -117,7 +256,7 @@ class Paper(ConfiguredBaseModel): 'value': 'Root Exudates Alter the Expression of Diverse ' 'Metabolic, Transport, Regulatory, and Stress Response ' 'Genes in Rhizosphere Pseudomonas.'}]} }) - authors: List[Author] = Field(default=..., description="""Authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', + authors: List[str] = Field(default=..., description="""Authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a list of authors (each as a ' 'separate entry).\n' @@ -138,7 +277,7 @@ class Paper(ConfiguredBaseModel): "found, use 'Not provided'."}}, 'domain_of': ['Paper'], 'examples': [{'value': '14 April 2021'}]} }) - experiments: List[Experiment] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', + experiments: List[str] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a list of experiments described ' 'in the paper.\n' @@ -161,6 +300,33 @@ class Author(NamedEntity): "none, return 'Not provided'."}}, 'domain_of': ['Author'], 'examples': [{'value': 'Olga V. Mavrodi'}]} }) + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v class Experiment(NamedEntity): @@ -190,7 +356,7 @@ class Experiment(NamedEntity): "none, 'Not provided'."}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Lab-based assay under 25°C'}]} }) - host: Optional[List[Host]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host', + host: Optional[List[str]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Identify the host organisms or systems ' 'used.\n' @@ -205,7 +371,7 @@ class Experiment(NamedEntity): "'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Plant'}]} }) - target_microbes: Optional[List[Microbe]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', + target_microbes: Optional[List[str]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a list of target microbes ' "studied. If none are mentioned, 'Not " @@ -226,7 +392,7 @@ class Experiment(NamedEntity): "If not stated, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'pH 7, 25°C'}]} }) - experimental_factors: Optional[List[ExperimentalFactor]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', + experimental_factors: Optional[List[str]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'List experimental factors or ' 'treatments.\n' @@ -331,6 +497,33 @@ class Experiment(NamedEntity): "If not stated, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': '48 h'}]} }) + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v class Host(NamedEntity): @@ -346,6 +539,33 @@ class Host(NamedEntity): "If not found, 'Not provided'.\n"}}, 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], 'examples': [{'value': 'Host name: B. distachyon'}]} }) + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v class Microbe(NamedEntity): @@ -360,6 +580,33 @@ class Microbe(NamedEntity): "If none, 'Not provided'.\n"}}, 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v class ExperimentalFactor(NamedEntity): @@ -375,11 +622,46 @@ class ExperimentalFactor(NamedEntity): "If absent, 'Not provided'.\n"}}, 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], 'examples': [{'value': 'Root exudates concentration'}]} }) + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v # Model rebuild # see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +ExtractionResult.model_rebuild() NamedEntity.model_rebuild() +CompoundExpression.model_rebuild() +Triple.model_rebuild() +TextWithTriples.model_rebuild() +TextWithEntity.model_rebuild() +RelationshipType.model_rebuild() +Publication.model_rebuild() +AnnotatorResult.model_rebuild() Paper.model_rebuild() Author.model_rebuild() Experiment.model_rebuild() diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 06a10475c..16895ceab 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -15,6 +15,7 @@ default_range: string imports: - linkml:types + - core ######################## # ENUMS @@ -37,8 +38,6 @@ enums: ######################## classes: - NamedEntity: - description: "A generic named entity for all non-root classes." Paper: tree_root: true @@ -50,8 +49,10 @@ classes: - value: "Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas." description: "From Mavrodi et al. 2021" annotations: - prompt: "Extract the paper's title. If not found, return 'Not provided'." - range: string + prompt: >- + Extract the paper's title. If not found, return 'Not provided'. + prompt.examples: + - value: "Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas." required: true authors: From 0c81c178feff8d034a036155ba6ac427aa43d3d9 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 11:08:21 -0500 Subject: [PATCH 04/25] Fix prompt example --- src/ontogpt/templates/biolog.py | 9 ++++++++- src/ontogpt/templates/biolog.yaml | 4 ++-- 2 files changed, 10 insertions(+), 3 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 0d4c2c663..e12928752 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -250,7 +250,14 @@ class Paper(ConfiguredBaseModel): study_title: str = Field(default=..., description="""The paper's title.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', 'annotations': {'prompt': {'tag': 'prompt', 'value': "Extract the paper's title. If not found, " - "return 'Not provided'."}}, + "return only 'Not provided'."}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Root Exudates Alter the ' + 'Expression of Diverse ' + 'Metabolic, Transport, ' + 'Regulatory, and Stress Response ' + 'Genes in Rhizosphere ' + 'Pseudomonas.'}}, 'domain_of': ['Paper'], 'examples': [{'description': 'From Mavrodi et al. 2021', 'value': 'Root Exudates Alter the Expression of Diverse ' diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 16895ceab..67f556db5 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -50,9 +50,9 @@ classes: description: "From Mavrodi et al. 2021" annotations: prompt: >- - Extract the paper's title. If not found, return 'Not provided'. + Extract the paper's title. If not found, return only 'Not provided'. prompt.examples: - - value: "Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas." + Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas. required: true authors: From 59b22a6761a050f31f082a6ba5779069ca9d3cb2 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 11:15:26 -0500 Subject: [PATCH 05/25] Extract full author list --- src/ontogpt/templates/biolog.py | 46 ++----------------------------- src/ontogpt/templates/biolog.yaml | 32 +++++++++++---------- 2 files changed, 20 insertions(+), 58 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index e12928752..4afcbf0ae 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -265,8 +265,8 @@ class Paper(ConfiguredBaseModel): 'Genes in Rhizosphere Pseudomonas.'}]} }) authors: List[str] = Field(default=..., description="""Authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract a list of authors (each as a ' - 'separate entry).\n' + 'value': 'Extract a list of authors, semicolon ' + 'delimited.\n' 'If multiple authors appear in one ' 'string, please split them carefully.\n' "If not found, use 'Not provided'.\n"}}, @@ -295,47 +295,6 @@ class Paper(ConfiguredBaseModel): 'expression'}]} }) -class Author(NamedEntity): - """ - An author of the paper. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the full name of this author. If ' - "none, return 'Not provided'."}}, - 'domain_of': ['Author'], - 'examples': [{'value': 'Olga V. Mavrodi'}]} }) - id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', - 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, - 'comments': ['this is populated during the grounding and normalization step'], - 'domain_of': ['NamedEntity', 'Publication']} }) - label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', - 'aliases': ['name'], - 'annotations': {'owl': {'tag': 'owl', - 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], - 'slot_uri': 'rdfs:label'} }) - original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', - 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, - 'comments': ['This is determined during grounding and normalization', - 'But is based on the full input text'], - 'domain_of': ['NamedEntity']} }) - - @field_validator('original_spans') - def pattern_original_spans(cls, v): - pattern=re.compile(r"^\d+:\d+$") - if isinstance(v,list): - for element in v: - if isinstance(v, str) and not pattern.match(element): - raise ValueError(f"Invalid original_spans format: {element}") - elif isinstance(v,str): - if not pattern.match(v): - raise ValueError(f"Invalid original_spans format: {v}") - return v - - class Experiment(NamedEntity): """ A single experiment from the paper. @@ -670,7 +629,6 @@ def pattern_original_spans(cls, v): Publication.model_rebuild() AnnotatorResult.model_rebuild() Paper.model_rebuild() -Author.model_rebuild() Experiment.model_rebuild() Host.model_rebuild() Microbe.model_rebuild() diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 67f556db5..0320a4730 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -62,10 +62,11 @@ classes: - value: "Janiece R. McWilliams" annotations: prompt: | - Extract a list of authors (each as a separate entry). + Extract a list of authors, semicolon delimited. If multiple authors appear in one string, please split them carefully. If not found, use 'Not provided'. - range: Author + # range: Author + range: string multivalued: true required: true @@ -99,18 +100,21 @@ classes: multivalued: true required: true - Author: - is_a: NamedEntity - description: "An author of the paper." - attributes: - name: - description: "Name of the author." - examples: - - value: "Olga V. Mavrodi" - annotations: - prompt: "Extract the full name of this author. If none, return 'Not provided'." - range: string - required: true + # Author class is redundant if name only, + # but can be useful if more author details are needed. + + # Author: + # is_a: NamedEntity + # description: "An author of the paper." + # attributes: + # name: + # description: "Name of the author." + # examples: + # - value: "Olga V. Mavrodi" + # annotations: + # prompt: "Extract the full name of this author. If none, return 'Not provided'." + # range: string + # required: true Experiment: is_a: NamedEntity From 0bb91e35cb229280aee6a29b6d13ddb6fbb48b7c Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 11:35:11 -0500 Subject: [PATCH 06/25] Restructure experiments slot --- src/ontogpt/templates/biolog.py | 56 ++++++++++++++++++++++++++++--- src/ontogpt/templates/biolog.yaml | 28 ++++++++++++++-- 2 files changed, 78 insertions(+), 6 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 4afcbf0ae..e4d493095 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -286,10 +286,58 @@ class Paper(ConfiguredBaseModel): 'examples': [{'value': '14 April 2021'}]} }) experiments: List[str] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract a list of experiments described ' - 'in the paper.\n' - "If nothing is mentioned, use 'Not " - "provided'.\n"}}, + 'value': 'Extract a list of brief descriptions of ' + 'experiments\n' + 'described in the paper. This list must ' + 'be semicolon-delimited.\n' + 'For each description, include all of the ' + 'following details if\n' + 'provided: experiment motivation (why it ' + 'was performed),\n' + 'experiment design (how it was performed, ' + 'including methods,\n' + 'tools, organisms, and chemicals used), ' + 'environment (the location\n' + 'and conditions in which the experiment ' + 'occurred, including\n' + 'metrics like temperature), all organisms ' + 'used and their\n' + 'high-level type (e.g., plant, animal), ' + 'the biological system\n' + 'under study (e.g., rhizosphere), ' + 'experimental conditions\n' + '(temperature, pH, etc.), and ' + 'experimental factors tested or ' + 'measured.\n' + 'Also note whether the experiment is a ' + 'Biolog experiment and if so,\n' + 'the type or subtype of experiment (e.g., ' + 'Phenotype MicroArray),\n' + 'the Biolog plates used (e.g., PM01), the ' + 'number of replicates, all\n' + 'key steps in the experimental protocol, ' + 'the types of data collected\n' + 'by the plate reader (e.g., OD, ' + 'respiration), the protocol for\n' + 'measuring optical density (e.g., OD600 ' + 'measured every 2 hours), \n' + 'the protocol for measuring respiration ' + '(e.g., Colorimetric change\n' + 'at 590 nm for formazan detection), the ' + 'instrument or equipment used\n' + '(e.g., OmniLog Phenotype MicroArray ' + 'System), the software used for\n' + 'data analysis (e.g., OmniLog Parametric ' + 'Analysis software), the\n' + 'incubation temperature (e.g., 25°C), the ' + 'total duration of incubation\n' + '(e.g., 48 h), and any other relevant ' + 'details.\n' + 'If nothing is mentioned regarding ' + "experiments, use 'Not provided'.\n" + 'Do not provide any details here not ' + 'related to experiments.\n' + 'Do not include newlines.\n'}}, 'domain_of': ['Paper'], 'examples': [{'value': 'Biolog Phenotype MicroArray; RNA-seq for gene ' 'expression'}]} }) diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 0320a4730..e8bf316bf 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -94,8 +94,32 @@ classes: - value: "Biolog Phenotype MicroArray; RNA-seq for gene expression" annotations: prompt: | - Extract a list of experiments described in the paper. - If nothing is mentioned, use 'Not provided'. + Extract a list of brief descriptions of experiments + described in the paper. This list must be semicolon-delimited. + For each description, include all of the following details if + provided: experiment motivation (why it was performed), + experiment design (how it was performed, including methods, + tools, organisms, and chemicals used), environment (the location + and conditions in which the experiment occurred, including + metrics like temperature), all organisms used and their + high-level type (e.g., plant, animal), the biological system + under study (e.g., rhizosphere), experimental conditions + (temperature, pH, etc.), and experimental factors tested or measured. + Also note whether the experiment is a Biolog experiment and if so, + the type or subtype of experiment (e.g., Phenotype MicroArray), + the Biolog plates used (e.g., PM01), the number of replicates, all + key steps in the experimental protocol, the types of data collected + by the plate reader (e.g., OD, respiration), the protocol for + measuring optical density (e.g., OD600 measured every 2 hours), + the protocol for measuring respiration (e.g., Colorimetric change + at 590 nm for formazan detection), the instrument or equipment used + (e.g., OmniLog Phenotype MicroArray System), the software used for + data analysis (e.g., OmniLog Parametric Analysis software), the + incubation temperature (e.g., 25°C), the total duration of incubation + (e.g., 48 h), and any other relevant details. + If nothing is mentioned regarding experiments, use 'Not provided'. + Do not provide any details here not related to experiments. + Do not include newlines. range: Experiment multivalued: true required: true From f694b9c45dd8cf1c52dcab574c6a7babe122f97b Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 11:46:27 -0500 Subject: [PATCH 07/25] Tuning slots in Experiment class --- src/ontogpt/templates/biolog.py | 75 ++++++++++++++++++++----------- src/ontogpt/templates/biolog.yaml | 53 +++++++++++++++------- 2 files changed, 86 insertions(+), 42 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index e4d493095..d829a32f1 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -284,7 +284,7 @@ class Paper(ConfiguredBaseModel): "found, use 'Not provided'."}}, 'domain_of': ['Paper'], 'examples': [{'value': '14 April 2021'}]} }) - experiments: List[str] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', + experiments: List[Experiment] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a list of brief descriptions of ' 'experiments\n' @@ -373,23 +373,25 @@ class Experiment(NamedEntity): host: Optional[List[str]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Identify the host organisms or systems ' - 'used.\n' - "If not mentioned, use 'Not " - "provided'.\n"}}, + 'used\n' + 'in a semicolon-delimited list.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Brachypodium distachyon; ' + 'Arabidopsis thaliana'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Brachypodium distachyon'}]} }) host_type: Optional[str] = Field(default=None, description="""Type of host (e.g., plant, animal).""", json_schema_extra = { "linkml_meta": {'alias': 'host_type', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Describe the host type if applicable ' "(e.g. 'plant'). If nothing specified, " - "'Not provided'.\n"}}, + "'Not provided'."}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Plant'}]} }) target_microbes: Optional[List[str]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a list of target microbes ' - "studied. If none are mentioned, 'Not " - "provided'.\n"}}, + 'studied in\n' + 'a semicolon-delimited list.\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) biological_system: Optional[str] = Field(default=None, description="""Biological system under study.""", json_schema_extra = { "linkml_meta": {'alias': 'biological_system', @@ -399,17 +401,19 @@ class Experiment(NamedEntity): "provided'."}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Root exudates system'}]} }) - conditions: Optional[str] = Field(default=None, description="""Experimental conditions (temperature, pH, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'conditions', + conditions: Optional[List[str]] = Field(default=None, description="""Experimental conditions (temperature, pH, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'conditions', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Identify relevant experimental ' - 'conditions (e.g., 25°C, pH 7).\n' + 'conditions in a\n' + 'semicolon-delimited list (e.g., 25°C; pH ' + '7).\n' "If not stated, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'pH 7, 25°C'}]} }) experimental_factors: Optional[List[str]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'List experimental factors or ' - 'treatments.\n' + 'value': 'List experimental factors or treatments\n' + 'in a semicolon-delimited list.\n' "If no info, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Root exudates presence vs. absence'}]} }) @@ -430,25 +434,37 @@ class Experiment(NamedEntity): 'examples': [{'value': 'Phenotype MicroArray'}]} }) plates: Optional[List[str]] = Field(default=None, description="""Biolog plates used (e.g., EcoPlate, GEN III, PM01, PM02A).""", json_schema_extra = { "linkml_meta": {'alias': 'plates', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the name(s) of any Biolog plates ' - '(e.g., PM01, PM02A).\n' - "If not mentioned, 'Not provided'.\n"}}, + 'value': 'Extract the name(s) of any Biolog ' + 'plates\n' + '(e.g., PM01, PM02A) in a ' + 'semicolon-delimited list.\n' + "If not mentioned, 'Not provided'.\n"}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'PM01; PM02A'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'PM01'}, {'value': 'PM02A'}]} }) replicates: Optional[str] = Field(default=None, description="""Number of replicates.""", json_schema_extra = { "linkml_meta": {'alias': 'replicates', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the number of replicates.\n' + 'value': 'Extract the number of replicates as a ' + 'single number.\n' "If not mentioned, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': '3'}]} }) protocol_steps: Optional[List[str]] = Field(default=None, description="""Key steps in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'protocol_steps', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'List the full protocol steps or key ' - 'experimental procedures, including\n' - 'inoculation details, incubation ' - 'times/temperatures, data recording ' - 'intervals, etc.\n' - "If not provided, 'Not provided'.\n"}}, + 'experimental procedures,\n' + 'including inoculation details, ' + 'incubation times/temperatures,\n' + 'data recording intervals, and data ' + 'analysis steps.\n' + 'This should be a semicolon-delimited ' + 'list.\n' + 'Do not include newlines.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '1) Grow cells overnight on LB ' + 'agar; 2) Harvest and suspend in ' + 'IF-0'}}, 'domain_of': ['Experiment'], 'examples': [{'value': '1) Grow cells overnight on LB agar 2) Harvest and ' 'suspend in IF-0 3) Adjust transmittance to 42% ' @@ -461,7 +477,9 @@ class Experiment(NamedEntity): plate_reader_types_of_data_collected: Optional[List[str]] = Field(default=None, description="""Data types collected (OD, respiration, formazan formation, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'plate_reader_types_of_data_collected', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Identify data types measured by the ' - 'plate reader (e.g. formazan, OD).\n' + 'plate reader\n' + '(e.g. formazan, OD) in a semi-colon ' + 'delimited list.\n' "If none, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Formazan production, recorded by OmniLog system'}]} }) @@ -484,16 +502,21 @@ class Experiment(NamedEntity): instrument_used: Optional[List[str]] = Field(default=None, description="""Instrument or equipment used for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'instrument_used', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract the name(s) of the ' - 'instrument/equipment used (e.g., OmniLog ' - 'system).\n' + 'instrument/equipment used\n' + '(e.g., OmniLog system) in a ' + 'semicolon-delimited list.\n' "If none, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'OmniLog Phenotype MicroArray System (Biolog)'}]} }) analysis_software: Optional[List[str]] = Field(default=None, description="""Software used to analyze or interpret data.""", json_schema_extra = { "linkml_meta": {'alias': 'analysis_software', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract any software or platform used ' - 'for data analysis.\n' - "If not mentioned, 'Not provided'.\n"}}, + 'value': 'Extract any software or platforms used ' + 'for data analysis\n' + 'in a semicolon-delimited list.\n' + "If not mentioned, 'Not provided'.\n"}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'OmniLog Parametric Analysis ' + 'software v1.20.02'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'OmniLog Parametric Analysis software v1.20.02 ' '(Biolog)'}]} }) diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index e8bf316bf..81edc1c43 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -121,6 +121,8 @@ classes: Do not provide any details here not related to experiments. Do not include newlines. range: Experiment + inlined: true + inlined_as_list: true multivalued: true required: true @@ -179,8 +181,10 @@ classes: - value: "Brachypodium distachyon" annotations: prompt: | - Identify the host organisms or systems used. - If not mentioned, use 'Not provided'. + Identify the host organisms or systems used + in a semicolon-delimited list. + prompt.examples: + Brachypodium distachyon; Arabidopsis thaliana range: Host multivalued: true required: false @@ -190,8 +194,9 @@ classes: examples: - value: "Plant" annotations: - prompt: | - Describe the host type if applicable (e.g. 'plant'). If nothing specified, 'Not provided'. + prompt: >- + Describe the host type if applicable (e.g. 'plant'). + If nothing specified, 'Not provided'. range: string required: false @@ -201,7 +206,8 @@ classes: - value: "Pseudomonas fluorescens" annotations: prompt: | - Extract a list of target microbes studied. If none are mentioned, 'Not provided'. + Extract a list of target microbes studied in + a semicolon-delimited list. range: Microbe multivalued: true required: false @@ -211,7 +217,9 @@ classes: examples: - value: "Root exudates system" annotations: - prompt: "Extract the focal biological system (e.g., rhizosphere). If none, 'Not provided'." + prompt: >- + Extract the focal biological system (e.g., rhizosphere). + If none, 'Not provided'. range: string required: false @@ -221,8 +229,10 @@ classes: - value: "pH 7, 25°C" annotations: prompt: | - Identify relevant experimental conditions (e.g., 25°C, pH 7). + Identify relevant experimental conditions in a + semicolon-delimited list (e.g., 25°C; pH 7). If not stated, 'Not provided'. + multivalued: true range: string required: false @@ -232,7 +242,8 @@ classes: - value: "Root exudates presence vs. absence" annotations: prompt: | - List experimental factors or treatments. + List experimental factors or treatments + in a semicolon-delimited list. If no info, 'Not provided'. range: ExperimentalFactor multivalued: true @@ -268,8 +279,10 @@ classes: - value: "PM02A" annotations: prompt: | - Extract the name(s) of any Biolog plates (e.g., PM01, PM02A). + Extract the name(s) of any Biolog plates + (e.g., PM01, PM02A) in a semicolon-delimited list. If not mentioned, 'Not provided'. + prompt.examples: PM01; PM02A range: string multivalued: true required: false @@ -280,7 +293,7 @@ classes: - value: "3" annotations: prompt: | - Extract the number of replicates. + Extract the number of replicates as a single number. If not mentioned, 'Not provided'. range: string required: false @@ -299,9 +312,13 @@ classes: 8) Data analysis with OmniLog Parametric Analysis software annotations: prompt: | - List the full protocol steps or key experimental procedures, including - inoculation details, incubation times/temperatures, data recording intervals, etc. - If not provided, 'Not provided'. + List the full protocol steps or key experimental procedures, + including inoculation details, incubation times/temperatures, + data recording intervals, and data analysis steps. + This should be a semicolon-delimited list. + Do not include newlines. + prompt.examples: + 1) Grow cells overnight on LB agar; 2) Harvest and suspend in IF-0 range: string multivalued: true required: false @@ -312,7 +329,8 @@ classes: - value: "Formazan production, recorded by OmniLog system" annotations: prompt: | - Identify data types measured by the plate reader (e.g. formazan, OD). + Identify data types measured by the plate reader + (e.g. formazan, OD) in a semi-colon delimited list. If none, 'Not provided'. range: string multivalued: true @@ -346,7 +364,8 @@ classes: - value: "OmniLog Phenotype MicroArray System (Biolog)" annotations: prompt: | - Extract the name(s) of the instrument/equipment used (e.g., OmniLog system). + Extract the name(s) of the instrument/equipment used + (e.g., OmniLog system) in a semicolon-delimited list. If none, 'Not provided'. range: string multivalued: true @@ -358,8 +377,10 @@ classes: - value: "OmniLog Parametric Analysis software v1.20.02 (Biolog)" annotations: prompt: | - Extract any software or platform used for data analysis. + Extract any software or platforms used for data analysis + in a semicolon-delimited list. If not mentioned, 'Not provided'. + prompt.examples: OmniLog Parametric Analysis software v1.20.02 range: string multivalued: true required: false From 7a6d63b8bad632faa59b669f62386b3d2fc7a904 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 11:59:52 -0500 Subject: [PATCH 08/25] Tuning for subclasses --- src/ontogpt/templates/biolog.py | 47 +++++++++++++------------------ src/ontogpt/templates/biolog.yaml | 29 ++++++++----------- 2 files changed, 30 insertions(+), 46 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index d829a32f1..a786e91b5 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -366,8 +366,7 @@ class Experiment(NamedEntity): 'with/without root exudates.'}]} }) environment: Optional[str] = Field(default=None, description="""Location or conditions in which the experiment occurred.""", json_schema_extra = { "linkml_meta": {'alias': 'environment', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the environment or setting. If ' - "none, 'Not provided'."}}, + 'value': 'Extract the environment or setting.'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Lab-based assay under 25°C'}]} }) host: Optional[List[str]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host', @@ -383,8 +382,7 @@ class Experiment(NamedEntity): host_type: Optional[str] = Field(default=None, description="""Type of host (e.g., plant, animal).""", json_schema_extra = { "linkml_meta": {'alias': 'host_type', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Describe the host type if applicable ' - "(e.g. 'plant'). If nothing specified, " - "'Not provided'."}}, + "(e.g. 'plant')."}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Plant'}]} }) target_microbes: Optional[List[str]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', @@ -397,8 +395,7 @@ class Experiment(NamedEntity): biological_system: Optional[str] = Field(default=None, description="""Biological system under study.""", json_schema_extra = { "linkml_meta": {'alias': 'biological_system', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract the focal biological system ' - "(e.g., rhizosphere). If none, 'Not " - "provided'."}}, + '(e.g., rhizosphere).'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Root exudates system'}]} }) conditions: Optional[List[str]] = Field(default=None, description="""Experimental conditions (temperature, pH, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'conditions', @@ -406,15 +403,13 @@ class Experiment(NamedEntity): 'value': 'Identify relevant experimental ' 'conditions in a\n' 'semicolon-delimited list (e.g., 25°C; pH ' - '7).\n' - "If not stated, 'Not provided'.\n"}}, + '7).\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'pH 7, 25°C'}]} }) experimental_factors: Optional[List[str]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'List experimental factors or treatments\n' - 'in a semicolon-delimited list.\n' - "If no info, 'Not provided'.\n"}}, + 'in a semicolon-delimited list.\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Root exudates presence vs. absence'}]} }) is_biolog_experiment: Optional[FlexibleBooleanEnum] = Field(default=None, description="""Whether the experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', @@ -512,8 +507,7 @@ class Experiment(NamedEntity): 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract any software or platforms used ' 'for data analysis\n' - 'in a semicolon-delimited list.\n' - "If not mentioned, 'Not provided'.\n"}, + 'in a semicolon-delimited list.\n'}, 'prompt.examples': {'tag': 'prompt.examples', 'value': 'OmniLog Parametric Analysis ' 'software v1.20.02'}}, @@ -523,15 +517,13 @@ class Experiment(NamedEntity): incubation_temperature: Optional[str] = Field(default=None, description="""Temperature used for incubating the plates or cultures.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_temperature', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract the incubation temperature if ' - 'provided (e.g. 25°C).\n' - "If not mentioned, 'Not provided'.\n"}}, + 'provided (e.g. 25°C).\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': '25°C'}]} }) incubation_duration: Optional[str] = Field(default=None, description="""Total duration of incubation (e.g., 48 h).""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_duration', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract how long the incubation lasted ' - '(e.g. 48 hours).\n' - "If not stated, 'Not provided'.\n"}}, + '(e.g. 48 hours).\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': '48 h'}]} }) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', @@ -565,17 +557,16 @@ def pattern_original_spans(cls, v): class Host(NamedEntity): """ - A host organism or system. + Detail of a host organism or system. """ linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Host.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + name: Optional[str] = Field(default=None, description="""Name of the host organism or system.""", json_schema_extra = { "linkml_meta": {'alias': 'name', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract details about this host ' - '(placeholder).\n' - "If not found, 'Not provided'.\n"}}, - 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], - 'examples': [{'value': 'Host name: B. distachyon'}]} }) + 'value': 'The name of the host organism or ' + 'system.\n'}}, + 'domain_of': ['Host', 'Microbe'], + 'examples': [{'value': 'B. distachyon'}]} }) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, 'comments': ['this is populated during the grounding and normalization step'], @@ -611,11 +602,11 @@ class Microbe(NamedEntity): """ linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Microbe.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + name: Optional[str] = Field(default=None, description="""Name of a microbial taxon.""", json_schema_extra = { "linkml_meta": {'alias': 'name', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract details about this microbe.\n' - "If none, 'Not provided'.\n"}}, - 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], + 'value': 'The name of the microbial species or ' + 'strain.\n'}}, + 'domain_of': ['Host', 'Microbe'], 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -657,7 +648,7 @@ class ExperimentalFactor(NamedEntity): 'value': 'Extract details about this experimental ' 'factor (e.g. concentration).\n' "If absent, 'Not provided'.\n"}}, - 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor'], + 'domain_of': ['ExperimentalFactor'], 'examples': [{'value': 'Root exudates concentration'}]} }) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 81edc1c43..23d073645 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -171,7 +171,7 @@ classes: examples: - value: "Lab-based assay under 25°C" annotations: - prompt: "Extract the environment or setting. If none, 'Not provided'." + prompt: Extract the environment or setting. range: string required: false @@ -196,7 +196,6 @@ classes: annotations: prompt: >- Describe the host type if applicable (e.g. 'plant'). - If nothing specified, 'Not provided'. range: string required: false @@ -219,7 +218,6 @@ classes: annotations: prompt: >- Extract the focal biological system (e.g., rhizosphere). - If none, 'Not provided'. range: string required: false @@ -231,7 +229,6 @@ classes: prompt: | Identify relevant experimental conditions in a semicolon-delimited list (e.g., 25°C; pH 7). - If not stated, 'Not provided'. multivalued: true range: string required: false @@ -244,7 +241,6 @@ classes: prompt: | List experimental factors or treatments in a semicolon-delimited list. - If no info, 'Not provided'. range: ExperimentalFactor multivalued: true required: false @@ -379,7 +375,6 @@ classes: prompt: | Extract any software or platforms used for data analysis in a semicolon-delimited list. - If not mentioned, 'Not provided'. prompt.examples: OmniLog Parametric Analysis software v1.20.02 range: string multivalued: true @@ -392,7 +387,6 @@ classes: annotations: prompt: | Extract the incubation temperature if provided (e.g. 25°C). - If not mentioned, 'Not provided'. range: string required: false @@ -403,37 +397,36 @@ classes: annotations: prompt: | Extract how long the incubation lasted (e.g. 48 hours). - If not stated, 'Not provided'. range: string required: false + # Placeholder, to be expanded Host: is_a: NamedEntity - description: "A host organism or system." + description: "Detail of a host organism or system." attributes: - placeholder_attribute: - description: "Placeholder attribute for Host." + name: + description: "Name of the host organism or system." examples: - - value: "Host name: B. distachyon" + - value: "B. distachyon" annotations: prompt: | - Extract details about this host (placeholder). - If not found, 'Not provided'. + The name of the host organism or system. range: string required: false + # Placeholder, to be expanded Microbe: is_a: NamedEntity description: "Information about a microbial species or strain." attributes: - placeholder_attribute: - description: "Placeholder attribute for Microbe." + name: + description: "Name of a microbial taxon." examples: - value: "Pseudomonas fluorescens" annotations: prompt: | - Extract details about this microbe. - If none, 'Not provided'. + The name of the microbial species or strain. range: string required: false From 0b60a23012277d0705defbf312dba9425c58246a Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 4 Mar 2025 12:17:39 -0500 Subject: [PATCH 09/25] Add description field to Experiment --- src/ontogpt/templates/biolog.py | 44 ++++++++----------------------- src/ontogpt/templates/biolog.yaml | 11 +++++--- 2 files changed, 19 insertions(+), 36 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index a786e91b5..805fd3b53 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -125,7 +125,7 @@ class NamedEntity(ConfiguredBaseModel): 'aliases': ['name'], 'annotations': {'owl': {'tag': 'owl', 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], + 'domain_of': ['NamedEntity', 'Experiment'], 'slot_uri': 'rdfs:label'} }) original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -202,7 +202,7 @@ class RelationshipType(NamedEntity): 'aliases': ['name'], 'annotations': {'owl': {'tag': 'owl', 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], + 'domain_of': ['NamedEntity', 'Experiment'], 'slot_uri': 'rdfs:label'} }) original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -343,12 +343,17 @@ class Paper(ConfiguredBaseModel): 'expression'}]} }) -class Experiment(NamedEntity): +class Experiment(ConfiguredBaseModel): """ A single experiment from the paper. """ linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + label: Optional[str] = Field(default=None, description="""A single sentence description of this experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'A single sentence description of this ' + 'experiment.\n'}}, + 'domain_of': ['NamedEntity', 'Experiment']} }) experiment_motivation: Optional[str] = Field(default=None, description="""Rationale for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_motivation', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Explain the motivation or purpose of the ' @@ -526,33 +531,6 @@ class Experiment(NamedEntity): '(e.g. 48 hours).\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': '48 h'}]} }) - id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', - 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, - 'comments': ['this is populated during the grounding and normalization step'], - 'domain_of': ['NamedEntity', 'Publication']} }) - label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', - 'aliases': ['name'], - 'annotations': {'owl': {'tag': 'owl', - 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], - 'slot_uri': 'rdfs:label'} }) - original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', - 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, - 'comments': ['This is determined during grounding and normalization', - 'But is based on the full input text'], - 'domain_of': ['NamedEntity']} }) - - @field_validator('original_spans') - def pattern_original_spans(cls, v): - pattern=re.compile(r"^\d+:\d+$") - if isinstance(v,list): - for element in v: - if isinstance(v, str) and not pattern.match(element): - raise ValueError(f"Invalid original_spans format: {element}") - elif isinstance(v,str): - if not pattern.match(v): - raise ValueError(f"Invalid original_spans format: {v}") - return v class Host(NamedEntity): @@ -575,7 +553,7 @@ class Host(NamedEntity): 'aliases': ['name'], 'annotations': {'owl': {'tag': 'owl', 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], + 'domain_of': ['NamedEntity', 'Experiment'], 'slot_uri': 'rdfs:label'} }) original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -616,7 +594,7 @@ class Microbe(NamedEntity): 'aliases': ['name'], 'annotations': {'owl': {'tag': 'owl', 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], + 'domain_of': ['NamedEntity', 'Experiment'], 'slot_uri': 'rdfs:label'} }) original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -658,7 +636,7 @@ class ExperimentalFactor(NamedEntity): 'aliases': ['name'], 'annotations': {'owl': {'tag': 'owl', 'value': 'AnnotationProperty, AnnotationAssertion'}}, - 'domain_of': ['NamedEntity'], + 'domain_of': ['NamedEntity', 'Experiment'], 'slot_uri': 'rdfs:label'} }) original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 23d073645..7ded619a3 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -121,8 +121,6 @@ classes: Do not provide any details here not related to experiments. Do not include newlines. range: Experiment - inlined: true - inlined_as_list: true multivalued: true required: true @@ -143,9 +141,16 @@ classes: # required: true Experiment: - is_a: NamedEntity description: "A single experiment from the paper." attributes: + + label: + description: >- + A single sentence description of this experiment. + annotations: + prompt: | + A single sentence description of this experiment. + experiment_motivation: description: "Rationale for the experiment." examples: From 098bf6ee27990c5530d47bfa8e03a1db8da16dd5 Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Tue, 4 Mar 2025 23:28:57 -0800 Subject: [PATCH 10/25] gpt 4.5 preview --- fmicb-12-651282__ontogpt_gpto4.5preview.txt | 3772 +++++++++++++++++++ 1 file changed, 3772 insertions(+) create mode 100644 fmicb-12-651282__ontogpt_gpto4.5preview.txt diff --git a/fmicb-12-651282__ontogpt_gpto4.5preview.txt b/fmicb-12-651282__ontogpt_gpto4.5preview.txt new file mode 100644 index 000000000..fadc88575 --- /dev/null +++ b/fmicb-12-651282__ontogpt_gpto4.5preview.txt @@ -0,0 +1,3772 @@ +The "poetry.dev-dependencies" section is deprecated and will be removed in a future version. Use "poetry.group.dev.dependencies" instead. +/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/pydantic/_internal/_config.py:345: UserWarning: Valid config keys have changed in V2: +* 'fields' has been removed + warnings.warn(message, UserWarning) +INFO:root:Logger root set to level 10 +INFO:root:Input file: /Users/marcin/Documents/VIMSS/ontology/LLMs/MATE-LLM/examples/mavrodi_etal/fmicb-12-651282.txt +INFO:root:Input text: ORIGINAL RESEARCH +published: 14 April 2021 +doi: 10.3389/fmicb.2021.651282 +Edited by: +Barbara Pivato, +Institut National de Recherche pour +l’agriculture, l’alimentation et +l’environnement (INRAE), France +Reviewed by: +Cara Helene Haney, +The University of British Columbia, +Canada +Xingang Zhou, +Northeast Agricultural University, +China +*Correspondence: +Dmitri V. Mavrodi +dmitri.mavrodi@usm.edu +Alex S. Flynt +alex.flynt@usm.edu +†These authors have contributed +equally to this work +Specialty section: +This article was submitted to +Microbe and Virus Interactions with +Plants, +a section of the journal +Frontiers in Microbiology +Received: 09 January 2021 +Accepted: 08 March 2021 +Published: 14 April 2021 +Citation: +Mavrodi OV, McWilliams JR, +Peter JO, Berim A, Hassan KA, +Elbourne LDH, LeTourneau MK, +Gang DR, Paulsen IT, Weller DM, +Thomashow LS, Flynt AS and +Mavrodi DV (2021) Root Exudates +Alter the Expression of Diverse +Metabolic, Transport, Regulatory, +and Stress Response Genes +in Rhizosphere Pseudomonas. +Front. Microbiol. 12:651282. +doi: 10.3389/fmicb.2021.651282 +Root Exudates Alter the Expression +of Diverse Metabolic, Transport, +Regulatory, and Stress Response +Genes in Rhizosphere Pseudomonas +Olga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2, +Karl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2, +Ian T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and +Dmitri V. Mavrodi1* +1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS, +United States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School +of Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular +Sciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics +and Quality Research Unit, Pullman, WA, United States +Plants live in association with microorganisms that positively influence plant +development, vigor, and fitness in response to pathogens and abiotic stressors. +The bulk of the plant microbiome is concentrated belowground at the plant rootsoil +interface. Plant roots secrete carbon-rich rhizodeposits containing primary and +secondary low molecular weight metabolites, lysates, and mucilages. These exudates +provide nutrients for soil microorganisms and modulate their affinity to host plants, but +molecular details of this process are largely unresolved. We addressed this gap by +focusing on the molecular dialog between eight well-characterized beneficial strains +of the Pseudomonas fluorescens group and Brachypodium distachyon, a model for +economically important food, feed, forage, and biomass crops of the grass family. We +collected and analyzed root exudates of B. distachyon and demonstrated the presence +of multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The +subsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many +of these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq +profiling of bacterial cultures amended with root exudates revealed changes in the +expression of genes encoding numerous catabolic and anabolic enzymes, transporters, +transcriptional regulators, stress response, and conserved hypothetical proteins. Almost +half of the differentially expressed genes mapped to the variable part of the strains’ +pangenome, reflecting the importance of the variable gene content in the adaptation of +P. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity +of cellular pathways and physiological responses underlying the establishment of +mutualistic interactions between these beneficial rhizobacteria and their plant hosts. +Keywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome +Frontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +INTRODUCTION +Plants are meta-organisms or holobionts that rely in part on +their microbiome for specific functions and traits. The ability +of the plant microbiome to influence plant development, vigor, +health, and fitness in response to abiotic stressors associated +with global climate change is documented by numerous studies +(Lugtenberg and Kamilova, 2009). There is mounting evidence +that plants actively recruit beneficial microbiomes, but many +aspects of this process are still very much a black box (Reinhold- +Hurek et al., 2015). The foundation for the differential affinity +of rhizobacteria toward host plants is built upon complex +chemical cross talk between microorganisms and plant roots. +Up to 40% of photosynthetically fixed carbon is released by +plant roots in the form of exudates and secretions, lysates, and +mucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990; +Badri and Vivanco, 2009). The release of these compounds is +actively controlled in response to environmental stimuli, and the +composition of root exudates varies greatly according to plant +species and physiological condition (Lynch, 1990; Nguyen, 2003; +Phillips et al., 2004; De-la-Pena et al., 2008). The presence and +composition of exudates strongly impact soil microorganisms, +which is consistent with the idea that plants actively select and +shape their root microbiota (Zolla et al., 2013). +Primary root exudates include simple and complex sugars, +amino acids, polypeptides and proteins, organic, aliphatic and +fatty acids, sterols, and phenolics (Nguyen, 2003; Badri and +Vivanco, 2009; Badri et al., 2009). These compounds serve as +carbon and energy sources for rhizobacteria, and the presence +of the intact corresponding catabolic pathways is essential +for competitive colonization of roots and disease suppression +(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and +Kamilova, 2009). Root exudates also contain numerous signal +molecules and secondary metabolites, the significance of which +is only now emerging (Walker et al., 2003; Bais et al., 2005, +2006). A handful of analyses of plant-induced gene expression +by transcriptional profiling in vitro (Mark et al., 2005) or in +the rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al., +2005; Matilla et al., 2007; Barret et al., 2009) have identified +multiple genes that are differentially regulated by exposure to +roots or root exudates. Bacterial pathways expressed during +rhizosphere colonization control utilization of plant-derived +metabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001; +Lugtenberg and Kamilova, 2009), motility and chemotaxis (de +Weert et al., 2002; Lugtenberg and Kamilova, 2009), phase +variation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002; +van den Broek et al., 2005), outer membrane integrity (de Weert +et al., 2006; Lugtenberg and Kamilova, 2009), and the ability +to sequester limiting resources (Raaijmakers et al., 1995) and +resist environmental stresses (Sarniguet et al., 1995; Miller and +Wood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In +its spatial and temporal properties, root colonization resembles +biofilm formation, and biofilm-related pathways also have been +implicated in adhesion to seeds and roots and rhizosphere +colonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003; +Yousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al., +2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally, +root exudates strongly affect the expression of diverse plant +growth promotion and biocontrol genes (Vacheron et al., 2013). +Over the past decade, the genomes of numerous rhizosphere +strains have been sequenced and analyzed, but functional +genomics studies of rhizosphere competence lag behind the +availability of sequence data. +This study explored the molecular dialog between the +model host plant Brachypodium distachyon and several wellcharacterized +rhizosphere strains of the Pseudomonas fluorescens +group. Brachypodium is a small annual grass originating in semiarid +regions of the Middle East that has emerged as a prime model +for economically important food, feed, forage, and biomass crops +of the grass family (Bevan et al., 2010; Schwartz et al., 2010; +Brkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The +biology, extensive collection of resources, and research tools +make B. distachyon an attractive model to investigate interactions +between plants and root-associated microbes. Pseudomonads +are ubiquitous Gram-negative g-proteobacteria that colonize +eukaryotic hosts and include both commensals and economically +important pathogens of plants and animals (Moore et al., +2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus +Pseudomonas currently comprises > 100 named species that have +been separated based on multilocus sequence analysis into 14 +species groups (Garrido-Sanz et al., 2016; Hesse et al., 2018). +The P. fluorescens group is the most diverse regarding both +the genetic distances within it, the number of species and the +large pangenome that makes up > 50% of the pangenome +of the genus as a whole (Loper et al., 2012). The group also +encompasses an unusually high proportion of strains that inhabit +the plant rhizosphere and possess plant growth promoting and +biocontrol properties. Naylor et al. (2017) profiled bacterial +communities associated with root tissues and rhizosphere of +18 different plant species of the Poaceae family. That study +identified Pseudomonas among taxa constituting the core grass +root microbiome and demonstrated that these bacteria were +enriched in C3 plants, including wheat, rye, barley, oat, and +Brachypodium. We confirmed the capacity of B. distachyon +Bd21 to serve as a host for rhizobacteria of the P. fluorescens +group in preliminary greenhouse assays with biocontrol strains +P. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens +Pf-5. Results of these experiments revealed that all strains +successfully established and colonized the roots of Brachypodium +(Supplementary Table 1). +In this study, we focused on eight well-studied strains +of the P. fluorescens complex that are supported by years +of studies, numerous refereed publications, and high-quality +genome sequences. By profiling transcriptomes of these strains +during growth in root exudates of B. distachyon, we revealed the +diversity of cellular pathways and physiological responses that +underlie the establishment of mutualistic interactions between +beneficial rhizobacteria and the host plant. Our results also +confirmed that root exudates contain carbohydrates, amino acids, +organic acids, and phenolics that serve as carbon and energy +sources for rhizobacteria. The root exudates also contained +osmoprotectants that may help microorganisms to persist in the +rhizosphere of drought-stressed plants. The diversity of microbial +genes perturbed by root exudates reflects the importance of +Frontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +the variable genome in adaptation of individual strains of +Pseudomonas to the rhizosphere lifestyle. +MATERIALS AND METHODS +Bacterial Strains Used in the Study +The eight Pseudomonas strains used for this study are +P. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens +SBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko +et al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller, +1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996), +P. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens +Pf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and +Stipanovic, 1980). The selected organisms have been studied +extensively for their role in biological control and plant +growth promotion (Supplementary Table 2). The strains were +maintained in the laboratory as frozen stocks (􀀀80 C) and +routinely cultured in King’s medium B (King et al., 1954) or +21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g +of Na2HPO4 2H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and +20 ml of a microelement solution (Smibert and Kreig, 1994; +Halverson and Firestone, 2000). +Propagation of Plants and Collection of +Root Exudates +B. distachyon Bd21 was established from seed material obtained +from the USDA-ARS Plant Germplasm Introduction and Testing +Research Unit (Pullman, WA, United States). Brachypodium +seeds were imbibed for 3 days at 4 C and sown in 7   7 cm +pots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture, +Agawam, MA, United States). Plants were grown in an IR- +89X (Percival Scientific, Perry, IA, United States) controlled +environment chamber retrofitted with 6500K and 3000K T5 54W +grow lights (Spectralux) under a 20-h light, 24 C/4-h dark, 18 C +cycle. Plants were watered and fertilized with Jack’s professional +water-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA, +United States). After 12 weeks and plant senescence, seeds +were collected, processed, and stored under desiccant and dark +conditions at room temperature. +To collect root exudates, seeds of B. distachyon Bd21 were +surface-sterilized, pregerminated, and placed in sterile 1 L widemouth +glass jars containing 113 g of 6-mm glass beads and +25 ml distilled water. Jars were covered with vented caps +and plants were grown hydroponically in an environmental +controlled growth chamber under conditions described above. +After 6 days, root exudates were extracted from individual +jars and their sterility was confirmed by spotting on nutrient +agar. Multiple batches of root exudates were collected, filtered +(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and +stored at –80 C. +Metabolomic Profiling of Root Exudates +Exudates were analyzed for primary metabolites at the Murdock +Metabolomics Laboratory at Washington State University +(Pullman, WA, United States). Freeze-dried residues were +suspended in 500 ml 50% aqueous acetonitrile and clarified by +centrifugation for 20 min at 21,000   g and 4 C. The liquid +chromatography mass spectrometry analysis was conducted +with a Synapt G2-S quadrupole-ion mobility spectrometrytime +of flight mass spectrometer system equipped with an +acquity ultra-performance liquid chromatograph (UPLC) and +an acquity photodiode array detector (all from Waters, Milford, +MA, United States). The exudate metabolites were separated +on a SeQuant ZIC-pHILIC HPLC column (2.1   100 mm, +3 mm) (Milllipore Sigma, Burlington, MA, United States) using +acetonitrile with 0.1% formic acid as solvent B and water +with 0.1% formic acid as solvent A at a flow rate of 400 ml +min􀀀1 and the following linear gradient extending over 14 min: +0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10% +B; 10 min, 80% B; and 14 min, 80% B. Mass spectra were +collected in positive ion mode over a range of m/z 50–1,200 +with a scan time of 0.2 s. The Q-TOF-MS source was at +3.0 kV and 120 C; the sampling cone at 40 V, desolvation +temperature was 250 C; cone gas and desolvation gas flow were +at 0 and 850 L h􀀀1, respectively. Leucine enkephalin was used +for post-acquisition mass correction. Target compounds were +visualized using selected ion chromatograms at 0.05 Da window +width. The compound identification was based on comparison +of chromatographic behavior and accurate masses to those of +authentic standards. +For gas chromatography, derivatization was carried out using +a modification of the procedure of Lee and Fiehn (2008). +The freeze-dried residues were suspended in 950 ml aqueous +methanol (84%, v/v) and clarified by centrifugation for 15 min +at 21,000   g at 4 C. The supernatants were spiked with 1 +mg of the internal standard salicylic acid-d6 (C/D/N Isotopes, +Quebec, Canada) and dried in vacuo. The dry residues were +suspended in 10 ml of O-methoxylamine hydrochloride (30 mg +ml􀀀1 in anhydrous pyridine, both from Millipore Sigma) and +incubated while mixing (1,000 RPM) for 90 min at 30 C. +Subsequently, samples were derivatized with 90 ml of MSTFA +with 1% TMCS (Thermo Fisher Scientific, Waltham, MA, +United States) for 30 min at 37 C. Gas chromatography-mass +spectroscopy analysis was performed using a Pegasus 4D timeof- +flight mass spectrometer (LECO, Saint Joseph MI) equipped +with a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A +oven (Agilent Technologies, Santa Clara, CA, United States). +The derivatization products were separated on a 30-m, 0.25 mm +i.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA, +United States) with an IntegraGuard precolumn using ultrapure +He at a constant flow of 0.9 ml min􀀀1 as carrier gas. The linear +thermal gradient started with a 1-min hold at 70 C, followed +by a ramp to 300 C at 10 C min􀀀1. The final temperature was +held for 5 min prior to returning to initial conditions. Mass +spectra were collected at 17 spectra s􀀀1. Peak identification was +conducted using the Fiehn primary metabolite library (Kind +et al., 2009) and an identity score cutoff of 700. Additionally, +authentic standards for a number of primary metabolites were +analyzed under identical conditions and the data used to compare +the chromatographic behavior. Peak alignment and spectrum +comparisons were carried out using the Statistical Compare +feature of ChromaTOF software (LECO). +Frontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +Isolation of RNA From Bacteria Cultured +in Root Exudates and RNA-Seq +The strains were pregrown overnight at 25 C on 21C-glucose +agar and then subcultured into 96-well microplates containing +liquid 21C-glucose medium amended with Brachypodium +exudates. The liquid medium was prepared by dissolving the +lyophilized root exudate material in an appropriate volume of +21C-glucose medium to concentrate root exudates 20-fold. The +growth medium was sterilized by passing it through a 0.22- +mm membrane filter. The control cultures were grown under +identical conditions in the absence of exudates. All treatments +were inoculated at OD600 of 0.1 and incubated for 20 to +22 h until cultures entered late-exponential growth phase at +25 C in an atmosphere of 15% oxygen [created by a ProOx +P110 oxygen controller (BioSpherix, Parish, NY, United States) +with a hypoxia C-chamber]. The cells were stabilized by the +addition RNAprotect reagent (QIAGEN, Germantown, MD, +United States) and total RNA was purified using a RNeasy Protect +Bacteria Mini Kit (QIAGEN) from three biological replicates of +each strain cultured under control conditions and in exudates. +The quality assessment of the extracted RNA samples was +performed with a NanoDrop OneC Spectrophotometer (Thermo +Fisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies) +and revealed A260/A280 and A260/A230 values of > 2.0 and a mean +RNA integrity numbers (RIN) value of > 9.2. +Three biological replicates of RNA samples were shipped on +dry ice to the DOE Joint Genome Institute (Walnut Creek, CA, +United States), where rRNA was depleted and stranded RNASeq +libraries were prepared, quantified by qPCR and sequenced +using a HiSeq 2500 instrument (Illumina). The fastq file reads +were filtered and processed with BBDuk1 to remove reads that +contained 1 or more “N” bases, had an average quality score +across the read less than 10 or had a minimum length < 51 bp +or 33% of the full read length. Reads mapped with BBMap (see +text footnote 2) to masked human, cat, dog, and mouse references +at 93% identity were removed. Another category of removed +sequences matched RNA spike-in, PhiX, common microbial +contaminants, and ribosomal RNAs. The processed reads from +each library were aligned to the reference genome using BBMap +with only unique mappings allowed (BAMs/directory). If a +read mapped to more than one location it was ignored. +featureCounts (Liao et al., 2014) was used to generate raw +gene counts, which were normalized to adjust for the length of +each gene and total number of reads mapped for each library. +The normalization formula used: n = [r/(l/1,000)]/(t/1,000,000), +where n = normalized read count for gene (G) for library (L); +r = raw read count for gene G for library L; l = gene G length; and +t = total reads mapped for library L. Raw gene counts were used to +evaluate the level of correlation between biological samples using +Pearson’s correlation. +Bioinformatic Analysis +Count tables generated by the JGI RNA-Seq pipeline were input +into DESeq2 (Love et al., 2014) to normalize and determine +1https://sourceforge.net/projects/bbmap/ +differential expression. Statistical significance was established +through DESeq2 by using three biological replicates for control +and root exudate conditions. Scatterplots were generated from +the DESeq2 data table outputs using ggplot2. Genes differentially +expressed between control and root exudate samples (log2 foldchanges +–2   to   2, adjusted p value   0.05) were used in +downstream analysis. The core genome and pangenome for the +Pseudomonas strains used in this study were computed using the +OthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic +Pangenome Accumulation v1.4.0 apps implemented in the U.S. +Department of Energy Systems Biology Knowledgebase (KBase) +(Arkin et al., 2018). Additional comparisons were conducted +with the PGAweb pangenome analysis pipeline (Chen et al., +2018). Differentially expressed genes were assigned to core, +non-core, and singleton parts of each strain’s proteome by +BLASTp with an E value cutoff of e-06, identity of 40% +and coverage of 60%. Functional annotation of differentially +expressed genes was carried out with the Blast2GO (Conesa +and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., +2018). Additional manual curation was performed using tools +implemented in the Integrated Microbial Genomes (IMG) +database (Markowitz et al., 2012), Pseudomonas Genome +Database (Winsor et al., 2009), Kyoto Encyclopedia of Genes and +Genomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3 +(Biomatters, Auckland, New Zealand). Metabolic functions +encoded by the differentially expressed genes were mapped +using iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses +were carried out by building multiple sequence alignments +with MAFFT v7.222 (Katoh and Standley, 2013) and inferring +neighbor-joining (NJ) phylogenies with Geneious Tree Builder. +The resultant phylogenetic trees were visualized with iTOL +(Letunic and Bork, 2016). Reproducibility of clades within the +inferred NJ trees was assessed by bootstrap resampling with +1,000 replicates. +Characterization of Carbon Source +Utilization With Biolog Phenotype +Microarrays +The utilization of carbon sources was analyzed using Phenotype +MicroArrays (Biolog, Hayward, CA, United States) as follows. +The bacteria were cultured overnight on Luria-Bertani agar +at 25 C, after which cells were harvested and suspended in +inoculating fluid (IF-0). The transmittance of the suspension +was adjusted to 42% using a Biolog turbidimeter. The cell +suspension was mixed with IF􀀀0 containing Dye Mix A +(Biolog) to achieve a final transmittance of 85%. One hundred +microliter aliquots of the adjusted cell suspension were +inoculated into PM01 and PM02A plates, which were then +incubated in an OmniLog Phenotype MicroArray System +(Biolog) at 25 C for 48 h. The formation of formazan was +recorded at 15 min intervals, and data were analyzed using +OmniLog Parametric Analysis software v1.20.02 (Biolog). +Relative growth of the studied strains was normalized to +growth on D-glucose and visualized using Heatmapper +(Babicki et al., 2016). +Frontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +Data Availability +Sequences generated in this project were deposited +under NCBI BioProject accession numbers PRJNA439743 +through PRJNA439790. +RESULTS +Metabolomic Profiling of Root Exudates +of B. distachyon +Metabolomics analysis of lyophilized root exudates revealed +the presence of numerous plant metabolites, 86 of which were +identified by matching their spectra to the LECO/Fiehn +Metabolomics library (Supplementary Table 3). These +metabolites included (i) carbohydrates and their derivatives +(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose, +and others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol, +galactinol, 2-deoxyerythritol, ribitol, threitol and +cellobitol); (iii) amino acids and derivatives (glutamine, +tyrosine, glutamic acid, asparagine, aspartic acid, valine, +phenylalanine, isoleucine, glycine, serine, proline, leucine, +tryptophan, cysteine, methionine, citrulline, and others); (iv) +organic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric, +fumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic, +malic, methylmalonic, nicotinic, quinic, succinic, threonic); +and (v) assorted metabolites including heterocyclic compounds, +phenolics, and biogenic amines, etc (3-hydroxypyridine, +maleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate, +5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid, +and urea). Results of the analysis also revealed that root exudates +of B. distachyon contain hydroxyectoine and the quaternary +amine (QA) glycine betaine (Supplementary Figure 1). +Phylogenetic and Pangenome Analyses +of Pseudomonas Strains Used in the +Study +We used a set of phylogenetic markers suggested by Mulet et al. +(2010) to investigate the relatedness of the eight strains used in +this study to distinct lineages recognized within the P. fluorescens +species complex. The multilocus sequence analysis based on +the concatenated sequences of the housekeeping genes rrs (16S +rRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with +strains 2-79 and SBW25) as a member of the P. fluorescens +subgroup (Figure 1). The rest of the strains clustered closely with +four additional subgroups of the P. fluorescens complex, namely +P. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1), +P. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of +the eight rhizosphere Pseudomonas strains varied in size by 1.43 +megabase (ranging from 5.65 to 7.07 Mb) and contained between +5,166 and 6,363 protein-coding genes (Figure 2A). The shared +gene content was characterized with OrthoMCL, which uses allagainst- +all BLASTp followed by the Markov Cluster algorithm to +identify protein groups shared between the compared genomes, +as well as groups representing species-specific gene expansion +families (Li et al., 2003). The pangenome analysis revealed a +core comprised of approximately 3,179 orthologs that were +shared among all strains and represented 50.0% to 61.5% +of each predicted proteome (Figures 2A,B). The non-core +pangenome contained genes shared by two or more (but not +all) strains and contained between 1,482 and 2,080 orthologs, +which corresponded to 28.7–36.3% of individual proteomes. The +rest of the predicted protein-coding genes were strain-specific +singletons that comprised 7.5% to 15.1% of the strain’s predicted +proteomes. In respect to divergence from the core genome, strain +Pf-5 was found to possess the highest proportion of unique genes +(n = 949) followed by 2-79 (n = 887). The entire pangenome of +the Pseudomonas strains encompassed over 12,000 homolog and +singleton gene families. +Further homolog family-based comparisons identified Q8r1- +96 and R1-43-08 as the most distantly related strains, with +3349 shared homologs (Supplementary Table 4A). Q8r1-96 and +Q2-87, which shared 4,489 homologs, were the most closely +related strains. The partitioning of homolog gene families into +the core, non-core, and singleton parts of the pangenome agreed +with phylogenetic relationships of the strains deduced from the +analysis of a selected subset of COGs (Clusters of Orthologous +Groups) (Figure 2C and Supplementary Tables 4B,C). The +COG-based phylogeny supported the multilocus sequence +analysis and revealed that the eight Pseudomonas strains form +three distinct clusters, the first of which contained 2-79, R1-30- +84, and SBW25. The second cluster included Q8r1-96 and Q2-87, +whereas the third encompassed strains 30-84, Pf-5, and Pf0-1. +Correlating the Composition of Root +Exudates With Metabolic Profiles of +Pseudomonas Strains +We used the Phenotype MicroArray PM1 and PM2 plates to +profile the eight Pseudomonas strains for the utilization of +190 different carbon sources. Results of the analysis identified +90 compounds that supported growth and clustered by their +intensities of utilization into three distinct groups (Figure 3). +Group I was comprised of 30 highly metabolized carbon +sources, which included several amino acids and intermediates of +glycolysis, pyruvate metabolism, and citrate cycle. Approximately +half of these compounds were catabolized by all eight strains, +and included several organic acids (fumaric, citric, gluconic, +malic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro, +Ala, and g-aminobutyric acid), carbohydrates (glucose, mannose, +and mannitol), and the purine nucleoside inosine. Group II +was composed of 44 chemically diverse carbon sources that +were variably utilized by the strains. These compounds were +carbohydrates, organic acids, amino acids, phenolics, and polyols, +and included known compatible solutes and intermediates of +metabolism of pentoses, galactose, starch, and sucrose. Group +III encompassed the rest of the Phenotype MicroArray test +panel and contained compounds that were not catabolized +by the tested strains. Among several notable exceptions were +a-hydroxyglutamic acid- g-lactone, putrescine, and itaconic, +citramalic, and succinamic acids, which supported the growth +of strains 2-79, 30-84, Pf-5, and SBW25. We further matched +the carbon metabolic profiles of the Pseudomonas strains against +the list of plant-derived metabolites from the root exudates of +Frontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the +P. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA +gyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor +method. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles). +B. distachyon Bd21. Interestingly, many carbon sources from +the Phenotype MicroArray panel were also present in the root +exudates of B. distachyon Bd21, and some of these compounds +(glucose, mannose, galactose, fructose, g-aminobutyric acid, +aspartic acid, citric acid, malic acid, fumaric acid, quinic acid, +alanine, glutamine, and glutamic acid) were catabolized by all +strains used in this study, while others (e.g., xylose, trehalose, +m-inositol) were actively utilized only by certain organisms +(Figure 3). The comparison of catabolic profiles across the +eight studied Pseudomonas strains revealed the presence of three +distinct clusters. The first cluster contained strains Q8r1-96 and +Q2-87, which consumed very similar sets of carbon sources, +as well as strain Pf0-1. The second cluster was composed of +2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster +was represented by a single strain, Pf-5. The overall similarity +of the catabolic profiles partially agreed with the separation +of the strains into different subgroups of the P. fluorescens +complex (see above). +Analysis of the RNA-seq Results +In order to understand the cellular responses of rhizosphere +Pseudomonas to plant exometabolites, we analyzed the +transcriptome changes in cultures grown in the presence of +root exudates. Under field conditions, rhizobacteria colonize +plant roots in the form of surface-attached microaerobic biofilms +(Hojberg et al., 1999). To mimic these conditions, the eight +Pseudomonas strains were grown statically at 72% air saturation +in 21C-glucose medium amended with root exudates and then +processed to extract total RNA (Supplementary Figure 2). +A total of 995 million raw sequencing reads were generated from +the RNA samples by using the Illumina HiSeq-2500 platform, +averaging 20.7 million reads per sample. The removal of lowquality +and rRNA sequences resulted in a total of 793 million +filtered reads that were mapped onto the eight Pseudomonas +genomes with a mean of 7.48 million mapped fragments per +genome. The differentially abundant transcripts were identified +by setting a p value of 0.05 (adjusted for multiple testing) and +Frontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all +eight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific +singletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of +the pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the +25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any +bias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939 +non-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied +strains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining +ortholog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node +represents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The +analysis was conducted in KBase (Arkin et al., 2018). +the log2 fold-change (FC) threshold    2.0 (Figure 4 and +Supplementary Tables 5–12). When compared with the control +conditions, an average of 204 genes per strain were differentially +expressed in the presence of root exudates, with the highest +(n = 425) and lowest (n = 112) numbers observed, respectively, +in SBW25 and Q2-87 (Figure 4). Overall, more genes were +induced than repressed in response to exudates, but the actual +numbers in each category varied substantially depending on the +identity of the Pseudomonas strain. In most strains, the bulk of +the differentially expressed genes was almost equally distributed +between the core (mean, 48.2%) and non-core (mean, 45.8%) +parts of the genome, whereas the strain-specific singleton genes +constituted on average only 5.9% (Figure 4B). One notable +exception was observed in Q8r1-96, where all differentially +expressed genes belonged to the core (73.8%) and non-core +(26.2%) parts of the genome. Another notable pattern was +observed in R1-43-08, where the majority of genes affected by +the presence of root exudate fell into the non-core category +(56.3%). The highest proportion of differentially expressed +singletons (11.3 and 10.4%, respectively) was identified in +strains SBW25 and Pf-5. +We further explored how the identified differentially +expressed genes were distributed across genomes of the eight +studied rhizosphere strains. The pairwise BLASTp comparisons +identified 2-79 and SBW25 as two strains that shared the highest +number of genes (n = 101) induced or repressed in response +to root exudates (Table 1). The second pair of strains with +a significant number of similar differentially expressed genes +(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1 +and 30-84, which shared 56 differentially expressed genes. These +patterns of shared genes were also observed when the results of +the pairwise BLASTp comparisons were converted into a binary +gene presence/absence matrix, which was then subjected to +cluster analysis using a UPGMA algorithm based on Sorensen’s +dissimilarity index or examined by non-metric multidimensional +scaling (NMDS) (Figure 5). +The differentially expressed Pseudomonas genes were +subjected to Blast2Go analysis and Gene Ontology (GO) +annotation (Figure 6). Metabolic process, catalytic activity, +and membrane were the most common annotation terms +across the three primary GO term categories (i.e., biological +process, molecular function, and cellular component). A total +of 1,694 GO terms was assigned to 805 upregulated genes, with +the majority of the GO terms related to molecular function +(682, 40.3%), followed by biological process (669, 39.5%), and +cellular component (343, 20.2%). In the 539 downregulated gene +category, 1,101 GO terms were assigned to biological process +(420, 38.1%), molecular function (417, 37.9%), and cellular +component (264, 24.0%). Within biological process, metabolic +process, cellular process, localization, response to stimulus, and +regulation were over-represented. Within molecular function, +the largest proportion was assigned to catalytic activity, binding, +and transporter activity categories. Within cellular component, +the majority were assigned to membrane, membrane part, cell, +and cell part categories. Across the eight strains, 37–42% of +differentially expressed genes had no Gene Ontology IDs and +encoded various conserved hypothetical proteins. +Functional Classification of Shared +Differentially Expressed Genes +The interrogation of RNA-seq data revealed multiple cellular +pathways that were differentially regulated in bacterial cultures +incubated with root exudates (Supplementary Figures 3, 4). +Frontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere +Pseudomonas strains used in the study. The hierarchical clustering analysis +was carried out using the average linkage method with Euclidean distances. +Carbon sources identified by red arrowheads were also detected in the sterile +root exudates of B. distachyon Bd21. +Although none of these differentially regulated pathways were +shared by all eight strains, the cross-strain comparisons revealed +several types of common and specific transcriptomic responses +that were elicited by the presence of plant exometabolites +(Table 2). The visual representation of core gene expression +patterns is provided in Supplementary Figure 5, which +shows heatmaps of expression profiles and p-adj values for +core genes shared by the studied strains. The figure is +accompanied by Supplementary Table 13 that lists predicted +functions of genes constituting the four distinct clusters +observed after hierarchical clustering of gene expression +values. The first category of shared differentially expressed +pathways functioned in the uptake and catabolism of selected +carbohydrates, quaternary ammonium compounds (QAs), and +phenolics. All strains except for R1-43-08, responded to root +exudates by inducing the fructose-specific phosphoenolpyruvate +(PEP)-dependent phosphotransferase system (PTSFru). The +components of this system are encoded by a conserved +operon and include the cytoplasmic polyprotein EI/HPr/EIIAFru +(FruB), the 1-phosphofructokinase FruK, and the fructosespecific +permease EIIBC (FruA) (Chavarria et al., 2016). The +PTSFru system functions by acquiring high-energy phosphates +from PEP and sequentially passing them, via the EI/HPr/EIIAFru +domains of FruB, to the EIIB component of FruA. The +phosphates are ultimately transferred by the EIIC transporter +to fructose yielding fructose 1-phosphate, which is channeled +into the central metabolic pathways through the action of the +phosphofructokinase FruK. +In all strains except for Q8r1-96 and Pf-5, the exposure to root +exudates resulted in the induction of two genes adjacent to the +fru cluster that encoded a Major Facilitator Superfamily (MFS) +transporter and an L-arabinonate dehydratase (Table 2). These +genes are predicted to participate in the uptake and catabolism +of L-arabinose, where L-arabinonate dehydratase plays an +important role by converting L-arabinonate to 2-dehydro-3- +deoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1- +43-08, and Q2-87, we also observed the induction of genes +encoding components of the AraFGH complex, an ATP-Binding +Cassette (ABC) superfamily transporter involved in the import of +arabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all +strains except SBW25 and R1-43-08 responded to the presence +of exudates by upregulating a conserved gene encoding an +aldose epimerase superfamily protein. Such enzymes equilibrate +alpha- and beta-anomers of aldoses and ensure that stereospecific +enzymes involved in the metabolism of free sugars do not act as +metabolic bottlenecks (Abayakoon et al., 2018). Although some +aldose epimerases have been linked to specific pathways, the +Pseudomonas gene identified in this study could not be assigned +to a particular metabolic process based on sequence analysis and +genomic location. +Several Pseudomonas strains responded to the presence +of root exudates by upregulating genes involved in the +uptake and catabolism of myo-inositol and possibly other +stereoisomers of inositol (Table 2). The upregulated catabolic +genes encode the dehydrogenase IolG, which oxidizes myoinositol +to its corresponding ketone, as well as IolE, IolD, IolB, +and IolC that collectively convert the 2-keto-myo-inositol to +acetyl-CoA and the glycolysis intermediate dihydroxyacetone +phosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1- +43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions +Frontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core +and non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established +threshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual +Pseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates. +also involved components of the putative inositol-specific +ABC transporter. The cross-genome comparisons revealed +that in all studied strains except for Pf0-1, components +of the myo-inositol utilization pathway were encoded +within a well-conserved gene cluster which, in addition to +catabolic and transport functions, also encodes a dedicated +transcriptional repressor. +All studied strains of Pseudomonas carry multiple genes +involved in scavenging the quaternary ammonium compounds +choline, glycine betaine (GB), carnitine, choline-O-sulfate, and +Frontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +TABLE 1 | The number of differentially expressed genes shared among the eight +studied strains of rhizosphere Pseudomonas. +Strain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5 +2–79 260 +SBW25 101 425 +R1-43-08 30 25 151 +Q8r1-96 32 39 21 145 +Q2-87 27 28 25 31 112 +30-84 27 23 24 32 28 136 +Pf0-1 38 50 29 29 50 56 205 +Pf-5 36 41 52 86 29 55 40 230 +The pairwise comparisons were conducted by BLASTp with the following cutoff +parameters: E-value < 1e-06, minimum percent identity > 40%, and minimum +percent coverage > 65%. The black diagonal cells show the number of differentially +expressed genes per strain. In other words, these are self comparison values. +sarcosine from the environment. Many of these genes were +differentially expressed, including those encoding parts of the +ABC transporter CbcXWV, which is predicted to function in +the uptake of choline under water-replete conditions (Table 2). +Among enzymes induced in the presence of root exudates were +the choline dehydrogenase BetA, which converts choline to +glycine betaine and a network of enzymes (i.e., the Rieske family +oxygenase GbcAB, the dimethyglycine demethylase DgcAB, and +the sarcosine oxidase SoxBDAG) that sequentially convert GB +to glycine. In 2-79 and SBW25, this group of differentially +regulated genes also included an AraC-family transcriptional +activator GbdR, which perceives intercellular levels of GB and +induces genes involved in the transport and catabolism of +glycine betaine and detoxification of the catabolic byproducts +(Hampel et al., 2014). +The last category of activated catabolic pathways included the +catechol branch of the b-ketoadipate pathway for the degradation +of aromatic compounds. In strains 30-84, Pf0-1, and Pf-5, +growth on root exudates resulted in upregulation of catechol- +1,2-dioxygenase, muconate cycloisomerase, and muconolactone +isomerase, which collectively cleave the catechol ring and +convert it to b-ketoadipate enol-lactone (Harwood and Parales, +1996). Finally, analysis of the P. synxantha 2-79 transcriptome +identified an induction of benABC genes encoding subunits +of benzoate 1,2-dioxygenase, an oxidoreductase that generates +catechol from benzoate. +In addition to various catabolic pathways, the exposure to root +exudates also induced several genes involved in the homeostasis +of copper (Table 2). Four of these genes form a conserved cluster +in genomes of the strains and encode the periplasmic coppersensing +two-component system CinRS, the plastocyanin/azurinlike +protein CinA, and the NADPH-dependent pre-Q0 reductase +CinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed +upregulation of a conserved operon encoding the multicopper +oxidase CopA, the periplasmic copper-binding protein CopC, the +inner membrane protein CopD, and outer membrane protein +CopB. In several Gram-negative bacteria, these Cop proteins are +thought to have dual functions and participate both in the uptake +of essential copper as well as in the sequestration of excess copper +in the periplasm and outer membrane. +The analysis of shared downregulated pathways revealed that +most of the strains respond to the presence of root exudates by +repressing genes involved in the uptake and catabolism of sulfur +compounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96, +Q2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB +operon responsible for the utilization of alkanesulfonates +as sulfur sources. The ssu operon is highly conserved in +fluorescent pseudomonads and encodes the FMNH2-dependent +monooxygenase SsuD and the NAD (P)H-dependent FMN +reductase SsuE, which together catalyze the desulfonation of +alkanesulfonates. Also, the ssu locus contains genes for the +molybdopterin-binding protein SsuF and the alkanesulfonatespecific +ABC-type transporter consisting of the sulfonate +substrate-binding protein SsuA, sulfonate permease protein +SsuC, and sulfonate transport ATP-binding protein SsuB. Finally, +in R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates +coincided with repression of the tauABCD operon, which +allows these strains to utilize taurine (2-aminoethanesulfonate) +as a sulfur source. The repressed tau genes encoded the +2-oxoglutarate-dependent taurine dioxygenase TauD and +substrate-binding, ATP-binding, and permease components of +the taurine-specific ABC transporter TauABC. +Other Differentially Expressed Pathways +In addition to their effect on several shared cellular pathways, +growth on root exudates resulted in the induction or repression of +numerous strain-specific genes. In closely related P. synxantha 2- +79 and P. fluorescens SBW25, we observed differential expression +of genes involved in energy metabolism, transport of amino +acids, and surface attachment (Supplementary Tables 5, 6). +Other notable differentially expressed pathways included 2– +79 gene clusters that encode enzymes for the catabolism of +trehalose, a prophage, and toxin/antitoxin system, as well as the +SBW25 operon predicted to control the synthesis of the capsular +exopolysaccharide colonic acid. The response of Pseudomonas sp. +R1-43-08 to root exudates also involved differential expression +of different energy metabolism pathways. In addition, we +observed the upregulation of genes involved in the uptake and +catabolism of xylose (also upregulated in 2–79) and repression +of enzymes for the biosynthesis of phenazine-1-carboxylic acid +and assimilation of inorganic sulfur and L-cysteine biosynthesis +(Supplementary Table 7). +The analysis of the Q8r1-96 transcriptome revealed +perturbation of different metabolic pathways including genes +encoding components of cytochrome C oxidase, transport +and catabolism of sorbitol/mannitol, metabolism of butanoic +acid, and biosynthesis of exopolysaccharides alginate and +poly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In +P. fluorescens Q2-87, we identified differential expression of +genes involved in metabolism of galactose, tryptophan, tyrosine, +glycine, serine, and threonine (Supplementary Table 9), +while in P. chlororaphis 30-84, growth on exudates activated +the biosynthesis of molybdopterin cofactor, catabolism of +galactonate and acetoin, and uptake and catabolism of putrescine +(Supplementary Table 10). The response of P. protegens Pf-5 to +root exudates involved upregulation of acetoin dehydrogenase, +which converts acetoin to acetaldehyde and acetyl-CoA, as +Frontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root +exudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis. +FIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of +B. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa +and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO +term. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO +terms. +well as pathways for the utilization of glycolate and putrescine +(Supplementary Table 11). Also induced were genes for +the production of pyrrolnitrin and PhlG hydrolase, which +modulate the metabolic loads attributed to the synthesis of +2,4-diacetylphloroglucinol. The differentially expressed genes of +P. fluorescens Pf0-1 included, among others, operons encoding +cytochrome C oxidase and enzymes for catabolism of malonic +acid (Supplementary Table 12). Yet another interesting +finding involved the induction of assorted genes acting in the +homeostasis of iron and defense against reactive oxygen species +(ROS). We observed activation of iron dicitrate transporters +(SBW25 and 30-84), genes for the biosynthesis of siderophores +ornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading +enzymes (2–79, 30–84), TonB siderophore receptors, and +components of the energy-transducing inner membrane +complex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially +expressed ROS defense pathways were represented by different +catalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and +Pf-5 and organic hydroperoxide resistance proteins in strains +SBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and +Pf0-1, the addition of exudates resulted in the upregulation of +peroxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic +and aromatic hydroperoxides. +DISCUSSION +Our analysis of B. distachyon root exudates revealed a complex +mix of primary and secondary metabolites, thus supporting +the view of the plant rhizosphere as a carbon-rich niche for +Frontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +TABLE 2 | The distribution and predicted functions of selected differentially expressed genesa. +Predicted function Strainb +2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 +Uptake and catabolism of fructose +D-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8) +1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7) +D-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2) +Uptake and catabolism of arabinose +MFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9) +L-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5) +Interconversion of alpha- and beta-anomers of aldoses +Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) +Uptake and catabolism of quaternary ammonium compounds +Choline dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1) +Transcriptional regulator GbdR 2756597125 (􀀀3.7) 649639087 (􀀀3.9) +Membrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3) +Hypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3) +Dimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2) +Dimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0) +Betaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7) +Betaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2) +Ser hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1) +Sarcosine oxidase, g subunit, SoxG 2597878304 (2.9) +Sarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4) +Sarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2) +Sarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3) +Betaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5) +ABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5) +Uptake and catabolism of myo-inositol +5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2) +2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1) +5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3) +2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4) +3D-(3,5/4)-trihydroxycyclohexane-1,2-dione +acylhydrolase, IolD +2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3) +Myo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9) +Inositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8) +Inositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5) +Inositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6) +(Continued) +Frontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +TABLE 2 | Continued +Predicted function Strainb +2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 +Uptake and catabolism of fructose +Catabolism of phenolics +Muconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8) +Muconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4) +Catechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5) +AraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2) +Benzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8) +Benzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8) +Benzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3) +Copper homeostasis +pre-Q0 reductase/7-cyano-7-deazaguanine +reductase CinQ +2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5) +Cupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4) +Heavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5) +Heavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0) +Copper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4) +Copper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0) +Copper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4) +Copper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2) +Conserved hypothetical protein +Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) +Uptake and catabolism of sulfonates +FMN-dependent monooxygenase SsuE 2756592254 (􀀀1.9) 637745334 (􀀀3.5) +Sulfonate substrate-binding protein SsuA 649639261 (􀀀2.0) 2756592253 (􀀀2.2) 2597878518 (􀀀2.6) 2597849636 (􀀀3.9) 637745333 (􀀀7.0) 637323272 (􀀀2.7) +FMN-dependent monooxygenase SsuD 649639260 (􀀀1.9) 2756592252 (􀀀2.3) 2597849637 (􀀀2.8) 637745332 (􀀀5.8) 637323271 (􀀀2.3) +Sulfonate permease protein SsuC 649639259 (􀀀2.4) 2756592251 (􀀀2.1) 2597878516 (􀀀1.9) 2597849638 (􀀀3.2) 637745331 (􀀀5.8) 637323270 (􀀀2.5) +Sulfonate transport ATP-binding protein SsuB 649639258 (􀀀2.5) 2756592250 (􀀀2.3) 2597849639 (􀀀2.5) 637745330 (􀀀5.6) 637323269 (􀀀2.7) +Molybdopterin binding protein SsuF 649639257 (􀀀3.1) 2756592249 (􀀀2.7) 2597878514 (􀀀2.9) 2597849641 (􀀀3.3) 637745339 (􀀀6.2) 637323268 (􀀀2.7) +Uptake and catabolism of taurine +Taurine substrate-binding protein TauA 2597854917 (􀀀2.0) 637740095 (􀀀4.1) +Taurine transport ATP-binding protein TauB 2756592398 (􀀀1.9) 637740094 (􀀀3.7) 637317614 (􀀀1.9) +Taurine permease protein TauC 2756592398 (􀀀1.9) 637740093 (􀀀3.9) 637317613 (􀀀2.0) +Taurine dioxygenase TauD 2597854920 (􀀀1.8) 637740092 (􀀀3.6) +aThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%. +bValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets). +Frontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +soil microorganisms. Our results were in agreement with a +recent report of 27 different sugars, amino acids, and organic +acids in Brachypodium exudates (Kawasaki et al., 2016). We +confirmed the presence of exometabolites identified in that +study, along with dozens of additional analytes that were +identified by matching their mass-spectra and retention indices +to the LECO/Fiehn Metabolomics library (Supplementary +Table 3). The complementation of the metabolomic analysis +with profiling of the bacteria by Biolog Phenotype MicroArrays +revealed that a substantial proportion of the characterized +exudate constituents were catabolized by a collection of eight +Pseudomonas strains from across the P. fluorescens group that is +known to form associations with plant roots. The amendment +of Pseudomonas cultures with root exudates caused changes in +the expression of multiple genes encoding catabolic and anabolic +enzymes, predicted transporters, transcriptional regulators, stress +response, and conserved hypothetical proteins. In most strains, +these differentially expressed genes were almost equally split +between the core and variable genome regions, mirroring +the substantial strain-to-strain variation in the genome size +and gene content within the P. fluorescens species complex +(Loper et al., 2012). +The analysis of transcriptome responses to root exudates +revealed several types of cellular pathways present in the strains +used in this study. The first category of such pathways was +involved in the catabolism of carbohydrates such as fructose, +arabinose, myo-inositol, xylose, trehalose, and galactose. Among +these catabolic traits, the ability to utilize fructose as a carbon +source is highly conserved among fluorescent pseudomonads. +In contrast, growth on arabinose, myo-inositol, xylose, and +trehalose is variably present and was traditionally used to +differentiate species and biovars within the P. fluorescens +group (Barrett et al., 1986). We speculate that such variably +distributed pathways contribute to the differential affinity of +pseudomonads toward host plants and/or to determine which +strains flourish in response to growing roots and changing +environments. Several independent studies have confirmed +the importance of carbohydrate catabolism pathways for the +biology of rhizosphere pseudomonads. For example, in vivo +expression technology (IVET) profiling of P. fluorescens SBW25 +identified xylose isomerase among genome regions essential +for the colonization of sugar beet seedlings (Liu et al., 2015), +whereas a genome-wide Tn-Seq screen of Pseudomonas simiae +identified genes for the catabolism of myo-inositol among traits +essential for the colonization of Arabidopsis thaliana roots +(Cole et al., 2017). +The response of rhizosphere Pseudomonas to Brachypodium +root exudates also involved pathways for the uptake and +metabolism of amino acids. We observed differential +expression of genes encoding the hydrophobic (HAAT) and +polar (PAAT) amino acid uptake transporters in strains +2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes +encoded enzymes for the catabolism of valine and glutamic +acid (2-79); metabolism of tryptophan, glycine, serine, and +threonine (Q2-87); and biosynthesis of methionine (Q8r1- +96). It is plausible that the abundance of amino acids in +root exudates is also linked to the repression of pathways +involved in the catabolism of sulfonates and taurine that +was observed in several strains (Table 2). Although the +preferred source of sulfur for P. fluorescens is unknown, in the +closely related P. aeruginosa, the sulfur starvation response +is triggered by the growth on any sulfur compound other +than sulfate, thiocyanate, and cysteine (Hummerjohann et al., +1998). This fact, together with the presence of cysteine and +cystine in the root exudates, suggest that root exudates of +Brachypodium may serve as an important source of sulfur for +rhizosphere Pseudomonas. These findings also agree well with +the reported scarcity of inorganic sulfate in the soil, and the +presence of sulfur mostly in the form of organic compounds, +including amino acids, proteins, sulfate esters, and sulfonates +(Autry and Fitzgerald, 1990). +Another interesting result of this study was the concerted +activation of copper and iron homeostasis pathways observed +in all of the Pseudomonas strains used in this work. In bacteria, +an excess of copper is toxic and triggers oxidative stress due to +the formation of free radicals, as well as disruption of protein +metalation and stability of iron-sulfur clusters (Bondarczuk +and Piotrowska-Seget, 2013). On the other hand, copper is an +essential trace element used as a cofactor in different enzymes. +Similarly, although elevated levels of iron cause redox stress, this +element is also found in active energy metabolism enzymes and is +crucial for bacterial growth (Andrews et al., 2003). The analysis +of metal homeostasis genes identified in this study suggests +that their induction was likely triggered by the deficiency of +copper and iron in bacterial cultures grown in the presence of +root exudates. We attribute this effect to the ability of some +components of root exudates to chelate soil metals. +Despite the abundance of iron in the soil, its bioavailability +is limited due to the low solubility of Fe (III) oxyhydrates +at neutral pH. The non-graminaceous plants circumvent this +problem by acidifying the rhizosphere and secreting flavins, +phenolics, and organic acids that chelate iron. The reduction +of these ferric chelates releases soluble ferrous iron taken up +by root cells (Kobayashi and Nishizawa, 2012). Graminaceous +plants, like Brachypodium, acquire iron by secreting into the +soil non-protein amino acids of the mugineic acid (MA) group, +which act as Fe (III)-chelating phytosiderophores. In addition to +iron, low-molecular-weight organic acids and phytosiderophores +bind other divalent and trivalent metals (including copper) and +contribute to heavy-metal tolerance in plants (Chen et al., 2017). +It is plausible that the presence of these plant exometabolites +is responsible for the deficit of iron and copper observed in +Pseudomonas cultures grown in the presence of root exudates. +These results further underscore the importance of diverse and +redundant metal homeostasis pathways found in genomes of the +P. fluorescens group for the ability of these organisms to colonize +and persist in the plant rhizosphere. +Recently, Klonowska et al. (2018) examined transcriptomic +responses of symbiotic nitrogen-fixing bacteria to root +exudates of the legume plant Mimosa pundica, which has +an unusual ability to support both alpha- (Rhizobium) and +beta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq, +the authors characterized genes involved in the perception of +root exudates in the nodulating bacteria Burkholderia phymatum +Frontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +STM815, Cupriavidus taiwanensis LMG19424, and Rhizobium +mesoamericanum STM3625. Interestingly, the analysis of +differentially expressed genes revealed induction of pathways +involved in the catabolism of fructose, xylose, myo-inositol, +and protocatechuate/catechol. Also upregulated were some +copper homeostasis, siderophore biosynthesis, and oxidative +stress genes. Finally, the analytical profiling of M. pundica +exudates revealed an overlap with Brachypodium in the types of +carbohydrates, amino acids, and organic acids present. These +findings suggest that differentially expressed genes shared by +multiple strains of the group P. fluorescens are not unique to +the Brachypodium-Pseudomonas system but represent a set +of conserved cellular pathways involved in the perception of +plant exometabolites by different clades of rhizosphere-dwelling +Proteobacteria. +Most strains included in this study were originally selected +based on the ability to colonize the rhizosphere and produce +secondary metabolites that alleviate the plant stress response +and/or inhibit soilborne pathogens. It has been suggested +that plant metabolites released into the rhizosphere affect +the biocontrol activity of plant-beneficial pseudomonads (de +Werra et al., 2011). We provide further support to this +hypothesis by demonstrating that in some strains, root exudates +modulate the expression of genes for the catabolism of +the plant growth-promoting metabolites acetoin and 2,3- +butanediol. The exposure to exudates also affected the expression +of genes for the synthesis of well-characterized antifungal +compounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4- +diacetylphloroglucinol. The modulatory effects were strainspecific, +suggesting significant differences in the regulatory +networks involved in the perception of plant signals and +regulation of the production of antibiotics and growthpromoting +metabolites. +The final significant finding of this study was the induction +of catabolism of quaternary amines (QAs) observed in multiple +strains of the P. fluorescens group during growth on root +exudates. This observation was supported by the detection of +glycine betaine in the root secretions of B. distachyon. The +presence of QAs in plant tissues and the capacity of these +metabolites to provide stress protection and nutrients to plant +pathogens and symbionts were reported before (Boncompagni +et al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our +study is among the first to highlight the potential importance of +these metabolites for rhizosphere interactions. Pseudomonads do +not synthesize QAs de novo but have evolved many pathways to +scavenge them from eukaryotic hosts, where these metabolites +are abundant due to the prominence of phosphatidylcholine +in cellular membranes. Strains of P. fluorescens carry genes +for the conversion of choline, carnitine, and glycine betaine to +glycine, as well as quaternary amine transporters of the BCCT +and ABC families that are also conserved in the opportunistic +human pathogen P. aeruginosa and the plant pathogen P. syringae +(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b). +In P. aeruginosa, choline catabolism genes are essential for +the ability of this pathogen to persist during lung infection +(Wargo, 2013a). Similarly, a P. syringae mutant deficient in +BetT, OpuC, and CbcXWV quaternary amine transporters had +reduced fitness during colonization of bean and soybean leaves +under greenhouse and field conditions (Chen et al., 2013). +Depending on water availability, P. aeruginosa and P. syringae +catabolize exogenously supplied QAs as carbon and nitrogen +sources or accumulate them as osmoprotectants (Chen et al., +2013; Wargo, 2013b). Our ongoing work in P. synxantha 2– +79 unraveled similar physiological responses and demonstrated +that QA transporters function differentially and redundantly +in the uptake of quaternary amines as nutrients (Pablo and +Mavrodi, unpublished). In contrast, under water stress, the QAs +choline, betaine, and carnitine are accumulated preferentially for +osmoprotection. Under drought stress, a 2–79 mutant devoid +of all known QA transporters was less competitive in the +colonization of the Brachypodium rhizosphere than its wild-type +parental strain. Interestingly, our metabolomic profiling of root +exudates also revealed proline, glutamine, and hydroxyectoine. +These metabolites act as compatible solutes in different groups +of microorganisms (Yancey et al., 1982; Empadinhas and da +Costa, 2008), suggesting an important role of root exudates +in the ability of Pseudomonas to persist in the rhizosphere of +drought-stressed plants. +DATA AVAILABILITY STATEMENT +The original contributions presented in the study are publicly +available. This data can be found here: NCBI BioProject accession +numbers PRJNA439743 through PRJNA439790. +AUTHOR CONTRIBUTIONS +DM, OM, and LT conceived the research project. OM and +JM collected root exudates. OM and DM cultured strains +and extracted total RNA. AB and DG performed metabolomic +analysis of root exudates. DM, JP, and AF analyzed RNA-seq +data. LE, KH, and IP conducted Biolog analyses. DM, AF, OM, +DW, and LT wrote the manuscript. All authors contributed to the +manuscript revision. +FUNDING +This study was funded by NSF grant IOS-1656872 and by an +award from the DOE Joint Genome Institute’s Community +Science Program. The authors also acknowledge support from +Australian Research Council Discovery grant (DP160103746) +and Mississippi INBRE, funded by an Institutional Development +Award (IDeA) from the National Institute of General +Medical Sciences of the National Institutes of Health under +grant P20GM103476. +SUPPLEMENTARY MATERIAL +The Supplementary Material for this article can be found online +at: https://www.frontiersin.org/articles/10.3389/fmicb.2021. +651282/full#supplementary-material +Frontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +REFERENCES +Abayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018). +Discovery and characterization of a sulfoquinovose mutarotase using kinetic +analysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383. +doi: 10.1042/bcj20170947 +Andrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron +homeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03) +00055-x +Arkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov, +S., et al. (2018). KBase: the United States department of energy systems biology +knowledgebase. Nat. Biotechnol. 36, 566–569. +Autry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil +organic sulfur. Biol. Fertil. Soils 10, 50–56. +Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al. +(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44, +W147–W153. +Badri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates. +Plant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x +Badri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere +chemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20, +642–650. doi: 10.1016/j.copbio.2009.09.014 +Bais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005). +Mediation of pathogen resistance by exudation of antimicrobials from roots. +Nature 434, 217–221. doi: 10.1038/nature03356 +Bais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of +root exudates in rhizosphere interations with plants and other organisms. Annu. +Rev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159 +Bangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic +locus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the +biological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe +Interact. 9, 83–90. +Barret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G., +and Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic +fungus Gaeumannomyces graminis var. tritici on gene expression of the +biocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe +Interact. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611 +Barrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas +fluorescens biovar V: its resolution into distinct component groups and the +relationship of these groups to other P. fluorescens biovars, to P. putida, and to +psychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol. +132, 2709–2721. doi: 10.1099/00221287-132-10-2709 +Bevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon +genomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21, +211–217. doi: 10.1016/j.copbio.2010.03.006 +Boncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999). +Occurrence of choline and glycine betaine uptake and metabolism in the family +Rhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65, +2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999 +Bondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper +resistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397– +405. doi: 10.1007/s10565-013-9262-1 +Brkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al. +(2011). Brachypodium as a model for the grasses: today and the future. Plant +Physiol. 157, 3–13. +Camacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type +4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens +WCS365. Dissertation, University of Leiden, Leiden. +Chavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A +metabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx +in Pseudomonas putida. mSystems 1:e00154-16. +Chen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread +plant-colonizing bacterial species Pseudomonas syringae detects and exploits an +extracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj. +12262 +Chen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb: +a web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi: +10.3389/fmicb.2018.01910 +Chen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal +acquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi. +2017.06.004 +Cole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M., +et al. (2017). Genome-wide identification of bacterial plant colonization genes. +PLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860 +Conesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional +analysis in plant genomics. Int. J. Plant Genomics 2008:619832. +Curl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag. +Darzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive +pathways explorer v3. Nucleic Acids Res. 46, W510–W513. +de Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al. +(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365 +plays a role in rhizosphere competence through maintaining the structure and +function of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi: +10.1111/j.1574-6941.2006.00158.x +de Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg, +G. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is +an important trait for tomato root colonization by Pseudomonas fluorescens. +Mol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11. +1173 +de Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand +microbe-derived compounds affect the expression of genes encoding +antifungal compounds in a pseudomonad with biocontrol activity. Appl. +Environ. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10 +Dekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998). +A site-specific recombinase is required for competitive root colonization by +Pseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056. +doi: 10.1073/pnas.95.12.7051 +De-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008). +Root-microbe communication through protein secretion. J. Biol. Chem. 283, +25247–25255. doi: 10.1074/jbc.m801967200 +Empadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms +in prokaryotes: distribution of compatible solutes. Int. Microbiol. 11, +151–161. +Espinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of +functions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182, +2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000 +Fuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in +Pseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365- +2958.2010.07250.x +Galvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline- +O-sulphate utilization from osmoprotection in Pseudomonas putida. Mol. +Microbiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x +Garrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R., +and Redondo-Nieto, M. (2016). Genomic and genetic diversity within the +Pseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal. +pone.0150183 +Halverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and +nonpermeating solutes on the fatty acid composition of Pseudomonas putida. +Appl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421. +2000 +Hampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and +Wargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas +aeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13 +Harwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and +the biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/ +annurev.micro.50.1.553 +Hesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018). +Genome-based evolutionary history of Pseudomonas spp. Environ. Microbiol. +20, 2142–2159. +Hinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003). +Transition from reversible to irreversible attachment during biofilm formation +by Pseudomonas fluorescens WCS365 requires an ABC transporter and a large +secreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003. +03615.x +Hojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999). +Oxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring +Frontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +the distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65, +4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999 +Hong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological +growth stages of Brachypodium distachyon: codification and description. Weed +Res. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x +Howell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum +induced damping-off of cotton seedlings by Pseudomonas fluorescens and +its antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto- +70-712 +Hummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz, +M. A. (1998). Regulation of the sulfate starvation response in Pseudomonas +aeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375– +1386. doi: 10.1099/00221287-144-5-1375 +Kabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam, +C. (2017). Physiological and biochemical responses involved in water deficit +tolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/ +journal.pone.0190284 +Kamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005). +Enrichment for enhanced competitive plant root tip colonizers selects for a new +class of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j. +1462-2920.2005.00889.x +Kanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008). +KEGG for linking genomes to life and the environment. Nucleic Acids Res. 36, +D480–D484. +Katoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment +software version 7: improvements in performance and usability. Mol. Biol. Evol. +30, 772–780. doi: 10.1093/molbev/mst010 +Kawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al. +(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium +distachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal. +pone.0164533 +Kind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al. +(2009). FiehnLib: mass spectral and retention index libraries for metabolomics +based on quadrupole and time-of-flight gas chromatography/mass +spectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522 +King, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the +demonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307. +Klonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018). +Transcriptomic profiling of Burkholderia phymatum STM815, Cupriavidus +taiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response +to Mimosa pudica root exudates illuminates the molecular basis of their +nodulation competitiveness and symbiotic evolutionary history.BMC Genomics +19:105. doi: 10.1186/s12864-018-4487-2 +Kobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and +regulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/ +annurev-arplant-042811-105522 +Kohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor +IolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193, +5155–5163. doi: 10.1128/jb.05371-11 +Lee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas +reinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7 +Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for +the display and annotation of phylogenetic and other trees. Nucleic Acids Res. +44, W242–W245. +Li, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of +ortholog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi: +10.1101/gr.1224503 +Liao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient +general purpose program for assigning sequence reads to genomic features. +Bioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656 +Liu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose +utilization by Pseudomonas fluorescens: overlapping genetic responses to xylose, +xylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi. +13142 +Loper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer, +B. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas +spp.: insights into diversity and inheritance of traits involved in multitrophic +interactions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784 +Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change +and dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550. +Lugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria. +Annu. Rev. Microbiol. 63, 541–556. +Lugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular +determinants of rhizosphere colonization by Pseudomonas. Annu. Rev. +Phytopathol. 39, 461–490. +Lynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch +(Chichester: JohnWiley & Sons), 177–206. +Mark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005). +Transcriptome profiling of bacterial responses to root exudates identifies genes +involved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454– +17459. doi: 10.1073/pnas.0506407102 +Markowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y., +et al. (2012). IMG: the integrated microbial genomes database and comparative +analysis system. Nucleic Acids Res. 40, D115–D122. +Martinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF, +the second largest Pseudomonas putida protein, contributes to plant root +colonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561. +doi: 10.1111/j.1365-2958.2010.07249.x +Matilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and +Ramos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving +forces of bacterial life in the rhizosphere. Genome Biol. 8:R179. +Miller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria. +Annu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101 +Moore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.- +L., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes, +eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt +(New York, NY: Springer), 646–703. +Mulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis +of the Pseudomonas species. Environ. Microbiol. 12, 1513–1530. +Naylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought +and host selection influence bacterial community dynamics in the grass root +microbiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118 +Nguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and +controls. Agronomie 23, 375–396. doi: 10.1051/agro:2003011 +Nielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface +interactions mediated by cellulose and a novel exopolysaccharide contribute +to Pseudomonas putida biofilm formation and fitness under water-limiting +conditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011. +02432.x +Parejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow, +L. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid +producing fluorescent Pseudomonas spp. from dryland cereal fields of central +Washington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012- +0015-0 +Phillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R. +(2004). Microbial products trigger amino acid exudation from plant roots. Plant +Physiol. 136, 2887–2894. doi: 10.1104/pp.104.044222 +Raaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek, +P. J., and Schippers, B. (1995). Utilization of heterologous siderophores and +rhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41, +126–135. doi: 10.1139/m95-017 +Raaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4- +diacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils. +Mol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144 +Rahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis, +J., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase +contains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/ +acschembio.7b00304 +Ramos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of +Pseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo +expression technology capture and identification of root-activated promoters. +J. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005 +Reinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T. +(2015). Roots shaping their microbiome: global hotspots for microbial activity. +Annu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10 +2342 +Frontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +Sanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and +Rivilla, R. (2002). Phenotypic selection and phase variation occur during +alfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184, +1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002 +Sarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995). +The sigma factor ss affects antibiotic production and biological control activity +of Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259. +doi: 10.1073/pnas.92.26.12255 +Schnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The +sigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance +towards desiccation and osmotic stress in the biocontrol agent Pseudomonas +fluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem. +67.12.5683-5693.2001 +Schroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006). +“Phytopathogenic pseudomonads and related plant-associated pseudomonads,” +in The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.- +H. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740. +doi: 10.1007/0-387-30746-x_23 +Schwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and +Amasino, R. M. (2010). Natural variation of flowering time and vernalization +responsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10. +1007/s12155-009-9069-3 +Silby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson, +R. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity +and plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51. +Silby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to +identify genes important in growth and survival of Pseudomonas fluorescens +Pf0-1 in soil: Discovery of expressed sequences with novel genetic organization. +J. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004 +Simons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A., +and Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere +colonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant +Microbe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600 +Simons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and +Lugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root +colonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe +Interact. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102 +Smibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods +for General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A. +Wood, and N. R. Kreig (Washington, DC: American Society of Microbiology), +607–654. +Thomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from +Pseudomonas fluorescens in biological control of Gaeumannomyces graminis +var. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988 +Thomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990). +Production of the antibiotic phenazine-1-carboxylic acid by fluorescent +Pseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56, +908–912. doi: 10.1128/aem.56.4.908-912.1990 +Tyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats, +W. G., et al. (2014). Selection and phenotypic characterization of a core +collection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25. +doi: 10.1186/1471-2229-14-25 +Vacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y., +Muller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system +functioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356 +van den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of +phenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol. +7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x +van Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity +of microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135. +doi: 10.1128/.61.2.121-135.1997 +Walker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M. +(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric. +Food Chem. 51, 2548–2554. +Wargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to +Pseudomonas aeruginosa survival during murine lung infection. PLoS One +8:e56850. doi: 10.1371/journal.pone.0056850 +Wargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine: +lessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112– +2120. +Whipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch +(Chichester: JohnWiley & Sons), 59–97. +Winsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock, +R. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly, +comprehensive comparisons of microbial genomes. Nucleic Acids Res. 37, +D483–D488. +Yahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The +Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. +Stackebrandt (New York, NY: Springer), 704–713. +Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982). +Living with water stress: evolution of osmolyte systems. Science 217, 1214–1222. +Ye, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web +tool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res. +46, W71–W75. +Yoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H., +et al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283, +10415–10424. +Yousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different, +overlapping mechanisms for colonization of abiotic and plant surfaces by +Pseudomonas putida. FEMS Microbiol. Lett. 288, 118–124. +Zboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere +colonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol. +J. 18, 3539–3554. +Zolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K., +et al. (2013). “Understanding root-microbiome interactions,” in Molecular +Microbial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John +Wiley & Sons), 745–754. +Conflict of Interest: The authors declare that the research was conducted in the +absence of any commercial or financial relationships that could be construed as a +potential conflict of interest. +Copyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne, +LeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an +open-access article distributed under the terms of the Creative Commons Attribution +License (CC BY). The use, distribution or reproduction in other forums is permitted, +provided the original author(s) and the copyright owner(s) are credited and that the +original publication in this journal is cited, in accordance with accepted academic +practice. No use, distribution or reproduction is permitted which does not comply +with these terms. +Frontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282 + + +INFO:ontogpt.io.template_loader:Loading custom schema from src/ontogpt/templates/biolog.yaml +INFO:linkml_runtime.utils.schemaview:Importing core as core from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates +INFO:linkml_runtime.utils.schemaview:Importing linkml:types as /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/linkml_runtime/linkml_model/model/schema/types from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=None +INFO:linkml_runtime.utils.schemaview:Importing core as core from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates +INFO:linkml_runtime.utils.schemaview:Importing linkml:types as /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/linkml_runtime/linkml_model/model/schema/types from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=None +INFO:ontogpt.io.template_loader:Getting class for template src/ontogpt/templates/biolog.yaml +INFO:root:Using template Paper +INFO:root:Using mappers (currently hardcoded) +INFO:ontogpt.clients.llm_client:Complete: engine=gpt-4.5-preview, prompt[100766]=From the text below, extract the following entities in the following format: + +study_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +DEBUG:LiteLLM:litellm.completion(api_key='', api_base=None, api_version=None, model='gpt-4.5-preview', messages=[{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +23:25:35 - LiteLLM:DEBUG: utils.py:296 - + +DEBUG:LiteLLM: + +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {} +DEBUG:LiteLLM:self.optional_params: {} +23:25:35 - LiteLLM:DEBUG: utils.py:296 - SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +DEBUG:LiteLLM:SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +23:25:35 - LiteLLM:DEBUG: utils.py:925 - INSIDE CHECKING SYNC CACHE +DEBUG:LiteLLM:INSIDE CHECKING SYNC CACHE +23:25:35 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:25:35 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:25:35 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:25:35 - LiteLLM:INFO: utils.py:2926 - +LiteLLM completion() model= gpt-4.5-preview; provider = openai +INFO:LiteLLM: +LiteLLM completion() model= gpt-4.5-preview; provider = openai +23:25:35 - LiteLLM:DEBUG: utils.py:2929 - +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +DEBUG:LiteLLM: +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +23:25:35 - LiteLLM:DEBUG: utils.py:2932 - +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +DEBUG:LiteLLM: +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +23:25:35 - LiteLLM:DEBUG: utils.py:296 - Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:self.optional_params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:httpx:load_ssl_context verify=True cert=None trust_env=True http2=False +DEBUG:httpx:load_verify_locations cafile='/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/certifi/cacert.pem' +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:614 - PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +DEBUG:LiteLLM:PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:624 -  + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:LiteLLM: + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:openai._base_client:Request options: {'method': 'post', 'url': '/chat/completions', 'headers': {'X-Stainless-Raw-Response': 'true'}, 'timeout': 600.0, 'files': None, 'json_data': {'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'model': 'gpt-4.5-preview', 'temperature': 1.0}, 'extra_json': {}} +DEBUG:openai._base_client:Sending HTTP Request: POST https://api.openai.com/v1/chat/completions +DEBUG:httpcore.connection:connect_tcp.started host='api.openai.com' port=443 local_address=None timeout=600.0 socket_options=None +DEBUG:httpcore.connection:connect_tcp.complete return_value= +DEBUG:httpcore.connection:start_tls.started ssl_context= server_hostname='api.openai.com' timeout=600.0 +DEBUG:httpcore.connection:start_tls.complete return_value= +DEBUG:httpcore.http11:send_request_headers.started request= +DEBUG:httpcore.http11:send_request_headers.complete +DEBUG:httpcore.http11:send_request_body.started request= +DEBUG:httpcore.http11:send_request_body.complete +DEBUG:httpcore.http11:receive_response_headers.started request= +DEBUG:httpcore.http11:receive_response_headers.complete return_value=(b'HTTP/1.1', 200, b'OK', [(b'Date', b'Wed, 05 Mar 2025 07:26:07 GMT'), (b'Content-Type', b'application/json'), (b'Transfer-Encoding', b'chunked'), (b'Connection', b'keep-alive'), (b'access-control-expose-headers', b'X-Request-ID'), (b'openai-organization', b'lawrence-berkeley-national-laboratory-8'), (b'openai-processing-ms', b'32096'), (b'openai-version', b'2020-10-01'), (b'x-ratelimit-limit-requests', b'10000'), (b'x-ratelimit-limit-tokens', b'2000000'), (b'x-ratelimit-remaining-requests', b'9999'), (b'x-ratelimit-remaining-tokens', b'1974681'), (b'x-ratelimit-reset-requests', b'6ms'), (b'x-ratelimit-reset-tokens', b'759ms'), (b'x-request-id', b'req_46b2bea89ddaddbe030dfd435c97dfa2'), (b'strict-transport-security', b'max-age=31536000; includeSubDomains; preload'), (b'cf-cache-status', b'DYNAMIC'), (b'Set-Cookie', b'__cf_bm=xWhtGZy4GkjfLHyiSPTezcmRYi.uvd.uXbx4Q4RzppU-1741159567-1.0.1.1-sOAa0JWsrXDlqfH0FF7KxXc__jUyHH4CoWDjTdrs03xP7R0RbK4IjCesizfy0YvXAH1o1bkXoyGY1JTx4kdcH8ni4FKVk1m.xv65nMtWwG4; path=/; expires=Wed, 05-Mar-25 07:56:07 GMT; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), (b'X-Content-Type-Options', b'nosniff'), (b'Set-Cookie', b'_cfuvid=oBEVFNLmxi3159bRP_cghDOOkjcknNYaVoFV5G375yo-1741159567673-0.0.1.1-604800000; path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), (b'Server', b'cloudflare'), (b'CF-RAY', b'91b7e158aba0ce4c-SJC'), (b'Content-Encoding', b'gzip'), (b'alt-svc', b'h3=":443"; ma=86400')]) +INFO:httpx:HTTP Request: POST https://api.openai.com/v1/chat/completions "HTTP/1.1 200 OK" +DEBUG:httpcore.http11:receive_response_body.started request= +DEBUG:httpcore.http11:receive_response_body.complete +DEBUG:httpcore.http11:response_closed.started +DEBUG:httpcore.http11:response_closed.complete +DEBUG:openai._base_client:HTTP Response: POST https://api.openai.com/v1/chat/completions "200 OK" Headers([('date', 'Wed, 05 Mar 2025 07:26:07 GMT'), ('content-type', 'application/json'), ('transfer-encoding', 'chunked'), ('connection', 'keep-alive'), ('access-control-expose-headers', 'X-Request-ID'), ('openai-organization', 'lawrence-berkeley-national-laboratory-8'), ('openai-processing-ms', '32096'), ('openai-version', '2020-10-01'), ('x-ratelimit-limit-requests', '10000'), ('x-ratelimit-limit-tokens', '2000000'), ('x-ratelimit-remaining-requests', '9999'), ('x-ratelimit-remaining-tokens', '1974681'), ('x-ratelimit-reset-requests', '6ms'), ('x-ratelimit-reset-tokens', '759ms'), ('x-request-id', 'req_46b2bea89ddaddbe030dfd435c97dfa2'), ('strict-transport-security', 'max-age=31536000; includeSubDomains; preload'), ('cf-cache-status', 'DYNAMIC'), ('set-cookie', '__cf_bm=xWhtGZy4GkjfLHyiSPTezcmRYi.uvd.uXbx4Q4RzppU-1741159567-1.0.1.1-sOAa0JWsrXDlqfH0FF7KxXc__jUyHH4CoWDjTdrs03xP7R0RbK4IjCesizfy0YvXAH1o1bkXoyGY1JTx4kdcH8ni4FKVk1m.xv65nMtWwG4; path=/; expires=Wed, 05-Mar-25 07:56:07 GMT; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), ('x-content-type-options', 'nosniff'), ('set-cookie', '_cfuvid=oBEVFNLmxi3159bRP_cghDOOkjcknNYaVoFV5G375yo-1741159567673-0.0.1.1-604800000; path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), ('server', 'cloudflare'), ('cf-ray', '91b7e158aba0ce4c-SJC'), ('content-encoding', 'gzip'), ('alt-svc', 'h3=":443"; ma=86400')]) +DEBUG:openai._base_client:request_id: req_46b2bea89ddaddbe030dfd435c97dfa2 +23:26:07 - LiteLLM:DEBUG: utils.py:296 - RAW RESPONSE: +{"id": "chatcmpl-B7dQFT52GJFSYAsipHD1MZCWpm98U", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas\nauthors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi\ndoi: 10.3389/fmicb.2021.651282\ndate: 14 April 2021\nexperiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C). Root exudates were extracted, sterilized with 0.22 \u00b5m filtration, lyophilized, and stored at \u201380\u00b0C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25\u00b0C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159535, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 687, "prompt_tokens": 30161, "total_tokens": 30848, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 30080}}} + + +DEBUG:LiteLLM:RAW RESPONSE: +{"id": "chatcmpl-B7dQFT52GJFSYAsipHD1MZCWpm98U", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas\nauthors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi\ndoi: 10.3389/fmicb.2021.651282\ndate: 14 April 2021\nexperiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C). Root exudates were extracted, sterilized with 0.22 \u00b5m filtration, lyophilized, and stored at \u201380\u00b0C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25\u00b0C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159535, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 687, "prompt_tokens": 30161, "total_tokens": 30848, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 30080}}} + + +23:26:07 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:26:07 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:26:07 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:26:07 - LiteLLM:INFO: utils.py:1026 - Wrapper: Completed Call, calling success_handler +INFO:LiteLLM:Wrapper: Completed Call, calling success_handler +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:982 - Logging Details LiteLLM-Success Call: Cache_hit=None +DEBUG:LiteLLM:Logging Details LiteLLM-Success Call: Cache_hit=None +23:26:07 - LiteLLM:DEBUG: cost_calculator.py:599 - completion_response response ms: None +DEBUG:LiteLLM:completion_response response ms: None +23:26:07 - LiteLLM:DEBUG: cost_calculator.py:599 - completion_response response ms: None +INFO:root:RAW TEXT: study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas +authors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi +doi: 10.3389/fmicb.2021.651282 +date: 14 April 2021 +experiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. +DEBUG:LiteLLM:completion_response response ms: None +INFO:root:PARSING LINE: study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas +DEBUG:root: FIELD: study_title +DEBUG:root:SLOT: study_title INL: None VALS: ['Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas'] +INFO:root:PARSING LINE: authors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi +DEBUG:root: FIELD: authors +DEBUG:root:SLOT: authors INL: None VALS: ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'] +INFO:root:PARSING LINE: doi: 10.3389/fmicb.2021.651282 +DEBUG:root: FIELD: doi +DEBUG:root:SLOT: doi INL: None VALS: ['10.3389/fmicb.2021.651282'] +INFO:root:PARSING LINE: date: 14 April 2021 +DEBUG:root: FIELD: date +DEBUG:root:SLOT: date INL: None VALS: ['14 April 2021'] +INFO:root:PARSING LINE: experiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. +DEBUG:root: FIELD: experiments +DEBUG:root:SLOT: experiments INL: True VALS: ['The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.'] +DEBUG:root: RECURSING ON SLOT: experiments, range=Experiment +INFO:ontogpt.clients.llm_client:Complete: engine=gpt-4.5-preview, prompt[5026]=From the text below, extract the following entities in the following format: + +label: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +DEBUG:LiteLLM:litellm.completion(api_key='', api_base=None, api_version=None, model='gpt-4.5-preview', messages=[{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +23:26:07 - LiteLLM:DEBUG: utils.py:296 - + +DEBUG:LiteLLM: + +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {} +DEBUG:LiteLLM:self.optional_params: {} +23:26:07 - LiteLLM:DEBUG: utils.py:296 - SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +DEBUG:LiteLLM:SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +23:26:07 - LiteLLM:DEBUG: utils.py:925 - INSIDE CHECKING SYNC CACHE +DEBUG:LiteLLM:INSIDE CHECKING SYNC CACHE +23:26:07 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:26:07 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:07 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:07 - LiteLLM:INFO: utils.py:2926 - +LiteLLM completion() model= gpt-4.5-preview; provider = openai +INFO:LiteLLM: +LiteLLM completion() model= gpt-4.5-preview; provider = openai +23:26:07 - LiteLLM:DEBUG: utils.py:2929 - +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +DEBUG:LiteLLM: +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +23:26:07 - LiteLLM:DEBUG: utils.py:2932 - +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +DEBUG:LiteLLM: +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +23:26:07 - LiteLLM:DEBUG: utils.py:296 - Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:self.optional_params: {'temperature': 1.0, 'extra_body': {}} +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:614 - PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +DEBUG:LiteLLM:PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:624 -  + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:LiteLLM: + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:openai._base_client:Request options: {'method': 'post', 'url': '/chat/completions', 'headers': {'X-Stainless-Raw-Response': 'true'}, 'timeout': 600.0, 'files': None, 'json_data': {'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'model': 'gpt-4.5-preview', 'temperature': 1.0}, 'extra_json': {}} +DEBUG:openai._base_client:Sending HTTP Request: POST https://api.openai.com/v1/chat/completions +DEBUG:httpcore.http11:send_request_headers.started request= +DEBUG:httpcore.http11:send_request_headers.complete +DEBUG:httpcore.http11:send_request_body.started request= +DEBUG:httpcore.http11:send_request_body.complete +DEBUG:httpcore.http11:receive_response_headers.started request= +DEBUG:httpcore.http11:receive_response_headers.complete return_value=(b'HTTP/1.1', 200, b'OK', [(b'Date', b'Wed, 05 Mar 2025 07:26:35 GMT'), (b'Content-Type', b'application/json'), (b'Transfer-Encoding', b'chunked'), (b'Connection', b'keep-alive'), (b'access-control-expose-headers', b'X-Request-ID'), (b'openai-organization', b'lawrence-berkeley-national-laboratory-8'), (b'openai-processing-ms', b'27531'), (b'openai-version', b'2020-10-01'), (b'x-ratelimit-limit-requests', b'10000'), (b'x-ratelimit-limit-tokens', b'2000000'), (b'x-ratelimit-remaining-requests', b'9999'), (b'x-ratelimit-remaining-tokens', b'1997008'), (b'x-ratelimit-reset-requests', b'6ms'), (b'x-ratelimit-reset-tokens', b'89ms'), (b'x-request-id', b'req_cd62a17c1834de22e9d897a211325165'), (b'strict-transport-security', b'max-age=31536000; includeSubDomains; preload'), (b'cf-cache-status', b'DYNAMIC'), (b'X-Content-Type-Options', b'nosniff'), (b'Server', b'cloudflare'), (b'CF-RAY', b'91b7e2224ecace4c-SJC'), (b'Content-Encoding', b'gzip'), (b'alt-svc', b'h3=":443"; ma=86400')]) +INFO:httpx:HTTP Request: POST https://api.openai.com/v1/chat/completions "HTTP/1.1 200 OK" +DEBUG:httpcore.http11:receive_response_body.started request= +DEBUG:httpcore.http11:receive_response_body.complete +DEBUG:httpcore.http11:response_closed.started +DEBUG:httpcore.http11:response_closed.complete +DEBUG:openai._base_client:HTTP Response: POST https://api.openai.com/v1/chat/completions "200 OK" Headers({'date': 'Wed, 05 Mar 2025 07:26:35 GMT', 'content-type': 'application/json', 'transfer-encoding': 'chunked', 'connection': 'keep-alive', 'access-control-expose-headers': 'X-Request-ID', 'openai-organization': 'lawrence-berkeley-national-laboratory-8', 'openai-processing-ms': '27531', 'openai-version': '2020-10-01', 'x-ratelimit-limit-requests': '10000', 'x-ratelimit-limit-tokens': '2000000', 'x-ratelimit-remaining-requests': '9999', 'x-ratelimit-remaining-tokens': '1997008', 'x-ratelimit-reset-requests': '6ms', 'x-ratelimit-reset-tokens': '89ms', 'x-request-id': 'req_cd62a17c1834de22e9d897a211325165', 'strict-transport-security': 'max-age=31536000; includeSubDomains; preload', 'cf-cache-status': 'DYNAMIC', 'x-content-type-options': 'nosniff', 'server': 'cloudflare', 'cf-ray': '91b7e2224ecace4c-SJC', 'content-encoding': 'gzip', 'alt-svc': 'h3=":443"; ma=86400'}) +DEBUG:openai._base_client:request_id: req_cd62a17c1834de22e9d897a211325165 +23:26:35 - LiteLLM:DEBUG: utils.py:296 - RAW RESPONSE: +{"id": "chatcmpl-B7dQlIq6tg5poLKl93TUpi5dfUF0M", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "label: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.\n\nexperiment_motivation: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.\n\nexperiment_design: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.\n\nenvironment: Controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C); microaerobic environment (15% O\u2082).\n\nhost: Brachypodium distachyon; Pseudomonas fluorescens-group strains\n\nhost_type: Plant; Bacteria\n\ntarget_microbes: Pseudomonas fluorescens\n\nbiological_system: Rhizosphere\n\nconditions: 20-hour light; 4-hour dark; 24\u00b0C daytime; 18\u00b0C nighttime; microaerobic (15% O\u2082)\n\nexperimental_factors: Presence of Brachypodium root exudates\n\nis_biolog_experiment: yes\n\ntype_of_biolog_experiment: Phenotype MicroArray\n\nplates: PM01; PM2A\n\nreplicates: 3\n\nprotocol_steps: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24\u00b0C, 4-hour dark 18\u00b0C); root exudates collected, sterilized via 0.22 \u00b5m filtration, lyophilized, stored at \u201380\u00b0C; root exudates analyzed by LC-MS and GC-MS; LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection; GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC; cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O\u2082) conditions; RNA extracted and depleted of rRNA; stranded library preparation and sequencing via Illumina HiSeq 2500; Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours; collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes; Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02; Transcriptome data analyzed using DESeq2 and comparative genomic tools.\n\nplate_reader_types_of_data_collected: Optical density; formazan (absorbance at 590 nm)\n\nod_protocol: Not provided\n\nrespiration_protocol: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25\u00b0C.\n\ninstrument_used: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer; Ultra Performance Liquid Chromatograph (UPLC); Pegasus 4D TOF Mass Spectrometer; Agilent 7890A GC; Illumina HiSeq 2500; OmniLog Phenotype MicroArray System\n\nanalysis_software: OmniLog Parametric Analysis software v1.20.02; DESeq2; OrthoMCL; COG-based phylogeny\n\nincubation_temperature: 25\u00b0C\n\nincubation_duration: 48 hours", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159567, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 818, "prompt_tokens": 1158, "total_tokens": 1976, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 0}}} + + +DEBUG:LiteLLM:RAW RESPONSE: +{"id": "chatcmpl-B7dQlIq6tg5poLKl93TUpi5dfUF0M", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "label: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.\n\nexperiment_motivation: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.\n\nexperiment_design: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.\n\nenvironment: Controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C); microaerobic environment (15% O\u2082).\n\nhost: Brachypodium distachyon; Pseudomonas fluorescens-group strains\n\nhost_type: Plant; Bacteria\n\ntarget_microbes: Pseudomonas fluorescens\n\nbiological_system: Rhizosphere\n\nconditions: 20-hour light; 4-hour dark; 24\u00b0C daytime; 18\u00b0C nighttime; microaerobic (15% O\u2082)\n\nexperimental_factors: Presence of Brachypodium root exudates\n\nis_biolog_experiment: yes\n\ntype_of_biolog_experiment: Phenotype MicroArray\n\nplates: PM01; PM2A\n\nreplicates: 3\n\nprotocol_steps: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24\u00b0C, 4-hour dark 18\u00b0C); root exudates collected, sterilized via 0.22 \u00b5m filtration, lyophilized, stored at \u201380\u00b0C; root exudates analyzed by LC-MS and GC-MS; LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection; GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC; cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O\u2082) conditions; RNA extracted and depleted of rRNA; stranded library preparation and sequencing via Illumina HiSeq 2500; Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours; collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes; Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02; Transcriptome data analyzed using DESeq2 and comparative genomic tools.\n\nplate_reader_types_of_data_collected: Optical density; formazan (absorbance at 590 nm)\n\nod_protocol: Not provided\n\nrespiration_protocol: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25\u00b0C.\n\ninstrument_used: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer; Ultra Performance Liquid Chromatograph (UPLC); Pegasus 4D TOF Mass Spectrometer; Agilent 7890A GC; Illumina HiSeq 2500; OmniLog Phenotype MicroArray System\n\nanalysis_software: OmniLog Parametric Analysis software v1.20.02; DESeq2; OrthoMCL; COG-based phylogeny\n\nincubation_temperature: 25\u00b0C\n\nincubation_duration: 48 hours", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159567, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 818, "prompt_tokens": 1158, "total_tokens": 1976, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 0}}} + + +23:26:35 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:26:35 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:35 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:35 - LiteLLM:INFO: utils.py:1026 - Wrapper: Completed Call, calling success_handler +INFO:LiteLLM:Wrapper: Completed Call, calling success_handler +23:26:35 - LiteLLM:DEBUG: cost_calculator.py:599 - completion_response response ms: None +DEBUG:LiteLLM:completion_response response ms: None +INFO:root:PARSING LINE: label: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates. +DEBUG:root: FIELD: label +DEBUG:root:SLOT: label INL: None VALS: ['This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.'] +INFO:root:PARSING LINE: experiment_motivation: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates. +DEBUG:root: FIELD: experiment_motivation +23:26:35 - LiteLLM:DEBUG: litellm_logging.py:982 - Logging Details LiteLLM-Success Call: Cache_hit=None +DEBUG:root:SLOT: experiment_motivation INL: None VALS: ['To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.'] +INFO:root:PARSING LINE: experiment_design: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays. +DEBUG:root: FIELD: experiment_design +DEBUG:LiteLLM:Logging Details LiteLLM-Success Call: Cache_hit=None +DEBUG:root:SLOT: experiment_design INL: None VALS: ['Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.'] +INFO:root:PARSING LINE: environment: Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂). +DEBUG:root: FIELD: environment +DEBUG:root:SLOT: environment INL: None VALS: ['Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).'] +INFO:root:PARSING LINE: host: Brachypodium distachyon; Pseudomonas fluorescens-group strains +DEBUG:root: FIELD: host +DEBUG:root:SLOT: host INL: False VALS: ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'] +INFO:root:PARSING LINE: host_type: Plant; Bacteria +DEBUG:root: FIELD: host_type +DEBUG:root:SLOT: host_type INL: None VALS: ['Plant; Bacteria'] +INFO:root:PARSING LINE: target_microbes: Pseudomonas fluorescens +DEBUG:root: FIELD: target_microbes +DEBUG:root:SLOT: target_microbes INL: False VALS: ['Pseudomonas fluorescens'] +INFO:root:PARSING LINE: biological_system: Rhizosphere +DEBUG:root: FIELD: biological_system +DEBUG:root:SLOT: biological_system INL: None VALS: ['Rhizosphere'] +INFO:root:PARSING LINE: conditions: 20-hour light; 4-hour dark; 24°C daytime; 18°C nighttime; microaerobic (15% O₂) +DEBUG:root: FIELD: conditions +DEBUG:root:SLOT: conditions INL: None VALS: ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'] +INFO:root:PARSING LINE: experimental_factors: Presence of Brachypodium root exudates +DEBUG:root: FIELD: experimental_factors +DEBUG:root:SLOT: experimental_factors INL: False VALS: ['Presence of Brachypodium root exudates'] +INFO:root:PARSING LINE: is_biolog_experiment: yes +DEBUG:root: FIELD: is_biolog_experiment +DEBUG:root:SLOT: is_biolog_experiment INL: None VALS: ['yes'] +INFO:root:PARSING LINE: type_of_biolog_experiment: Phenotype MicroArray +DEBUG:root: FIELD: type_of_biolog_experiment +DEBUG:root:SLOT: type_of_biolog_experiment INL: None VALS: ['Phenotype MicroArray'] +INFO:root:PARSING LINE: plates: PM01; PM2A +DEBUG:root: FIELD: plates +DEBUG:root:SLOT: plates INL: None VALS: ['PM01', 'PM2A'] +INFO:root:PARSING LINE: replicates: 3 +DEBUG:root: FIELD: replicates +DEBUG:root:SLOT: replicates INL: None VALS: ['3'] +INFO:root:PARSING LINE: protocol_steps: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C); root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C; root exudates analyzed by LC-MS and GC-MS; LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection; GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC; cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions; RNA extracted and depleted of rRNA; stranded library preparation and sequencing via Illumina HiSeq 2500; Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours; collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes; Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02; Transcriptome data analyzed using DESeq2 and comparative genomic tools. +DEBUG:root: FIELD: protocol_steps +DEBUG:root:SLOT: protocol_steps INL: None VALS: ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'] +INFO:root:PARSING LINE: plate_reader_types_of_data_collected: Optical density; formazan (absorbance at 590 nm) +DEBUG:root: FIELD: plate_reader_types_of_data_collected +DEBUG:root:SLOT: plate_reader_types_of_data_collected INL: None VALS: ['Optical density', 'formazan (absorbance at 590 nm)'] +INFO:root:PARSING LINE: od_protocol: Not provided +DEBUG:root: FIELD: od_protocol +DEBUG:root:SLOT: od_protocol INL: None VALS: ['Not provided'] +INFO:root:PARSING LINE: respiration_protocol: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C. +DEBUG:root: FIELD: respiration_protocol +DEBUG:root:SLOT: respiration_protocol INL: None VALS: ['Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.'] +INFO:root:PARSING LINE: instrument_used: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer; Ultra Performance Liquid Chromatograph (UPLC); Pegasus 4D TOF Mass Spectrometer; Agilent 7890A GC; Illumina HiSeq 2500; OmniLog Phenotype MicroArray System +DEBUG:root: FIELD: instrument_used +DEBUG:root:SLOT: instrument_used INL: None VALS: ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'] +INFO:root:PARSING LINE: analysis_software: OmniLog Parametric Analysis software v1.20.02; DESeq2; OrthoMCL; COG-based phylogeny +DEBUG:root: FIELD: analysis_software +DEBUG:root:SLOT: analysis_software INL: None VALS: ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'] +INFO:root:PARSING LINE: incubation_temperature: 25°C +DEBUG:root: FIELD: incubation_temperature +DEBUG:root:SLOT: incubation_temperature INL: None VALS: ['25°C'] +INFO:root:PARSING LINE: incubation_duration: 48 hours +DEBUG:root: FIELD: incubation_duration +DEBUG:root:SLOT: incubation_duration INL: None VALS: ['48 hours'] +DEBUG:root:RAW: {'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [{'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'}]} +DEBUG:root:Grounding annotation object {'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [{'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'}]} +DEBUG:root:FIELD: study_title SLOT: study_title +DEBUG:root: VAL: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas +DEBUG:root:FIELD: authors SLOT: authors +DEBUG:root: VAL: Olga V. Mavrodi +DEBUG:root: VAL: Janiece R. McWilliams +DEBUG:root: VAL: Jacob O. Peter +DEBUG:root: VAL: Anna Berim +DEBUG:root: VAL: Karl A. Hassan +DEBUG:root: VAL: Liam D. H. Elbourne +DEBUG:root: VAL: Melissa K. LeTourneau +DEBUG:root: VAL: David R. Gang +DEBUG:root: VAL: Ian T. Paulsen +DEBUG:root: VAL: David M. Weller +DEBUG:root: VAL: Linda S. Thomashow +DEBUG:root: VAL: Alex S. Flynt +DEBUG:root: VAL: Dmitri V. Mavrodi +DEBUG:root:FIELD: doi SLOT: doi +DEBUG:root: VAL: 10.3389/fmicb.2021.651282 +DEBUG:root:FIELD: date SLOT: date +DEBUG:root: VAL: 14 April 2021 +DEBUG:root:FIELD: experiments SLOT: experiments +DEBUG:root: VAL: {'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +DEBUG:root:Grounding annotation object {'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +DEBUG:root:FIELD: label SLOT: label +DEBUG:root: VAL: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates. +DEBUG:root:FIELD: experiment_motivation SLOT: experiment_motivation +DEBUG:root: VAL: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates. +DEBUG:root:FIELD: experiment_design SLOT: experiment_design +DEBUG:root: VAL: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays. +DEBUG:root:FIELD: environment SLOT: environment +DEBUG:root: VAL: Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂). +DEBUG:root:FIELD: host SLOT: host +DEBUG:root: VAL: Brachypodium distachyon +INFO:ontogpt.engines.knowledge_engine:GROUNDING Brachypodium distachyon using Host +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Brachypodium distachyon to Host +DEBUG:root: VAL: Pseudomonas fluorescens-group strains +INFO:ontogpt.engines.knowledge_engine:GROUNDING Pseudomonas fluorescens-group strains using Host +INFO:ontogpt.engines.knowledge_engine:Singularized Pseudomonas fluorescens-group strains to pseudomonas fluorescens-group strain +INFO:ontogpt.engines.knowledge_engine:GROUNDING pseudomonas fluorescens-group strain using Host +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Pseudomonas fluorescens-group strains to Host +DEBUG:root:FIELD: host_type SLOT: host_type +DEBUG:root: VAL: Plant; Bacteria +DEBUG:root:FIELD: target_microbes SLOT: target_microbes +DEBUG:root: VAL: Pseudomonas fluorescens +INFO:ontogpt.engines.knowledge_engine:GROUNDING Pseudomonas fluorescens using Microbe +INFO:ontogpt.engines.knowledge_engine:Singularized Pseudomonas fluorescens to pseudomonas fluorescen +INFO:ontogpt.engines.knowledge_engine:GROUNDING pseudomonas fluorescen using Microbe +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Pseudomonas fluorescens to Microbe +DEBUG:root:FIELD: biological_system SLOT: biological_system +DEBUG:root: VAL: Rhizosphere +DEBUG:root:FIELD: conditions SLOT: conditions +DEBUG:root: VAL: 20-hour light +DEBUG:root: VAL: 4-hour dark +DEBUG:root: VAL: 24°C daytime +DEBUG:root: VAL: 18°C nighttime +DEBUG:root: VAL: microaerobic (15% O₂) +DEBUG:root:FIELD: experimental_factors SLOT: experimental_factors +DEBUG:root: VAL: Presence of Brachypodium root exudates +INFO:ontogpt.engines.knowledge_engine:GROUNDING Presence of Brachypodium root exudates using ExperimentalFactor +INFO:ontogpt.engines.knowledge_engine:Singularized Presence of Brachypodium root exudates to presence of brachypodium root exudate +INFO:ontogpt.engines.knowledge_engine:GROUNDING presence of brachypodium root exudate using ExperimentalFactor +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Presence of Brachypodium root exudates to ExperimentalFactor +DEBUG:root:FIELD: is_biolog_experiment SLOT: is_biolog_experiment +DEBUG:root: VAL: yes +INFO:root:Looking for yes in FlexibleBooleanEnum +DEBUG:root:FIELD: type_of_biolog_experiment SLOT: type_of_biolog_experiment +DEBUG:root: VAL: Phenotype MicroArray +DEBUG:root:FIELD: plates SLOT: plates +DEBUG:root: VAL: PM01 +DEBUG:root: VAL: PM2A +DEBUG:root:FIELD: replicates SLOT: replicates +DEBUG:root: VAL: 3 +DEBUG:root:FIELD: protocol_steps SLOT: protocol_steps +DEBUG:root: VAL: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C) +DEBUG:root: VAL: root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C +DEBUG:root: VAL: root exudates analyzed by LC-MS and GC-MS +DEBUG:root: VAL: LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection +DEBUG:root: VAL: GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC +DEBUG:root: VAL: cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions +DEBUG:root: VAL: RNA extracted and depleted of rRNA +DEBUG:root: VAL: stranded library preparation and sequencing via Illumina HiSeq 2500 +DEBUG:root: VAL: Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours +DEBUG:root: VAL: collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes +DEBUG:root: VAL: Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02 +DEBUG:root: VAL: Transcriptome data analyzed using DESeq2 and comparative genomic tools. +DEBUG:root:FIELD: plate_reader_types_of_data_collected SLOT: plate_reader_types_of_data_collected +DEBUG:root: VAL: Optical density +DEBUG:root: VAL: formazan (absorbance at 590 nm) +DEBUG:root:FIELD: od_protocol SLOT: od_protocol +DEBUG:root: VAL: Not provided +DEBUG:root:FIELD: respiration_protocol SLOT: respiration_protocol +DEBUG:root: VAL: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C. +DEBUG:root:FIELD: instrument_used SLOT: instrument_used +DEBUG:root: VAL: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer +DEBUG:root: VAL: Ultra Performance Liquid Chromatograph (UPLC) +DEBUG:root: VAL: Pegasus 4D TOF Mass Spectrometer +DEBUG:root: VAL: Agilent 7890A GC +DEBUG:root: VAL: Illumina HiSeq 2500 +DEBUG:root: VAL: OmniLog Phenotype MicroArray System +DEBUG:root:FIELD: analysis_software SLOT: analysis_software +DEBUG:root: VAL: OmniLog Parametric Analysis software v1.20.02 +DEBUG:root: VAL: DESeq2 +DEBUG:root: VAL: OrthoMCL +DEBUG:root: VAL: COG-based phylogeny +DEBUG:root:FIELD: incubation_temperature SLOT: incubation_temperature +DEBUG:root: VAL: 25°C +DEBUG:root:FIELD: incubation_duration SLOT: incubation_duration +DEBUG:root: VAL: 48 hours +DEBUG:root:Creating object from dict {'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['AUTO:Pseudomonas%20fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +INFO:root:{'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['AUTO:Pseudomonas%20fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +DEBUG:root:Creating object from dict {'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [Experiment(label='This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', experiment_motivation='To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', experiment_design='Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', environment='Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', host=['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], host_type='Plant; Bacteria', target_microbes=['AUTO:Pseudomonas%20fluorescens'], biological_system='Rhizosphere', conditions=['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], experimental_factors=['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], is_biolog_experiment='yes', type_of_biolog_experiment='Phenotype MicroArray', plates=['PM01', 'PM2A'], replicates='3', protocol_steps=['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], plate_reader_types_of_data_collected=['Optical density', 'formazan (absorbance at 590 nm)'], od_protocol='Not provided', respiration_protocol='Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', instrument_used=['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], analysis_software=['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], incubation_temperature='25°C', incubation_duration='48 hours')]} +INFO:root:{'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [Experiment(label='This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', experiment_motivation='To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', experiment_design='Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', environment='Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', host=['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], host_type='Plant; Bacteria', target_microbes=['AUTO:Pseudomonas%20fluorescens'], biological_system='Rhizosphere', conditions=['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], experimental_factors=['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], is_biolog_experiment='yes', type_of_biolog_experiment='Phenotype MicroArray', plates=['PM01', 'PM2A'], replicates='3', protocol_steps=['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], plate_reader_types_of_data_collected=['Optical density', 'formazan (absorbance at 590 nm)'], od_protocol='Not provided', respiration_protocol='Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', instrument_used=['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], analysis_software=['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], incubation_temperature='25°C', incubation_duration='48 hours')]} +INFO:root:Output format: yaml +--- +input_text: |+ + ORIGINAL RESEARCH + published: 14 April 2021 + doi: 10.3389/fmicb.2021.651282 + Edited by: + Barbara Pivato, + Institut National de Recherche pour + l’agriculture, l’alimentation et + l’environnement (INRAE), France + Reviewed by: + Cara Helene Haney, + The University of British Columbia, + Canada + Xingang Zhou, + Northeast Agricultural University, + China + *Correspondence: + Dmitri V. Mavrodi + dmitri.mavrodi@usm.edu + Alex S. Flynt + alex.flynt@usm.edu + †These authors have contributed + equally to this work + Specialty section: + This article was submitted to + Microbe and Virus Interactions with + Plants, + a section of the journal + Frontiers in Microbiology + Received: 09 January 2021 + Accepted: 08 March 2021 + Published: 14 April 2021 + Citation: + Mavrodi OV, McWilliams JR, + Peter JO, Berim A, Hassan KA, + Elbourne LDH, LeTourneau MK, + Gang DR, Paulsen IT, Weller DM, + Thomashow LS, Flynt AS and + Mavrodi DV (2021) Root Exudates + Alter the Expression of Diverse + Metabolic, Transport, Regulatory, + and Stress Response Genes + in Rhizosphere Pseudomonas. + Front. Microbiol. 12:651282. + doi: 10.3389/fmicb.2021.651282 + Root Exudates Alter the Expression + of Diverse Metabolic, Transport, + Regulatory, and Stress Response + Genes in Rhizosphere Pseudomonas + Olga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2, + Karl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2, + Ian T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and + Dmitri V. Mavrodi1* + 1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS, + United States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School + of Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular + Sciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics + and Quality Research Unit, Pullman, WA, United States + Plants live in association with microorganisms that positively influence plant + development, vigor, and fitness in response to pathogens and abiotic stressors. + The bulk of the plant microbiome is concentrated belowground at the plant rootsoil + interface. Plant roots secrete carbon-rich rhizodeposits containing primary and + secondary low molecular weight metabolites, lysates, and mucilages. These exudates + provide nutrients for soil microorganisms and modulate their affinity to host plants, but + molecular details of this process are largely unresolved. We addressed this gap by + focusing on the molecular dialog between eight well-characterized beneficial strains + of the Pseudomonas fluorescens group and Brachypodium distachyon, a model for + economically important food, feed, forage, and biomass crops of the grass family. We + collected and analyzed root exudates of B. distachyon and demonstrated the presence + of multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The + subsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many + of these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq + profiling of bacterial cultures amended with root exudates revealed changes in the + expression of genes encoding numerous catabolic and anabolic enzymes, transporters, + transcriptional regulators, stress response, and conserved hypothetical proteins. Almost + half of the differentially expressed genes mapped to the variable part of the strains’ + pangenome, reflecting the importance of the variable gene content in the adaptation of + P. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity + of cellular pathways and physiological responses underlying the establishment of + mutualistic interactions between these beneficial rhizobacteria and their plant hosts. + Keywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome + Frontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + INTRODUCTION + Plants are meta-organisms or holobionts that rely in part on + their microbiome for specific functions and traits. The ability + of the plant microbiome to influence plant development, vigor, + health, and fitness in response to abiotic stressors associated + with global climate change is documented by numerous studies + (Lugtenberg and Kamilova, 2009). There is mounting evidence + that plants actively recruit beneficial microbiomes, but many + aspects of this process are still very much a black box (Reinhold- + Hurek et al., 2015). The foundation for the differential affinity + of rhizobacteria toward host plants is built upon complex + chemical cross talk between microorganisms and plant roots. + Up to 40% of photosynthetically fixed carbon is released by + plant roots in the form of exudates and secretions, lysates, and + mucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990; + Badri and Vivanco, 2009). The release of these compounds is + actively controlled in response to environmental stimuli, and the + composition of root exudates varies greatly according to plant + species and physiological condition (Lynch, 1990; Nguyen, 2003; + Phillips et al., 2004; De-la-Pena et al., 2008). The presence and + composition of exudates strongly impact soil microorganisms, + which is consistent with the idea that plants actively select and + shape their root microbiota (Zolla et al., 2013). + Primary root exudates include simple and complex sugars, + amino acids, polypeptides and proteins, organic, aliphatic and + fatty acids, sterols, and phenolics (Nguyen, 2003; Badri and + Vivanco, 2009; Badri et al., 2009). These compounds serve as + carbon and energy sources for rhizobacteria, and the presence + of the intact corresponding catabolic pathways is essential + for competitive colonization of roots and disease suppression + (Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and + Kamilova, 2009). Root exudates also contain numerous signal + molecules and secondary metabolites, the significance of which + is only now emerging (Walker et al., 2003; Bais et al., 2005, + 2006). A handful of analyses of plant-induced gene expression + by transcriptional profiling in vitro (Mark et al., 2005) or in + the rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al., + 2005; Matilla et al., 2007; Barret et al., 2009) have identified + multiple genes that are differentially regulated by exposure to + roots or root exudates. Bacterial pathways expressed during + rhizosphere colonization control utilization of plant-derived + metabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001; + Lugtenberg and Kamilova, 2009), motility and chemotaxis (de + Weert et al., 2002; Lugtenberg and Kamilova, 2009), phase + variation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002; + van den Broek et al., 2005), outer membrane integrity (de Weert + et al., 2006; Lugtenberg and Kamilova, 2009), and the ability + to sequester limiting resources (Raaijmakers et al., 1995) and + resist environmental stresses (Sarniguet et al., 1995; Miller and + Wood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In + its spatial and temporal properties, root colonization resembles + biofilm formation, and biofilm-related pathways also have been + implicated in adhesion to seeds and roots and rhizosphere + colonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003; + Yousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al., + 2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally, + root exudates strongly affect the expression of diverse plant + growth promotion and biocontrol genes (Vacheron et al., 2013). + Over the past decade, the genomes of numerous rhizosphere + strains have been sequenced and analyzed, but functional + genomics studies of rhizosphere competence lag behind the + availability of sequence data. + This study explored the molecular dialog between the + model host plant Brachypodium distachyon and several wellcharacterized + rhizosphere strains of the Pseudomonas fluorescens + group. Brachypodium is a small annual grass originating in semiarid + regions of the Middle East that has emerged as a prime model + for economically important food, feed, forage, and biomass crops + of the grass family (Bevan et al., 2010; Schwartz et al., 2010; + Brkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The + biology, extensive collection of resources, and research tools + make B. distachyon an attractive model to investigate interactions + between plants and root-associated microbes. Pseudomonads + are ubiquitous Gram-negative g-proteobacteria that colonize + eukaryotic hosts and include both commensals and economically + important pathogens of plants and animals (Moore et al., + 2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus + Pseudomonas currently comprises > 100 named species that have + been separated based on multilocus sequence analysis into 14 + species groups (Garrido-Sanz et al., 2016; Hesse et al., 2018). + The P. fluorescens group is the most diverse regarding both + the genetic distances within it, the number of species and the + large pangenome that makes up > 50% of the pangenome + of the genus as a whole (Loper et al., 2012). The group also + encompasses an unusually high proportion of strains that inhabit + the plant rhizosphere and possess plant growth promoting and + biocontrol properties. Naylor et al. (2017) profiled bacterial + communities associated with root tissues and rhizosphere of + 18 different plant species of the Poaceae family. That study + identified Pseudomonas among taxa constituting the core grass + root microbiome and demonstrated that these bacteria were + enriched in C3 plants, including wheat, rye, barley, oat, and + Brachypodium. We confirmed the capacity of B. distachyon + Bd21 to serve as a host for rhizobacteria of the P. fluorescens + group in preliminary greenhouse assays with biocontrol strains + P. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens + Pf-5. Results of these experiments revealed that all strains + successfully established and colonized the roots of Brachypodium + (Supplementary Table 1). + In this study, we focused on eight well-studied strains + of the P. fluorescens complex that are supported by years + of studies, numerous refereed publications, and high-quality + genome sequences. By profiling transcriptomes of these strains + during growth in root exudates of B. distachyon, we revealed the + diversity of cellular pathways and physiological responses that + underlie the establishment of mutualistic interactions between + beneficial rhizobacteria and the host plant. Our results also + confirmed that root exudates contain carbohydrates, amino acids, + organic acids, and phenolics that serve as carbon and energy + sources for rhizobacteria. The root exudates also contained + osmoprotectants that may help microorganisms to persist in the + rhizosphere of drought-stressed plants. The diversity of microbial + genes perturbed by root exudates reflects the importance of + Frontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + the variable genome in adaptation of individual strains of + Pseudomonas to the rhizosphere lifestyle. + MATERIALS AND METHODS + Bacterial Strains Used in the Study + The eight Pseudomonas strains used for this study are + P. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens + SBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko + et al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller, + 1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996), + P. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens + Pf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and + Stipanovic, 1980). The selected organisms have been studied + extensively for their role in biological control and plant + growth promotion (Supplementary Table 2). The strains were + maintained in the laboratory as frozen stocks (􀀀80 C) and + routinely cultured in King’s medium B (King et al., 1954) or + 21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g + of Na2HPO4 2H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and + 20 ml of a microelement solution (Smibert and Kreig, 1994; + Halverson and Firestone, 2000). + Propagation of Plants and Collection of + Root Exudates + B. distachyon Bd21 was established from seed material obtained + from the USDA-ARS Plant Germplasm Introduction and Testing + Research Unit (Pullman, WA, United States). Brachypodium + seeds were imbibed for 3 days at 4 C and sown in 7   7 cm + pots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture, + Agawam, MA, United States). Plants were grown in an IR- + 89X (Percival Scientific, Perry, IA, United States) controlled + environment chamber retrofitted with 6500K and 3000K T5 54W + grow lights (Spectralux) under a 20-h light, 24 C/4-h dark, 18 C + cycle. Plants were watered and fertilized with Jack’s professional + water-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA, + United States). After 12 weeks and plant senescence, seeds + were collected, processed, and stored under desiccant and dark + conditions at room temperature. + To collect root exudates, seeds of B. distachyon Bd21 were + surface-sterilized, pregerminated, and placed in sterile 1 L widemouth + glass jars containing 113 g of 6-mm glass beads and + 25 ml distilled water. Jars were covered with vented caps + and plants were grown hydroponically in an environmental + controlled growth chamber under conditions described above. + After 6 days, root exudates were extracted from individual + jars and their sterility was confirmed by spotting on nutrient + agar. Multiple batches of root exudates were collected, filtered + (0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and + stored at –80 C. + Metabolomic Profiling of Root Exudates + Exudates were analyzed for primary metabolites at the Murdock + Metabolomics Laboratory at Washington State University + (Pullman, WA, United States). Freeze-dried residues were + suspended in 500 ml 50% aqueous acetonitrile and clarified by + centrifugation for 20 min at 21,000   g and 4 C. The liquid + chromatography mass spectrometry analysis was conducted + with a Synapt G2-S quadrupole-ion mobility spectrometrytime + of flight mass spectrometer system equipped with an + acquity ultra-performance liquid chromatograph (UPLC) and + an acquity photodiode array detector (all from Waters, Milford, + MA, United States). The exudate metabolites were separated + on a SeQuant ZIC-pHILIC HPLC column (2.1   100 mm, + 3 mm) (Milllipore Sigma, Burlington, MA, United States) using + acetonitrile with 0.1% formic acid as solvent B and water + with 0.1% formic acid as solvent A at a flow rate of 400 ml + min􀀀1 and the following linear gradient extending over 14 min: + 0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10% + B; 10 min, 80% B; and 14 min, 80% B. Mass spectra were + collected in positive ion mode over a range of m/z 50–1,200 + with a scan time of 0.2 s. The Q-TOF-MS source was at + 3.0 kV and 120 C; the sampling cone at 40 V, desolvation + temperature was 250 C; cone gas and desolvation gas flow were + at 0 and 850 L h􀀀1, respectively. Leucine enkephalin was used + for post-acquisition mass correction. Target compounds were + visualized using selected ion chromatograms at 0.05 Da window + width. The compound identification was based on comparison + of chromatographic behavior and accurate masses to those of + authentic standards. + For gas chromatography, derivatization was carried out using + a modification of the procedure of Lee and Fiehn (2008). + The freeze-dried residues were suspended in 950 ml aqueous + methanol (84%, v/v) and clarified by centrifugation for 15 min + at 21,000   g at 4 C. The supernatants were spiked with 1 + mg of the internal standard salicylic acid-d6 (C/D/N Isotopes, + Quebec, Canada) and dried in vacuo. The dry residues were + suspended in 10 ml of O-methoxylamine hydrochloride (30 mg + ml􀀀1 in anhydrous pyridine, both from Millipore Sigma) and + incubated while mixing (1,000 RPM) for 90 min at 30 C. + Subsequently, samples were derivatized with 90 ml of MSTFA + with 1% TMCS (Thermo Fisher Scientific, Waltham, MA, + United States) for 30 min at 37 C. Gas chromatography-mass + spectroscopy analysis was performed using a Pegasus 4D timeof- + flight mass spectrometer (LECO, Saint Joseph MI) equipped + with a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A + oven (Agilent Technologies, Santa Clara, CA, United States). + The derivatization products were separated on a 30-m, 0.25 mm + i.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA, + United States) with an IntegraGuard precolumn using ultrapure + He at a constant flow of 0.9 ml min􀀀1 as carrier gas. The linear + thermal gradient started with a 1-min hold at 70 C, followed + by a ramp to 300 C at 10 C min􀀀1. The final temperature was + held for 5 min prior to returning to initial conditions. Mass + spectra were collected at 17 spectra s􀀀1. Peak identification was + conducted using the Fiehn primary metabolite library (Kind + et al., 2009) and an identity score cutoff of 700. Additionally, + authentic standards for a number of primary metabolites were + analyzed under identical conditions and the data used to compare + the chromatographic behavior. Peak alignment and spectrum + comparisons were carried out using the Statistical Compare + feature of ChromaTOF software (LECO). + Frontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + Isolation of RNA From Bacteria Cultured + in Root Exudates and RNA-Seq + The strains were pregrown overnight at 25 C on 21C-glucose + agar and then subcultured into 96-well microplates containing + liquid 21C-glucose medium amended with Brachypodium + exudates. The liquid medium was prepared by dissolving the + lyophilized root exudate material in an appropriate volume of + 21C-glucose medium to concentrate root exudates 20-fold. The + growth medium was sterilized by passing it through a 0.22- + mm membrane filter. The control cultures were grown under + identical conditions in the absence of exudates. All treatments + were inoculated at OD600 of 0.1 and incubated for 20 to + 22 h until cultures entered late-exponential growth phase at + 25 C in an atmosphere of 15% oxygen [created by a ProOx + P110 oxygen controller (BioSpherix, Parish, NY, United States) + with a hypoxia C-chamber]. The cells were stabilized by the + addition RNAprotect reagent (QIAGEN, Germantown, MD, + United States) and total RNA was purified using a RNeasy Protect + Bacteria Mini Kit (QIAGEN) from three biological replicates of + each strain cultured under control conditions and in exudates. + The quality assessment of the extracted RNA samples was + performed with a NanoDrop OneC Spectrophotometer (Thermo + Fisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies) + and revealed A260/A280 and A260/A230 values of > 2.0 and a mean + RNA integrity numbers (RIN) value of > 9.2. + Three biological replicates of RNA samples were shipped on + dry ice to the DOE Joint Genome Institute (Walnut Creek, CA, + United States), where rRNA was depleted and stranded RNASeq + libraries were prepared, quantified by qPCR and sequenced + using a HiSeq 2500 instrument (Illumina). The fastq file reads + were filtered and processed with BBDuk1 to remove reads that + contained 1 or more “N” bases, had an average quality score + across the read less than 10 or had a minimum length < 51 bp + or 33% of the full read length. Reads mapped with BBMap (see + text footnote 2) to masked human, cat, dog, and mouse references + at 93% identity were removed. Another category of removed + sequences matched RNA spike-in, PhiX, common microbial + contaminants, and ribosomal RNAs. The processed reads from + each library were aligned to the reference genome using BBMap + with only unique mappings allowed (BAMs/directory). If a + read mapped to more than one location it was ignored. + featureCounts (Liao et al., 2014) was used to generate raw + gene counts, which were normalized to adjust for the length of + each gene and total number of reads mapped for each library. + The normalization formula used: n = [r/(l/1,000)]/(t/1,000,000), + where n = normalized read count for gene (G) for library (L); + r = raw read count for gene G for library L; l = gene G length; and + t = total reads mapped for library L. Raw gene counts were used to + evaluate the level of correlation between biological samples using + Pearson’s correlation. + Bioinformatic Analysis + Count tables generated by the JGI RNA-Seq pipeline were input + into DESeq2 (Love et al., 2014) to normalize and determine + 1https://sourceforge.net/projects/bbmap/ + differential expression. Statistical significance was established + through DESeq2 by using three biological replicates for control + and root exudate conditions. Scatterplots were generated from + the DESeq2 data table outputs using ggplot2. Genes differentially + expressed between control and root exudate samples (log2 foldchanges + –2   to   2, adjusted p value   0.05) were used in + downstream analysis. The core genome and pangenome for the + Pseudomonas strains used in this study were computed using the + OthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic + Pangenome Accumulation v1.4.0 apps implemented in the U.S. + Department of Energy Systems Biology Knowledgebase (KBase) + (Arkin et al., 2018). Additional comparisons were conducted + with the PGAweb pangenome analysis pipeline (Chen et al., + 2018). Differentially expressed genes were assigned to core, + non-core, and singleton parts of each strain’s proteome by + BLASTp with an E value cutoff of e-06, identity of 40% + and coverage of 60%. Functional annotation of differentially + expressed genes was carried out with the Blast2GO (Conesa + and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., + 2018). Additional manual curation was performed using tools + implemented in the Integrated Microbial Genomes (IMG) + database (Markowitz et al., 2012), Pseudomonas Genome + Database (Winsor et al., 2009), Kyoto Encyclopedia of Genes and + Genomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3 + (Biomatters, Auckland, New Zealand). Metabolic functions + encoded by the differentially expressed genes were mapped + using iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses + were carried out by building multiple sequence alignments + with MAFFT v7.222 (Katoh and Standley, 2013) and inferring + neighbor-joining (NJ) phylogenies with Geneious Tree Builder. + The resultant phylogenetic trees were visualized with iTOL + (Letunic and Bork, 2016). Reproducibility of clades within the + inferred NJ trees was assessed by bootstrap resampling with + 1,000 replicates. + Characterization of Carbon Source + Utilization With Biolog Phenotype + Microarrays + The utilization of carbon sources was analyzed using Phenotype + MicroArrays (Biolog, Hayward, CA, United States) as follows. + The bacteria were cultured overnight on Luria-Bertani agar + at 25 C, after which cells were harvested and suspended in + inoculating fluid (IF-0). The transmittance of the suspension + was adjusted to 42% using a Biolog turbidimeter. The cell + suspension was mixed with IF􀀀0 containing Dye Mix A + (Biolog) to achieve a final transmittance of 85%. One hundred + microliter aliquots of the adjusted cell suspension were + inoculated into PM01 and PM02A plates, which were then + incubated in an OmniLog Phenotype MicroArray System + (Biolog) at 25 C for 48 h. The formation of formazan was + recorded at 15 min intervals, and data were analyzed using + OmniLog Parametric Analysis software v1.20.02 (Biolog). + Relative growth of the studied strains was normalized to + growth on D-glucose and visualized using Heatmapper + (Babicki et al., 2016). + Frontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + Data Availability + Sequences generated in this project were deposited + under NCBI BioProject accession numbers PRJNA439743 + through PRJNA439790. + RESULTS + Metabolomic Profiling of Root Exudates + of B. distachyon + Metabolomics analysis of lyophilized root exudates revealed + the presence of numerous plant metabolites, 86 of which were + identified by matching their spectra to the LECO/Fiehn + Metabolomics library (Supplementary Table 3). These + metabolites included (i) carbohydrates and their derivatives + (glucose, fructose, xylose, sucrose, trehalose, maltose, galactose, + and others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol, + galactinol, 2-deoxyerythritol, ribitol, threitol and + cellobitol); (iii) amino acids and derivatives (glutamine, + tyrosine, glutamic acid, asparagine, aspartic acid, valine, + phenylalanine, isoleucine, glycine, serine, proline, leucine, + tryptophan, cysteine, methionine, citrulline, and others); (iv) + organic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric, + fumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic, + malic, methylmalonic, nicotinic, quinic, succinic, threonic); + and (v) assorted metabolites including heterocyclic compounds, + phenolics, and biogenic amines, etc (3-hydroxypyridine, + maleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate, + 5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid, + and urea). Results of the analysis also revealed that root exudates + of B. distachyon contain hydroxyectoine and the quaternary + amine (QA) glycine betaine (Supplementary Figure 1). + Phylogenetic and Pangenome Analyses + of Pseudomonas Strains Used in the + Study + We used a set of phylogenetic markers suggested by Mulet et al. + (2010) to investigate the relatedness of the eight strains used in + this study to distinct lineages recognized within the P. fluorescens + species complex. The multilocus sequence analysis based on + the concatenated sequences of the housekeeping genes rrs (16S + rRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with + strains 2-79 and SBW25) as a member of the P. fluorescens + subgroup (Figure 1). The rest of the strains clustered closely with + four additional subgroups of the P. fluorescens complex, namely + P. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1), + P. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of + the eight rhizosphere Pseudomonas strains varied in size by 1.43 + megabase (ranging from 5.65 to 7.07 Mb) and contained between + 5,166 and 6,363 protein-coding genes (Figure 2A). The shared + gene content was characterized with OrthoMCL, which uses allagainst- + all BLASTp followed by the Markov Cluster algorithm to + identify protein groups shared between the compared genomes, + as well as groups representing species-specific gene expansion + families (Li et al., 2003). The pangenome analysis revealed a + core comprised of approximately 3,179 orthologs that were + shared among all strains and represented 50.0% to 61.5% + of each predicted proteome (Figures 2A,B). The non-core + pangenome contained genes shared by two or more (but not + all) strains and contained between 1,482 and 2,080 orthologs, + which corresponded to 28.7–36.3% of individual proteomes. The + rest of the predicted protein-coding genes were strain-specific + singletons that comprised 7.5% to 15.1% of the strain’s predicted + proteomes. In respect to divergence from the core genome, strain + Pf-5 was found to possess the highest proportion of unique genes + (n = 949) followed by 2-79 (n = 887). The entire pangenome of + the Pseudomonas strains encompassed over 12,000 homolog and + singleton gene families. + Further homolog family-based comparisons identified Q8r1- + 96 and R1-43-08 as the most distantly related strains, with + 3349 shared homologs (Supplementary Table 4A). Q8r1-96 and + Q2-87, which shared 4,489 homologs, were the most closely + related strains. The partitioning of homolog gene families into + the core, non-core, and singleton parts of the pangenome agreed + with phylogenetic relationships of the strains deduced from the + analysis of a selected subset of COGs (Clusters of Orthologous + Groups) (Figure 2C and Supplementary Tables 4B,C). The + COG-based phylogeny supported the multilocus sequence + analysis and revealed that the eight Pseudomonas strains form + three distinct clusters, the first of which contained 2-79, R1-30- + 84, and SBW25. The second cluster included Q8r1-96 and Q2-87, + whereas the third encompassed strains 30-84, Pf-5, and Pf0-1. + Correlating the Composition of Root + Exudates With Metabolic Profiles of + Pseudomonas Strains + We used the Phenotype MicroArray PM1 and PM2 plates to + profile the eight Pseudomonas strains for the utilization of + 190 different carbon sources. Results of the analysis identified + 90 compounds that supported growth and clustered by their + intensities of utilization into three distinct groups (Figure 3). + Group I was comprised of 30 highly metabolized carbon + sources, which included several amino acids and intermediates of + glycolysis, pyruvate metabolism, and citrate cycle. Approximately + half of these compounds were catabolized by all eight strains, + and included several organic acids (fumaric, citric, gluconic, + malic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro, + Ala, and g-aminobutyric acid), carbohydrates (glucose, mannose, + and mannitol), and the purine nucleoside inosine. Group II + was composed of 44 chemically diverse carbon sources that + were variably utilized by the strains. These compounds were + carbohydrates, organic acids, amino acids, phenolics, and polyols, + and included known compatible solutes and intermediates of + metabolism of pentoses, galactose, starch, and sucrose. Group + III encompassed the rest of the Phenotype MicroArray test + panel and contained compounds that were not catabolized + by the tested strains. Among several notable exceptions were + a-hydroxyglutamic acid- g-lactone, putrescine, and itaconic, + citramalic, and succinamic acids, which supported the growth + of strains 2-79, 30-84, Pf-5, and SBW25. We further matched + the carbon metabolic profiles of the Pseudomonas strains against + the list of plant-derived metabolites from the root exudates of + Frontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the + P. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA + gyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor + method. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles). + B. distachyon Bd21. Interestingly, many carbon sources from + the Phenotype MicroArray panel were also present in the root + exudates of B. distachyon Bd21, and some of these compounds + (glucose, mannose, galactose, fructose, g-aminobutyric acid, + aspartic acid, citric acid, malic acid, fumaric acid, quinic acid, + alanine, glutamine, and glutamic acid) were catabolized by all + strains used in this study, while others (e.g., xylose, trehalose, + m-inositol) were actively utilized only by certain organisms + (Figure 3). The comparison of catabolic profiles across the + eight studied Pseudomonas strains revealed the presence of three + distinct clusters. The first cluster contained strains Q8r1-96 and + Q2-87, which consumed very similar sets of carbon sources, + as well as strain Pf0-1. The second cluster was composed of + 2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster + was represented by a single strain, Pf-5. The overall similarity + of the catabolic profiles partially agreed with the separation + of the strains into different subgroups of the P. fluorescens + complex (see above). + Analysis of the RNA-seq Results + In order to understand the cellular responses of rhizosphere + Pseudomonas to plant exometabolites, we analyzed the + transcriptome changes in cultures grown in the presence of + root exudates. Under field conditions, rhizobacteria colonize + plant roots in the form of surface-attached microaerobic biofilms + (Hojberg et al., 1999). To mimic these conditions, the eight + Pseudomonas strains were grown statically at 72% air saturation + in 21C-glucose medium amended with root exudates and then + processed to extract total RNA (Supplementary Figure 2). + A total of 995 million raw sequencing reads were generated from + the RNA samples by using the Illumina HiSeq-2500 platform, + averaging 20.7 million reads per sample. The removal of lowquality + and rRNA sequences resulted in a total of 793 million + filtered reads that were mapped onto the eight Pseudomonas + genomes with a mean of 7.48 million mapped fragments per + genome. The differentially abundant transcripts were identified + by setting a p value of 0.05 (adjusted for multiple testing) and + Frontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all + eight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific + singletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of + the pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the + 25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any + bias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939 + non-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied + strains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining + ortholog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node + represents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The + analysis was conducted in KBase (Arkin et al., 2018). + the log2 fold-change (FC) threshold    2.0 (Figure 4 and + Supplementary Tables 5–12). When compared with the control + conditions, an average of 204 genes per strain were differentially + expressed in the presence of root exudates, with the highest + (n = 425) and lowest (n = 112) numbers observed, respectively, + in SBW25 and Q2-87 (Figure 4). Overall, more genes were + induced than repressed in response to exudates, but the actual + numbers in each category varied substantially depending on the + identity of the Pseudomonas strain. In most strains, the bulk of + the differentially expressed genes was almost equally distributed + between the core (mean, 48.2%) and non-core (mean, 45.8%) + parts of the genome, whereas the strain-specific singleton genes + constituted on average only 5.9% (Figure 4B). One notable + exception was observed in Q8r1-96, where all differentially + expressed genes belonged to the core (73.8%) and non-core + (26.2%) parts of the genome. Another notable pattern was + observed in R1-43-08, where the majority of genes affected by + the presence of root exudate fell into the non-core category + (56.3%). The highest proportion of differentially expressed + singletons (11.3 and 10.4%, respectively) was identified in + strains SBW25 and Pf-5. + We further explored how the identified differentially + expressed genes were distributed across genomes of the eight + studied rhizosphere strains. The pairwise BLASTp comparisons + identified 2-79 and SBW25 as two strains that shared the highest + number of genes (n = 101) induced or repressed in response + to root exudates (Table 1). The second pair of strains with + a significant number of similar differentially expressed genes + (n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1 + and 30-84, which shared 56 differentially expressed genes. These + patterns of shared genes were also observed when the results of + the pairwise BLASTp comparisons were converted into a binary + gene presence/absence matrix, which was then subjected to + cluster analysis using a UPGMA algorithm based on Sorensen’s + dissimilarity index or examined by non-metric multidimensional + scaling (NMDS) (Figure 5). + The differentially expressed Pseudomonas genes were + subjected to Blast2Go analysis and Gene Ontology (GO) + annotation (Figure 6). Metabolic process, catalytic activity, + and membrane were the most common annotation terms + across the three primary GO term categories (i.e., biological + process, molecular function, and cellular component). A total + of 1,694 GO terms was assigned to 805 upregulated genes, with + the majority of the GO terms related to molecular function + (682, 40.3%), followed by biological process (669, 39.5%), and + cellular component (343, 20.2%). In the 539 downregulated gene + category, 1,101 GO terms were assigned to biological process + (420, 38.1%), molecular function (417, 37.9%), and cellular + component (264, 24.0%). Within biological process, metabolic + process, cellular process, localization, response to stimulus, and + regulation were over-represented. Within molecular function, + the largest proportion was assigned to catalytic activity, binding, + and transporter activity categories. Within cellular component, + the majority were assigned to membrane, membrane part, cell, + and cell part categories. Across the eight strains, 37–42% of + differentially expressed genes had no Gene Ontology IDs and + encoded various conserved hypothetical proteins. + Functional Classification of Shared + Differentially Expressed Genes + The interrogation of RNA-seq data revealed multiple cellular + pathways that were differentially regulated in bacterial cultures + incubated with root exudates (Supplementary Figures 3, 4). + Frontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere + Pseudomonas strains used in the study. The hierarchical clustering analysis + was carried out using the average linkage method with Euclidean distances. + Carbon sources identified by red arrowheads were also detected in the sterile + root exudates of B. distachyon Bd21. + Although none of these differentially regulated pathways were + shared by all eight strains, the cross-strain comparisons revealed + several types of common and specific transcriptomic responses + that were elicited by the presence of plant exometabolites + (Table 2). The visual representation of core gene expression + patterns is provided in Supplementary Figure 5, which + shows heatmaps of expression profiles and p-adj values for + core genes shared by the studied strains. The figure is + accompanied by Supplementary Table 13 that lists predicted + functions of genes constituting the four distinct clusters + observed after hierarchical clustering of gene expression + values. The first category of shared differentially expressed + pathways functioned in the uptake and catabolism of selected + carbohydrates, quaternary ammonium compounds (QAs), and + phenolics. All strains except for R1-43-08, responded to root + exudates by inducing the fructose-specific phosphoenolpyruvate + (PEP)-dependent phosphotransferase system (PTSFru). The + components of this system are encoded by a conserved + operon and include the cytoplasmic polyprotein EI/HPr/EIIAFru + (FruB), the 1-phosphofructokinase FruK, and the fructosespecific + permease EIIBC (FruA) (Chavarria et al., 2016). The + PTSFru system functions by acquiring high-energy phosphates + from PEP and sequentially passing them, via the EI/HPr/EIIAFru + domains of FruB, to the EIIB component of FruA. The + phosphates are ultimately transferred by the EIIC transporter + to fructose yielding fructose 1-phosphate, which is channeled + into the central metabolic pathways through the action of the + phosphofructokinase FruK. + In all strains except for Q8r1-96 and Pf-5, the exposure to root + exudates resulted in the induction of two genes adjacent to the + fru cluster that encoded a Major Facilitator Superfamily (MFS) + transporter and an L-arabinonate dehydratase (Table 2). These + genes are predicted to participate in the uptake and catabolism + of L-arabinose, where L-arabinonate dehydratase plays an + important role by converting L-arabinonate to 2-dehydro-3- + deoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1- + 43-08, and Q2-87, we also observed the induction of genes + encoding components of the AraFGH complex, an ATP-Binding + Cassette (ABC) superfamily transporter involved in the import of + arabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all + strains except SBW25 and R1-43-08 responded to the presence + of exudates by upregulating a conserved gene encoding an + aldose epimerase superfamily protein. Such enzymes equilibrate + alpha- and beta-anomers of aldoses and ensure that stereospecific + enzymes involved in the metabolism of free sugars do not act as + metabolic bottlenecks (Abayakoon et al., 2018). Although some + aldose epimerases have been linked to specific pathways, the + Pseudomonas gene identified in this study could not be assigned + to a particular metabolic process based on sequence analysis and + genomic location. + Several Pseudomonas strains responded to the presence + of root exudates by upregulating genes involved in the + uptake and catabolism of myo-inositol and possibly other + stereoisomers of inositol (Table 2). The upregulated catabolic + genes encode the dehydrogenase IolG, which oxidizes myoinositol + to its corresponding ketone, as well as IolE, IolD, IolB, + and IolC that collectively convert the 2-keto-myo-inositol to + acetyl-CoA and the glycolysis intermediate dihydroxyacetone + phosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1- + 43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions + Frontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core + and non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established + threshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual + Pseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates. + also involved components of the putative inositol-specific + ABC transporter. The cross-genome comparisons revealed + that in all studied strains except for Pf0-1, components + of the myo-inositol utilization pathway were encoded + within a well-conserved gene cluster which, in addition to + catabolic and transport functions, also encodes a dedicated + transcriptional repressor. + All studied strains of Pseudomonas carry multiple genes + involved in scavenging the quaternary ammonium compounds + choline, glycine betaine (GB), carnitine, choline-O-sulfate, and + Frontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + TABLE 1 | The number of differentially expressed genes shared among the eight + studied strains of rhizosphere Pseudomonas. + Strain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5 + 2–79 260 + SBW25 101 425 + R1-43-08 30 25 151 + Q8r1-96 32 39 21 145 + Q2-87 27 28 25 31 112 + 30-84 27 23 24 32 28 136 + Pf0-1 38 50 29 29 50 56 205 + Pf-5 36 41 52 86 29 55 40 230 + The pairwise comparisons were conducted by BLASTp with the following cutoff + parameters: E-value < 1e-06, minimum percent identity > 40%, and minimum + percent coverage > 65%. The black diagonal cells show the number of differentially + expressed genes per strain. In other words, these are self comparison values. + sarcosine from the environment. Many of these genes were + differentially expressed, including those encoding parts of the + ABC transporter CbcXWV, which is predicted to function in + the uptake of choline under water-replete conditions (Table 2). + Among enzymes induced in the presence of root exudates were + the choline dehydrogenase BetA, which converts choline to + glycine betaine and a network of enzymes (i.e., the Rieske family + oxygenase GbcAB, the dimethyglycine demethylase DgcAB, and + the sarcosine oxidase SoxBDAG) that sequentially convert GB + to glycine. In 2-79 and SBW25, this group of differentially + regulated genes also included an AraC-family transcriptional + activator GbdR, which perceives intercellular levels of GB and + induces genes involved in the transport and catabolism of + glycine betaine and detoxification of the catabolic byproducts + (Hampel et al., 2014). + The last category of activated catabolic pathways included the + catechol branch of the b-ketoadipate pathway for the degradation + of aromatic compounds. In strains 30-84, Pf0-1, and Pf-5, + growth on root exudates resulted in upregulation of catechol- + 1,2-dioxygenase, muconate cycloisomerase, and muconolactone + isomerase, which collectively cleave the catechol ring and + convert it to b-ketoadipate enol-lactone (Harwood and Parales, + 1996). Finally, analysis of the P. synxantha 2-79 transcriptome + identified an induction of benABC genes encoding subunits + of benzoate 1,2-dioxygenase, an oxidoreductase that generates + catechol from benzoate. + In addition to various catabolic pathways, the exposure to root + exudates also induced several genes involved in the homeostasis + of copper (Table 2). Four of these genes form a conserved cluster + in genomes of the strains and encode the periplasmic coppersensing + two-component system CinRS, the plastocyanin/azurinlike + protein CinA, and the NADPH-dependent pre-Q0 reductase + CinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed + upregulation of a conserved operon encoding the multicopper + oxidase CopA, the periplasmic copper-binding protein CopC, the + inner membrane protein CopD, and outer membrane protein + CopB. In several Gram-negative bacteria, these Cop proteins are + thought to have dual functions and participate both in the uptake + of essential copper as well as in the sequestration of excess copper + in the periplasm and outer membrane. + The analysis of shared downregulated pathways revealed that + most of the strains respond to the presence of root exudates by + repressing genes involved in the uptake and catabolism of sulfur + compounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96, + Q2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB + operon responsible for the utilization of alkanesulfonates + as sulfur sources. The ssu operon is highly conserved in + fluorescent pseudomonads and encodes the FMNH2-dependent + monooxygenase SsuD and the NAD (P)H-dependent FMN + reductase SsuE, which together catalyze the desulfonation of + alkanesulfonates. Also, the ssu locus contains genes for the + molybdopterin-binding protein SsuF and the alkanesulfonatespecific + ABC-type transporter consisting of the sulfonate + substrate-binding protein SsuA, sulfonate permease protein + SsuC, and sulfonate transport ATP-binding protein SsuB. Finally, + in R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates + coincided with repression of the tauABCD operon, which + allows these strains to utilize taurine (2-aminoethanesulfonate) + as a sulfur source. The repressed tau genes encoded the + 2-oxoglutarate-dependent taurine dioxygenase TauD and + substrate-binding, ATP-binding, and permease components of + the taurine-specific ABC transporter TauABC. + Other Differentially Expressed Pathways + In addition to their effect on several shared cellular pathways, + growth on root exudates resulted in the induction or repression of + numerous strain-specific genes. In closely related P. synxantha 2- + 79 and P. fluorescens SBW25, we observed differential expression + of genes involved in energy metabolism, transport of amino + acids, and surface attachment (Supplementary Tables 5, 6). + Other notable differentially expressed pathways included 2– + 79 gene clusters that encode enzymes for the catabolism of + trehalose, a prophage, and toxin/antitoxin system, as well as the + SBW25 operon predicted to control the synthesis of the capsular + exopolysaccharide colonic acid. The response of Pseudomonas sp. + R1-43-08 to root exudates also involved differential expression + of different energy metabolism pathways. In addition, we + observed the upregulation of genes involved in the uptake and + catabolism of xylose (also upregulated in 2–79) and repression + of enzymes for the biosynthesis of phenazine-1-carboxylic acid + and assimilation of inorganic sulfur and L-cysteine biosynthesis + (Supplementary Table 7). + The analysis of the Q8r1-96 transcriptome revealed + perturbation of different metabolic pathways including genes + encoding components of cytochrome C oxidase, transport + and catabolism of sorbitol/mannitol, metabolism of butanoic + acid, and biosynthesis of exopolysaccharides alginate and + poly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In + P. fluorescens Q2-87, we identified differential expression of + genes involved in metabolism of galactose, tryptophan, tyrosine, + glycine, serine, and threonine (Supplementary Table 9), + while in P. chlororaphis 30-84, growth on exudates activated + the biosynthesis of molybdopterin cofactor, catabolism of + galactonate and acetoin, and uptake and catabolism of putrescine + (Supplementary Table 10). The response of P. protegens Pf-5 to + root exudates involved upregulation of acetoin dehydrogenase, + which converts acetoin to acetaldehyde and acetyl-CoA, as + Frontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root + exudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis. + FIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of + B. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa + and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO + term. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO + terms. + well as pathways for the utilization of glycolate and putrescine + (Supplementary Table 11). Also induced were genes for + the production of pyrrolnitrin and PhlG hydrolase, which + modulate the metabolic loads attributed to the synthesis of + 2,4-diacetylphloroglucinol. The differentially expressed genes of + P. fluorescens Pf0-1 included, among others, operons encoding + cytochrome C oxidase and enzymes for catabolism of malonic + acid (Supplementary Table 12). Yet another interesting + finding involved the induction of assorted genes acting in the + homeostasis of iron and defense against reactive oxygen species + (ROS). We observed activation of iron dicitrate transporters + (SBW25 and 30-84), genes for the biosynthesis of siderophores + ornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading + enzymes (2–79, 30–84), TonB siderophore receptors, and + components of the energy-transducing inner membrane + complex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially + expressed ROS defense pathways were represented by different + catalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and + Pf-5 and organic hydroperoxide resistance proteins in strains + SBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and + Pf0-1, the addition of exudates resulted in the upregulation of + peroxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic + and aromatic hydroperoxides. + DISCUSSION + Our analysis of B. distachyon root exudates revealed a complex + mix of primary and secondary metabolites, thus supporting + the view of the plant rhizosphere as a carbon-rich niche for + Frontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + TABLE 2 | The distribution and predicted functions of selected differentially expressed genesa. + Predicted function Strainb + 2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 + Uptake and catabolism of fructose + D-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8) + 1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7) + D-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2) + Uptake and catabolism of arabinose + MFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9) + L-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5) + Interconversion of alpha- and beta-anomers of aldoses + Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) + Uptake and catabolism of quaternary ammonium compounds + Choline dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1) + Transcriptional regulator GbdR 2756597125 (􀀀3.7) 649639087 (􀀀3.9) + Membrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3) + Hypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3) + Dimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2) + Dimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0) + Betaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7) + Betaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2) + Ser hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1) + Sarcosine oxidase, g subunit, SoxG 2597878304 (2.9) + Sarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4) + Sarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2) + Sarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3) + Betaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5) + ABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5) + Uptake and catabolism of myo-inositol + 5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2) + 2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1) + 5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3) + 2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4) + 3D-(3,5/4)-trihydroxycyclohexane-1,2-dione + acylhydrolase, IolD + 2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3) + Myo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9) + Inositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8) + Inositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5) + Inositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6) + (Continued) + Frontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + TABLE 2 | Continued + Predicted function Strainb + 2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 + Uptake and catabolism of fructose + Catabolism of phenolics + Muconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8) + Muconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4) + Catechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5) + AraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2) + Benzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8) + Benzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8) + Benzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3) + Copper homeostasis + pre-Q0 reductase/7-cyano-7-deazaguanine + reductase CinQ + 2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5) + Cupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4) + Heavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5) + Heavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0) + Copper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4) + Copper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0) + Copper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4) + Copper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2) + Conserved hypothetical protein + Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) + Uptake and catabolism of sulfonates + FMN-dependent monooxygenase SsuE 2756592254 (􀀀1.9) 637745334 (􀀀3.5) + Sulfonate substrate-binding protein SsuA 649639261 (􀀀2.0) 2756592253 (􀀀2.2) 2597878518 (􀀀2.6) 2597849636 (􀀀3.9) 637745333 (􀀀7.0) 637323272 (􀀀2.7) + FMN-dependent monooxygenase SsuD 649639260 (􀀀1.9) 2756592252 (􀀀2.3) 2597849637 (􀀀2.8) 637745332 (􀀀5.8) 637323271 (􀀀2.3) + Sulfonate permease protein SsuC 649639259 (􀀀2.4) 2756592251 (􀀀2.1) 2597878516 (􀀀1.9) 2597849638 (􀀀3.2) 637745331 (􀀀5.8) 637323270 (􀀀2.5) + Sulfonate transport ATP-binding protein SsuB 649639258 (􀀀2.5) 2756592250 (􀀀2.3) 2597849639 (􀀀2.5) 637745330 (􀀀5.6) 637323269 (􀀀2.7) + Molybdopterin binding protein SsuF 649639257 (􀀀3.1) 2756592249 (􀀀2.7) 2597878514 (􀀀2.9) 2597849641 (􀀀3.3) 637745339 (􀀀6.2) 637323268 (􀀀2.7) + Uptake and catabolism of taurine + Taurine substrate-binding protein TauA 2597854917 (􀀀2.0) 637740095 (􀀀4.1) + Taurine transport ATP-binding protein TauB 2756592398 (􀀀1.9) 637740094 (􀀀3.7) 637317614 (􀀀1.9) + Taurine permease protein TauC 2756592398 (􀀀1.9) 637740093 (􀀀3.9) 637317613 (􀀀2.0) + Taurine dioxygenase TauD 2597854920 (􀀀1.8) 637740092 (􀀀3.6) + aThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%. + bValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets). + Frontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + soil microorganisms. Our results were in agreement with a + recent report of 27 different sugars, amino acids, and organic + acids in Brachypodium exudates (Kawasaki et al., 2016). We + confirmed the presence of exometabolites identified in that + study, along with dozens of additional analytes that were + identified by matching their mass-spectra and retention indices + to the LECO/Fiehn Metabolomics library (Supplementary + Table 3). The complementation of the metabolomic analysis + with profiling of the bacteria by Biolog Phenotype MicroArrays + revealed that a substantial proportion of the characterized + exudate constituents were catabolized by a collection of eight + Pseudomonas strains from across the P. fluorescens group that is + known to form associations with plant roots. The amendment + of Pseudomonas cultures with root exudates caused changes in + the expression of multiple genes encoding catabolic and anabolic + enzymes, predicted transporters, transcriptional regulators, stress + response, and conserved hypothetical proteins. In most strains, + these differentially expressed genes were almost equally split + between the core and variable genome regions, mirroring + the substantial strain-to-strain variation in the genome size + and gene content within the P. fluorescens species complex + (Loper et al., 2012). + The analysis of transcriptome responses to root exudates + revealed several types of cellular pathways present in the strains + used in this study. The first category of such pathways was + involved in the catabolism of carbohydrates such as fructose, + arabinose, myo-inositol, xylose, trehalose, and galactose. Among + these catabolic traits, the ability to utilize fructose as a carbon + source is highly conserved among fluorescent pseudomonads. + In contrast, growth on arabinose, myo-inositol, xylose, and + trehalose is variably present and was traditionally used to + differentiate species and biovars within the P. fluorescens + group (Barrett et al., 1986). We speculate that such variably + distributed pathways contribute to the differential affinity of + pseudomonads toward host plants and/or to determine which + strains flourish in response to growing roots and changing + environments. Several independent studies have confirmed + the importance of carbohydrate catabolism pathways for the + biology of rhizosphere pseudomonads. For example, in vivo + expression technology (IVET) profiling of P. fluorescens SBW25 + identified xylose isomerase among genome regions essential + for the colonization of sugar beet seedlings (Liu et al., 2015), + whereas a genome-wide Tn-Seq screen of Pseudomonas simiae + identified genes for the catabolism of myo-inositol among traits + essential for the colonization of Arabidopsis thaliana roots + (Cole et al., 2017). + The response of rhizosphere Pseudomonas to Brachypodium + root exudates also involved pathways for the uptake and + metabolism of amino acids. We observed differential + expression of genes encoding the hydrophobic (HAAT) and + polar (PAAT) amino acid uptake transporters in strains + 2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes + encoded enzymes for the catabolism of valine and glutamic + acid (2-79); metabolism of tryptophan, glycine, serine, and + threonine (Q2-87); and biosynthesis of methionine (Q8r1- + 96). It is plausible that the abundance of amino acids in + root exudates is also linked to the repression of pathways + involved in the catabolism of sulfonates and taurine that + was observed in several strains (Table 2). Although the + preferred source of sulfur for P. fluorescens is unknown, in the + closely related P. aeruginosa, the sulfur starvation response + is triggered by the growth on any sulfur compound other + than sulfate, thiocyanate, and cysteine (Hummerjohann et al., + 1998). This fact, together with the presence of cysteine and + cystine in the root exudates, suggest that root exudates of + Brachypodium may serve as an important source of sulfur for + rhizosphere Pseudomonas. These findings also agree well with + the reported scarcity of inorganic sulfate in the soil, and the + presence of sulfur mostly in the form of organic compounds, + including amino acids, proteins, sulfate esters, and sulfonates + (Autry and Fitzgerald, 1990). + Another interesting result of this study was the concerted + activation of copper and iron homeostasis pathways observed + in all of the Pseudomonas strains used in this work. In bacteria, + an excess of copper is toxic and triggers oxidative stress due to + the formation of free radicals, as well as disruption of protein + metalation and stability of iron-sulfur clusters (Bondarczuk + and Piotrowska-Seget, 2013). On the other hand, copper is an + essential trace element used as a cofactor in different enzymes. + Similarly, although elevated levels of iron cause redox stress, this + element is also found in active energy metabolism enzymes and is + crucial for bacterial growth (Andrews et al., 2003). The analysis + of metal homeostasis genes identified in this study suggests + that their induction was likely triggered by the deficiency of + copper and iron in bacterial cultures grown in the presence of + root exudates. We attribute this effect to the ability of some + components of root exudates to chelate soil metals. + Despite the abundance of iron in the soil, its bioavailability + is limited due to the low solubility of Fe (III) oxyhydrates + at neutral pH. The non-graminaceous plants circumvent this + problem by acidifying the rhizosphere and secreting flavins, + phenolics, and organic acids that chelate iron. The reduction + of these ferric chelates releases soluble ferrous iron taken up + by root cells (Kobayashi and Nishizawa, 2012). Graminaceous + plants, like Brachypodium, acquire iron by secreting into the + soil non-protein amino acids of the mugineic acid (MA) group, + which act as Fe (III)-chelating phytosiderophores. In addition to + iron, low-molecular-weight organic acids and phytosiderophores + bind other divalent and trivalent metals (including copper) and + contribute to heavy-metal tolerance in plants (Chen et al., 2017). + It is plausible that the presence of these plant exometabolites + is responsible for the deficit of iron and copper observed in + Pseudomonas cultures grown in the presence of root exudates. + These results further underscore the importance of diverse and + redundant metal homeostasis pathways found in genomes of the + P. fluorescens group for the ability of these organisms to colonize + and persist in the plant rhizosphere. + Recently, Klonowska et al. (2018) examined transcriptomic + responses of symbiotic nitrogen-fixing bacteria to root + exudates of the legume plant Mimosa pundica, which has + an unusual ability to support both alpha- (Rhizobium) and + beta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq, + the authors characterized genes involved in the perception of + root exudates in the nodulating bacteria Burkholderia phymatum + Frontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + STM815, Cupriavidus taiwanensis LMG19424, and Rhizobium + mesoamericanum STM3625. Interestingly, the analysis of + differentially expressed genes revealed induction of pathways + involved in the catabolism of fructose, xylose, myo-inositol, + and protocatechuate/catechol. Also upregulated were some + copper homeostasis, siderophore biosynthesis, and oxidative + stress genes. Finally, the analytical profiling of M. pundica + exudates revealed an overlap with Brachypodium in the types of + carbohydrates, amino acids, and organic acids present. These + findings suggest that differentially expressed genes shared by + multiple strains of the group P. fluorescens are not unique to + the Brachypodium-Pseudomonas system but represent a set + of conserved cellular pathways involved in the perception of + plant exometabolites by different clades of rhizosphere-dwelling + Proteobacteria. + Most strains included in this study were originally selected + based on the ability to colonize the rhizosphere and produce + secondary metabolites that alleviate the plant stress response + and/or inhibit soilborne pathogens. It has been suggested + that plant metabolites released into the rhizosphere affect + the biocontrol activity of plant-beneficial pseudomonads (de + Werra et al., 2011). We provide further support to this + hypothesis by demonstrating that in some strains, root exudates + modulate the expression of genes for the catabolism of + the plant growth-promoting metabolites acetoin and 2,3- + butanediol. The exposure to exudates also affected the expression + of genes for the synthesis of well-characterized antifungal + compounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4- + diacetylphloroglucinol. The modulatory effects were strainspecific, + suggesting significant differences in the regulatory + networks involved in the perception of plant signals and + regulation of the production of antibiotics and growthpromoting + metabolites. + The final significant finding of this study was the induction + of catabolism of quaternary amines (QAs) observed in multiple + strains of the P. fluorescens group during growth on root + exudates. This observation was supported by the detection of + glycine betaine in the root secretions of B. distachyon. The + presence of QAs in plant tissues and the capacity of these + metabolites to provide stress protection and nutrients to plant + pathogens and symbionts were reported before (Boncompagni + et al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our + study is among the first to highlight the potential importance of + these metabolites for rhizosphere interactions. Pseudomonads do + not synthesize QAs de novo but have evolved many pathways to + scavenge them from eukaryotic hosts, where these metabolites + are abundant due to the prominence of phosphatidylcholine + in cellular membranes. Strains of P. fluorescens carry genes + for the conversion of choline, carnitine, and glycine betaine to + glycine, as well as quaternary amine transporters of the BCCT + and ABC families that are also conserved in the opportunistic + human pathogen P. aeruginosa and the plant pathogen P. syringae + (Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b). + In P. aeruginosa, choline catabolism genes are essential for + the ability of this pathogen to persist during lung infection + (Wargo, 2013a). Similarly, a P. syringae mutant deficient in + BetT, OpuC, and CbcXWV quaternary amine transporters had + reduced fitness during colonization of bean and soybean leaves + under greenhouse and field conditions (Chen et al., 2013). + Depending on water availability, P. aeruginosa and P. syringae + catabolize exogenously supplied QAs as carbon and nitrogen + sources or accumulate them as osmoprotectants (Chen et al., + 2013; Wargo, 2013b). Our ongoing work in P. synxantha 2– + 79 unraveled similar physiological responses and demonstrated + that QA transporters function differentially and redundantly + in the uptake of quaternary amines as nutrients (Pablo and + Mavrodi, unpublished). In contrast, under water stress, the QAs + choline, betaine, and carnitine are accumulated preferentially for + osmoprotection. Under drought stress, a 2–79 mutant devoid + of all known QA transporters was less competitive in the + colonization of the Brachypodium rhizosphere than its wild-type + parental strain. Interestingly, our metabolomic profiling of root + exudates also revealed proline, glutamine, and hydroxyectoine. + These metabolites act as compatible solutes in different groups + of microorganisms (Yancey et al., 1982; Empadinhas and da + Costa, 2008), suggesting an important role of root exudates + in the ability of Pseudomonas to persist in the rhizosphere of + drought-stressed plants. + DATA AVAILABILITY STATEMENT + The original contributions presented in the study are publicly + available. This data can be found here: NCBI BioProject accession + numbers PRJNA439743 through PRJNA439790. + AUTHOR CONTRIBUTIONS + DM, OM, and LT conceived the research project. OM and + JM collected root exudates. OM and DM cultured strains + and extracted total RNA. AB and DG performed metabolomic + analysis of root exudates. DM, JP, and AF analyzed RNA-seq + data. LE, KH, and IP conducted Biolog analyses. DM, AF, OM, + DW, and LT wrote the manuscript. All authors contributed to the + manuscript revision. + FUNDING + This study was funded by NSF grant IOS-1656872 and by an + award from the DOE Joint Genome Institute’s Community + Science Program. The authors also acknowledge support from + Australian Research Council Discovery grant (DP160103746) + and Mississippi INBRE, funded by an Institutional Development + Award (IDeA) from the National Institute of General + Medical Sciences of the National Institutes of Health under + grant P20GM103476. + SUPPLEMENTARY MATERIAL + The Supplementary Material for this article can be found online + at: https://www.frontiersin.org/articles/10.3389/fmicb.2021. + 651282/full#supplementary-material + Frontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + REFERENCES + Abayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018). + Discovery and characterization of a sulfoquinovose mutarotase using kinetic + analysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383. + doi: 10.1042/bcj20170947 + Andrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron + homeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03) + 00055-x + Arkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov, + S., et al. (2018). KBase: the United States department of energy systems biology + knowledgebase. Nat. Biotechnol. 36, 566–569. + Autry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil + organic sulfur. Biol. Fertil. Soils 10, 50–56. + Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al. + (2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44, + W147–W153. + Badri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates. + Plant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x + Badri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere + chemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20, + 642–650. doi: 10.1016/j.copbio.2009.09.014 + Bais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005). + Mediation of pathogen resistance by exudation of antimicrobials from roots. + Nature 434, 217–221. doi: 10.1038/nature03356 + Bais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of + root exudates in rhizosphere interations with plants and other organisms. Annu. + Rev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159 + Bangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic + locus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the + biological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe + Interact. 9, 83–90. + Barret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G., + and Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic + fungus Gaeumannomyces graminis var. tritici on gene expression of the + biocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe + Interact. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611 + Barrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas + fluorescens biovar V: its resolution into distinct component groups and the + relationship of these groups to other P. fluorescens biovars, to P. putida, and to + psychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol. + 132, 2709–2721. doi: 10.1099/00221287-132-10-2709 + Bevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon + genomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21, + 211–217. doi: 10.1016/j.copbio.2010.03.006 + Boncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999). + Occurrence of choline and glycine betaine uptake and metabolism in the family + Rhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65, + 2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999 + Bondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper + resistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397– + 405. doi: 10.1007/s10565-013-9262-1 + Brkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al. + (2011). Brachypodium as a model for the grasses: today and the future. Plant + Physiol. 157, 3–13. + Camacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type + 4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens + WCS365. Dissertation, University of Leiden, Leiden. + Chavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A + metabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx + in Pseudomonas putida. mSystems 1:e00154-16. + Chen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread + plant-colonizing bacterial species Pseudomonas syringae detects and exploits an + extracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj. + 12262 + Chen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb: + a web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi: + 10.3389/fmicb.2018.01910 + Chen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal + acquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi. + 2017.06.004 + Cole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M., + et al. (2017). Genome-wide identification of bacterial plant colonization genes. + PLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860 + Conesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional + analysis in plant genomics. Int. J. Plant Genomics 2008:619832. + Curl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag. + Darzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive + pathways explorer v3. Nucleic Acids Res. 46, W510–W513. + de Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al. + (2006). The two-component colR/S system of Pseudomonas fluorescens WCS365 + plays a role in rhizosphere competence through maintaining the structure and + function of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi: + 10.1111/j.1574-6941.2006.00158.x + de Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg, + G. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is + an important trait for tomato root colonization by Pseudomonas fluorescens. + Mol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11. + 1173 + de Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand + microbe-derived compounds affect the expression of genes encoding + antifungal compounds in a pseudomonad with biocontrol activity. Appl. + Environ. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10 + Dekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998). + A site-specific recombinase is required for competitive root colonization by + Pseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056. + doi: 10.1073/pnas.95.12.7051 + De-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008). + Root-microbe communication through protein secretion. J. Biol. Chem. 283, + 25247–25255. doi: 10.1074/jbc.m801967200 + Empadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms + in prokaryotes: distribution of compatible solutes. Int. Microbiol. 11, + 151–161. + Espinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of + functions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182, + 2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000 + Fuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in + Pseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365- + 2958.2010.07250.x + Galvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline- + O-sulphate utilization from osmoprotection in Pseudomonas putida. Mol. + Microbiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x + Garrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R., + and Redondo-Nieto, M. (2016). Genomic and genetic diversity within the + Pseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal. + pone.0150183 + Halverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and + nonpermeating solutes on the fatty acid composition of Pseudomonas putida. + Appl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421. + 2000 + Hampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and + Wargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas + aeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13 + Harwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and + the biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/ + annurev.micro.50.1.553 + Hesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018). + Genome-based evolutionary history of Pseudomonas spp. Environ. Microbiol. + 20, 2142–2159. + Hinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003). + Transition from reversible to irreversible attachment during biofilm formation + by Pseudomonas fluorescens WCS365 requires an ABC transporter and a large + secreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003. + 03615.x + Hojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999). + Oxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring + Frontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + the distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65, + 4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999 + Hong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological + growth stages of Brachypodium distachyon: codification and description. Weed + Res. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x + Howell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum + induced damping-off of cotton seedlings by Pseudomonas fluorescens and + its antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto- + 70-712 + Hummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz, + M. A. (1998). Regulation of the sulfate starvation response in Pseudomonas + aeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375– + 1386. doi: 10.1099/00221287-144-5-1375 + Kabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam, + C. (2017). Physiological and biochemical responses involved in water deficit + tolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/ + journal.pone.0190284 + Kamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005). + Enrichment for enhanced competitive plant root tip colonizers selects for a new + class of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j. + 1462-2920.2005.00889.x + Kanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008). + KEGG for linking genomes to life and the environment. Nucleic Acids Res. 36, + D480–D484. + Katoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment + software version 7: improvements in performance and usability. Mol. Biol. Evol. + 30, 772–780. doi: 10.1093/molbev/mst010 + Kawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al. + (2016). Microbiome and exudates of the root and rhizosphere of Brachypodium + distachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal. + pone.0164533 + Kind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al. + (2009). FiehnLib: mass spectral and retention index libraries for metabolomics + based on quadrupole and time-of-flight gas chromatography/mass + spectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522 + King, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the + demonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307. + Klonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018). + Transcriptomic profiling of Burkholderia phymatum STM815, Cupriavidus + taiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response + to Mimosa pudica root exudates illuminates the molecular basis of their + nodulation competitiveness and symbiotic evolutionary history.BMC Genomics + 19:105. doi: 10.1186/s12864-018-4487-2 + Kobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and + regulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/ + annurev-arplant-042811-105522 + Kohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor + IolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193, + 5155–5163. doi: 10.1128/jb.05371-11 + Lee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas + reinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7 + Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for + the display and annotation of phylogenetic and other trees. Nucleic Acids Res. + 44, W242–W245. + Li, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of + ortholog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi: + 10.1101/gr.1224503 + Liao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient + general purpose program for assigning sequence reads to genomic features. + Bioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656 + Liu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose + utilization by Pseudomonas fluorescens: overlapping genetic responses to xylose, + xylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi. + 13142 + Loper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer, + B. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas + spp.: insights into diversity and inheritance of traits involved in multitrophic + interactions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784 + Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change + and dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550. + Lugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria. + Annu. Rev. Microbiol. 63, 541–556. + Lugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular + determinants of rhizosphere colonization by Pseudomonas. Annu. Rev. + Phytopathol. 39, 461–490. + Lynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch + (Chichester: JohnWiley & Sons), 177–206. + Mark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005). + Transcriptome profiling of bacterial responses to root exudates identifies genes + involved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454– + 17459. doi: 10.1073/pnas.0506407102 + Markowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y., + et al. (2012). IMG: the integrated microbial genomes database and comparative + analysis system. Nucleic Acids Res. 40, D115–D122. + Martinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF, + the second largest Pseudomonas putida protein, contributes to plant root + colonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561. + doi: 10.1111/j.1365-2958.2010.07249.x + Matilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and + Ramos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving + forces of bacterial life in the rhizosphere. Genome Biol. 8:R179. + Miller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria. + Annu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101 + Moore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.- + L., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes, + eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt + (New York, NY: Springer), 646–703. + Mulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis + of the Pseudomonas species. Environ. Microbiol. 12, 1513–1530. + Naylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought + and host selection influence bacterial community dynamics in the grass root + microbiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118 + Nguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and + controls. Agronomie 23, 375–396. doi: 10.1051/agro:2003011 + Nielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface + interactions mediated by cellulose and a novel exopolysaccharide contribute + to Pseudomonas putida biofilm formation and fitness under water-limiting + conditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011. + 02432.x + Parejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow, + L. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid + producing fluorescent Pseudomonas spp. from dryland cereal fields of central + Washington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012- + 0015-0 + Phillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R. + (2004). Microbial products trigger amino acid exudation from plant roots. Plant + Physiol. 136, 2887–2894. doi: 10.1104/pp.104.044222 + Raaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek, + P. J., and Schippers, B. (1995). Utilization of heterologous siderophores and + rhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41, + 126–135. doi: 10.1139/m95-017 + Raaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4- + diacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils. + Mol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144 + Rahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis, + J., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase + contains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/ + acschembio.7b00304 + Ramos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of + Pseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo + expression technology capture and identification of root-activated promoters. + J. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005 + Reinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T. + (2015). Roots shaping their microbiome: global hotspots for microbial activity. + Annu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10 + 2342 + Frontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + Sanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and + Rivilla, R. (2002). Phenotypic selection and phase variation occur during + alfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184, + 1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002 + Sarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995). + The sigma factor ss affects antibiotic production and biological control activity + of Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259. + doi: 10.1073/pnas.92.26.12255 + Schnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The + sigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance + towards desiccation and osmotic stress in the biocontrol agent Pseudomonas + fluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem. + 67.12.5683-5693.2001 + Schroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006). + “Phytopathogenic pseudomonads and related plant-associated pseudomonads,” + in The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.- + H. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740. + doi: 10.1007/0-387-30746-x_23 + Schwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and + Amasino, R. M. (2010). Natural variation of flowering time and vernalization + responsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10. + 1007/s12155-009-9069-3 + Silby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson, + R. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity + and plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51. + Silby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to + identify genes important in growth and survival of Pseudomonas fluorescens + Pf0-1 in soil: Discovery of expressed sequences with novel genetic organization. + J. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004 + Simons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A., + and Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere + colonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant + Microbe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600 + Simons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and + Lugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root + colonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe + Interact. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102 + Smibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods + for General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A. + Wood, and N. R. Kreig (Washington, DC: American Society of Microbiology), + 607–654. + Thomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from + Pseudomonas fluorescens in biological control of Gaeumannomyces graminis + var. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988 + Thomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990). + Production of the antibiotic phenazine-1-carboxylic acid by fluorescent + Pseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56, + 908–912. doi: 10.1128/aem.56.4.908-912.1990 + Tyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats, + W. G., et al. (2014). Selection and phenotypic characterization of a core + collection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25. + doi: 10.1186/1471-2229-14-25 + Vacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y., + Muller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system + functioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356 + van den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of + phenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol. + 7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x + van Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity + of microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135. + doi: 10.1128/.61.2.121-135.1997 + Walker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M. + (2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric. + Food Chem. 51, 2548–2554. + Wargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to + Pseudomonas aeruginosa survival during murine lung infection. PLoS One + 8:e56850. doi: 10.1371/journal.pone.0056850 + Wargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine: + lessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112– + 2120. + Whipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch + (Chichester: JohnWiley & Sons), 59–97. + Winsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock, + R. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly, + comprehensive comparisons of microbial genomes. Nucleic Acids Res. 37, + D483–D488. + Yahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The + Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. + Stackebrandt (New York, NY: Springer), 704–713. + Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982). + Living with water stress: evolution of osmolyte systems. Science 217, 1214–1222. + Ye, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web + tool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res. + 46, W71–W75. + Yoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H., + et al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283, + 10415–10424. + Yousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different, + overlapping mechanisms for colonization of abiotic and plant surfaces by + Pseudomonas putida. FEMS Microbiol. Lett. 288, 118–124. + Zboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere + colonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol. + J. 18, 3539–3554. + Zolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K., + et al. (2013). “Understanding root-microbiome interactions,” in Molecular + Microbial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John + Wiley & Sons), 745–754. + Conflict of Interest: The authors declare that the research was conducted in the + absence of any commercial or financial relationships that could be construed as a + potential conflict of interest. + Copyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne, + LeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an + open-access article distributed under the terms of the Creative Commons Attribution + License (CC BY). The use, distribution or reproduction in other forums is permitted, + provided the original author(s) and the copyright owner(s) are credited and that the + original publication in this journal is cited, in accordance with accepted academic + practice. No use, distribution or reproduction is permitted which does not comply + with these terms. + Frontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282 + +raw_completion_output: |- + study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas + authors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi + doi: 10.3389/fmicb.2021.651282 + date: 14 April 2021 + experiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. +prompt: |+ + From the text below, extract the following entities in the following format: + + label: + experiment_motivation: + experiment_design: + environment: + host: + host_type: + target_microbes: + biological_system: + conditions: + experimental_factors: + is_biolog_experiment: + type_of_biolog_experiment: + plates: + replicates: + protocol_steps: + plate_reader_types_of_data_collected: + od_protocol: + respiration_protocol: + instrument_used: + analysis_software: + incubation_temperature: + incubation_duration: + + + Text: + The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. + + === + +extracted_object: + study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, + Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas + authors: + - Olga V. Mavrodi + - Janiece R. McWilliams + - Jacob O. Peter + - Anna Berim + - Karl A. Hassan + - Liam D. H. Elbourne + - Melissa K. LeTourneau + - David R. Gang + - Ian T. Paulsen + - David M. Weller + - Linda S. Thomashow + - Alex S. Flynt + - Dmitri V. Mavrodi + doi: 10.3389/fmicb.2021.651282 + date: 14 April 2021 + experiments: + - label: This experiment investigates the molecular communication between beneficial + rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon + by analyzing bacterial transcriptional and metabolic responses to plant root + exudates. + experiment_motivation: To examine molecular communication between beneficial + rhizosphere bacteria and their host plant, specifically focusing on bacterial + responses to plant-derived metabolites present within root exudates. + experiment_design: Surface-sterilized Brachypodium distachyon plants were grown + hydroponically to collect root exudates, which were characterized by LC-MS + and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with + root exudates under controlled microaerobic conditions, followed by transcriptomic + analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays. + environment: Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); + microaerobic environment (15% O₂). + host: + - AUTO:Brachypodium%20distachyon + - AUTO:Pseudomonas%20fluorescens-group%20strains + host_type: Plant; Bacteria + target_microbes: + - AUTO:Pseudomonas%20fluorescens + biological_system: Rhizosphere + conditions: + - 20-hour light + - 4-hour dark + - 24°C daytime + - 18°C nighttime + - microaerobic (15% O₂) + experimental_factors: + - AUTO:Presence%20of%20Brachypodium%20root%20exudates + is_biolog_experiment: yes + type_of_biolog_experiment: Phenotype MicroArray + plates: + - PM01 + - PM2A + replicates: '3' + protocol_steps: + - Surface-sterilized plants grown hydroponically in sterile jars filled with + glass beads and distilled water under controlled conditions (20-hour light + 24°C, 4-hour dark 18°C) + - root exudates collected, sterilized via 0.22 µm filtration, lyophilized, + stored at –80°C + - root exudates analyzed by LC-MS and GC-MS + - LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass + Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode + array detection + - GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to + Agilent 7890A GC + - cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown + on medium with root exudates under microaerobic (15% O₂) conditions + - RNA extracted and depleted of rRNA + - stranded library preparation and sequencing via Illumina HiSeq 2500 + - Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation + in OmniLog Phenotype MicroArray System at 25°C for 48 hours + - collection of optical density and colorimetric respiration data (formazan + production, absorbance at 590 nm) every 15 minutes + - Biolog data analysis performed using OmniLog Parametric Analysis software + v1.20.02 + - Transcriptome data analyzed using DESeq2 and comparative genomic tools. + plate_reader_types_of_data_collected: + - Optical density + - formazan (absorbance at 590 nm) + od_protocol: Not provided + respiration_protocol: Colorimetric respiration measured by formazan production + (absorbance detection at 590 nm), collected every 15 minutes over a period + of 48 hours at 25°C. + instrument_used: + - Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer + - Ultra Performance Liquid Chromatograph (UPLC) + - Pegasus 4D TOF Mass Spectrometer + - Agilent 7890A GC + - Illumina HiSeq 2500 + - OmniLog Phenotype MicroArray System + analysis_software: + - OmniLog Parametric Analysis software v1.20.02 + - DESeq2 + - OrthoMCL + - COG-based phylogeny + incubation_temperature: 25°C + incubation_duration: 48 hours +named_entities: + - id: AUTO:Brachypodium%20distachyon + label: Brachypodium distachyon + original_spans: + - 2671:2693 + - 7905:7927 + - 75924:75946 + - 82382:82404 + - 84035:84057 + - 92737:92759 + - 94848:94870 + - id: AUTO:Pseudomonas%20fluorescens-group%20strains + label: Pseudomonas fluorescens-group strains + - id: AUTO:Pseudomonas%20fluorescens + label: Pseudomonas fluorescens + original_spans: + - 2637:2659 + - 7986:8008 + - 75071:75093 + - 75391:75413 + - 75571:75593 + - 76860:76882 + - 78450:78472 + - 78876:78898 + - 79419:79441 + - 80612:80634 + - 81674:81696 + - 81944:81966 + - 82617:82639 + - 86158:86180 + - 91547:91569 + - 91814:91836 + - 92145:92167 + - 93005:93027 + - 93180:93202 + - 93811:93833 + - 94250:94272 + - id: AUTO:Presence%20of%20Brachypodium%20root%20exudates + label: Presence of Brachypodium root exudates From 7441b0cd6808ecbb6f4e24288dafa76b1c1c5492 Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Tue, 4 Mar 2025 23:30:39 -0800 Subject: [PATCH 11/25] gpt 4.5 preview --- fmicb-12-651282__ontogpt_gpt4.5preview.txt | 3772 ++++++++++++++++++++ 1 file changed, 3772 insertions(+) create mode 100644 fmicb-12-651282__ontogpt_gpt4.5preview.txt diff --git a/fmicb-12-651282__ontogpt_gpt4.5preview.txt b/fmicb-12-651282__ontogpt_gpt4.5preview.txt new file mode 100644 index 000000000..fadc88575 --- /dev/null +++ b/fmicb-12-651282__ontogpt_gpt4.5preview.txt @@ -0,0 +1,3772 @@ +The "poetry.dev-dependencies" section is deprecated and will be removed in a future version. Use "poetry.group.dev.dependencies" instead. +/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/pydantic/_internal/_config.py:345: UserWarning: Valid config keys have changed in V2: +* 'fields' has been removed + warnings.warn(message, UserWarning) +INFO:root:Logger root set to level 10 +INFO:root:Input file: /Users/marcin/Documents/VIMSS/ontology/LLMs/MATE-LLM/examples/mavrodi_etal/fmicb-12-651282.txt +INFO:root:Input text: ORIGINAL RESEARCH +published: 14 April 2021 +doi: 10.3389/fmicb.2021.651282 +Edited by: +Barbara Pivato, +Institut National de Recherche pour +l’agriculture, l’alimentation et +l’environnement (INRAE), France +Reviewed by: +Cara Helene Haney, +The University of British Columbia, +Canada +Xingang Zhou, +Northeast Agricultural University, +China +*Correspondence: +Dmitri V. Mavrodi +dmitri.mavrodi@usm.edu +Alex S. Flynt +alex.flynt@usm.edu +†These authors have contributed +equally to this work +Specialty section: +This article was submitted to +Microbe and Virus Interactions with +Plants, +a section of the journal +Frontiers in Microbiology +Received: 09 January 2021 +Accepted: 08 March 2021 +Published: 14 April 2021 +Citation: +Mavrodi OV, McWilliams JR, +Peter JO, Berim A, Hassan KA, +Elbourne LDH, LeTourneau MK, +Gang DR, Paulsen IT, Weller DM, +Thomashow LS, Flynt AS and +Mavrodi DV (2021) Root Exudates +Alter the Expression of Diverse +Metabolic, Transport, Regulatory, +and Stress Response Genes +in Rhizosphere Pseudomonas. +Front. Microbiol. 12:651282. +doi: 10.3389/fmicb.2021.651282 +Root Exudates Alter the Expression +of Diverse Metabolic, Transport, +Regulatory, and Stress Response +Genes in Rhizosphere Pseudomonas +Olga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2, +Karl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2, +Ian T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and +Dmitri V. Mavrodi1* +1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS, +United States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School +of Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular +Sciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics +and Quality Research Unit, Pullman, WA, United States +Plants live in association with microorganisms that positively influence plant +development, vigor, and fitness in response to pathogens and abiotic stressors. +The bulk of the plant microbiome is concentrated belowground at the plant rootsoil +interface. Plant roots secrete carbon-rich rhizodeposits containing primary and +secondary low molecular weight metabolites, lysates, and mucilages. These exudates +provide nutrients for soil microorganisms and modulate their affinity to host plants, but +molecular details of this process are largely unresolved. We addressed this gap by +focusing on the molecular dialog between eight well-characterized beneficial strains +of the Pseudomonas fluorescens group and Brachypodium distachyon, a model for +economically important food, feed, forage, and biomass crops of the grass family. We +collected and analyzed root exudates of B. distachyon and demonstrated the presence +of multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The +subsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many +of these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq +profiling of bacterial cultures amended with root exudates revealed changes in the +expression of genes encoding numerous catabolic and anabolic enzymes, transporters, +transcriptional regulators, stress response, and conserved hypothetical proteins. Almost +half of the differentially expressed genes mapped to the variable part of the strains’ +pangenome, reflecting the importance of the variable gene content in the adaptation of +P. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity +of cellular pathways and physiological responses underlying the establishment of +mutualistic interactions between these beneficial rhizobacteria and their plant hosts. +Keywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome +Frontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +INTRODUCTION +Plants are meta-organisms or holobionts that rely in part on +their microbiome for specific functions and traits. The ability +of the plant microbiome to influence plant development, vigor, +health, and fitness in response to abiotic stressors associated +with global climate change is documented by numerous studies +(Lugtenberg and Kamilova, 2009). There is mounting evidence +that plants actively recruit beneficial microbiomes, but many +aspects of this process are still very much a black box (Reinhold- +Hurek et al., 2015). The foundation for the differential affinity +of rhizobacteria toward host plants is built upon complex +chemical cross talk between microorganisms and plant roots. +Up to 40% of photosynthetically fixed carbon is released by +plant roots in the form of exudates and secretions, lysates, and +mucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990; +Badri and Vivanco, 2009). The release of these compounds is +actively controlled in response to environmental stimuli, and the +composition of root exudates varies greatly according to plant +species and physiological condition (Lynch, 1990; Nguyen, 2003; +Phillips et al., 2004; De-la-Pena et al., 2008). The presence and +composition of exudates strongly impact soil microorganisms, +which is consistent with the idea that plants actively select and +shape their root microbiota (Zolla et al., 2013). +Primary root exudates include simple and complex sugars, +amino acids, polypeptides and proteins, organic, aliphatic and +fatty acids, sterols, and phenolics (Nguyen, 2003; Badri and +Vivanco, 2009; Badri et al., 2009). These compounds serve as +carbon and energy sources for rhizobacteria, and the presence +of the intact corresponding catabolic pathways is essential +for competitive colonization of roots and disease suppression +(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and +Kamilova, 2009). Root exudates also contain numerous signal +molecules and secondary metabolites, the significance of which +is only now emerging (Walker et al., 2003; Bais et al., 2005, +2006). A handful of analyses of plant-induced gene expression +by transcriptional profiling in vitro (Mark et al., 2005) or in +the rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al., +2005; Matilla et al., 2007; Barret et al., 2009) have identified +multiple genes that are differentially regulated by exposure to +roots or root exudates. Bacterial pathways expressed during +rhizosphere colonization control utilization of plant-derived +metabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001; +Lugtenberg and Kamilova, 2009), motility and chemotaxis (de +Weert et al., 2002; Lugtenberg and Kamilova, 2009), phase +variation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002; +van den Broek et al., 2005), outer membrane integrity (de Weert +et al., 2006; Lugtenberg and Kamilova, 2009), and the ability +to sequester limiting resources (Raaijmakers et al., 1995) and +resist environmental stresses (Sarniguet et al., 1995; Miller and +Wood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In +its spatial and temporal properties, root colonization resembles +biofilm formation, and biofilm-related pathways also have been +implicated in adhesion to seeds and roots and rhizosphere +colonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003; +Yousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al., +2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally, +root exudates strongly affect the expression of diverse plant +growth promotion and biocontrol genes (Vacheron et al., 2013). +Over the past decade, the genomes of numerous rhizosphere +strains have been sequenced and analyzed, but functional +genomics studies of rhizosphere competence lag behind the +availability of sequence data. +This study explored the molecular dialog between the +model host plant Brachypodium distachyon and several wellcharacterized +rhizosphere strains of the Pseudomonas fluorescens +group. Brachypodium is a small annual grass originating in semiarid +regions of the Middle East that has emerged as a prime model +for economically important food, feed, forage, and biomass crops +of the grass family (Bevan et al., 2010; Schwartz et al., 2010; +Brkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The +biology, extensive collection of resources, and research tools +make B. distachyon an attractive model to investigate interactions +between plants and root-associated microbes. Pseudomonads +are ubiquitous Gram-negative g-proteobacteria that colonize +eukaryotic hosts and include both commensals and economically +important pathogens of plants and animals (Moore et al., +2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus +Pseudomonas currently comprises > 100 named species that have +been separated based on multilocus sequence analysis into 14 +species groups (Garrido-Sanz et al., 2016; Hesse et al., 2018). +The P. fluorescens group is the most diverse regarding both +the genetic distances within it, the number of species and the +large pangenome that makes up > 50% of the pangenome +of the genus as a whole (Loper et al., 2012). The group also +encompasses an unusually high proportion of strains that inhabit +the plant rhizosphere and possess plant growth promoting and +biocontrol properties. Naylor et al. (2017) profiled bacterial +communities associated with root tissues and rhizosphere of +18 different plant species of the Poaceae family. That study +identified Pseudomonas among taxa constituting the core grass +root microbiome and demonstrated that these bacteria were +enriched in C3 plants, including wheat, rye, barley, oat, and +Brachypodium. We confirmed the capacity of B. distachyon +Bd21 to serve as a host for rhizobacteria of the P. fluorescens +group in preliminary greenhouse assays with biocontrol strains +P. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens +Pf-5. Results of these experiments revealed that all strains +successfully established and colonized the roots of Brachypodium +(Supplementary Table 1). +In this study, we focused on eight well-studied strains +of the P. fluorescens complex that are supported by years +of studies, numerous refereed publications, and high-quality +genome sequences. By profiling transcriptomes of these strains +during growth in root exudates of B. distachyon, we revealed the +diversity of cellular pathways and physiological responses that +underlie the establishment of mutualistic interactions between +beneficial rhizobacteria and the host plant. Our results also +confirmed that root exudates contain carbohydrates, amino acids, +organic acids, and phenolics that serve as carbon and energy +sources for rhizobacteria. The root exudates also contained +osmoprotectants that may help microorganisms to persist in the +rhizosphere of drought-stressed plants. The diversity of microbial +genes perturbed by root exudates reflects the importance of +Frontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +the variable genome in adaptation of individual strains of +Pseudomonas to the rhizosphere lifestyle. +MATERIALS AND METHODS +Bacterial Strains Used in the Study +The eight Pseudomonas strains used for this study are +P. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens +SBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko +et al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller, +1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996), +P. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens +Pf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and +Stipanovic, 1980). The selected organisms have been studied +extensively for their role in biological control and plant +growth promotion (Supplementary Table 2). The strains were +maintained in the laboratory as frozen stocks (􀀀80 C) and +routinely cultured in King’s medium B (King et al., 1954) or +21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g +of Na2HPO4 2H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and +20 ml of a microelement solution (Smibert and Kreig, 1994; +Halverson and Firestone, 2000). +Propagation of Plants and Collection of +Root Exudates +B. distachyon Bd21 was established from seed material obtained +from the USDA-ARS Plant Germplasm Introduction and Testing +Research Unit (Pullman, WA, United States). Brachypodium +seeds were imbibed for 3 days at 4 C and sown in 7   7 cm +pots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture, +Agawam, MA, United States). Plants were grown in an IR- +89X (Percival Scientific, Perry, IA, United States) controlled +environment chamber retrofitted with 6500K and 3000K T5 54W +grow lights (Spectralux) under a 20-h light, 24 C/4-h dark, 18 C +cycle. Plants were watered and fertilized with Jack’s professional +water-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA, +United States). After 12 weeks and plant senescence, seeds +were collected, processed, and stored under desiccant and dark +conditions at room temperature. +To collect root exudates, seeds of B. distachyon Bd21 were +surface-sterilized, pregerminated, and placed in sterile 1 L widemouth +glass jars containing 113 g of 6-mm glass beads and +25 ml distilled water. Jars were covered with vented caps +and plants were grown hydroponically in an environmental +controlled growth chamber under conditions described above. +After 6 days, root exudates were extracted from individual +jars and their sterility was confirmed by spotting on nutrient +agar. Multiple batches of root exudates were collected, filtered +(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and +stored at –80 C. +Metabolomic Profiling of Root Exudates +Exudates were analyzed for primary metabolites at the Murdock +Metabolomics Laboratory at Washington State University +(Pullman, WA, United States). Freeze-dried residues were +suspended in 500 ml 50% aqueous acetonitrile and clarified by +centrifugation for 20 min at 21,000   g and 4 C. The liquid +chromatography mass spectrometry analysis was conducted +with a Synapt G2-S quadrupole-ion mobility spectrometrytime +of flight mass spectrometer system equipped with an +acquity ultra-performance liquid chromatograph (UPLC) and +an acquity photodiode array detector (all from Waters, Milford, +MA, United States). The exudate metabolites were separated +on a SeQuant ZIC-pHILIC HPLC column (2.1   100 mm, +3 mm) (Milllipore Sigma, Burlington, MA, United States) using +acetonitrile with 0.1% formic acid as solvent B and water +with 0.1% formic acid as solvent A at a flow rate of 400 ml +min􀀀1 and the following linear gradient extending over 14 min: +0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10% +B; 10 min, 80% B; and 14 min, 80% B. Mass spectra were +collected in positive ion mode over a range of m/z 50–1,200 +with a scan time of 0.2 s. The Q-TOF-MS source was at +3.0 kV and 120 C; the sampling cone at 40 V, desolvation +temperature was 250 C; cone gas and desolvation gas flow were +at 0 and 850 L h􀀀1, respectively. Leucine enkephalin was used +for post-acquisition mass correction. Target compounds were +visualized using selected ion chromatograms at 0.05 Da window +width. The compound identification was based on comparison +of chromatographic behavior and accurate masses to those of +authentic standards. +For gas chromatography, derivatization was carried out using +a modification of the procedure of Lee and Fiehn (2008). +The freeze-dried residues were suspended in 950 ml aqueous +methanol (84%, v/v) and clarified by centrifugation for 15 min +at 21,000   g at 4 C. The supernatants were spiked with 1 +mg of the internal standard salicylic acid-d6 (C/D/N Isotopes, +Quebec, Canada) and dried in vacuo. The dry residues were +suspended in 10 ml of O-methoxylamine hydrochloride (30 mg +ml􀀀1 in anhydrous pyridine, both from Millipore Sigma) and +incubated while mixing (1,000 RPM) for 90 min at 30 C. +Subsequently, samples were derivatized with 90 ml of MSTFA +with 1% TMCS (Thermo Fisher Scientific, Waltham, MA, +United States) for 30 min at 37 C. Gas chromatography-mass +spectroscopy analysis was performed using a Pegasus 4D timeof- +flight mass spectrometer (LECO, Saint Joseph MI) equipped +with a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A +oven (Agilent Technologies, Santa Clara, CA, United States). +The derivatization products were separated on a 30-m, 0.25 mm +i.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA, +United States) with an IntegraGuard precolumn using ultrapure +He at a constant flow of 0.9 ml min􀀀1 as carrier gas. The linear +thermal gradient started with a 1-min hold at 70 C, followed +by a ramp to 300 C at 10 C min􀀀1. The final temperature was +held for 5 min prior to returning to initial conditions. Mass +spectra were collected at 17 spectra s􀀀1. Peak identification was +conducted using the Fiehn primary metabolite library (Kind +et al., 2009) and an identity score cutoff of 700. Additionally, +authentic standards for a number of primary metabolites were +analyzed under identical conditions and the data used to compare +the chromatographic behavior. Peak alignment and spectrum +comparisons were carried out using the Statistical Compare +feature of ChromaTOF software (LECO). +Frontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +Isolation of RNA From Bacteria Cultured +in Root Exudates and RNA-Seq +The strains were pregrown overnight at 25 C on 21C-glucose +agar and then subcultured into 96-well microplates containing +liquid 21C-glucose medium amended with Brachypodium +exudates. The liquid medium was prepared by dissolving the +lyophilized root exudate material in an appropriate volume of +21C-glucose medium to concentrate root exudates 20-fold. The +growth medium was sterilized by passing it through a 0.22- +mm membrane filter. The control cultures were grown under +identical conditions in the absence of exudates. All treatments +were inoculated at OD600 of 0.1 and incubated for 20 to +22 h until cultures entered late-exponential growth phase at +25 C in an atmosphere of 15% oxygen [created by a ProOx +P110 oxygen controller (BioSpherix, Parish, NY, United States) +with a hypoxia C-chamber]. The cells were stabilized by the +addition RNAprotect reagent (QIAGEN, Germantown, MD, +United States) and total RNA was purified using a RNeasy Protect +Bacteria Mini Kit (QIAGEN) from three biological replicates of +each strain cultured under control conditions and in exudates. +The quality assessment of the extracted RNA samples was +performed with a NanoDrop OneC Spectrophotometer (Thermo +Fisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies) +and revealed A260/A280 and A260/A230 values of > 2.0 and a mean +RNA integrity numbers (RIN) value of > 9.2. +Three biological replicates of RNA samples were shipped on +dry ice to the DOE Joint Genome Institute (Walnut Creek, CA, +United States), where rRNA was depleted and stranded RNASeq +libraries were prepared, quantified by qPCR and sequenced +using a HiSeq 2500 instrument (Illumina). The fastq file reads +were filtered and processed with BBDuk1 to remove reads that +contained 1 or more “N” bases, had an average quality score +across the read less than 10 or had a minimum length < 51 bp +or 33% of the full read length. Reads mapped with BBMap (see +text footnote 2) to masked human, cat, dog, and mouse references +at 93% identity were removed. Another category of removed +sequences matched RNA spike-in, PhiX, common microbial +contaminants, and ribosomal RNAs. The processed reads from +each library were aligned to the reference genome using BBMap +with only unique mappings allowed (BAMs/directory). If a +read mapped to more than one location it was ignored. +featureCounts (Liao et al., 2014) was used to generate raw +gene counts, which were normalized to adjust for the length of +each gene and total number of reads mapped for each library. +The normalization formula used: n = [r/(l/1,000)]/(t/1,000,000), +where n = normalized read count for gene (G) for library (L); +r = raw read count for gene G for library L; l = gene G length; and +t = total reads mapped for library L. Raw gene counts were used to +evaluate the level of correlation between biological samples using +Pearson’s correlation. +Bioinformatic Analysis +Count tables generated by the JGI RNA-Seq pipeline were input +into DESeq2 (Love et al., 2014) to normalize and determine +1https://sourceforge.net/projects/bbmap/ +differential expression. Statistical significance was established +through DESeq2 by using three biological replicates for control +and root exudate conditions. Scatterplots were generated from +the DESeq2 data table outputs using ggplot2. Genes differentially +expressed between control and root exudate samples (log2 foldchanges +–2   to   2, adjusted p value   0.05) were used in +downstream analysis. The core genome and pangenome for the +Pseudomonas strains used in this study were computed using the +OthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic +Pangenome Accumulation v1.4.0 apps implemented in the U.S. +Department of Energy Systems Biology Knowledgebase (KBase) +(Arkin et al., 2018). Additional comparisons were conducted +with the PGAweb pangenome analysis pipeline (Chen et al., +2018). Differentially expressed genes were assigned to core, +non-core, and singleton parts of each strain’s proteome by +BLASTp with an E value cutoff of e-06, identity of 40% +and coverage of 60%. Functional annotation of differentially +expressed genes was carried out with the Blast2GO (Conesa +and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., +2018). Additional manual curation was performed using tools +implemented in the Integrated Microbial Genomes (IMG) +database (Markowitz et al., 2012), Pseudomonas Genome +Database (Winsor et al., 2009), Kyoto Encyclopedia of Genes and +Genomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3 +(Biomatters, Auckland, New Zealand). Metabolic functions +encoded by the differentially expressed genes were mapped +using iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses +were carried out by building multiple sequence alignments +with MAFFT v7.222 (Katoh and Standley, 2013) and inferring +neighbor-joining (NJ) phylogenies with Geneious Tree Builder. +The resultant phylogenetic trees were visualized with iTOL +(Letunic and Bork, 2016). Reproducibility of clades within the +inferred NJ trees was assessed by bootstrap resampling with +1,000 replicates. +Characterization of Carbon Source +Utilization With Biolog Phenotype +Microarrays +The utilization of carbon sources was analyzed using Phenotype +MicroArrays (Biolog, Hayward, CA, United States) as follows. +The bacteria were cultured overnight on Luria-Bertani agar +at 25 C, after which cells were harvested and suspended in +inoculating fluid (IF-0). The transmittance of the suspension +was adjusted to 42% using a Biolog turbidimeter. The cell +suspension was mixed with IF􀀀0 containing Dye Mix A +(Biolog) to achieve a final transmittance of 85%. One hundred +microliter aliquots of the adjusted cell suspension were +inoculated into PM01 and PM02A plates, which were then +incubated in an OmniLog Phenotype MicroArray System +(Biolog) at 25 C for 48 h. The formation of formazan was +recorded at 15 min intervals, and data were analyzed using +OmniLog Parametric Analysis software v1.20.02 (Biolog). +Relative growth of the studied strains was normalized to +growth on D-glucose and visualized using Heatmapper +(Babicki et al., 2016). +Frontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +Data Availability +Sequences generated in this project were deposited +under NCBI BioProject accession numbers PRJNA439743 +through PRJNA439790. +RESULTS +Metabolomic Profiling of Root Exudates +of B. distachyon +Metabolomics analysis of lyophilized root exudates revealed +the presence of numerous plant metabolites, 86 of which were +identified by matching their spectra to the LECO/Fiehn +Metabolomics library (Supplementary Table 3). These +metabolites included (i) carbohydrates and their derivatives +(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose, +and others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol, +galactinol, 2-deoxyerythritol, ribitol, threitol and +cellobitol); (iii) amino acids and derivatives (glutamine, +tyrosine, glutamic acid, asparagine, aspartic acid, valine, +phenylalanine, isoleucine, glycine, serine, proline, leucine, +tryptophan, cysteine, methionine, citrulline, and others); (iv) +organic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric, +fumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic, +malic, methylmalonic, nicotinic, quinic, succinic, threonic); +and (v) assorted metabolites including heterocyclic compounds, +phenolics, and biogenic amines, etc (3-hydroxypyridine, +maleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate, +5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid, +and urea). Results of the analysis also revealed that root exudates +of B. distachyon contain hydroxyectoine and the quaternary +amine (QA) glycine betaine (Supplementary Figure 1). +Phylogenetic and Pangenome Analyses +of Pseudomonas Strains Used in the +Study +We used a set of phylogenetic markers suggested by Mulet et al. +(2010) to investigate the relatedness of the eight strains used in +this study to distinct lineages recognized within the P. fluorescens +species complex. The multilocus sequence analysis based on +the concatenated sequences of the housekeeping genes rrs (16S +rRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with +strains 2-79 and SBW25) as a member of the P. fluorescens +subgroup (Figure 1). The rest of the strains clustered closely with +four additional subgroups of the P. fluorescens complex, namely +P. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1), +P. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of +the eight rhizosphere Pseudomonas strains varied in size by 1.43 +megabase (ranging from 5.65 to 7.07 Mb) and contained between +5,166 and 6,363 protein-coding genes (Figure 2A). The shared +gene content was characterized with OrthoMCL, which uses allagainst- +all BLASTp followed by the Markov Cluster algorithm to +identify protein groups shared between the compared genomes, +as well as groups representing species-specific gene expansion +families (Li et al., 2003). The pangenome analysis revealed a +core comprised of approximately 3,179 orthologs that were +shared among all strains and represented 50.0% to 61.5% +of each predicted proteome (Figures 2A,B). The non-core +pangenome contained genes shared by two or more (but not +all) strains and contained between 1,482 and 2,080 orthologs, +which corresponded to 28.7–36.3% of individual proteomes. The +rest of the predicted protein-coding genes were strain-specific +singletons that comprised 7.5% to 15.1% of the strain’s predicted +proteomes. In respect to divergence from the core genome, strain +Pf-5 was found to possess the highest proportion of unique genes +(n = 949) followed by 2-79 (n = 887). The entire pangenome of +the Pseudomonas strains encompassed over 12,000 homolog and +singleton gene families. +Further homolog family-based comparisons identified Q8r1- +96 and R1-43-08 as the most distantly related strains, with +3349 shared homologs (Supplementary Table 4A). Q8r1-96 and +Q2-87, which shared 4,489 homologs, were the most closely +related strains. The partitioning of homolog gene families into +the core, non-core, and singleton parts of the pangenome agreed +with phylogenetic relationships of the strains deduced from the +analysis of a selected subset of COGs (Clusters of Orthologous +Groups) (Figure 2C and Supplementary Tables 4B,C). The +COG-based phylogeny supported the multilocus sequence +analysis and revealed that the eight Pseudomonas strains form +three distinct clusters, the first of which contained 2-79, R1-30- +84, and SBW25. The second cluster included Q8r1-96 and Q2-87, +whereas the third encompassed strains 30-84, Pf-5, and Pf0-1. +Correlating the Composition of Root +Exudates With Metabolic Profiles of +Pseudomonas Strains +We used the Phenotype MicroArray PM1 and PM2 plates to +profile the eight Pseudomonas strains for the utilization of +190 different carbon sources. Results of the analysis identified +90 compounds that supported growth and clustered by their +intensities of utilization into three distinct groups (Figure 3). +Group I was comprised of 30 highly metabolized carbon +sources, which included several amino acids and intermediates of +glycolysis, pyruvate metabolism, and citrate cycle. Approximately +half of these compounds were catabolized by all eight strains, +and included several organic acids (fumaric, citric, gluconic, +malic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro, +Ala, and g-aminobutyric acid), carbohydrates (glucose, mannose, +and mannitol), and the purine nucleoside inosine. Group II +was composed of 44 chemically diverse carbon sources that +were variably utilized by the strains. These compounds were +carbohydrates, organic acids, amino acids, phenolics, and polyols, +and included known compatible solutes and intermediates of +metabolism of pentoses, galactose, starch, and sucrose. Group +III encompassed the rest of the Phenotype MicroArray test +panel and contained compounds that were not catabolized +by the tested strains. Among several notable exceptions were +a-hydroxyglutamic acid- g-lactone, putrescine, and itaconic, +citramalic, and succinamic acids, which supported the growth +of strains 2-79, 30-84, Pf-5, and SBW25. We further matched +the carbon metabolic profiles of the Pseudomonas strains against +the list of plant-derived metabolites from the root exudates of +Frontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the +P. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA +gyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor +method. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles). +B. distachyon Bd21. Interestingly, many carbon sources from +the Phenotype MicroArray panel were also present in the root +exudates of B. distachyon Bd21, and some of these compounds +(glucose, mannose, galactose, fructose, g-aminobutyric acid, +aspartic acid, citric acid, malic acid, fumaric acid, quinic acid, +alanine, glutamine, and glutamic acid) were catabolized by all +strains used in this study, while others (e.g., xylose, trehalose, +m-inositol) were actively utilized only by certain organisms +(Figure 3). The comparison of catabolic profiles across the +eight studied Pseudomonas strains revealed the presence of three +distinct clusters. The first cluster contained strains Q8r1-96 and +Q2-87, which consumed very similar sets of carbon sources, +as well as strain Pf0-1. The second cluster was composed of +2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster +was represented by a single strain, Pf-5. The overall similarity +of the catabolic profiles partially agreed with the separation +of the strains into different subgroups of the P. fluorescens +complex (see above). +Analysis of the RNA-seq Results +In order to understand the cellular responses of rhizosphere +Pseudomonas to plant exometabolites, we analyzed the +transcriptome changes in cultures grown in the presence of +root exudates. Under field conditions, rhizobacteria colonize +plant roots in the form of surface-attached microaerobic biofilms +(Hojberg et al., 1999). To mimic these conditions, the eight +Pseudomonas strains were grown statically at 72% air saturation +in 21C-glucose medium amended with root exudates and then +processed to extract total RNA (Supplementary Figure 2). +A total of 995 million raw sequencing reads were generated from +the RNA samples by using the Illumina HiSeq-2500 platform, +averaging 20.7 million reads per sample. The removal of lowquality +and rRNA sequences resulted in a total of 793 million +filtered reads that were mapped onto the eight Pseudomonas +genomes with a mean of 7.48 million mapped fragments per +genome. The differentially abundant transcripts were identified +by setting a p value of 0.05 (adjusted for multiple testing) and +Frontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all +eight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific +singletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of +the pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the +25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any +bias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939 +non-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied +strains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining +ortholog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node +represents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The +analysis was conducted in KBase (Arkin et al., 2018). +the log2 fold-change (FC) threshold    2.0 (Figure 4 and +Supplementary Tables 5–12). When compared with the control +conditions, an average of 204 genes per strain were differentially +expressed in the presence of root exudates, with the highest +(n = 425) and lowest (n = 112) numbers observed, respectively, +in SBW25 and Q2-87 (Figure 4). Overall, more genes were +induced than repressed in response to exudates, but the actual +numbers in each category varied substantially depending on the +identity of the Pseudomonas strain. In most strains, the bulk of +the differentially expressed genes was almost equally distributed +between the core (mean, 48.2%) and non-core (mean, 45.8%) +parts of the genome, whereas the strain-specific singleton genes +constituted on average only 5.9% (Figure 4B). One notable +exception was observed in Q8r1-96, where all differentially +expressed genes belonged to the core (73.8%) and non-core +(26.2%) parts of the genome. Another notable pattern was +observed in R1-43-08, where the majority of genes affected by +the presence of root exudate fell into the non-core category +(56.3%). The highest proportion of differentially expressed +singletons (11.3 and 10.4%, respectively) was identified in +strains SBW25 and Pf-5. +We further explored how the identified differentially +expressed genes were distributed across genomes of the eight +studied rhizosphere strains. The pairwise BLASTp comparisons +identified 2-79 and SBW25 as two strains that shared the highest +number of genes (n = 101) induced or repressed in response +to root exudates (Table 1). The second pair of strains with +a significant number of similar differentially expressed genes +(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1 +and 30-84, which shared 56 differentially expressed genes. These +patterns of shared genes were also observed when the results of +the pairwise BLASTp comparisons were converted into a binary +gene presence/absence matrix, which was then subjected to +cluster analysis using a UPGMA algorithm based on Sorensen’s +dissimilarity index or examined by non-metric multidimensional +scaling (NMDS) (Figure 5). +The differentially expressed Pseudomonas genes were +subjected to Blast2Go analysis and Gene Ontology (GO) +annotation (Figure 6). Metabolic process, catalytic activity, +and membrane were the most common annotation terms +across the three primary GO term categories (i.e., biological +process, molecular function, and cellular component). A total +of 1,694 GO terms was assigned to 805 upregulated genes, with +the majority of the GO terms related to molecular function +(682, 40.3%), followed by biological process (669, 39.5%), and +cellular component (343, 20.2%). In the 539 downregulated gene +category, 1,101 GO terms were assigned to biological process +(420, 38.1%), molecular function (417, 37.9%), and cellular +component (264, 24.0%). Within biological process, metabolic +process, cellular process, localization, response to stimulus, and +regulation were over-represented. Within molecular function, +the largest proportion was assigned to catalytic activity, binding, +and transporter activity categories. Within cellular component, +the majority were assigned to membrane, membrane part, cell, +and cell part categories. Across the eight strains, 37–42% of +differentially expressed genes had no Gene Ontology IDs and +encoded various conserved hypothetical proteins. +Functional Classification of Shared +Differentially Expressed Genes +The interrogation of RNA-seq data revealed multiple cellular +pathways that were differentially regulated in bacterial cultures +incubated with root exudates (Supplementary Figures 3, 4). +Frontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere +Pseudomonas strains used in the study. The hierarchical clustering analysis +was carried out using the average linkage method with Euclidean distances. +Carbon sources identified by red arrowheads were also detected in the sterile +root exudates of B. distachyon Bd21. +Although none of these differentially regulated pathways were +shared by all eight strains, the cross-strain comparisons revealed +several types of common and specific transcriptomic responses +that were elicited by the presence of plant exometabolites +(Table 2). The visual representation of core gene expression +patterns is provided in Supplementary Figure 5, which +shows heatmaps of expression profiles and p-adj values for +core genes shared by the studied strains. The figure is +accompanied by Supplementary Table 13 that lists predicted +functions of genes constituting the four distinct clusters +observed after hierarchical clustering of gene expression +values. The first category of shared differentially expressed +pathways functioned in the uptake and catabolism of selected +carbohydrates, quaternary ammonium compounds (QAs), and +phenolics. All strains except for R1-43-08, responded to root +exudates by inducing the fructose-specific phosphoenolpyruvate +(PEP)-dependent phosphotransferase system (PTSFru). The +components of this system are encoded by a conserved +operon and include the cytoplasmic polyprotein EI/HPr/EIIAFru +(FruB), the 1-phosphofructokinase FruK, and the fructosespecific +permease EIIBC (FruA) (Chavarria et al., 2016). The +PTSFru system functions by acquiring high-energy phosphates +from PEP and sequentially passing them, via the EI/HPr/EIIAFru +domains of FruB, to the EIIB component of FruA. The +phosphates are ultimately transferred by the EIIC transporter +to fructose yielding fructose 1-phosphate, which is channeled +into the central metabolic pathways through the action of the +phosphofructokinase FruK. +In all strains except for Q8r1-96 and Pf-5, the exposure to root +exudates resulted in the induction of two genes adjacent to the +fru cluster that encoded a Major Facilitator Superfamily (MFS) +transporter and an L-arabinonate dehydratase (Table 2). These +genes are predicted to participate in the uptake and catabolism +of L-arabinose, where L-arabinonate dehydratase plays an +important role by converting L-arabinonate to 2-dehydro-3- +deoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1- +43-08, and Q2-87, we also observed the induction of genes +encoding components of the AraFGH complex, an ATP-Binding +Cassette (ABC) superfamily transporter involved in the import of +arabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all +strains except SBW25 and R1-43-08 responded to the presence +of exudates by upregulating a conserved gene encoding an +aldose epimerase superfamily protein. Such enzymes equilibrate +alpha- and beta-anomers of aldoses and ensure that stereospecific +enzymes involved in the metabolism of free sugars do not act as +metabolic bottlenecks (Abayakoon et al., 2018). Although some +aldose epimerases have been linked to specific pathways, the +Pseudomonas gene identified in this study could not be assigned +to a particular metabolic process based on sequence analysis and +genomic location. +Several Pseudomonas strains responded to the presence +of root exudates by upregulating genes involved in the +uptake and catabolism of myo-inositol and possibly other +stereoisomers of inositol (Table 2). The upregulated catabolic +genes encode the dehydrogenase IolG, which oxidizes myoinositol +to its corresponding ketone, as well as IolE, IolD, IolB, +and IolC that collectively convert the 2-keto-myo-inositol to +acetyl-CoA and the glycolysis intermediate dihydroxyacetone +phosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1- +43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions +Frontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core +and non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established +threshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual +Pseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates. +also involved components of the putative inositol-specific +ABC transporter. The cross-genome comparisons revealed +that in all studied strains except for Pf0-1, components +of the myo-inositol utilization pathway were encoded +within a well-conserved gene cluster which, in addition to +catabolic and transport functions, also encodes a dedicated +transcriptional repressor. +All studied strains of Pseudomonas carry multiple genes +involved in scavenging the quaternary ammonium compounds +choline, glycine betaine (GB), carnitine, choline-O-sulfate, and +Frontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +TABLE 1 | The number of differentially expressed genes shared among the eight +studied strains of rhizosphere Pseudomonas. +Strain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5 +2–79 260 +SBW25 101 425 +R1-43-08 30 25 151 +Q8r1-96 32 39 21 145 +Q2-87 27 28 25 31 112 +30-84 27 23 24 32 28 136 +Pf0-1 38 50 29 29 50 56 205 +Pf-5 36 41 52 86 29 55 40 230 +The pairwise comparisons were conducted by BLASTp with the following cutoff +parameters: E-value < 1e-06, minimum percent identity > 40%, and minimum +percent coverage > 65%. The black diagonal cells show the number of differentially +expressed genes per strain. In other words, these are self comparison values. +sarcosine from the environment. Many of these genes were +differentially expressed, including those encoding parts of the +ABC transporter CbcXWV, which is predicted to function in +the uptake of choline under water-replete conditions (Table 2). +Among enzymes induced in the presence of root exudates were +the choline dehydrogenase BetA, which converts choline to +glycine betaine and a network of enzymes (i.e., the Rieske family +oxygenase GbcAB, the dimethyglycine demethylase DgcAB, and +the sarcosine oxidase SoxBDAG) that sequentially convert GB +to glycine. In 2-79 and SBW25, this group of differentially +regulated genes also included an AraC-family transcriptional +activator GbdR, which perceives intercellular levels of GB and +induces genes involved in the transport and catabolism of +glycine betaine and detoxification of the catabolic byproducts +(Hampel et al., 2014). +The last category of activated catabolic pathways included the +catechol branch of the b-ketoadipate pathway for the degradation +of aromatic compounds. In strains 30-84, Pf0-1, and Pf-5, +growth on root exudates resulted in upregulation of catechol- +1,2-dioxygenase, muconate cycloisomerase, and muconolactone +isomerase, which collectively cleave the catechol ring and +convert it to b-ketoadipate enol-lactone (Harwood and Parales, +1996). Finally, analysis of the P. synxantha 2-79 transcriptome +identified an induction of benABC genes encoding subunits +of benzoate 1,2-dioxygenase, an oxidoreductase that generates +catechol from benzoate. +In addition to various catabolic pathways, the exposure to root +exudates also induced several genes involved in the homeostasis +of copper (Table 2). Four of these genes form a conserved cluster +in genomes of the strains and encode the periplasmic coppersensing +two-component system CinRS, the plastocyanin/azurinlike +protein CinA, and the NADPH-dependent pre-Q0 reductase +CinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed +upregulation of a conserved operon encoding the multicopper +oxidase CopA, the periplasmic copper-binding protein CopC, the +inner membrane protein CopD, and outer membrane protein +CopB. In several Gram-negative bacteria, these Cop proteins are +thought to have dual functions and participate both in the uptake +of essential copper as well as in the sequestration of excess copper +in the periplasm and outer membrane. +The analysis of shared downregulated pathways revealed that +most of the strains respond to the presence of root exudates by +repressing genes involved in the uptake and catabolism of sulfur +compounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96, +Q2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB +operon responsible for the utilization of alkanesulfonates +as sulfur sources. The ssu operon is highly conserved in +fluorescent pseudomonads and encodes the FMNH2-dependent +monooxygenase SsuD and the NAD (P)H-dependent FMN +reductase SsuE, which together catalyze the desulfonation of +alkanesulfonates. Also, the ssu locus contains genes for the +molybdopterin-binding protein SsuF and the alkanesulfonatespecific +ABC-type transporter consisting of the sulfonate +substrate-binding protein SsuA, sulfonate permease protein +SsuC, and sulfonate transport ATP-binding protein SsuB. Finally, +in R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates +coincided with repression of the tauABCD operon, which +allows these strains to utilize taurine (2-aminoethanesulfonate) +as a sulfur source. The repressed tau genes encoded the +2-oxoglutarate-dependent taurine dioxygenase TauD and +substrate-binding, ATP-binding, and permease components of +the taurine-specific ABC transporter TauABC. +Other Differentially Expressed Pathways +In addition to their effect on several shared cellular pathways, +growth on root exudates resulted in the induction or repression of +numerous strain-specific genes. In closely related P. synxantha 2- +79 and P. fluorescens SBW25, we observed differential expression +of genes involved in energy metabolism, transport of amino +acids, and surface attachment (Supplementary Tables 5, 6). +Other notable differentially expressed pathways included 2– +79 gene clusters that encode enzymes for the catabolism of +trehalose, a prophage, and toxin/antitoxin system, as well as the +SBW25 operon predicted to control the synthesis of the capsular +exopolysaccharide colonic acid. The response of Pseudomonas sp. +R1-43-08 to root exudates also involved differential expression +of different energy metabolism pathways. In addition, we +observed the upregulation of genes involved in the uptake and +catabolism of xylose (also upregulated in 2–79) and repression +of enzymes for the biosynthesis of phenazine-1-carboxylic acid +and assimilation of inorganic sulfur and L-cysteine biosynthesis +(Supplementary Table 7). +The analysis of the Q8r1-96 transcriptome revealed +perturbation of different metabolic pathways including genes +encoding components of cytochrome C oxidase, transport +and catabolism of sorbitol/mannitol, metabolism of butanoic +acid, and biosynthesis of exopolysaccharides alginate and +poly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In +P. fluorescens Q2-87, we identified differential expression of +genes involved in metabolism of galactose, tryptophan, tyrosine, +glycine, serine, and threonine (Supplementary Table 9), +while in P. chlororaphis 30-84, growth on exudates activated +the biosynthesis of molybdopterin cofactor, catabolism of +galactonate and acetoin, and uptake and catabolism of putrescine +(Supplementary Table 10). The response of P. protegens Pf-5 to +root exudates involved upregulation of acetoin dehydrogenase, +which converts acetoin to acetaldehyde and acetyl-CoA, as +Frontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +FIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root +exudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis. +FIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of +B. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa +and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO +term. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO +terms. +well as pathways for the utilization of glycolate and putrescine +(Supplementary Table 11). Also induced were genes for +the production of pyrrolnitrin and PhlG hydrolase, which +modulate the metabolic loads attributed to the synthesis of +2,4-diacetylphloroglucinol. The differentially expressed genes of +P. fluorescens Pf0-1 included, among others, operons encoding +cytochrome C oxidase and enzymes for catabolism of malonic +acid (Supplementary Table 12). Yet another interesting +finding involved the induction of assorted genes acting in the +homeostasis of iron and defense against reactive oxygen species +(ROS). We observed activation of iron dicitrate transporters +(SBW25 and 30-84), genes for the biosynthesis of siderophores +ornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading +enzymes (2–79, 30–84), TonB siderophore receptors, and +components of the energy-transducing inner membrane +complex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially +expressed ROS defense pathways were represented by different +catalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and +Pf-5 and organic hydroperoxide resistance proteins in strains +SBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and +Pf0-1, the addition of exudates resulted in the upregulation of +peroxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic +and aromatic hydroperoxides. +DISCUSSION +Our analysis of B. distachyon root exudates revealed a complex +mix of primary and secondary metabolites, thus supporting +the view of the plant rhizosphere as a carbon-rich niche for +Frontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +TABLE 2 | The distribution and predicted functions of selected differentially expressed genesa. +Predicted function Strainb +2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 +Uptake and catabolism of fructose +D-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8) +1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7) +D-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2) +Uptake and catabolism of arabinose +MFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9) +L-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5) +Interconversion of alpha- and beta-anomers of aldoses +Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) +Uptake and catabolism of quaternary ammonium compounds +Choline dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1) +Transcriptional regulator GbdR 2756597125 (􀀀3.7) 649639087 (􀀀3.9) +Membrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3) +Hypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3) +Dimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2) +Dimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0) +Betaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7) +Betaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2) +Ser hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1) +Sarcosine oxidase, g subunit, SoxG 2597878304 (2.9) +Sarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4) +Sarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2) +Sarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3) +Betaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5) +ABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5) +Uptake and catabolism of myo-inositol +5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2) +2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1) +5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3) +2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4) +3D-(3,5/4)-trihydroxycyclohexane-1,2-dione +acylhydrolase, IolD +2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3) +Myo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9) +Inositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8) +Inositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5) +Inositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6) +(Continued) +Frontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +TABLE 2 | Continued +Predicted function Strainb +2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 +Uptake and catabolism of fructose +Catabolism of phenolics +Muconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8) +Muconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4) +Catechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5) +AraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2) +Benzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8) +Benzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8) +Benzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3) +Copper homeostasis +pre-Q0 reductase/7-cyano-7-deazaguanine +reductase CinQ +2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5) +Cupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4) +Heavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5) +Heavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0) +Copper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4) +Copper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0) +Copper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4) +Copper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2) +Conserved hypothetical protein +Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) +Uptake and catabolism of sulfonates +FMN-dependent monooxygenase SsuE 2756592254 (􀀀1.9) 637745334 (􀀀3.5) +Sulfonate substrate-binding protein SsuA 649639261 (􀀀2.0) 2756592253 (􀀀2.2) 2597878518 (􀀀2.6) 2597849636 (􀀀3.9) 637745333 (􀀀7.0) 637323272 (􀀀2.7) +FMN-dependent monooxygenase SsuD 649639260 (􀀀1.9) 2756592252 (􀀀2.3) 2597849637 (􀀀2.8) 637745332 (􀀀5.8) 637323271 (􀀀2.3) +Sulfonate permease protein SsuC 649639259 (􀀀2.4) 2756592251 (􀀀2.1) 2597878516 (􀀀1.9) 2597849638 (􀀀3.2) 637745331 (􀀀5.8) 637323270 (􀀀2.5) +Sulfonate transport ATP-binding protein SsuB 649639258 (􀀀2.5) 2756592250 (􀀀2.3) 2597849639 (􀀀2.5) 637745330 (􀀀5.6) 637323269 (􀀀2.7) +Molybdopterin binding protein SsuF 649639257 (􀀀3.1) 2756592249 (􀀀2.7) 2597878514 (􀀀2.9) 2597849641 (􀀀3.3) 637745339 (􀀀6.2) 637323268 (􀀀2.7) +Uptake and catabolism of taurine +Taurine substrate-binding protein TauA 2597854917 (􀀀2.0) 637740095 (􀀀4.1) +Taurine transport ATP-binding protein TauB 2756592398 (􀀀1.9) 637740094 (􀀀3.7) 637317614 (􀀀1.9) +Taurine permease protein TauC 2756592398 (􀀀1.9) 637740093 (􀀀3.9) 637317613 (􀀀2.0) +Taurine dioxygenase TauD 2597854920 (􀀀1.8) 637740092 (􀀀3.6) +aThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%. +bValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets). +Frontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +soil microorganisms. Our results were in agreement with a +recent report of 27 different sugars, amino acids, and organic +acids in Brachypodium exudates (Kawasaki et al., 2016). We +confirmed the presence of exometabolites identified in that +study, along with dozens of additional analytes that were +identified by matching their mass-spectra and retention indices +to the LECO/Fiehn Metabolomics library (Supplementary +Table 3). The complementation of the metabolomic analysis +with profiling of the bacteria by Biolog Phenotype MicroArrays +revealed that a substantial proportion of the characterized +exudate constituents were catabolized by a collection of eight +Pseudomonas strains from across the P. fluorescens group that is +known to form associations with plant roots. The amendment +of Pseudomonas cultures with root exudates caused changes in +the expression of multiple genes encoding catabolic and anabolic +enzymes, predicted transporters, transcriptional regulators, stress +response, and conserved hypothetical proteins. In most strains, +these differentially expressed genes were almost equally split +between the core and variable genome regions, mirroring +the substantial strain-to-strain variation in the genome size +and gene content within the P. fluorescens species complex +(Loper et al., 2012). +The analysis of transcriptome responses to root exudates +revealed several types of cellular pathways present in the strains +used in this study. The first category of such pathways was +involved in the catabolism of carbohydrates such as fructose, +arabinose, myo-inositol, xylose, trehalose, and galactose. Among +these catabolic traits, the ability to utilize fructose as a carbon +source is highly conserved among fluorescent pseudomonads. +In contrast, growth on arabinose, myo-inositol, xylose, and +trehalose is variably present and was traditionally used to +differentiate species and biovars within the P. fluorescens +group (Barrett et al., 1986). We speculate that such variably +distributed pathways contribute to the differential affinity of +pseudomonads toward host plants and/or to determine which +strains flourish in response to growing roots and changing +environments. Several independent studies have confirmed +the importance of carbohydrate catabolism pathways for the +biology of rhizosphere pseudomonads. For example, in vivo +expression technology (IVET) profiling of P. fluorescens SBW25 +identified xylose isomerase among genome regions essential +for the colonization of sugar beet seedlings (Liu et al., 2015), +whereas a genome-wide Tn-Seq screen of Pseudomonas simiae +identified genes for the catabolism of myo-inositol among traits +essential for the colonization of Arabidopsis thaliana roots +(Cole et al., 2017). +The response of rhizosphere Pseudomonas to Brachypodium +root exudates also involved pathways for the uptake and +metabolism of amino acids. We observed differential +expression of genes encoding the hydrophobic (HAAT) and +polar (PAAT) amino acid uptake transporters in strains +2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes +encoded enzymes for the catabolism of valine and glutamic +acid (2-79); metabolism of tryptophan, glycine, serine, and +threonine (Q2-87); and biosynthesis of methionine (Q8r1- +96). It is plausible that the abundance of amino acids in +root exudates is also linked to the repression of pathways +involved in the catabolism of sulfonates and taurine that +was observed in several strains (Table 2). Although the +preferred source of sulfur for P. fluorescens is unknown, in the +closely related P. aeruginosa, the sulfur starvation response +is triggered by the growth on any sulfur compound other +than sulfate, thiocyanate, and cysteine (Hummerjohann et al., +1998). This fact, together with the presence of cysteine and +cystine in the root exudates, suggest that root exudates of +Brachypodium may serve as an important source of sulfur for +rhizosphere Pseudomonas. These findings also agree well with +the reported scarcity of inorganic sulfate in the soil, and the +presence of sulfur mostly in the form of organic compounds, +including amino acids, proteins, sulfate esters, and sulfonates +(Autry and Fitzgerald, 1990). +Another interesting result of this study was the concerted +activation of copper and iron homeostasis pathways observed +in all of the Pseudomonas strains used in this work. In bacteria, +an excess of copper is toxic and triggers oxidative stress due to +the formation of free radicals, as well as disruption of protein +metalation and stability of iron-sulfur clusters (Bondarczuk +and Piotrowska-Seget, 2013). On the other hand, copper is an +essential trace element used as a cofactor in different enzymes. +Similarly, although elevated levels of iron cause redox stress, this +element is also found in active energy metabolism enzymes and is +crucial for bacterial growth (Andrews et al., 2003). The analysis +of metal homeostasis genes identified in this study suggests +that their induction was likely triggered by the deficiency of +copper and iron in bacterial cultures grown in the presence of +root exudates. We attribute this effect to the ability of some +components of root exudates to chelate soil metals. +Despite the abundance of iron in the soil, its bioavailability +is limited due to the low solubility of Fe (III) oxyhydrates +at neutral pH. The non-graminaceous plants circumvent this +problem by acidifying the rhizosphere and secreting flavins, +phenolics, and organic acids that chelate iron. The reduction +of these ferric chelates releases soluble ferrous iron taken up +by root cells (Kobayashi and Nishizawa, 2012). Graminaceous +plants, like Brachypodium, acquire iron by secreting into the +soil non-protein amino acids of the mugineic acid (MA) group, +which act as Fe (III)-chelating phytosiderophores. In addition to +iron, low-molecular-weight organic acids and phytosiderophores +bind other divalent and trivalent metals (including copper) and +contribute to heavy-metal tolerance in plants (Chen et al., 2017). +It is plausible that the presence of these plant exometabolites +is responsible for the deficit of iron and copper observed in +Pseudomonas cultures grown in the presence of root exudates. +These results further underscore the importance of diverse and +redundant metal homeostasis pathways found in genomes of the +P. fluorescens group for the ability of these organisms to colonize +and persist in the plant rhizosphere. +Recently, Klonowska et al. (2018) examined transcriptomic +responses of symbiotic nitrogen-fixing bacteria to root +exudates of the legume plant Mimosa pundica, which has +an unusual ability to support both alpha- (Rhizobium) and +beta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq, +the authors characterized genes involved in the perception of +root exudates in the nodulating bacteria Burkholderia phymatum +Frontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +STM815, Cupriavidus taiwanensis LMG19424, and Rhizobium +mesoamericanum STM3625. Interestingly, the analysis of +differentially expressed genes revealed induction of pathways +involved in the catabolism of fructose, xylose, myo-inositol, +and protocatechuate/catechol. Also upregulated were some +copper homeostasis, siderophore biosynthesis, and oxidative +stress genes. Finally, the analytical profiling of M. pundica +exudates revealed an overlap with Brachypodium in the types of +carbohydrates, amino acids, and organic acids present. These +findings suggest that differentially expressed genes shared by +multiple strains of the group P. fluorescens are not unique to +the Brachypodium-Pseudomonas system but represent a set +of conserved cellular pathways involved in the perception of +plant exometabolites by different clades of rhizosphere-dwelling +Proteobacteria. +Most strains included in this study were originally selected +based on the ability to colonize the rhizosphere and produce +secondary metabolites that alleviate the plant stress response +and/or inhibit soilborne pathogens. It has been suggested +that plant metabolites released into the rhizosphere affect +the biocontrol activity of plant-beneficial pseudomonads (de +Werra et al., 2011). We provide further support to this +hypothesis by demonstrating that in some strains, root exudates +modulate the expression of genes for the catabolism of +the plant growth-promoting metabolites acetoin and 2,3- +butanediol. The exposure to exudates also affected the expression +of genes for the synthesis of well-characterized antifungal +compounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4- +diacetylphloroglucinol. The modulatory effects were strainspecific, +suggesting significant differences in the regulatory +networks involved in the perception of plant signals and +regulation of the production of antibiotics and growthpromoting +metabolites. +The final significant finding of this study was the induction +of catabolism of quaternary amines (QAs) observed in multiple +strains of the P. fluorescens group during growth on root +exudates. This observation was supported by the detection of +glycine betaine in the root secretions of B. distachyon. The +presence of QAs in plant tissues and the capacity of these +metabolites to provide stress protection and nutrients to plant +pathogens and symbionts were reported before (Boncompagni +et al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our +study is among the first to highlight the potential importance of +these metabolites for rhizosphere interactions. Pseudomonads do +not synthesize QAs de novo but have evolved many pathways to +scavenge them from eukaryotic hosts, where these metabolites +are abundant due to the prominence of phosphatidylcholine +in cellular membranes. Strains of P. fluorescens carry genes +for the conversion of choline, carnitine, and glycine betaine to +glycine, as well as quaternary amine transporters of the BCCT +and ABC families that are also conserved in the opportunistic +human pathogen P. aeruginosa and the plant pathogen P. syringae +(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b). +In P. aeruginosa, choline catabolism genes are essential for +the ability of this pathogen to persist during lung infection +(Wargo, 2013a). Similarly, a P. syringae mutant deficient in +BetT, OpuC, and CbcXWV quaternary amine transporters had +reduced fitness during colonization of bean and soybean leaves +under greenhouse and field conditions (Chen et al., 2013). +Depending on water availability, P. aeruginosa and P. syringae +catabolize exogenously supplied QAs as carbon and nitrogen +sources or accumulate them as osmoprotectants (Chen et al., +2013; Wargo, 2013b). Our ongoing work in P. synxantha 2– +79 unraveled similar physiological responses and demonstrated +that QA transporters function differentially and redundantly +in the uptake of quaternary amines as nutrients (Pablo and +Mavrodi, unpublished). In contrast, under water stress, the QAs +choline, betaine, and carnitine are accumulated preferentially for +osmoprotection. Under drought stress, a 2–79 mutant devoid +of all known QA transporters was less competitive in the +colonization of the Brachypodium rhizosphere than its wild-type +parental strain. Interestingly, our metabolomic profiling of root +exudates also revealed proline, glutamine, and hydroxyectoine. +These metabolites act as compatible solutes in different groups +of microorganisms (Yancey et al., 1982; Empadinhas and da +Costa, 2008), suggesting an important role of root exudates +in the ability of Pseudomonas to persist in the rhizosphere of +drought-stressed plants. +DATA AVAILABILITY STATEMENT +The original contributions presented in the study are publicly +available. This data can be found here: NCBI BioProject accession +numbers PRJNA439743 through PRJNA439790. +AUTHOR CONTRIBUTIONS +DM, OM, and LT conceived the research project. OM and +JM collected root exudates. OM and DM cultured strains +and extracted total RNA. AB and DG performed metabolomic +analysis of root exudates. DM, JP, and AF analyzed RNA-seq +data. LE, KH, and IP conducted Biolog analyses. DM, AF, OM, +DW, and LT wrote the manuscript. All authors contributed to the +manuscript revision. +FUNDING +This study was funded by NSF grant IOS-1656872 and by an +award from the DOE Joint Genome Institute’s Community +Science Program. The authors also acknowledge support from +Australian Research Council Discovery grant (DP160103746) +and Mississippi INBRE, funded by an Institutional Development +Award (IDeA) from the National Institute of General +Medical Sciences of the National Institutes of Health under +grant P20GM103476. +SUPPLEMENTARY MATERIAL +The Supplementary Material for this article can be found online +at: https://www.frontiersin.org/articles/10.3389/fmicb.2021. +651282/full#supplementary-material +Frontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +REFERENCES +Abayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018). +Discovery and characterization of a sulfoquinovose mutarotase using kinetic +analysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383. +doi: 10.1042/bcj20170947 +Andrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron +homeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03) +00055-x +Arkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov, +S., et al. (2018). KBase: the United States department of energy systems biology +knowledgebase. Nat. Biotechnol. 36, 566–569. +Autry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil +organic sulfur. Biol. Fertil. Soils 10, 50–56. +Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al. +(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44, +W147–W153. +Badri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates. +Plant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x +Badri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere +chemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20, +642–650. doi: 10.1016/j.copbio.2009.09.014 +Bais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005). +Mediation of pathogen resistance by exudation of antimicrobials from roots. +Nature 434, 217–221. doi: 10.1038/nature03356 +Bais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of +root exudates in rhizosphere interations with plants and other organisms. Annu. +Rev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159 +Bangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic +locus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the +biological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe +Interact. 9, 83–90. +Barret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G., +and Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic +fungus Gaeumannomyces graminis var. tritici on gene expression of the +biocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe +Interact. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611 +Barrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas +fluorescens biovar V: its resolution into distinct component groups and the +relationship of these groups to other P. fluorescens biovars, to P. putida, and to +psychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol. +132, 2709–2721. doi: 10.1099/00221287-132-10-2709 +Bevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon +genomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21, +211–217. doi: 10.1016/j.copbio.2010.03.006 +Boncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999). +Occurrence of choline and glycine betaine uptake and metabolism in the family +Rhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65, +2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999 +Bondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper +resistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397– +405. doi: 10.1007/s10565-013-9262-1 +Brkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al. +(2011). Brachypodium as a model for the grasses: today and the future. Plant +Physiol. 157, 3–13. +Camacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type +4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens +WCS365. Dissertation, University of Leiden, Leiden. +Chavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A +metabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx +in Pseudomonas putida. mSystems 1:e00154-16. +Chen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread +plant-colonizing bacterial species Pseudomonas syringae detects and exploits an +extracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj. +12262 +Chen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb: +a web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi: +10.3389/fmicb.2018.01910 +Chen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal +acquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi. +2017.06.004 +Cole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M., +et al. (2017). Genome-wide identification of bacterial plant colonization genes. +PLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860 +Conesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional +analysis in plant genomics. Int. J. Plant Genomics 2008:619832. +Curl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag. +Darzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive +pathways explorer v3. Nucleic Acids Res. 46, W510–W513. +de Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al. +(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365 +plays a role in rhizosphere competence through maintaining the structure and +function of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi: +10.1111/j.1574-6941.2006.00158.x +de Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg, +G. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is +an important trait for tomato root colonization by Pseudomonas fluorescens. +Mol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11. +1173 +de Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand +microbe-derived compounds affect the expression of genes encoding +antifungal compounds in a pseudomonad with biocontrol activity. Appl. +Environ. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10 +Dekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998). +A site-specific recombinase is required for competitive root colonization by +Pseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056. +doi: 10.1073/pnas.95.12.7051 +De-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008). +Root-microbe communication through protein secretion. J. Biol. Chem. 283, +25247–25255. doi: 10.1074/jbc.m801967200 +Empadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms +in prokaryotes: distribution of compatible solutes. Int. Microbiol. 11, +151–161. +Espinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of +functions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182, +2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000 +Fuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in +Pseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365- +2958.2010.07250.x +Galvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline- +O-sulphate utilization from osmoprotection in Pseudomonas putida. Mol. +Microbiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x +Garrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R., +and Redondo-Nieto, M. (2016). Genomic and genetic diversity within the +Pseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal. +pone.0150183 +Halverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and +nonpermeating solutes on the fatty acid composition of Pseudomonas putida. +Appl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421. +2000 +Hampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and +Wargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas +aeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13 +Harwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and +the biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/ +annurev.micro.50.1.553 +Hesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018). +Genome-based evolutionary history of Pseudomonas spp. Environ. Microbiol. +20, 2142–2159. +Hinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003). +Transition from reversible to irreversible attachment during biofilm formation +by Pseudomonas fluorescens WCS365 requires an ABC transporter and a large +secreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003. +03615.x +Hojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999). +Oxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring +Frontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +the distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65, +4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999 +Hong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological +growth stages of Brachypodium distachyon: codification and description. Weed +Res. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x +Howell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum +induced damping-off of cotton seedlings by Pseudomonas fluorescens and +its antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto- +70-712 +Hummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz, +M. A. (1998). Regulation of the sulfate starvation response in Pseudomonas +aeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375– +1386. doi: 10.1099/00221287-144-5-1375 +Kabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam, +C. (2017). Physiological and biochemical responses involved in water deficit +tolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/ +journal.pone.0190284 +Kamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005). +Enrichment for enhanced competitive plant root tip colonizers selects for a new +class of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j. +1462-2920.2005.00889.x +Kanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008). +KEGG for linking genomes to life and the environment. Nucleic Acids Res. 36, +D480–D484. +Katoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment +software version 7: improvements in performance and usability. Mol. Biol. Evol. +30, 772–780. doi: 10.1093/molbev/mst010 +Kawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al. +(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium +distachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal. +pone.0164533 +Kind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al. +(2009). FiehnLib: mass spectral and retention index libraries for metabolomics +based on quadrupole and time-of-flight gas chromatography/mass +spectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522 +King, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the +demonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307. +Klonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018). +Transcriptomic profiling of Burkholderia phymatum STM815, Cupriavidus +taiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response +to Mimosa pudica root exudates illuminates the molecular basis of their +nodulation competitiveness and symbiotic evolutionary history.BMC Genomics +19:105. doi: 10.1186/s12864-018-4487-2 +Kobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and +regulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/ +annurev-arplant-042811-105522 +Kohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor +IolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193, +5155–5163. doi: 10.1128/jb.05371-11 +Lee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas +reinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7 +Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for +the display and annotation of phylogenetic and other trees. Nucleic Acids Res. +44, W242–W245. +Li, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of +ortholog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi: +10.1101/gr.1224503 +Liao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient +general purpose program for assigning sequence reads to genomic features. +Bioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656 +Liu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose +utilization by Pseudomonas fluorescens: overlapping genetic responses to xylose, +xylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi. +13142 +Loper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer, +B. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas +spp.: insights into diversity and inheritance of traits involved in multitrophic +interactions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784 +Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change +and dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550. +Lugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria. +Annu. Rev. Microbiol. 63, 541–556. +Lugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular +determinants of rhizosphere colonization by Pseudomonas. Annu. Rev. +Phytopathol. 39, 461–490. +Lynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch +(Chichester: JohnWiley & Sons), 177–206. +Mark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005). +Transcriptome profiling of bacterial responses to root exudates identifies genes +involved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454– +17459. doi: 10.1073/pnas.0506407102 +Markowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y., +et al. (2012). IMG: the integrated microbial genomes database and comparative +analysis system. Nucleic Acids Res. 40, D115–D122. +Martinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF, +the second largest Pseudomonas putida protein, contributes to plant root +colonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561. +doi: 10.1111/j.1365-2958.2010.07249.x +Matilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and +Ramos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving +forces of bacterial life in the rhizosphere. Genome Biol. 8:R179. +Miller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria. +Annu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101 +Moore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.- +L., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes, +eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt +(New York, NY: Springer), 646–703. +Mulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis +of the Pseudomonas species. Environ. Microbiol. 12, 1513–1530. +Naylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought +and host selection influence bacterial community dynamics in the grass root +microbiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118 +Nguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and +controls. Agronomie 23, 375–396. doi: 10.1051/agro:2003011 +Nielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface +interactions mediated by cellulose and a novel exopolysaccharide contribute +to Pseudomonas putida biofilm formation and fitness under water-limiting +conditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011. +02432.x +Parejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow, +L. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid +producing fluorescent Pseudomonas spp. from dryland cereal fields of central +Washington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012- +0015-0 +Phillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R. +(2004). Microbial products trigger amino acid exudation from plant roots. Plant +Physiol. 136, 2887–2894. doi: 10.1104/pp.104.044222 +Raaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek, +P. J., and Schippers, B. (1995). Utilization of heterologous siderophores and +rhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41, +126–135. doi: 10.1139/m95-017 +Raaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4- +diacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils. +Mol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144 +Rahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis, +J., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase +contains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/ +acschembio.7b00304 +Ramos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of +Pseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo +expression technology capture and identification of root-activated promoters. +J. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005 +Reinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T. +(2015). Roots shaping their microbiome: global hotspots for microbial activity. +Annu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10 +2342 +Frontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282 +Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome +Sanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and +Rivilla, R. (2002). Phenotypic selection and phase variation occur during +alfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184, +1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002 +Sarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995). +The sigma factor ss affects antibiotic production and biological control activity +of Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259. +doi: 10.1073/pnas.92.26.12255 +Schnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The +sigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance +towards desiccation and osmotic stress in the biocontrol agent Pseudomonas +fluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem. +67.12.5683-5693.2001 +Schroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006). +“Phytopathogenic pseudomonads and related plant-associated pseudomonads,” +in The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.- +H. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740. +doi: 10.1007/0-387-30746-x_23 +Schwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and +Amasino, R. M. (2010). Natural variation of flowering time and vernalization +responsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10. +1007/s12155-009-9069-3 +Silby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson, +R. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity +and plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51. +Silby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to +identify genes important in growth and survival of Pseudomonas fluorescens +Pf0-1 in soil: Discovery of expressed sequences with novel genetic organization. +J. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004 +Simons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A., +and Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere +colonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant +Microbe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600 +Simons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and +Lugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root +colonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe +Interact. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102 +Smibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods +for General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A. +Wood, and N. R. Kreig (Washington, DC: American Society of Microbiology), +607–654. +Thomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from +Pseudomonas fluorescens in biological control of Gaeumannomyces graminis +var. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988 +Thomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990). +Production of the antibiotic phenazine-1-carboxylic acid by fluorescent +Pseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56, +908–912. doi: 10.1128/aem.56.4.908-912.1990 +Tyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats, +W. G., et al. (2014). Selection and phenotypic characterization of a core +collection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25. +doi: 10.1186/1471-2229-14-25 +Vacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y., +Muller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system +functioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356 +van den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of +phenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol. +7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x +van Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity +of microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135. +doi: 10.1128/.61.2.121-135.1997 +Walker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M. +(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric. +Food Chem. 51, 2548–2554. +Wargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to +Pseudomonas aeruginosa survival during murine lung infection. PLoS One +8:e56850. doi: 10.1371/journal.pone.0056850 +Wargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine: +lessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112– +2120. +Whipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch +(Chichester: JohnWiley & Sons), 59–97. +Winsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock, +R. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly, +comprehensive comparisons of microbial genomes. Nucleic Acids Res. 37, +D483–D488. +Yahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The +Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. +Stackebrandt (New York, NY: Springer), 704–713. +Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982). +Living with water stress: evolution of osmolyte systems. Science 217, 1214–1222. +Ye, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web +tool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res. +46, W71–W75. +Yoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H., +et al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283, +10415–10424. +Yousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different, +overlapping mechanisms for colonization of abiotic and plant surfaces by +Pseudomonas putida. FEMS Microbiol. Lett. 288, 118–124. +Zboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere +colonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol. +J. 18, 3539–3554. +Zolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K., +et al. (2013). “Understanding root-microbiome interactions,” in Molecular +Microbial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John +Wiley & Sons), 745–754. +Conflict of Interest: The authors declare that the research was conducted in the +absence of any commercial or financial relationships that could be construed as a +potential conflict of interest. +Copyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne, +LeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an +open-access article distributed under the terms of the Creative Commons Attribution +License (CC BY). The use, distribution or reproduction in other forums is permitted, +provided the original author(s) and the copyright owner(s) are credited and that the +original publication in this journal is cited, in accordance with accepted academic +practice. No use, distribution or reproduction is permitted which does not comply +with these terms. +Frontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282 + + +INFO:ontogpt.io.template_loader:Loading custom schema from src/ontogpt/templates/biolog.yaml +INFO:linkml_runtime.utils.schemaview:Importing core as core from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates +INFO:linkml_runtime.utils.schemaview:Importing linkml:types as /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/linkml_runtime/linkml_model/model/schema/types from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=None +INFO:linkml_runtime.utils.schemaview:Importing core as core from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates +INFO:linkml_runtime.utils.schemaview:Importing linkml:types as /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/linkml_runtime/linkml_model/model/schema/types from source /Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml; base_dir=None +INFO:ontogpt.io.template_loader:Getting class for template src/ontogpt/templates/biolog.yaml +INFO:root:Using template Paper +INFO:root:Using mappers (currently hardcoded) +INFO:ontogpt.clients.llm_client:Complete: engine=gpt-4.5-preview, prompt[100766]=From the text below, extract the following entities in the following format: + +study_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +DEBUG:LiteLLM:litellm.completion(api_key='', api_base=None, api_version=None, model='gpt-4.5-preview', messages=[{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +23:25:35 - LiteLLM:DEBUG: utils.py:296 - + +DEBUG:LiteLLM: + +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {} +DEBUG:LiteLLM:self.optional_params: {} +23:25:35 - LiteLLM:DEBUG: utils.py:296 - SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +DEBUG:LiteLLM:SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +23:25:35 - LiteLLM:DEBUG: utils.py:925 - INSIDE CHECKING SYNC CACHE +DEBUG:LiteLLM:INSIDE CHECKING SYNC CACHE +23:25:35 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:25:35 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:25:35 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:25:35 - LiteLLM:INFO: utils.py:2926 - +LiteLLM completion() model= gpt-4.5-preview; provider = openai +INFO:LiteLLM: +LiteLLM completion() model= gpt-4.5-preview; provider = openai +23:25:35 - LiteLLM:DEBUG: utils.py:2929 - +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +DEBUG:LiteLLM: +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +23:25:35 - LiteLLM:DEBUG: utils.py:2932 - +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +DEBUG:LiteLLM: +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +23:25:35 - LiteLLM:DEBUG: utils.py:296 - Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:self.optional_params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:httpx:load_ssl_context verify=True cert=None trust_env=True http2=False +DEBUG:httpx:load_verify_locations cafile='/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/venv/lib/python3.10/site-packages/certifi/cacert.pem' +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:614 - PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +DEBUG:LiteLLM:PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +23:25:35 - LiteLLM:DEBUG: litellm_logging.py:624 -  + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:LiteLLM: + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:openai._base_client:Request options: {'method': 'post', 'url': '/chat/completions', 'headers': {'X-Stainless-Raw-Response': 'true'}, 'timeout': 600.0, 'files': None, 'json_data': {'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}], 'model': 'gpt-4.5-preview', 'temperature': 1.0}, 'extra_json': {}} +DEBUG:openai._base_client:Sending HTTP Request: POST https://api.openai.com/v1/chat/completions +DEBUG:httpcore.connection:connect_tcp.started host='api.openai.com' port=443 local_address=None timeout=600.0 socket_options=None +DEBUG:httpcore.connection:connect_tcp.complete return_value= +DEBUG:httpcore.connection:start_tls.started ssl_context= server_hostname='api.openai.com' timeout=600.0 +DEBUG:httpcore.connection:start_tls.complete return_value= +DEBUG:httpcore.http11:send_request_headers.started request= +DEBUG:httpcore.http11:send_request_headers.complete +DEBUG:httpcore.http11:send_request_body.started request= +DEBUG:httpcore.http11:send_request_body.complete +DEBUG:httpcore.http11:receive_response_headers.started request= +DEBUG:httpcore.http11:receive_response_headers.complete return_value=(b'HTTP/1.1', 200, b'OK', [(b'Date', b'Wed, 05 Mar 2025 07:26:07 GMT'), (b'Content-Type', b'application/json'), (b'Transfer-Encoding', b'chunked'), (b'Connection', b'keep-alive'), (b'access-control-expose-headers', b'X-Request-ID'), (b'openai-organization', b'lawrence-berkeley-national-laboratory-8'), (b'openai-processing-ms', b'32096'), (b'openai-version', b'2020-10-01'), (b'x-ratelimit-limit-requests', b'10000'), (b'x-ratelimit-limit-tokens', b'2000000'), (b'x-ratelimit-remaining-requests', b'9999'), (b'x-ratelimit-remaining-tokens', b'1974681'), (b'x-ratelimit-reset-requests', b'6ms'), (b'x-ratelimit-reset-tokens', b'759ms'), (b'x-request-id', b'req_46b2bea89ddaddbe030dfd435c97dfa2'), (b'strict-transport-security', b'max-age=31536000; includeSubDomains; preload'), (b'cf-cache-status', b'DYNAMIC'), (b'Set-Cookie', b'__cf_bm=xWhtGZy4GkjfLHyiSPTezcmRYi.uvd.uXbx4Q4RzppU-1741159567-1.0.1.1-sOAa0JWsrXDlqfH0FF7KxXc__jUyHH4CoWDjTdrs03xP7R0RbK4IjCesizfy0YvXAH1o1bkXoyGY1JTx4kdcH8ni4FKVk1m.xv65nMtWwG4; path=/; expires=Wed, 05-Mar-25 07:56:07 GMT; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), (b'X-Content-Type-Options', b'nosniff'), (b'Set-Cookie', b'_cfuvid=oBEVFNLmxi3159bRP_cghDOOkjcknNYaVoFV5G375yo-1741159567673-0.0.1.1-604800000; path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), (b'Server', b'cloudflare'), (b'CF-RAY', b'91b7e158aba0ce4c-SJC'), (b'Content-Encoding', b'gzip'), (b'alt-svc', b'h3=":443"; ma=86400')]) +INFO:httpx:HTTP Request: POST https://api.openai.com/v1/chat/completions "HTTP/1.1 200 OK" +DEBUG:httpcore.http11:receive_response_body.started request= +DEBUG:httpcore.http11:receive_response_body.complete +DEBUG:httpcore.http11:response_closed.started +DEBUG:httpcore.http11:response_closed.complete +DEBUG:openai._base_client:HTTP Response: POST https://api.openai.com/v1/chat/completions "200 OK" Headers([('date', 'Wed, 05 Mar 2025 07:26:07 GMT'), ('content-type', 'application/json'), ('transfer-encoding', 'chunked'), ('connection', 'keep-alive'), ('access-control-expose-headers', 'X-Request-ID'), ('openai-organization', 'lawrence-berkeley-national-laboratory-8'), ('openai-processing-ms', '32096'), ('openai-version', '2020-10-01'), ('x-ratelimit-limit-requests', '10000'), ('x-ratelimit-limit-tokens', '2000000'), ('x-ratelimit-remaining-requests', '9999'), ('x-ratelimit-remaining-tokens', '1974681'), ('x-ratelimit-reset-requests', '6ms'), ('x-ratelimit-reset-tokens', '759ms'), ('x-request-id', 'req_46b2bea89ddaddbe030dfd435c97dfa2'), ('strict-transport-security', 'max-age=31536000; includeSubDomains; preload'), ('cf-cache-status', 'DYNAMIC'), ('set-cookie', '__cf_bm=xWhtGZy4GkjfLHyiSPTezcmRYi.uvd.uXbx4Q4RzppU-1741159567-1.0.1.1-sOAa0JWsrXDlqfH0FF7KxXc__jUyHH4CoWDjTdrs03xP7R0RbK4IjCesizfy0YvXAH1o1bkXoyGY1JTx4kdcH8ni4FKVk1m.xv65nMtWwG4; path=/; expires=Wed, 05-Mar-25 07:56:07 GMT; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), ('x-content-type-options', 'nosniff'), ('set-cookie', '_cfuvid=oBEVFNLmxi3159bRP_cghDOOkjcknNYaVoFV5G375yo-1741159567673-0.0.1.1-604800000; path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None'), ('server', 'cloudflare'), ('cf-ray', '91b7e158aba0ce4c-SJC'), ('content-encoding', 'gzip'), ('alt-svc', 'h3=":443"; ma=86400')]) +DEBUG:openai._base_client:request_id: req_46b2bea89ddaddbe030dfd435c97dfa2 +23:26:07 - LiteLLM:DEBUG: utils.py:296 - RAW RESPONSE: +{"id": "chatcmpl-B7dQFT52GJFSYAsipHD1MZCWpm98U", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas\nauthors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi\ndoi: 10.3389/fmicb.2021.651282\ndate: 14 April 2021\nexperiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C). Root exudates were extracted, sterilized with 0.22 \u00b5m filtration, lyophilized, and stored at \u201380\u00b0C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25\u00b0C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159535, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 687, "prompt_tokens": 30161, "total_tokens": 30848, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 30080}}} + + +DEBUG:LiteLLM:RAW RESPONSE: +{"id": "chatcmpl-B7dQFT52GJFSYAsipHD1MZCWpm98U", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas\nauthors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi\ndoi: 10.3389/fmicb.2021.651282\ndate: 14 April 2021\nexperiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C). Root exudates were extracted, sterilized with 0.22 \u00b5m filtration, lyophilized, and stored at \u201380\u00b0C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25\u00b0C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159535, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 687, "prompt_tokens": 30161, "total_tokens": 30848, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 30080}}} + + +23:26:07 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nstudy_title: \nauthors: \ndoi: \ndate: \nexperiments: \n\n\nText:\nORIGINAL RESEARCH\npublished: 14 April 2021\ndoi: 10.3389/fmicb.2021.651282\nEdited by:\nBarbara Pivato,\nInstitut National de Recherche pour\nl’agriculture, l’alimentation et\nl’environnement (INRAE), France\nReviewed by:\nCara Helene Haney,\nThe University of British Columbia,\nCanada\nXingang Zhou,\nNortheast Agricultural University,\nChina\n*Correspondence:\nDmitri V. Mavrodi\ndmitri.mavrodi@usm.edu\nAlex S. Flynt\nalex.flynt@usm.edu\n†These authors have contributed\nequally to this work\nSpecialty section:\nThis article was submitted to\nMicrobe and Virus Interactions with\nPlants,\na section of the journal\nFrontiers in Microbiology\nReceived: 09 January 2021\nAccepted: 08 March 2021\nPublished: 14 April 2021\nCitation:\nMavrodi OV, McWilliams JR,\nPeter JO, Berim A, Hassan KA,\nElbourne LDH, LeTourneau MK,\nGang DR, Paulsen IT, Weller DM,\nThomashow LS, Flynt AS and\nMavrodi DV (2021) Root Exudates\nAlter the Expression of Diverse\nMetabolic, Transport, Regulatory,\nand Stress Response Genes\nin Rhizosphere Pseudomonas.\nFront. Microbiol. 12:651282.\ndoi: 10.3389/fmicb.2021.651282\nRoot Exudates Alter the Expression\nof Diverse Metabolic, Transport,\nRegulatory, and Stress Response\nGenes in Rhizosphere Pseudomonas\nOlga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2,\nKarl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2,\nIan T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and\nDmitri V. Mavrodi1*\n1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS,\nUnited States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School\nof Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular\nSciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics\nand Quality Research Unit, Pullman, WA, United States\nPlants live in association with microorganisms that positively influence plant\ndevelopment, vigor, and fitness in response to pathogens and abiotic stressors.\nThe bulk of the plant microbiome is concentrated belowground at the plant rootsoil\ninterface. Plant roots secrete carbon-rich rhizodeposits containing primary and\nsecondary low molecular weight metabolites, lysates, and mucilages. These exudates\nprovide nutrients for soil microorganisms and modulate their affinity to host plants, but\nmolecular details of this process are largely unresolved. We addressed this gap by\nfocusing on the molecular dialog between eight well-characterized beneficial strains\nof the Pseudomonas fluorescens group and Brachypodium distachyon, a model for\neconomically important food, feed, forage, and biomass crops of the grass family. We\ncollected and analyzed root exudates of B. distachyon and demonstrated the presence\nof multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The\nsubsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many\nof these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq\nprofiling of bacterial cultures amended with root exudates revealed changes in the\nexpression of genes encoding numerous catabolic and anabolic enzymes, transporters,\ntranscriptional regulators, stress response, and conserved hypothetical proteins. Almost\nhalf of the differentially expressed genes mapped to the variable part of the strains’\npangenome, reflecting the importance of the variable gene content in the adaptation of\nP. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity\nof cellular pathways and physiological responses underlying the establishment of\nmutualistic interactions between these beneficial rhizobacteria and their plant hosts.\nKeywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome\nFrontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nINTRODUCTION\nPlants are meta-organisms or holobionts that rely in part on\ntheir microbiome for specific functions and traits. The ability\nof the plant microbiome to influence plant development, vigor,\nhealth, and fitness in response to abiotic stressors associated\nwith global climate change is documented by numerous studies\n(Lugtenberg and Kamilova, 2009). There is mounting evidence\nthat plants actively recruit beneficial microbiomes, but many\naspects of this process are still very much a black box (Reinhold-\nHurek et al., 2015). The foundation for the differential affinity\nof rhizobacteria toward host plants is built upon complex\nchemical cross talk between microorganisms and plant roots.\nUp to 40% of photosynthetically fixed carbon is released by\nplant roots in the form of exudates and secretions, lysates, and\nmucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990;\nBadri and Vivanco, 2009). The release of these compounds is\nactively controlled in response to environmental stimuli, and the\ncomposition of root exudates varies greatly according to plant\nspecies and physiological condition (Lynch, 1990; Nguyen, 2003;\nPhillips et al., 2004; De-la-Pena et al., 2008). The presence and\ncomposition of exudates strongly impact soil microorganisms,\nwhich is consistent with the idea that plants actively select and\nshape their root microbiota (Zolla et al., 2013).\nPrimary root exudates include simple and complex sugars,\namino acids, polypeptides and proteins, organic, aliphatic and\nfatty acids, sterols, and phenolics (Nguyen, 2003; Badri and\nVivanco, 2009; Badri et al., 2009). These compounds serve as\ncarbon and energy sources for rhizobacteria, and the presence\nof the intact corresponding catabolic pathways is essential\nfor competitive colonization of roots and disease suppression\n(Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and\nKamilova, 2009). Root exudates also contain numerous signal\nmolecules and secondary metabolites, the significance of which\nis only now emerging (Walker et al., 2003; Bais et al., 2005,\n2006). A handful of analyses of plant-induced gene expression\nby transcriptional profiling in vitro (Mark et al., 2005) or in\nthe rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al.,\n2005; Matilla et al., 2007; Barret et al., 2009) have identified\nmultiple genes that are differentially regulated by exposure to\nroots or root exudates. Bacterial pathways expressed during\nrhizosphere colonization control utilization of plant-derived\nmetabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001;\nLugtenberg and Kamilova, 2009), motility and chemotaxis (de\nWeert et al., 2002; Lugtenberg and Kamilova, 2009), phase\nvariation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002;\nvan den Broek et al., 2005), outer membrane integrity (de Weert\net al., 2006; Lugtenberg and Kamilova, 2009), and the ability\nto sequester limiting resources (Raaijmakers et al., 1995) and\nresist environmental stresses (Sarniguet et al., 1995; Miller and\nWood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In\nits spatial and temporal properties, root colonization resembles\nbiofilm formation, and biofilm-related pathways also have been\nimplicated in adhesion to seeds and roots and rhizosphere\ncolonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003;\nYousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al.,\n2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally,\nroot exudates strongly affect the expression of diverse plant\ngrowth promotion and biocontrol genes (Vacheron et al., 2013).\nOver the past decade, the genomes of numerous rhizosphere\nstrains have been sequenced and analyzed, but functional\ngenomics studies of rhizosphere competence lag behind the\navailability of sequence data.\nThis study explored the molecular dialog between the\nmodel host plant Brachypodium distachyon and several wellcharacterized\nrhizosphere strains of the Pseudomonas fluorescens\ngroup. Brachypodium is a small annual grass originating in semiarid\nregions of the Middle East that has emerged as a prime model\nfor economically important food, feed, forage, and biomass crops\nof the grass family (Bevan et al., 2010; Schwartz et al., 2010;\nBrkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The\nbiology, extensive collection of resources, and research tools\nmake B. distachyon an attractive model to investigate interactions\nbetween plants and root-associated microbes. Pseudomonads\nare ubiquitous Gram-negative g-proteobacteria that colonize\neukaryotic hosts and include both commensals and economically\nimportant pathogens of plants and animals (Moore et al.,\n2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus\nPseudomonas currently comprises > 100 named species that have\nbeen separated based on multilocus sequence analysis into 14\nspecies groups (Garrido-Sanz et al., 2016; Hesse et al., 2018).\nThe P. fluorescens group is the most diverse regarding both\nthe genetic distances within it, the number of species and the\nlarge pangenome that makes up > 50% of the pangenome\nof the genus as a whole (Loper et al., 2012). The group also\nencompasses an unusually high proportion of strains that inhabit\nthe plant rhizosphere and possess plant growth promoting and\nbiocontrol properties. Naylor et al. (2017) profiled bacterial\ncommunities associated with root tissues and rhizosphere of\n18 different plant species of the Poaceae family. That study\nidentified Pseudomonas among taxa constituting the core grass\nroot microbiome and demonstrated that these bacteria were\nenriched in C3 plants, including wheat, rye, barley, oat, and\nBrachypodium. We confirmed the capacity of B. distachyon\nBd21 to serve as a host for rhizobacteria of the P. fluorescens\ngroup in preliminary greenhouse assays with biocontrol strains\nP. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens\nPf-5. Results of these experiments revealed that all strains\nsuccessfully established and colonized the roots of Brachypodium\n(Supplementary Table 1).\nIn this study, we focused on eight well-studied strains\nof the P. fluorescens complex that are supported by years\nof studies, numerous refereed publications, and high-quality\ngenome sequences. By profiling transcriptomes of these strains\nduring growth in root exudates of B. distachyon, we revealed the\ndiversity of cellular pathways and physiological responses that\nunderlie the establishment of mutualistic interactions between\nbeneficial rhizobacteria and the host plant. Our results also\nconfirmed that root exudates contain carbohydrates, amino acids,\norganic acids, and phenolics that serve as carbon and energy\nsources for rhizobacteria. The root exudates also contained\nosmoprotectants that may help microorganisms to persist in the\nrhizosphere of drought-stressed plants. The diversity of microbial\ngenes perturbed by root exudates reflects the importance of\nFrontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe variable genome in adaptation of individual strains of\nPseudomonas to the rhizosphere lifestyle.\nMATERIALS AND METHODS\nBacterial Strains Used in the Study\nThe eight Pseudomonas strains used for this study are\nP. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens\nSBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko\net al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller,\n1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996),\nP. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens\nPf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and\nStipanovic, 1980). The selected organisms have been studied\nextensively for their role in biological control and plant\ngrowth promotion (Supplementary Table 2). The strains were\nmaintained in the laboratory as frozen stocks (\U0010000080\u2000C) and\nroutinely cultured in King’s medium B (King et al., 1954) or\n21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g\nof Na2HPO4\u20002H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and\n20 ml of a microelement solution (Smibert and Kreig, 1994;\nHalverson and Firestone, 2000).\nPropagation of Plants and Collection of\nRoot Exudates\nB. distachyon Bd21 was established from seed material obtained\nfrom the USDA-ARS Plant Germplasm Introduction and Testing\nResearch Unit (Pullman, WA, United States). Brachypodium\nseeds were imbibed for 3 days at 4\u2000C and sown in 7 \u2000 7 cm\npots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture,\nAgawam, MA, United States). Plants were grown in an IR-\n89X (Percival Scientific, Perry, IA, United States) controlled\nenvironment chamber retrofitted with 6500K and 3000K T5 54W\ngrow lights (Spectralux) under a 20-h light, 24\u2000C/4-h dark, 18\u2000C\ncycle. Plants were watered and fertilized with Jack’s professional\nwater-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA,\nUnited States). After 12 weeks and plant senescence, seeds\nwere collected, processed, and stored under desiccant and dark\nconditions at room temperature.\nTo collect root exudates, seeds of B. distachyon Bd21 were\nsurface-sterilized, pregerminated, and placed in sterile 1 L widemouth\nglass jars containing 113 g of 6-mm glass beads and\n25 ml distilled water. Jars were covered with vented caps\nand plants were grown hydroponically in an environmental\ncontrolled growth chamber under conditions described above.\nAfter 6 days, root exudates were extracted from individual\njars and their sterility was confirmed by spotting on nutrient\nagar. Multiple batches of root exudates were collected, filtered\n(0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and\nstored at –80\u2000C.\nMetabolomic Profiling of Root Exudates\nExudates were analyzed for primary metabolites at the Murdock\nMetabolomics Laboratory at Washington State University\n(Pullman, WA, United States). Freeze-dried residues were\nsuspended in 500 ml 50% aqueous acetonitrile and clarified by\ncentrifugation for 20 min at 21,000 \u2000 g and 4\u2000C. The liquid\nchromatography mass spectrometry analysis was conducted\nwith a Synapt G2-S quadrupole-ion mobility spectrometrytime\nof flight mass spectrometer system equipped with an\nacquity ultra-performance liquid chromatograph (UPLC) and\nan acquity photodiode array detector (all from Waters, Milford,\nMA, United States). The exudate metabolites were separated\non a SeQuant ZIC-pHILIC HPLC column (2.1 \u2000 100 mm,\n3 mm) (Milllipore Sigma, Burlington, MA, United States) using\nacetonitrile with 0.1% formic acid as solvent B and water\nwith 0.1% formic acid as solvent A at a flow rate of 400 ml\nmin\U001000001 and the following linear gradient extending over 14 min:\n0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10%\nB; 10 min, 80% B; and 14 min, 80% B. Mass spectra were\ncollected in positive ion mode over a range of m/z 50–1,200\nwith a scan time of 0.2 s. The Q-TOF-MS source was at\n3.0 kV and 120\u2000C; the sampling cone at 40 V, desolvation\ntemperature was 250\u2000C; cone gas and desolvation gas flow were\nat 0 and 850 L h\U001000001, respectively. Leucine enkephalin was used\nfor post-acquisition mass correction. Target compounds were\nvisualized using selected ion chromatograms at 0.05 Da window\nwidth. The compound identification was based on comparison\nof chromatographic behavior and accurate masses to those of\nauthentic standards.\nFor gas chromatography, derivatization was carried out using\na modification of the procedure of Lee and Fiehn (2008).\nThe freeze-dried residues were suspended in 950 ml aqueous\nmethanol (84%, v/v) and clarified by centrifugation for 15 min\nat 21,000 \u2000 g at 4\u2000C. The supernatants were spiked with 1\nmg of the internal standard salicylic acid-d6 (C/D/N Isotopes,\nQuebec, Canada) and dried in vacuo. The dry residues were\nsuspended in 10 ml of O-methoxylamine hydrochloride (30 mg\nml\U001000001 in anhydrous pyridine, both from Millipore Sigma) and\nincubated while mixing (1,000 RPM) for 90 min at 30\u2000C.\nSubsequently, samples were derivatized with 90 ml of MSTFA\nwith 1% TMCS (Thermo Fisher Scientific, Waltham, MA,\nUnited States) for 30 min at 37\u2000C. Gas chromatography-mass\nspectroscopy analysis was performed using a Pegasus 4D timeof-\nflight mass spectrometer (LECO, Saint Joseph MI) equipped\nwith a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A\noven (Agilent Technologies, Santa Clara, CA, United States).\nThe derivatization products were separated on a 30-m, 0.25 mm\ni.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA,\nUnited States) with an IntegraGuard precolumn using ultrapure\nHe at a constant flow of 0.9 ml min\U001000001 as carrier gas. The linear\nthermal gradient started with a 1-min hold at 70\u2000C, followed\nby a ramp to 300\u2000C at 10\u2000C min\U001000001. The final temperature was\nheld for 5 min prior to returning to initial conditions. Mass\nspectra were collected at 17 spectra s\U001000001. Peak identification was\nconducted using the Fiehn primary metabolite library (Kind\net al., 2009) and an identity score cutoff of 700. Additionally,\nauthentic standards for a number of primary metabolites were\nanalyzed under identical conditions and the data used to compare\nthe chromatographic behavior. Peak alignment and spectrum\ncomparisons were carried out using the Statistical Compare\nfeature of ChromaTOF software (LECO).\nFrontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nIsolation of RNA From Bacteria Cultured\nin Root Exudates and RNA-Seq\nThe strains were pregrown overnight at 25\u2000C on 21C-glucose\nagar and then subcultured into 96-well microplates containing\nliquid 21C-glucose medium amended with Brachypodium\nexudates. The liquid medium was prepared by dissolving the\nlyophilized root exudate material in an appropriate volume of\n21C-glucose medium to concentrate root exudates 20-fold. The\ngrowth medium was sterilized by passing it through a 0.22-\nmm membrane filter. The control cultures were grown under\nidentical conditions in the absence of exudates. All treatments\nwere inoculated at OD600 of 0.1 and incubated for 20 to\n22 h until cultures entered late-exponential growth phase at\n25\u2000C in an atmosphere of 15% oxygen [created by a ProOx\nP110 oxygen controller (BioSpherix, Parish, NY, United States)\nwith a hypoxia C-chamber]. The cells were stabilized by the\naddition RNAprotect reagent (QIAGEN, Germantown, MD,\nUnited States) and total RNA was purified using a RNeasy Protect\nBacteria Mini Kit (QIAGEN) from three biological replicates of\neach strain cultured under control conditions and in exudates.\nThe quality assessment of the extracted RNA samples was\nperformed with a NanoDrop OneC Spectrophotometer (Thermo\nFisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies)\nand revealed A260/A280 and A260/A230 values of > 2.0 and a mean\nRNA integrity numbers (RIN) value of > 9.2.\nThree biological replicates of RNA samples were shipped on\ndry ice to the DOE Joint Genome Institute (Walnut Creek, CA,\nUnited States), where rRNA was depleted and stranded RNASeq\nlibraries were prepared, quantified by qPCR and sequenced\nusing a HiSeq 2500 instrument (Illumina). The fastq file reads\nwere filtered and processed with BBDuk1 to remove reads that\ncontained 1 or more “N” bases, had an average quality score\nacross the read less than 10 or had a minimum length < 51 bp\nor 33% of the full read length. Reads mapped with BBMap (see\ntext footnote 2) to masked human, cat, dog, and mouse references\nat 93% identity were removed. Another category of removed\nsequences matched RNA spike-in, PhiX, common microbial\ncontaminants, and ribosomal RNAs. The processed reads from\neach library were aligned to the reference genome using BBMap\nwith only unique mappings allowed (BAMs/directory). If a\nread mapped to more than one location it was ignored.\nfeatureCounts (Liao et al., 2014) was used to generate raw\ngene counts, which were normalized to adjust for the length of\neach gene and total number of reads mapped for each library.\nThe normalization formula used: n = [r/(l/1,000)]/(t/1,000,000),\nwhere n = normalized read count for gene (G) for library (L);\nr = raw read count for gene G for library L; l = gene G length; and\nt = total reads mapped for library L. Raw gene counts were used to\nevaluate the level of correlation between biological samples using\nPearson’s correlation.\nBioinformatic Analysis\nCount tables generated by the JGI RNA-Seq pipeline were input\ninto DESeq2 (Love et al., 2014) to normalize and determine\n1https://sourceforge.net/projects/bbmap/\ndifferential expression. Statistical significance was established\nthrough DESeq2 by using three biological replicates for control\nand root exudate conditions. Scatterplots were generated from\nthe DESeq2 data table outputs using ggplot2. Genes differentially\nexpressed between control and root exudate samples (log2 foldchanges\n–2 \u2000 to \u2000 2, adjusted p value \u2000 0.05) were used in\ndownstream analysis. The core genome and pangenome for the\nPseudomonas strains used in this study were computed using the\nOthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic\nPangenome Accumulation v1.4.0 apps implemented in the U.S.\nDepartment of Energy Systems Biology Knowledgebase (KBase)\n(Arkin et al., 2018). Additional comparisons were conducted\nwith the PGAweb pangenome analysis pipeline (Chen et al.,\n2018). Differentially expressed genes were assigned to core,\nnon-core, and singleton parts of each strain’s proteome by\nBLASTp with an E value cutoff of e-06, identity of 40%\nand coverage of 60%. Functional annotation of differentially\nexpressed genes was carried out with the Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al.,\n2018). Additional manual curation was performed using tools\nimplemented in the Integrated Microbial Genomes (IMG)\ndatabase (Markowitz et al., 2012), Pseudomonas Genome\nDatabase (Winsor et al., 2009), Kyoto Encyclopedia of Genes and\nGenomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3\n(Biomatters, Auckland, New Zealand). Metabolic functions\nencoded by the differentially expressed genes were mapped\nusing iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses\nwere carried out by building multiple sequence alignments\nwith MAFFT v7.222 (Katoh and Standley, 2013) and inferring\nneighbor-joining (NJ) phylogenies with Geneious Tree Builder.\nThe resultant phylogenetic trees were visualized with iTOL\n(Letunic and Bork, 2016). Reproducibility of clades within the\ninferred NJ trees was assessed by bootstrap resampling with\n1,000 replicates.\nCharacterization of Carbon Source\nUtilization With Biolog Phenotype\nMicroarrays\nThe utilization of carbon sources was analyzed using Phenotype\nMicroArrays (Biolog, Hayward, CA, United States) as follows.\nThe bacteria were cultured overnight on Luria-Bertani agar\nat 25\u2000C, after which cells were harvested and suspended in\ninoculating fluid (IF-0). The transmittance of the suspension\nwas adjusted to 42% using a Biolog turbidimeter. The cell\nsuspension was mixed with IF\U001000000 containing Dye Mix A\n(Biolog) to achieve a final transmittance of 85%. One hundred\nmicroliter aliquots of the adjusted cell suspension were\ninoculated into PM01 and PM02A plates, which were then\nincubated in an OmniLog Phenotype MicroArray System\n(Biolog) at 25\u2000C for 48 h. The formation of formazan was\nrecorded at 15 min intervals, and data were analyzed using\nOmniLog Parametric Analysis software v1.20.02 (Biolog).\nRelative growth of the studied strains was normalized to\ngrowth on D-glucose and visualized using Heatmapper\n(Babicki et al., 2016).\nFrontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nData Availability\nSequences generated in this project were deposited\nunder NCBI BioProject accession numbers PRJNA439743\nthrough PRJNA439790.\nRESULTS\nMetabolomic Profiling of Root Exudates\nof B. distachyon\nMetabolomics analysis of lyophilized root exudates revealed\nthe presence of numerous plant metabolites, 86 of which were\nidentified by matching their spectra to the LECO/Fiehn\nMetabolomics library (Supplementary Table 3). These\nmetabolites included (i) carbohydrates and their derivatives\n(glucose, fructose, xylose, sucrose, trehalose, maltose, galactose,\nand others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol,\ngalactinol, 2-deoxyerythritol, ribitol, threitol and\ncellobitol); (iii) amino acids and derivatives (glutamine,\ntyrosine, glutamic acid, asparagine, aspartic acid, valine,\nphenylalanine, isoleucine, glycine, serine, proline, leucine,\ntryptophan, cysteine, methionine, citrulline, and others); (iv)\norganic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric,\nfumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic,\nmalic, methylmalonic, nicotinic, quinic, succinic, threonic);\nand (v) assorted metabolites including heterocyclic compounds,\nphenolics, and biogenic amines, etc (3-hydroxypyridine,\nmaleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate,\n5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid,\nand urea). Results of the analysis also revealed that root exudates\nof B. distachyon contain hydroxyectoine and the quaternary\namine (QA) glycine betaine (Supplementary Figure 1).\nPhylogenetic and Pangenome Analyses\nof Pseudomonas Strains Used in the\nStudy\nWe used a set of phylogenetic markers suggested by Mulet et al.\n(2010) to investigate the relatedness of the eight strains used in\nthis study to distinct lineages recognized within the P. fluorescens\nspecies complex. The multilocus sequence analysis based on\nthe concatenated sequences of the housekeeping genes rrs (16S\nrRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with\nstrains 2-79 and SBW25) as a member of the P. fluorescens\nsubgroup (Figure 1). The rest of the strains clustered closely with\nfour additional subgroups of the P. fluorescens complex, namely\nP. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1),\nP. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of\nthe eight rhizosphere Pseudomonas strains varied in size by 1.43\nmegabase (ranging from 5.65 to 7.07 Mb) and contained between\n5,166 and 6,363 protein-coding genes (Figure 2A). The shared\ngene content was characterized with OrthoMCL, which uses allagainst-\nall BLASTp followed by the Markov Cluster algorithm to\nidentify protein groups shared between the compared genomes,\nas well as groups representing species-specific gene expansion\nfamilies (Li et al., 2003). The pangenome analysis revealed a\ncore comprised of approximately 3,179 orthologs that were\nshared among all strains and represented 50.0% to 61.5%\nof each predicted proteome (Figures 2A,B). The non-core\npangenome contained genes shared by two or more (but not\nall) strains and contained between 1,482 and 2,080 orthologs,\nwhich corresponded to 28.7–36.3% of individual proteomes. The\nrest of the predicted protein-coding genes were strain-specific\nsingletons that comprised 7.5% to 15.1% of the strain’s predicted\nproteomes. In respect to divergence from the core genome, strain\nPf-5 was found to possess the highest proportion of unique genes\n(n = 949) followed by 2-79 (n = 887). The entire pangenome of\nthe Pseudomonas strains encompassed over 12,000 homolog and\nsingleton gene families.\nFurther homolog family-based comparisons identified Q8r1-\n96 and R1-43-08 as the most distantly related strains, with\n3349 shared homologs (Supplementary Table 4A). Q8r1-96 and\nQ2-87, which shared 4,489 homologs, were the most closely\nrelated strains. The partitioning of homolog gene families into\nthe core, non-core, and singleton parts of the pangenome agreed\nwith phylogenetic relationships of the strains deduced from the\nanalysis of a selected subset of COGs (Clusters of Orthologous\nGroups) (Figure 2C and Supplementary Tables 4B,C). The\nCOG-based phylogeny supported the multilocus sequence\nanalysis and revealed that the eight Pseudomonas strains form\nthree distinct clusters, the first of which contained 2-79, R1-30-\n84, and SBW25. The second cluster included Q8r1-96 and Q2-87,\nwhereas the third encompassed strains 30-84, Pf-5, and Pf0-1.\nCorrelating the Composition of Root\nExudates With Metabolic Profiles of\nPseudomonas Strains\nWe used the Phenotype MicroArray PM1 and PM2 plates to\nprofile the eight Pseudomonas strains for the utilization of\n190 different carbon sources. Results of the analysis identified\n90 compounds that supported growth and clustered by their\nintensities of utilization into three distinct groups (Figure 3).\nGroup I was comprised of 30 highly metabolized carbon\nsources, which included several amino acids and intermediates of\nglycolysis, pyruvate metabolism, and citrate cycle. Approximately\nhalf of these compounds were catabolized by all eight strains,\nand included several organic acids (fumaric, citric, gluconic,\nmalic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro,\nAla, and g-aminobutyric acid), carbohydrates (glucose, mannose,\nand mannitol), and the purine nucleoside inosine. Group II\nwas composed of 44 chemically diverse carbon sources that\nwere variably utilized by the strains. These compounds were\ncarbohydrates, organic acids, amino acids, phenolics, and polyols,\nand included known compatible solutes and intermediates of\nmetabolism of pentoses, galactose, starch, and sucrose. Group\nIII encompassed the rest of the Phenotype MicroArray test\npanel and contained compounds that were not catabolized\nby the tested strains. Among several notable exceptions were\na-hydroxyglutamic acid- g-lactone, putrescine, and itaconic,\ncitramalic, and succinamic acids, which supported the growth\nof strains 2-79, 30-84, Pf-5, and SBW25. We further matched\nthe carbon metabolic profiles of the Pseudomonas strains against\nthe list of plant-derived metabolites from the root exudates of\nFrontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the\nP. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA\ngyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor\nmethod. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles).\nB. distachyon Bd21. Interestingly, many carbon sources from\nthe Phenotype MicroArray panel were also present in the root\nexudates of B. distachyon Bd21, and some of these compounds\n(glucose, mannose, galactose, fructose, g-aminobutyric acid,\naspartic acid, citric acid, malic acid, fumaric acid, quinic acid,\nalanine, glutamine, and glutamic acid) were catabolized by all\nstrains used in this study, while others (e.g., xylose, trehalose,\nm-inositol) were actively utilized only by certain organisms\n(Figure 3). The comparison of catabolic profiles across the\neight studied Pseudomonas strains revealed the presence of three\ndistinct clusters. The first cluster contained strains Q8r1-96 and\nQ2-87, which consumed very similar sets of carbon sources,\nas well as strain Pf0-1. The second cluster was composed of\n2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster\nwas represented by a single strain, Pf-5. The overall similarity\nof the catabolic profiles partially agreed with the separation\nof the strains into different subgroups of the P. fluorescens\ncomplex (see above).\nAnalysis of the RNA-seq Results\nIn order to understand the cellular responses of rhizosphere\nPseudomonas to plant exometabolites, we analyzed the\ntranscriptome changes in cultures grown in the presence of\nroot exudates. Under field conditions, rhizobacteria colonize\nplant roots in the form of surface-attached microaerobic biofilms\n(Hojberg et al., 1999). To mimic these conditions, the eight\nPseudomonas strains were grown statically at 72% air saturation\nin 21C-glucose medium amended with root exudates and then\nprocessed to extract total RNA (Supplementary Figure 2).\nA total of 995 million raw sequencing reads were generated from\nthe RNA samples by using the Illumina HiSeq-2500 platform,\naveraging 20.7 million reads per sample. The removal of lowquality\nand rRNA sequences resulted in a total of 793 million\nfiltered reads that were mapped onto the eight Pseudomonas\ngenomes with a mean of 7.48 million mapped fragments per\ngenome. The differentially abundant transcripts were identified\nby setting a p value of 0.05 (adjusted for multiple testing) and\nFrontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all\neight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific\nsingletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of\nthe pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the\n25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any\nbias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939\nnon-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied\nstrains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining\northolog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node\nrepresents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The\nanalysis was conducted in KBase (Arkin et al., 2018).\nthe log2 fold-change (FC) threshold \u2000\u2000 2.0 (Figure 4 and\nSupplementary Tables 5–12). When compared with the control\nconditions, an average of 204 genes per strain were differentially\nexpressed in the presence of root exudates, with the highest\n(n = 425) and lowest (n = 112) numbers observed, respectively,\nin SBW25 and Q2-87 (Figure 4). Overall, more genes were\ninduced than repressed in response to exudates, but the actual\nnumbers in each category varied substantially depending on the\nidentity of the Pseudomonas strain. In most strains, the bulk of\nthe differentially expressed genes was almost equally distributed\nbetween the core (mean, 48.2%) and non-core (mean, 45.8%)\nparts of the genome, whereas the strain-specific singleton genes\nconstituted on average only 5.9% (Figure 4B). One notable\nexception was observed in Q8r1-96, where all differentially\nexpressed genes belonged to the core (73.8%) and non-core\n(26.2%) parts of the genome. Another notable pattern was\nobserved in R1-43-08, where the majority of genes affected by\nthe presence of root exudate fell into the non-core category\n(56.3%). The highest proportion of differentially expressed\nsingletons (11.3 and 10.4%, respectively) was identified in\nstrains SBW25 and Pf-5.\nWe further explored how the identified differentially\nexpressed genes were distributed across genomes of the eight\nstudied rhizosphere strains. The pairwise BLASTp comparisons\nidentified 2-79 and SBW25 as two strains that shared the highest\nnumber of genes (n = 101) induced or repressed in response\nto root exudates (Table 1). The second pair of strains with\na significant number of similar differentially expressed genes\n(n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1\nand 30-84, which shared 56 differentially expressed genes. These\npatterns of shared genes were also observed when the results of\nthe pairwise BLASTp comparisons were converted into a binary\ngene presence/absence matrix, which was then subjected to\ncluster analysis using a UPGMA algorithm based on Sorensen’s\ndissimilarity index or examined by non-metric multidimensional\nscaling (NMDS) (Figure 5).\nThe differentially expressed Pseudomonas genes were\nsubjected to Blast2Go analysis and Gene Ontology (GO)\nannotation (Figure 6). Metabolic process, catalytic activity,\nand membrane were the most common annotation terms\nacross the three primary GO term categories (i.e., biological\nprocess, molecular function, and cellular component). A total\nof 1,694 GO terms was assigned to 805 upregulated genes, with\nthe majority of the GO terms related to molecular function\n(682, 40.3%), followed by biological process (669, 39.5%), and\ncellular component (343, 20.2%). In the 539 downregulated gene\ncategory, 1,101 GO terms were assigned to biological process\n(420, 38.1%), molecular function (417, 37.9%), and cellular\ncomponent (264, 24.0%). Within biological process, metabolic\nprocess, cellular process, localization, response to stimulus, and\nregulation were over-represented. Within molecular function,\nthe largest proportion was assigned to catalytic activity, binding,\nand transporter activity categories. Within cellular component,\nthe majority were assigned to membrane, membrane part, cell,\nand cell part categories. Across the eight strains, 37–42% of\ndifferentially expressed genes had no Gene Ontology IDs and\nencoded various conserved hypothetical proteins.\nFunctional Classification of Shared\nDifferentially Expressed Genes\nThe interrogation of RNA-seq data revealed multiple cellular\npathways that were differentially regulated in bacterial cultures\nincubated with root exudates (Supplementary Figures 3, 4).\nFrontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere\nPseudomonas strains used in the study. The hierarchical clustering analysis\nwas carried out using the average linkage method with Euclidean distances.\nCarbon sources identified by red arrowheads were also detected in the sterile\nroot exudates of B. distachyon Bd21.\nAlthough none of these differentially regulated pathways were\nshared by all eight strains, the cross-strain comparisons revealed\nseveral types of common and specific transcriptomic responses\nthat were elicited by the presence of plant exometabolites\n(Table 2). The visual representation of core gene expression\npatterns is provided in Supplementary Figure 5, which\nshows heatmaps of expression profiles and p-adj values for\ncore genes shared by the studied strains. The figure is\naccompanied by Supplementary Table 13 that lists predicted\nfunctions of genes constituting the four distinct clusters\nobserved after hierarchical clustering of gene expression\nvalues. The first category of shared differentially expressed\npathways functioned in the uptake and catabolism of selected\ncarbohydrates, quaternary ammonium compounds (QAs), and\nphenolics. All strains except for R1-43-08, responded to root\nexudates by inducing the fructose-specific phosphoenolpyruvate\n(PEP)-dependent phosphotransferase system (PTSFru). The\ncomponents of this system are encoded by a conserved\noperon and include the cytoplasmic polyprotein EI/HPr/EIIAFru\n(FruB), the 1-phosphofructokinase FruK, and the fructosespecific\npermease EIIBC (FruA) (Chavarria et al., 2016). The\nPTSFru system functions by acquiring high-energy phosphates\nfrom PEP and sequentially passing them, via the EI/HPr/EIIAFru\ndomains of FruB, to the EIIB component of FruA. The\nphosphates are ultimately transferred by the EIIC transporter\nto fructose yielding fructose 1-phosphate, which is channeled\ninto the central metabolic pathways through the action of the\nphosphofructokinase FruK.\nIn all strains except for Q8r1-96 and Pf-5, the exposure to root\nexudates resulted in the induction of two genes adjacent to the\nfru cluster that encoded a Major Facilitator Superfamily (MFS)\ntransporter and an L-arabinonate dehydratase (Table 2). These\ngenes are predicted to participate in the uptake and catabolism\nof L-arabinose, where L-arabinonate dehydratase plays an\nimportant role by converting L-arabinonate to 2-dehydro-3-\ndeoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1-\n43-08, and Q2-87, we also observed the induction of genes\nencoding components of the AraFGH complex, an ATP-Binding\nCassette (ABC) superfamily transporter involved in the import of\narabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all\nstrains except SBW25 and R1-43-08 responded to the presence\nof exudates by upregulating a conserved gene encoding an\naldose epimerase superfamily protein. Such enzymes equilibrate\nalpha- and beta-anomers of aldoses and ensure that stereospecific\nenzymes involved in the metabolism of free sugars do not act as\nmetabolic bottlenecks (Abayakoon et al., 2018). Although some\naldose epimerases have been linked to specific pathways, the\nPseudomonas gene identified in this study could not be assigned\nto a particular metabolic process based on sequence analysis and\ngenomic location.\nSeveral Pseudomonas strains responded to the presence\nof root exudates by upregulating genes involved in the\nuptake and catabolism of myo-inositol and possibly other\nstereoisomers of inositol (Table 2). The upregulated catabolic\ngenes encode the dehydrogenase IolG, which oxidizes myoinositol\nto its corresponding ketone, as well as IolE, IolD, IolB,\nand IolC that collectively convert the 2-keto-myo-inositol to\nacetyl-CoA and the glycolysis intermediate dihydroxyacetone\nphosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1-\n43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions\nFrontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core\nand non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established\nthreshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual\nPseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates.\nalso involved components of the putative inositol-specific\nABC transporter. The cross-genome comparisons revealed\nthat in all studied strains except for Pf0-1, components\nof the myo-inositol utilization pathway were encoded\nwithin a well-conserved gene cluster which, in addition to\ncatabolic and transport functions, also encodes a dedicated\ntranscriptional repressor.\nAll studied strains of Pseudomonas carry multiple genes\ninvolved in scavenging the quaternary ammonium compounds\ncholine, glycine betaine (GB), carnitine, choline-O-sulfate, and\nFrontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 1 | The number of differentially expressed genes shared among the eight\nstudied strains of rhizosphere Pseudomonas.\nStrain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5\n2–79 260\nSBW25 101 425\nR1-43-08 30 25 151\nQ8r1-96 32 39 21 145\nQ2-87 27 28 25 31 112\n30-84 27 23 24 32 28 136\nPf0-1 38 50 29 29 50 56 205\nPf-5 36 41 52 86 29 55 40 230\nThe pairwise comparisons were conducted by BLASTp with the following cutoff\nparameters: E-value < 1e-06, minimum percent identity > 40%, and minimum\npercent coverage > 65%. The black diagonal cells show the number of differentially\nexpressed genes per strain. In other words, these are self comparison values.\nsarcosine from the environment. Many of these genes were\ndifferentially expressed, including those encoding parts of the\nABC transporter CbcXWV, which is predicted to function in\nthe uptake of choline under water-replete conditions (Table 2).\nAmong enzymes induced in the presence of root exudates were\nthe choline dehydrogenase BetA, which converts choline to\nglycine betaine and a network of enzymes (i.e., the Rieske family\noxygenase GbcAB, the dimethyglycine demethylase DgcAB, and\nthe sarcosine oxidase SoxBDAG) that sequentially convert GB\nto glycine. In 2-79 and SBW25, this group of differentially\nregulated genes also included an AraC-family transcriptional\nactivator GbdR, which perceives intercellular levels of GB and\ninduces genes involved in the transport and catabolism of\nglycine betaine and detoxification of the catabolic byproducts\n(Hampel et al., 2014).\nThe last category of activated catabolic pathways included the\ncatechol branch of the b-ketoadipate pathway for the degradation\nof aromatic compounds. In strains 30-84, Pf0-1, and Pf-5,\ngrowth on root exudates resulted in upregulation of catechol-\n1,2-dioxygenase, muconate cycloisomerase, and muconolactone\nisomerase, which collectively cleave the catechol ring and\nconvert it to b-ketoadipate enol-lactone (Harwood and Parales,\n1996). Finally, analysis of the P. synxantha 2-79 transcriptome\nidentified an induction of benABC genes encoding subunits\nof benzoate 1,2-dioxygenase, an oxidoreductase that generates\ncatechol from benzoate.\nIn addition to various catabolic pathways, the exposure to root\nexudates also induced several genes involved in the homeostasis\nof copper (Table 2). Four of these genes form a conserved cluster\nin genomes of the strains and encode the periplasmic coppersensing\ntwo-component system CinRS, the plastocyanin/azurinlike\nprotein CinA, and the NADPH-dependent pre-Q0 reductase\nCinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed\nupregulation of a conserved operon encoding the multicopper\noxidase CopA, the periplasmic copper-binding protein CopC, the\ninner membrane protein CopD, and outer membrane protein\nCopB. In several Gram-negative bacteria, these Cop proteins are\nthought to have dual functions and participate both in the uptake\nof essential copper as well as in the sequestration of excess copper\nin the periplasm and outer membrane.\nThe analysis of shared downregulated pathways revealed that\nmost of the strains respond to the presence of root exudates by\nrepressing genes involved in the uptake and catabolism of sulfur\ncompounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96,\nQ2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB\noperon responsible for the utilization of alkanesulfonates\nas sulfur sources. The ssu operon is highly conserved in\nfluorescent pseudomonads and encodes the FMNH2-dependent\nmonooxygenase SsuD and the NAD (P)H-dependent FMN\nreductase SsuE, which together catalyze the desulfonation of\nalkanesulfonates. Also, the ssu locus contains genes for the\nmolybdopterin-binding protein SsuF and the alkanesulfonatespecific\nABC-type transporter consisting of the sulfonate\nsubstrate-binding protein SsuA, sulfonate permease protein\nSsuC, and sulfonate transport ATP-binding protein SsuB. Finally,\nin R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates\ncoincided with repression of the tauABCD operon, which\nallows these strains to utilize taurine (2-aminoethanesulfonate)\nas a sulfur source. The repressed tau genes encoded the\n2-oxoglutarate-dependent taurine dioxygenase TauD and\nsubstrate-binding, ATP-binding, and permease components of\nthe taurine-specific ABC transporter TauABC.\nOther Differentially Expressed Pathways\nIn addition to their effect on several shared cellular pathways,\ngrowth on root exudates resulted in the induction or repression of\nnumerous strain-specific genes. In closely related P. synxantha 2-\n79 and P. fluorescens SBW25, we observed differential expression\nof genes involved in energy metabolism, transport of amino\nacids, and surface attachment (Supplementary Tables 5, 6).\nOther notable differentially expressed pathways included 2–\n79 gene clusters that encode enzymes for the catabolism of\ntrehalose, a prophage, and toxin/antitoxin system, as well as the\nSBW25 operon predicted to control the synthesis of the capsular\nexopolysaccharide colonic acid. The response of Pseudomonas sp.\nR1-43-08 to root exudates also involved differential expression\nof different energy metabolism pathways. In addition, we\nobserved the upregulation of genes involved in the uptake and\ncatabolism of xylose (also upregulated in 2–79) and repression\nof enzymes for the biosynthesis of phenazine-1-carboxylic acid\nand assimilation of inorganic sulfur and L-cysteine biosynthesis\n(Supplementary Table 7).\nThe analysis of the Q8r1-96 transcriptome revealed\nperturbation of different metabolic pathways including genes\nencoding components of cytochrome C oxidase, transport\nand catabolism of sorbitol/mannitol, metabolism of butanoic\nacid, and biosynthesis of exopolysaccharides alginate and\npoly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In\nP. fluorescens Q2-87, we identified differential expression of\ngenes involved in metabolism of galactose, tryptophan, tyrosine,\nglycine, serine, and threonine (Supplementary Table 9),\nwhile in P. chlororaphis 30-84, growth on exudates activated\nthe biosynthesis of molybdopterin cofactor, catabolism of\ngalactonate and acetoin, and uptake and catabolism of putrescine\n(Supplementary Table 10). The response of P. protegens Pf-5 to\nroot exudates involved upregulation of acetoin dehydrogenase,\nwhich converts acetoin to acetaldehyde and acetyl-CoA, as\nFrontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nFIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root\nexudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis.\nFIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of\nB. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa\nand Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO\nterm. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO\nterms.\nwell as pathways for the utilization of glycolate and putrescine\n(Supplementary Table 11). Also induced were genes for\nthe production of pyrrolnitrin and PhlG hydrolase, which\nmodulate the metabolic loads attributed to the synthesis of\n2,4-diacetylphloroglucinol. The differentially expressed genes of\nP. fluorescens Pf0-1 included, among others, operons encoding\ncytochrome C oxidase and enzymes for catabolism of malonic\nacid (Supplementary Table 12). Yet another interesting\nfinding involved the induction of assorted genes acting in the\nhomeostasis of iron and defense against reactive oxygen species\n(ROS). We observed activation of iron dicitrate transporters\n(SBW25 and 30-84), genes for the biosynthesis of siderophores\nornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading\nenzymes (2–79, 30–84), TonB siderophore receptors, and\ncomponents of the energy-transducing inner membrane\ncomplex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially\nexpressed ROS defense pathways were represented by different\ncatalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and\nPf-5 and organic hydroperoxide resistance proteins in strains\nSBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and\nPf0-1, the addition of exudates resulted in the upregulation of\nperoxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic\nand aromatic hydroperoxides.\nDISCUSSION\nOur analysis of B. distachyon root exudates revealed a complex\nmix of primary and secondary metabolites, thus supporting\nthe view of the plant rhizosphere as a carbon-rich niche for\nFrontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | The distribution and predicted functions of selected differentially expressed genesa.\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nD-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8)\n1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7)\nD-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2)\nUptake and catabolism of arabinose\nMFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9)\nL-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5)\nInterconversion of alpha- and beta-anomers of aldoses\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of quaternary ammonium compounds\nCholine dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1)\nTranscriptional regulator GbdR 2756597125 (\U001000003.7) 649639087 (\U001000003.9)\nMembrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3)\nHypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3)\nDimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2)\nDimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0)\nBetaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7)\nBetaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2)\nSer hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1)\nSarcosine oxidase, g subunit, SoxG 2597878304 (2.9)\nSarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4)\nSarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2)\nSarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3)\nBetaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5)\nABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5)\nUptake and catabolism of myo-inositol\n5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2)\n2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1)\n5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3)\n2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4)\n3D-(3,5/4)-trihydroxycyclohexane-1,2-dione\nacylhydrolase, IolD\n2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3)\nMyo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9)\nInositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8)\nInositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5)\nInositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6)\n(Continued)\nFrontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nTABLE 2 | Continued\nPredicted function Strainb\n2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5\nUptake and catabolism of fructose\nCatabolism of phenolics\nMuconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8)\nMuconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4)\nCatechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5)\nAraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2)\nBenzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8)\nBenzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8)\nBenzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3)\nCopper homeostasis\npre-Q0 reductase/7-cyano-7-deazaguanine\nreductase CinQ\n2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5)\nCupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4)\nHeavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5)\nHeavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0)\nCopper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4)\nCopper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0)\nCopper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4)\nCopper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2)\nConserved hypothetical protein\nAldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5)\nUptake and catabolism of sulfonates\nFMN-dependent monooxygenase SsuE 2756592254 (\U001000001.9) 637745334 (\U001000003.5)\nSulfonate substrate-binding protein SsuA 649639261 (\U001000002.0) 2756592253 (\U001000002.2) 2597878518 (\U001000002.6) 2597849636 (\U001000003.9) 637745333 (\U001000007.0) 637323272 (\U001000002.7)\nFMN-dependent monooxygenase SsuD 649639260 (\U001000001.9) 2756592252 (\U001000002.3) 2597849637 (\U001000002.8) 637745332 (\U001000005.8) 637323271 (\U001000002.3)\nSulfonate permease protein SsuC 649639259 (\U001000002.4) 2756592251 (\U001000002.1) 2597878516 (\U001000001.9) 2597849638 (\U001000003.2) 637745331 (\U001000005.8) 637323270 (\U001000002.5)\nSulfonate transport ATP-binding protein SsuB 649639258 (\U001000002.5) 2756592250 (\U001000002.3) 2597849639 (\U001000002.5) 637745330 (\U001000005.6) 637323269 (\U001000002.7)\nMolybdopterin binding protein SsuF 649639257 (\U001000003.1) 2756592249 (\U001000002.7) 2597878514 (\U001000002.9) 2597849641 (\U001000003.3) 637745339 (\U001000006.2) 637323268 (\U001000002.7)\nUptake and catabolism of taurine\nTaurine substrate-binding protein TauA 2597854917 (\U001000002.0) 637740095 (\U001000004.1)\nTaurine transport ATP-binding protein TauB 2756592398 (\U001000001.9) 637740094 (\U001000003.7) 637317614 (\U001000001.9)\nTaurine permease protein TauC 2756592398 (\U001000001.9) 637740093 (\U001000003.9) 637317613 (\U001000002.0)\nTaurine dioxygenase TauD 2597854920 (\U001000001.8) 637740092 (\U001000003.6)\naThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%.\nbValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets).\nFrontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nsoil microorganisms. Our results were in agreement with a\nrecent report of 27 different sugars, amino acids, and organic\nacids in Brachypodium exudates (Kawasaki et al., 2016). We\nconfirmed the presence of exometabolites identified in that\nstudy, along with dozens of additional analytes that were\nidentified by matching their mass-spectra and retention indices\nto the LECO/Fiehn Metabolomics library (Supplementary\nTable 3). The complementation of the metabolomic analysis\nwith profiling of the bacteria by Biolog Phenotype MicroArrays\nrevealed that a substantial proportion of the characterized\nexudate constituents were catabolized by a collection of eight\nPseudomonas strains from across the P. fluorescens group that is\nknown to form associations with plant roots. The amendment\nof Pseudomonas cultures with root exudates caused changes in\nthe expression of multiple genes encoding catabolic and anabolic\nenzymes, predicted transporters, transcriptional regulators, stress\nresponse, and conserved hypothetical proteins. In most strains,\nthese differentially expressed genes were almost equally split\nbetween the core and variable genome regions, mirroring\nthe substantial strain-to-strain variation in the genome size\nand gene content within the P. fluorescens species complex\n(Loper et al., 2012).\nThe analysis of transcriptome responses to root exudates\nrevealed several types of cellular pathways present in the strains\nused in this study. The first category of such pathways was\ninvolved in the catabolism of carbohydrates such as fructose,\narabinose, myo-inositol, xylose, trehalose, and galactose. Among\nthese catabolic traits, the ability to utilize fructose as a carbon\nsource is highly conserved among fluorescent pseudomonads.\nIn contrast, growth on arabinose, myo-inositol, xylose, and\ntrehalose is variably present and was traditionally used to\ndifferentiate species and biovars within the P. fluorescens\ngroup (Barrett et al., 1986). We speculate that such variably\ndistributed pathways contribute to the differential affinity of\npseudomonads toward host plants and/or to determine which\nstrains flourish in response to growing roots and changing\nenvironments. Several independent studies have confirmed\nthe importance of carbohydrate catabolism pathways for the\nbiology of rhizosphere pseudomonads. For example, in vivo\nexpression technology (IVET) profiling of P. fluorescens SBW25\nidentified xylose isomerase among genome regions essential\nfor the colonization of sugar beet seedlings (Liu et al., 2015),\nwhereas a genome-wide Tn-Seq screen of Pseudomonas simiae\nidentified genes for the catabolism of myo-inositol among traits\nessential for the colonization of Arabidopsis thaliana roots\n(Cole et al., 2017).\nThe response of rhizosphere Pseudomonas to Brachypodium\nroot exudates also involved pathways for the uptake and\nmetabolism of amino acids. We observed differential\nexpression of genes encoding the hydrophobic (HAAT) and\npolar (PAAT) amino acid uptake transporters in strains\n2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes\nencoded enzymes for the catabolism of valine and glutamic\nacid (2-79); metabolism of tryptophan, glycine, serine, and\nthreonine (Q2-87); and biosynthesis of methionine (Q8r1-\n96). It is plausible that the abundance of amino acids in\nroot exudates is also linked to the repression of pathways\ninvolved in the catabolism of sulfonates and taurine that\nwas observed in several strains (Table 2). Although the\npreferred source of sulfur for P. fluorescens is unknown, in the\nclosely related P. aeruginosa, the sulfur starvation response\nis triggered by the growth on any sulfur compound other\nthan sulfate, thiocyanate, and cysteine (Hummerjohann et al.,\n1998). This fact, together with the presence of cysteine and\ncystine in the root exudates, suggest that root exudates of\nBrachypodium may serve as an important source of sulfur for\nrhizosphere Pseudomonas. These findings also agree well with\nthe reported scarcity of inorganic sulfate in the soil, and the\npresence of sulfur mostly in the form of organic compounds,\nincluding amino acids, proteins, sulfate esters, and sulfonates\n(Autry and Fitzgerald, 1990).\nAnother interesting result of this study was the concerted\nactivation of copper and iron homeostasis pathways observed\nin all of the Pseudomonas strains used in this work. In bacteria,\nan excess of copper is toxic and triggers oxidative stress due to\nthe formation of free radicals, as well as disruption of protein\nmetalation and stability of iron-sulfur clusters (Bondarczuk\nand Piotrowska-Seget, 2013). On the other hand, copper is an\nessential trace element used as a cofactor in different enzymes.\nSimilarly, although elevated levels of iron cause redox stress, this\nelement is also found in active energy metabolism enzymes and is\ncrucial for bacterial growth (Andrews et al., 2003). The analysis\nof metal homeostasis genes identified in this study suggests\nthat their induction was likely triggered by the deficiency of\ncopper and iron in bacterial cultures grown in the presence of\nroot exudates. We attribute this effect to the ability of some\ncomponents of root exudates to chelate soil metals.\nDespite the abundance of iron in the soil, its bioavailability\nis limited due to the low solubility of Fe (III) oxyhydrates\nat neutral pH. The non-graminaceous plants circumvent this\nproblem by acidifying the rhizosphere and secreting flavins,\nphenolics, and organic acids that chelate iron. The reduction\nof these ferric chelates releases soluble ferrous iron taken up\nby root cells (Kobayashi and Nishizawa, 2012). Graminaceous\nplants, like Brachypodium, acquire iron by secreting into the\nsoil non-protein amino acids of the mugineic acid (MA) group,\nwhich act as Fe (III)-chelating phytosiderophores. In addition to\niron, low-molecular-weight organic acids and phytosiderophores\nbind other divalent and trivalent metals (including copper) and\ncontribute to heavy-metal tolerance in plants (Chen et al., 2017).\nIt is plausible that the presence of these plant exometabolites\nis responsible for the deficit of iron and copper observed in\nPseudomonas cultures grown in the presence of root exudates.\nThese results further underscore the importance of diverse and\nredundant metal homeostasis pathways found in genomes of the\nP. fluorescens group for the ability of these organisms to colonize\nand persist in the plant rhizosphere.\nRecently, Klonowska et al. (2018) examined transcriptomic\nresponses of symbiotic nitrogen-fixing bacteria to root\nexudates of the legume plant Mimosa pundica, which has\nan unusual ability to support both alpha- (Rhizobium) and\nbeta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq,\nthe authors characterized genes involved in the perception of\nroot exudates in the nodulating bacteria Burkholderia phymatum\nFrontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSTM815, Cupriavidus taiwanensis LMG19424, and Rhizobium\nmesoamericanum STM3625. Interestingly, the analysis of\ndifferentially expressed genes revealed induction of pathways\ninvolved in the catabolism of fructose, xylose, myo-inositol,\nand protocatechuate/catechol. Also upregulated were some\ncopper homeostasis, siderophore biosynthesis, and oxidative\nstress genes. Finally, the analytical profiling of M. pundica\nexudates revealed an overlap with Brachypodium in the types of\ncarbohydrates, amino acids, and organic acids present. These\nfindings suggest that differentially expressed genes shared by\nmultiple strains of the group P. fluorescens are not unique to\nthe Brachypodium-Pseudomonas system but represent a set\nof conserved cellular pathways involved in the perception of\nplant exometabolites by different clades of rhizosphere-dwelling\nProteobacteria.\nMost strains included in this study were originally selected\nbased on the ability to colonize the rhizosphere and produce\nsecondary metabolites that alleviate the plant stress response\nand/or inhibit soilborne pathogens. It has been suggested\nthat plant metabolites released into the rhizosphere affect\nthe biocontrol activity of plant-beneficial pseudomonads (de\nWerra et al., 2011). We provide further support to this\nhypothesis by demonstrating that in some strains, root exudates\nmodulate the expression of genes for the catabolism of\nthe plant growth-promoting metabolites acetoin and 2,3-\nbutanediol. The exposure to exudates also affected the expression\nof genes for the synthesis of well-characterized antifungal\ncompounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4-\ndiacetylphloroglucinol. The modulatory effects were strainspecific,\nsuggesting significant differences in the regulatory\nnetworks involved in the perception of plant signals and\nregulation of the production of antibiotics and growthpromoting\nmetabolites.\nThe final significant finding of this study was the induction\nof catabolism of quaternary amines (QAs) observed in multiple\nstrains of the P. fluorescens group during growth on root\nexudates. This observation was supported by the detection of\nglycine betaine in the root secretions of B. distachyon. The\npresence of QAs in plant tissues and the capacity of these\nmetabolites to provide stress protection and nutrients to plant\npathogens and symbionts were reported before (Boncompagni\net al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our\nstudy is among the first to highlight the potential importance of\nthese metabolites for rhizosphere interactions. Pseudomonads do\nnot synthesize QAs de novo but have evolved many pathways to\nscavenge them from eukaryotic hosts, where these metabolites\nare abundant due to the prominence of phosphatidylcholine\nin cellular membranes. Strains of P. fluorescens carry genes\nfor the conversion of choline, carnitine, and glycine betaine to\nglycine, as well as quaternary amine transporters of the BCCT\nand ABC families that are also conserved in the opportunistic\nhuman pathogen P. aeruginosa and the plant pathogen P. syringae\n(Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b).\nIn P. aeruginosa, choline catabolism genes are essential for\nthe ability of this pathogen to persist during lung infection\n(Wargo, 2013a). Similarly, a P. syringae mutant deficient in\nBetT, OpuC, and CbcXWV quaternary amine transporters had\nreduced fitness during colonization of bean and soybean leaves\nunder greenhouse and field conditions (Chen et al., 2013).\nDepending on water availability, P. aeruginosa and P. syringae\ncatabolize exogenously supplied QAs as carbon and nitrogen\nsources or accumulate them as osmoprotectants (Chen et al.,\n2013; Wargo, 2013b). Our ongoing work in P. synxantha 2–\n79 unraveled similar physiological responses and demonstrated\nthat QA transporters function differentially and redundantly\nin the uptake of quaternary amines as nutrients (Pablo and\nMavrodi, unpublished). In contrast, under water stress, the QAs\ncholine, betaine, and carnitine are accumulated preferentially for\nosmoprotection. Under drought stress, a 2–79 mutant devoid\nof all known QA transporters was less competitive in the\ncolonization of the Brachypodium rhizosphere than its wild-type\nparental strain. Interestingly, our metabolomic profiling of root\nexudates also revealed proline, glutamine, and hydroxyectoine.\nThese metabolites act as compatible solutes in different groups\nof microorganisms (Yancey et al., 1982; Empadinhas and da\nCosta, 2008), suggesting an important role of root exudates\nin the ability of Pseudomonas to persist in the rhizosphere of\ndrought-stressed plants.\nDATA AVAILABILITY STATEMENT\nThe original contributions presented in the study are publicly\navailable. This data can be found here: NCBI BioProject accession\nnumbers PRJNA439743 through PRJNA439790.\nAUTHOR CONTRIBUTIONS\nDM, OM, and LT conceived the research project. OM and\nJM collected root exudates. OM and DM cultured strains\nand extracted total RNA. AB and DG performed metabolomic\nanalysis of root exudates. DM, JP, and AF analyzed RNA-seq\ndata. LE, KH, and IP conducted Biolog analyses. DM, AF, OM,\nDW, and LT wrote the manuscript. All authors contributed to the\nmanuscript revision.\nFUNDING\nThis study was funded by NSF grant IOS-1656872 and by an\naward from the DOE Joint Genome Institute’s Community\nScience Program. The authors also acknowledge support from\nAustralian Research Council Discovery grant (DP160103746)\nand Mississippi INBRE, funded by an Institutional Development\nAward (IDeA) from the National Institute of General\nMedical Sciences of the National Institutes of Health under\ngrant P20GM103476.\nSUPPLEMENTARY MATERIAL\nThe Supplementary Material for this article can be found online\nat: https://www.frontiersin.org/articles/10.3389/fmicb.2021.\n651282/full#supplementary-material\nFrontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nREFERENCES\nAbayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018).\nDiscovery and characterization of a sulfoquinovose mutarotase using kinetic\nanalysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383.\ndoi: 10.1042/bcj20170947\nAndrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron\nhomeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03)\n00055-x\nArkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov,\nS., et al. (2018). KBase: the United States department of energy systems biology\nknowledgebase. Nat. Biotechnol. 36, 566–569.\nAutry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil\norganic sulfur. Biol. Fertil. Soils 10, 50–56.\nBabicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al.\n(2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44,\nW147–W153.\nBadri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates.\nPlant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x\nBadri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere\nchemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20,\n642–650. doi: 10.1016/j.copbio.2009.09.014\nBais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005).\nMediation of pathogen resistance by exudation of antimicrobials from roots.\nNature 434, 217–221. doi: 10.1038/nature03356\nBais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of\nroot exudates in rhizosphere interations with plants and other organisms. Annu.\nRev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159\nBangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic\nlocus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the\nbiological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe\nInteract. 9, 83–90.\nBarret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G.,\nand Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic\nfungus Gaeumannomyces graminis var. tritici on gene expression of the\nbiocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe\nInteract. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611\nBarrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas\nfluorescens biovar V: its resolution into distinct component groups and the\nrelationship of these groups to other P. fluorescens biovars, to P. putida, and to\npsychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol.\n132, 2709–2721. doi: 10.1099/00221287-132-10-2709\nBevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon\ngenomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21,\n211–217. doi: 10.1016/j.copbio.2010.03.006\nBoncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999).\nOccurrence of choline and glycine betaine uptake and metabolism in the family\nRhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65,\n2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999\nBondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper\nresistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397–\n405. doi: 10.1007/s10565-013-9262-1\nBrkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al.\n(2011). Brachypodium as a model for the grasses: today and the future. Plant\nPhysiol. 157, 3–13.\nCamacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type\n4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens\nWCS365. Dissertation, University of Leiden, Leiden.\nChavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A\nmetabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx\nin Pseudomonas putida. mSystems 1:e00154-16.\nChen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread\nplant-colonizing bacterial species Pseudomonas syringae detects and exploits an\nextracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj.\n12262\nChen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb:\na web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi:\n10.3389/fmicb.2018.01910\nChen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal\nacquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi.\n2017.06.004\nCole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M.,\net al. (2017). Genome-wide identification of bacterial plant colonization genes.\nPLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860\nConesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional\nanalysis in plant genomics. Int. J. Plant Genomics 2008:619832.\nCurl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag.\nDarzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive\npathways explorer v3. Nucleic Acids Res. 46, W510–W513.\nde Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al.\n(2006). The two-component colR/S system of Pseudomonas fluorescens WCS365\nplays a role in rhizosphere competence through maintaining the structure and\nfunction of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi:\n10.1111/j.1574-6941.2006.00158.x\nde Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg,\nG. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is\nan important trait for tomato root colonization by Pseudomonas fluorescens.\nMol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11.\n1173\nde Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand\nmicrobe-derived compounds affect the expression of genes encoding\nantifungal compounds in a pseudomonad with biocontrol activity. Appl.\nEnviron. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10\nDekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998).\nA site-specific recombinase is required for competitive root colonization by\nPseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056.\ndoi: 10.1073/pnas.95.12.7051\nDe-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008).\nRoot-microbe communication through protein secretion. J. Biol. Chem. 283,\n25247–25255. doi: 10.1074/jbc.m801967200\nEmpadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms\nin prokaryotes: distribution of compatible solutes. Int. Microbiol. 11,\n151–161.\nEspinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of\nfunctions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182,\n2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000\nFuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in\nPseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365-\n2958.2010.07250.x\nGalvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline-\nO-sulphate utilization from osmoprotection in Pseudomonas putida. Mol.\nMicrobiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x\nGarrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R.,\nand Redondo-Nieto, M. (2016). Genomic and genetic diversity within the\nPseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal.\npone.0150183\nHalverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and\nnonpermeating solutes on the fatty acid composition of Pseudomonas putida.\nAppl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421.\n2000\nHampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and\nWargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas\naeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13\nHarwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and\nthe biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/\nannurev.micro.50.1.553\nHesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018).\nGenome-based evolutionary history of Pseudomonas spp. Environ. Microbiol.\n20, 2142–2159.\nHinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003).\nTransition from reversible to irreversible attachment during biofilm formation\nby Pseudomonas fluorescens WCS365 requires an ABC transporter and a large\nsecreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003.\n03615.x\nHojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999).\nOxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring\nFrontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nthe distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65,\n4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999\nHong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological\ngrowth stages of Brachypodium distachyon: codification and description. Weed\nRes. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x\nHowell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum\ninduced damping-off of cotton seedlings by Pseudomonas fluorescens and\nits antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto-\n70-712\nHummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz,\nM. A. (1998). Regulation of the sulfate starvation response in Pseudomonas\naeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375–\n1386. doi: 10.1099/00221287-144-5-1375\nKabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam,\nC. (2017). Physiological and biochemical responses involved in water deficit\ntolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/\njournal.pone.0190284\nKamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005).\nEnrichment for enhanced competitive plant root tip colonizers selects for a new\nclass of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j.\n1462-2920.2005.00889.x\nKanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008).\nKEGG for linking genomes to life and the environment. Nucleic Acids Res. 36,\nD480–D484.\nKatoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment\nsoftware version 7: improvements in performance and usability. Mol. Biol. Evol.\n30, 772–780. doi: 10.1093/molbev/mst010\nKawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al.\n(2016). Microbiome and exudates of the root and rhizosphere of Brachypodium\ndistachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal.\npone.0164533\nKind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al.\n(2009). FiehnLib: mass spectral and retention index libraries for metabolomics\nbased on quadrupole and time-of-flight gas chromatography/mass\nspectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522\nKing, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the\ndemonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307.\nKlonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018).\nTranscriptomic profiling of Burkholderia phymatum STM815, Cupriavidus\ntaiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response\nto Mimosa pudica root exudates illuminates the molecular basis of their\nnodulation competitiveness and symbiotic evolutionary history.BMC Genomics\n19:105. doi: 10.1186/s12864-018-4487-2\nKobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and\nregulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/\nannurev-arplant-042811-105522\nKohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor\nIolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193,\n5155–5163. doi: 10.1128/jb.05371-11\nLee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas\nreinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7\nLetunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for\nthe display and annotation of phylogenetic and other trees. Nucleic Acids Res.\n44, W242–W245.\nLi, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of\northolog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi:\n10.1101/gr.1224503\nLiao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient\ngeneral purpose program for assigning sequence reads to genomic features.\nBioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656\nLiu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose\nutilization by Pseudomonas fluorescens: overlapping genetic responses to xylose,\nxylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi.\n13142\nLoper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer,\nB. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas\nspp.: insights into diversity and inheritance of traits involved in multitrophic\ninteractions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784\nLove, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change\nand dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550.\nLugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria.\nAnnu. Rev. Microbiol. 63, 541–556.\nLugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular\ndeterminants of rhizosphere colonization by Pseudomonas. Annu. Rev.\nPhytopathol. 39, 461–490.\nLynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 177–206.\nMark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005).\nTranscriptome profiling of bacterial responses to root exudates identifies genes\ninvolved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454–\n17459. doi: 10.1073/pnas.0506407102\nMarkowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y.,\net al. (2012). IMG: the integrated microbial genomes database and comparative\nanalysis system. Nucleic Acids Res. 40, D115–D122.\nMartinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF,\nthe second largest Pseudomonas putida protein, contributes to plant root\ncolonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561.\ndoi: 10.1111/j.1365-2958.2010.07249.x\nMatilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and\nRamos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving\nforces of bacterial life in the rhizosphere. Genome Biol. 8:R179.\nMiller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria.\nAnnu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101\nMoore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.-\nL., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes,\neds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt\n(New York, NY: Springer), 646–703.\nMulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis\nof the Pseudomonas species. Environ. Microbiol. 12, 1513–1530.\nNaylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought\nand host selection influence bacterial community dynamics in the grass root\nmicrobiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118\nNguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and\ncontrols. Agronomie 23, 375–396. doi: 10.1051/agro:2003011\nNielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface\ninteractions mediated by cellulose and a novel exopolysaccharide contribute\nto Pseudomonas putida biofilm formation and fitness under water-limiting\nconditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011.\n02432.x\nParejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow,\nL. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid\nproducing fluorescent Pseudomonas spp. from dryland cereal fields of central\nWashington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012-\n0015-0\nPhillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R.\n(2004). Microbial products trigger amino acid exudation from plant roots. Plant\nPhysiol. 136, 2887–2894. doi: 10.1104/pp.104.044222\nRaaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek,\nP. J., and Schippers, B. (1995). Utilization of heterologous siderophores and\nrhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41,\n126–135. doi: 10.1139/m95-017\nRaaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4-\ndiacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils.\nMol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144\nRahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis,\nJ., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase\ncontains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/\nacschembio.7b00304\nRamos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of\nPseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo\nexpression technology capture and identification of root-activated promoters.\nJ. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005\nReinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T.\n(2015). Roots shaping their microbiome: global hotspots for microbial activity.\nAnnu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10\n2342\nFrontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282\nMavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome\nSanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and\nRivilla, R. (2002). Phenotypic selection and phase variation occur during\nalfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184,\n1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002\nSarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995).\nThe sigma factor ss affects antibiotic production and biological control activity\nof Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259.\ndoi: 10.1073/pnas.92.26.12255\nSchnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The\nsigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance\ntowards desiccation and osmotic stress in the biocontrol agent Pseudomonas\nfluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem.\n67.12.5683-5693.2001\nSchroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006).\n“Phytopathogenic pseudomonads and related plant-associated pseudomonads,”\nin The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-\nH. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740.\ndoi: 10.1007/0-387-30746-x_23\nSchwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and\nAmasino, R. M. (2010). Natural variation of flowering time and vernalization\nresponsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10.\n1007/s12155-009-9069-3\nSilby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson,\nR. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity\nand plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51.\nSilby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to\nidentify genes important in growth and survival of Pseudomonas fluorescens\nPf0-1 in soil: Discovery of expressed sequences with novel genetic organization.\nJ. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004\nSimons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A.,\nand Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere\ncolonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant\nMicrobe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600\nSimons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and\nLugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root\ncolonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe\nInteract. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102\nSmibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods\nfor General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A.\nWood, and N. R. Kreig (Washington, DC: American Society of Microbiology),\n607–654.\nThomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from\nPseudomonas fluorescens in biological control of Gaeumannomyces graminis\nvar. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988\nThomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990).\nProduction of the antibiotic phenazine-1-carboxylic acid by fluorescent\nPseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56,\n908–912. doi: 10.1128/aem.56.4.908-912.1990\nTyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats,\nW. G., et al. (2014). Selection and phenotypic characterization of a core\ncollection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25.\ndoi: 10.1186/1471-2229-14-25\nVacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y.,\nMuller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system\nfunctioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356\nvan den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of\nphenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol.\n7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x\nvan Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity\nof microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135.\ndoi: 10.1128/.61.2.121-135.1997\nWalker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M.\n(2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric.\nFood Chem. 51, 2548–2554.\nWargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to\nPseudomonas aeruginosa survival during murine lung infection. PLoS One\n8:e56850. doi: 10.1371/journal.pone.0056850\nWargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine:\nlessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112–\n2120.\nWhipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch\n(Chichester: JohnWiley & Sons), 59–97.\nWinsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock,\nR. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly,\ncomprehensive comparisons of microbial genomes. Nucleic Acids Res. 37,\nD483–D488.\nYahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The\nProkaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E.\nStackebrandt (New York, NY: Springer), 704–713.\nYancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982).\nLiving with water stress: evolution of osmolyte systems. Science 217, 1214–1222.\nYe, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web\ntool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res.\n46, W71–W75.\nYoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H.,\net al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283,\n10415–10424.\nYousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different,\noverlapping mechanisms for colonization of abiotic and plant surfaces by\nPseudomonas putida. FEMS Microbiol. Lett. 288, 118–124.\nZboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere\ncolonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol.\nJ. 18, 3539–3554.\nZolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K.,\net al. (2013). “Understanding root-microbiome interactions,” in Molecular\nMicrobial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John\nWiley & Sons), 745–754.\nConflict of Interest: The authors declare that the research was conducted in the\nabsence of any commercial or financial relationships that could be construed as a\npotential conflict of interest.\nCopyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne,\nLeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an\nopen-access article distributed under the terms of the Creative Commons Attribution\nLicense (CC BY). The use, distribution or reproduction in other forums is permitted,\nprovided the original author(s) and the copyright owner(s) are credited and that the\noriginal publication in this journal is cited, in accordance with accepted academic\npractice. No use, distribution or reproduction is permitted which does not comply\nwith these terms.\nFrontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282\n\n\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:26:07 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Hashed cache key (SHA-256): b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:26:07 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +DEBUG:LiteLLM:Final hashed key: b1f7eb6221460714470c72292331f6af0ae09f3ca1a603d9f73dce5b5b6515c2 +23:26:07 - LiteLLM:INFO: utils.py:1026 - Wrapper: Completed Call, calling success_handler +INFO:LiteLLM:Wrapper: Completed Call, calling success_handler +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:982 - Logging Details LiteLLM-Success Call: Cache_hit=None +DEBUG:LiteLLM:Logging Details LiteLLM-Success Call: Cache_hit=None +23:26:07 - LiteLLM:DEBUG: cost_calculator.py:599 - completion_response response ms: None +DEBUG:LiteLLM:completion_response response ms: None +23:26:07 - LiteLLM:DEBUG: cost_calculator.py:599 - completion_response response ms: None +INFO:root:RAW TEXT: study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas +authors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi +doi: 10.3389/fmicb.2021.651282 +date: 14 April 2021 +experiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. +DEBUG:LiteLLM:completion_response response ms: None +INFO:root:PARSING LINE: study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas +DEBUG:root: FIELD: study_title +DEBUG:root:SLOT: study_title INL: None VALS: ['Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas'] +INFO:root:PARSING LINE: authors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi +DEBUG:root: FIELD: authors +DEBUG:root:SLOT: authors INL: None VALS: ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'] +INFO:root:PARSING LINE: doi: 10.3389/fmicb.2021.651282 +DEBUG:root: FIELD: doi +DEBUG:root:SLOT: doi INL: None VALS: ['10.3389/fmicb.2021.651282'] +INFO:root:PARSING LINE: date: 14 April 2021 +DEBUG:root: FIELD: date +DEBUG:root:SLOT: date INL: None VALS: ['14 April 2021'] +INFO:root:PARSING LINE: experiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. +DEBUG:root: FIELD: experiments +DEBUG:root:SLOT: experiments INL: True VALS: ['The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.'] +DEBUG:root: RECURSING ON SLOT: experiments, range=Experiment +INFO:ontogpt.clients.llm_client:Complete: engine=gpt-4.5-preview, prompt[5026]=From the text below, extract the following entities in the following format: + +label: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +DEBUG:LiteLLM:litellm.completion(api_key='', api_base=None, api_version=None, model='gpt-4.5-preview', messages=[{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], temperature=1.0, caching=True, custom_llm_provider='openai') +23:26:07 - LiteLLM:DEBUG: utils.py:296 - + +DEBUG:LiteLLM: + +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {} +DEBUG:LiteLLM:self.optional_params: {} +23:26:07 - LiteLLM:DEBUG: utils.py:296 - SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +DEBUG:LiteLLM:SYNC kwargs[caching]: True; litellm.cache: ; kwargs.get('cache')['no-cache']: False +23:26:07 - LiteLLM:DEBUG: utils.py:925 - INSIDE CHECKING SYNC CACHE +DEBUG:LiteLLM:INSIDE CHECKING SYNC CACHE +23:26:07 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:26:07 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:07 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:07 - LiteLLM:INFO: utils.py:2926 - +LiteLLM completion() model= gpt-4.5-preview; provider = openai +INFO:LiteLLM: +LiteLLM completion() model= gpt-4.5-preview; provider = openai +23:26:07 - LiteLLM:DEBUG: utils.py:2929 - +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +DEBUG:LiteLLM: +LiteLLM: Params passed to completion() {'functions': None, 'function_call': None, 'temperature': 1.0, 'top_p': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'additional_drop_params': None, 'custom_llm_provider': 'openai', 'drop_params': None, 'model': 'gpt-4.5-preview', 'n': None} +23:26:07 - LiteLLM:DEBUG: utils.py:2932 - +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +DEBUG:LiteLLM: +LiteLLM: Non-Default params passed to completion() {'temperature': 1.0} +23:26:07 - LiteLLM:DEBUG: utils.py:296 - Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:Final returned optional params: {'temperature': 1.0, 'extra_body': {}} +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:363 - self.optional_params: {'temperature': 1.0, 'extra_body': {}} +DEBUG:LiteLLM:self.optional_params: {'temperature': 1.0, 'extra_body': {}} +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:614 - PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +DEBUG:LiteLLM:PRE-API-CALL ADDITIONAL ARGS: {'headers': None, 'api_base': ParseResult(scheme='https', userinfo='', host='api.openai.com', port=None, path='/v1/', query=None, fragment=None), 'acompletion': False, 'complete_input_dict': {'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}} +23:26:07 - LiteLLM:DEBUG: litellm_logging.py:624 -  + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:LiteLLM: + +POST Request Sent from LiteLLM: +curl -X POST \ +https://api.openai.com/v1/ \ +-d '{'model': 'gpt-4.5-preview', 'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'temperature': 1.0, 'extra_body': {}}' + + +DEBUG:openai._base_client:Request options: {'method': 'post', 'url': '/chat/completions', 'headers': {'X-Stainless-Raw-Response': 'true'}, 'timeout': 600.0, 'files': None, 'json_data': {'messages': [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}], 'model': 'gpt-4.5-preview', 'temperature': 1.0}, 'extra_json': {}} +DEBUG:openai._base_client:Sending HTTP Request: POST https://api.openai.com/v1/chat/completions +DEBUG:httpcore.http11:send_request_headers.started request= +DEBUG:httpcore.http11:send_request_headers.complete +DEBUG:httpcore.http11:send_request_body.started request= +DEBUG:httpcore.http11:send_request_body.complete +DEBUG:httpcore.http11:receive_response_headers.started request= +DEBUG:httpcore.http11:receive_response_headers.complete return_value=(b'HTTP/1.1', 200, b'OK', [(b'Date', b'Wed, 05 Mar 2025 07:26:35 GMT'), (b'Content-Type', b'application/json'), (b'Transfer-Encoding', b'chunked'), (b'Connection', b'keep-alive'), (b'access-control-expose-headers', b'X-Request-ID'), (b'openai-organization', b'lawrence-berkeley-national-laboratory-8'), (b'openai-processing-ms', b'27531'), (b'openai-version', b'2020-10-01'), (b'x-ratelimit-limit-requests', b'10000'), (b'x-ratelimit-limit-tokens', b'2000000'), (b'x-ratelimit-remaining-requests', b'9999'), (b'x-ratelimit-remaining-tokens', b'1997008'), (b'x-ratelimit-reset-requests', b'6ms'), (b'x-ratelimit-reset-tokens', b'89ms'), (b'x-request-id', b'req_cd62a17c1834de22e9d897a211325165'), (b'strict-transport-security', b'max-age=31536000; includeSubDomains; preload'), (b'cf-cache-status', b'DYNAMIC'), (b'X-Content-Type-Options', b'nosniff'), (b'Server', b'cloudflare'), (b'CF-RAY', b'91b7e2224ecace4c-SJC'), (b'Content-Encoding', b'gzip'), (b'alt-svc', b'h3=":443"; ma=86400')]) +INFO:httpx:HTTP Request: POST https://api.openai.com/v1/chat/completions "HTTP/1.1 200 OK" +DEBUG:httpcore.http11:receive_response_body.started request= +DEBUG:httpcore.http11:receive_response_body.complete +DEBUG:httpcore.http11:response_closed.started +DEBUG:httpcore.http11:response_closed.complete +DEBUG:openai._base_client:HTTP Response: POST https://api.openai.com/v1/chat/completions "200 OK" Headers({'date': 'Wed, 05 Mar 2025 07:26:35 GMT', 'content-type': 'application/json', 'transfer-encoding': 'chunked', 'connection': 'keep-alive', 'access-control-expose-headers': 'X-Request-ID', 'openai-organization': 'lawrence-berkeley-national-laboratory-8', 'openai-processing-ms': '27531', 'openai-version': '2020-10-01', 'x-ratelimit-limit-requests': '10000', 'x-ratelimit-limit-tokens': '2000000', 'x-ratelimit-remaining-requests': '9999', 'x-ratelimit-remaining-tokens': '1997008', 'x-ratelimit-reset-requests': '6ms', 'x-ratelimit-reset-tokens': '89ms', 'x-request-id': 'req_cd62a17c1834de22e9d897a211325165', 'strict-transport-security': 'max-age=31536000; includeSubDomains; preload', 'cf-cache-status': 'DYNAMIC', 'x-content-type-options': 'nosniff', 'server': 'cloudflare', 'cf-ray': '91b7e2224ecace4c-SJC', 'content-encoding': 'gzip', 'alt-svc': 'h3=":443"; ma=86400'}) +DEBUG:openai._base_client:request_id: req_cd62a17c1834de22e9d897a211325165 +23:26:35 - LiteLLM:DEBUG: utils.py:296 - RAW RESPONSE: +{"id": "chatcmpl-B7dQlIq6tg5poLKl93TUpi5dfUF0M", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "label: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.\n\nexperiment_motivation: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.\n\nexperiment_design: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.\n\nenvironment: Controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C); microaerobic environment (15% O\u2082).\n\nhost: Brachypodium distachyon; Pseudomonas fluorescens-group strains\n\nhost_type: Plant; Bacteria\n\ntarget_microbes: Pseudomonas fluorescens\n\nbiological_system: Rhizosphere\n\nconditions: 20-hour light; 4-hour dark; 24\u00b0C daytime; 18\u00b0C nighttime; microaerobic (15% O\u2082)\n\nexperimental_factors: Presence of Brachypodium root exudates\n\nis_biolog_experiment: yes\n\ntype_of_biolog_experiment: Phenotype MicroArray\n\nplates: PM01; PM2A\n\nreplicates: 3\n\nprotocol_steps: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24\u00b0C, 4-hour dark 18\u00b0C); root exudates collected, sterilized via 0.22 \u00b5m filtration, lyophilized, stored at \u201380\u00b0C; root exudates analyzed by LC-MS and GC-MS; LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection; GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC; cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O\u2082) conditions; RNA extracted and depleted of rRNA; stranded library preparation and sequencing via Illumina HiSeq 2500; Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours; collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes; Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02; Transcriptome data analyzed using DESeq2 and comparative genomic tools.\n\nplate_reader_types_of_data_collected: Optical density; formazan (absorbance at 590 nm)\n\nod_protocol: Not provided\n\nrespiration_protocol: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25\u00b0C.\n\ninstrument_used: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer; Ultra Performance Liquid Chromatograph (UPLC); Pegasus 4D TOF Mass Spectrometer; Agilent 7890A GC; Illumina HiSeq 2500; OmniLog Phenotype MicroArray System\n\nanalysis_software: OmniLog Parametric Analysis software v1.20.02; DESeq2; OrthoMCL; COG-based phylogeny\n\nincubation_temperature: 25\u00b0C\n\nincubation_duration: 48 hours", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159567, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 818, "prompt_tokens": 1158, "total_tokens": 1976, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 0}}} + + +DEBUG:LiteLLM:RAW RESPONSE: +{"id": "chatcmpl-B7dQlIq6tg5poLKl93TUpi5dfUF0M", "choices": [{"finish_reason": "stop", "index": 0, "logprobs": null, "message": {"content": "label: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.\n\nexperiment_motivation: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.\n\nexperiment_design: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.\n\nenvironment: Controlled environment (20-hour light at 24\u00b0C, 4-hour dark at 18\u00b0C); microaerobic environment (15% O\u2082).\n\nhost: Brachypodium distachyon; Pseudomonas fluorescens-group strains\n\nhost_type: Plant; Bacteria\n\ntarget_microbes: Pseudomonas fluorescens\n\nbiological_system: Rhizosphere\n\nconditions: 20-hour light; 4-hour dark; 24\u00b0C daytime; 18\u00b0C nighttime; microaerobic (15% O\u2082)\n\nexperimental_factors: Presence of Brachypodium root exudates\n\nis_biolog_experiment: yes\n\ntype_of_biolog_experiment: Phenotype MicroArray\n\nplates: PM01; PM2A\n\nreplicates: 3\n\nprotocol_steps: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24\u00b0C, 4-hour dark 18\u00b0C); root exudates collected, sterilized via 0.22 \u00b5m filtration, lyophilized, stored at \u201380\u00b0C; root exudates analyzed by LC-MS and GC-MS; LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection; GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC; cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O\u2082) conditions; RNA extracted and depleted of rRNA; stranded library preparation and sequencing via Illumina HiSeq 2500; Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25\u00b0C for 48 hours; collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes; Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02; Transcriptome data analyzed using DESeq2 and comparative genomic tools.\n\nplate_reader_types_of_data_collected: Optical density; formazan (absorbance at 590 nm)\n\nod_protocol: Not provided\n\nrespiration_protocol: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25\u00b0C.\n\ninstrument_used: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer; Ultra Performance Liquid Chromatograph (UPLC); Pegasus 4D TOF Mass Spectrometer; Agilent 7890A GC; Illumina HiSeq 2500; OmniLog Phenotype MicroArray System\n\nanalysis_software: OmniLog Parametric Analysis software v1.20.02; DESeq2; OrthoMCL; COG-based phylogeny\n\nincubation_temperature: 25\u00b0C\n\nincubation_duration: 48 hours", "refusal": null, "role": "assistant", "audio": null, "function_call": null, "tool_calls": null}}], "created": 1741159567, "model": "gpt-4.5-preview-2025-02-27", "object": "chat.completion", "service_tier": "default", "system_fingerprint": null, "usage": {"completion_tokens": 818, "prompt_tokens": 1158, "total_tokens": 1976, "completion_tokens_details": {"accepted_prediction_tokens": 0, "audio_tokens": 0, "reasoning_tokens": 0, "rejected_prediction_tokens": 0}, "prompt_tokens_details": {"audio_tokens": 0, "cached_tokens": 0}}} + + +23:26:35 - LiteLLM:DEBUG: caching.py:268 - +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +DEBUG:LiteLLM: +Created cache key: model: gpt-4.5-previewmessages: [{'content': "From the text below, extract the following entities in the following format:\n\nlabel: \nexperiment_motivation: \nexperiment_design: \nenvironment: \nhost: \nhost_type: \ntarget_microbes: \nbiological_system: \nconditions: \nexperimental_factors: \nis_biolog_experiment: \ntype_of_biolog_experiment: \nplates: \nreplicates: \nprotocol_steps: \nplate_reader_types_of_data_collected: \nod_protocol: \nrespiration_protocol: \ninstrument_used: \nanalysis_software: \nincubation_temperature: \nincubation_duration: \n\n\nText:\nThe experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization.\n\n===\n\n", 'role': 'user'}]temperature: 1.0 +23:26:35 - LiteLLM:DEBUG: caching.py:445 - Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Hashed cache key (SHA-256): fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:35 - LiteLLM:DEBUG: caching.py:462 - Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +DEBUG:LiteLLM:Final hashed key: fe70325c30c52969b0030ddafdc5914313d260ea2120330381f82b42ad828a85 +23:26:35 - LiteLLM:INFO: utils.py:1026 - Wrapper: Completed Call, calling success_handler +INFO:LiteLLM:Wrapper: Completed Call, calling success_handler +23:26:35 - LiteLLM:DEBUG: cost_calculator.py:599 - completion_response response ms: None +DEBUG:LiteLLM:completion_response response ms: None +INFO:root:PARSING LINE: label: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates. +DEBUG:root: FIELD: label +DEBUG:root:SLOT: label INL: None VALS: ['This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.'] +INFO:root:PARSING LINE: experiment_motivation: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates. +DEBUG:root: FIELD: experiment_motivation +23:26:35 - LiteLLM:DEBUG: litellm_logging.py:982 - Logging Details LiteLLM-Success Call: Cache_hit=None +DEBUG:root:SLOT: experiment_motivation INL: None VALS: ['To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.'] +INFO:root:PARSING LINE: experiment_design: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays. +DEBUG:root: FIELD: experiment_design +DEBUG:LiteLLM:Logging Details LiteLLM-Success Call: Cache_hit=None +DEBUG:root:SLOT: experiment_design INL: None VALS: ['Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.'] +INFO:root:PARSING LINE: environment: Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂). +DEBUG:root: FIELD: environment +DEBUG:root:SLOT: environment INL: None VALS: ['Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).'] +INFO:root:PARSING LINE: host: Brachypodium distachyon; Pseudomonas fluorescens-group strains +DEBUG:root: FIELD: host +DEBUG:root:SLOT: host INL: False VALS: ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'] +INFO:root:PARSING LINE: host_type: Plant; Bacteria +DEBUG:root: FIELD: host_type +DEBUG:root:SLOT: host_type INL: None VALS: ['Plant; Bacteria'] +INFO:root:PARSING LINE: target_microbes: Pseudomonas fluorescens +DEBUG:root: FIELD: target_microbes +DEBUG:root:SLOT: target_microbes INL: False VALS: ['Pseudomonas fluorescens'] +INFO:root:PARSING LINE: biological_system: Rhizosphere +DEBUG:root: FIELD: biological_system +DEBUG:root:SLOT: biological_system INL: None VALS: ['Rhizosphere'] +INFO:root:PARSING LINE: conditions: 20-hour light; 4-hour dark; 24°C daytime; 18°C nighttime; microaerobic (15% O₂) +DEBUG:root: FIELD: conditions +DEBUG:root:SLOT: conditions INL: None VALS: ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'] +INFO:root:PARSING LINE: experimental_factors: Presence of Brachypodium root exudates +DEBUG:root: FIELD: experimental_factors +DEBUG:root:SLOT: experimental_factors INL: False VALS: ['Presence of Brachypodium root exudates'] +INFO:root:PARSING LINE: is_biolog_experiment: yes +DEBUG:root: FIELD: is_biolog_experiment +DEBUG:root:SLOT: is_biolog_experiment INL: None VALS: ['yes'] +INFO:root:PARSING LINE: type_of_biolog_experiment: Phenotype MicroArray +DEBUG:root: FIELD: type_of_biolog_experiment +DEBUG:root:SLOT: type_of_biolog_experiment INL: None VALS: ['Phenotype MicroArray'] +INFO:root:PARSING LINE: plates: PM01; PM2A +DEBUG:root: FIELD: plates +DEBUG:root:SLOT: plates INL: None VALS: ['PM01', 'PM2A'] +INFO:root:PARSING LINE: replicates: 3 +DEBUG:root: FIELD: replicates +DEBUG:root:SLOT: replicates INL: None VALS: ['3'] +INFO:root:PARSING LINE: protocol_steps: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C); root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C; root exudates analyzed by LC-MS and GC-MS; LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection; GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC; cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions; RNA extracted and depleted of rRNA; stranded library preparation and sequencing via Illumina HiSeq 2500; Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours; collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes; Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02; Transcriptome data analyzed using DESeq2 and comparative genomic tools. +DEBUG:root: FIELD: protocol_steps +DEBUG:root:SLOT: protocol_steps INL: None VALS: ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'] +INFO:root:PARSING LINE: plate_reader_types_of_data_collected: Optical density; formazan (absorbance at 590 nm) +DEBUG:root: FIELD: plate_reader_types_of_data_collected +DEBUG:root:SLOT: plate_reader_types_of_data_collected INL: None VALS: ['Optical density', 'formazan (absorbance at 590 nm)'] +INFO:root:PARSING LINE: od_protocol: Not provided +DEBUG:root: FIELD: od_protocol +DEBUG:root:SLOT: od_protocol INL: None VALS: ['Not provided'] +INFO:root:PARSING LINE: respiration_protocol: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C. +DEBUG:root: FIELD: respiration_protocol +DEBUG:root:SLOT: respiration_protocol INL: None VALS: ['Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.'] +INFO:root:PARSING LINE: instrument_used: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer; Ultra Performance Liquid Chromatograph (UPLC); Pegasus 4D TOF Mass Spectrometer; Agilent 7890A GC; Illumina HiSeq 2500; OmniLog Phenotype MicroArray System +DEBUG:root: FIELD: instrument_used +DEBUG:root:SLOT: instrument_used INL: None VALS: ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'] +INFO:root:PARSING LINE: analysis_software: OmniLog Parametric Analysis software v1.20.02; DESeq2; OrthoMCL; COG-based phylogeny +DEBUG:root: FIELD: analysis_software +DEBUG:root:SLOT: analysis_software INL: None VALS: ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'] +INFO:root:PARSING LINE: incubation_temperature: 25°C +DEBUG:root: FIELD: incubation_temperature +DEBUG:root:SLOT: incubation_temperature INL: None VALS: ['25°C'] +INFO:root:PARSING LINE: incubation_duration: 48 hours +DEBUG:root: FIELD: incubation_duration +DEBUG:root:SLOT: incubation_duration INL: None VALS: ['48 hours'] +DEBUG:root:RAW: {'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [{'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'}]} +DEBUG:root:Grounding annotation object {'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [{'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'}]} +DEBUG:root:FIELD: study_title SLOT: study_title +DEBUG:root: VAL: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas +DEBUG:root:FIELD: authors SLOT: authors +DEBUG:root: VAL: Olga V. Mavrodi +DEBUG:root: VAL: Janiece R. McWilliams +DEBUG:root: VAL: Jacob O. Peter +DEBUG:root: VAL: Anna Berim +DEBUG:root: VAL: Karl A. Hassan +DEBUG:root: VAL: Liam D. H. Elbourne +DEBUG:root: VAL: Melissa K. LeTourneau +DEBUG:root: VAL: David R. Gang +DEBUG:root: VAL: Ian T. Paulsen +DEBUG:root: VAL: David M. Weller +DEBUG:root: VAL: Linda S. Thomashow +DEBUG:root: VAL: Alex S. Flynt +DEBUG:root: VAL: Dmitri V. Mavrodi +DEBUG:root:FIELD: doi SLOT: doi +DEBUG:root: VAL: 10.3389/fmicb.2021.651282 +DEBUG:root:FIELD: date SLOT: date +DEBUG:root: VAL: 14 April 2021 +DEBUG:root:FIELD: experiments SLOT: experiments +DEBUG:root: VAL: {'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +DEBUG:root:Grounding annotation object {'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['Brachypodium distachyon', 'Pseudomonas fluorescens-group strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['Pseudomonas fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['Presence of Brachypodium root exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +DEBUG:root:FIELD: label SLOT: label +DEBUG:root: VAL: This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates. +DEBUG:root:FIELD: experiment_motivation SLOT: experiment_motivation +DEBUG:root: VAL: To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates. +DEBUG:root:FIELD: experiment_design SLOT: experiment_design +DEBUG:root: VAL: Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays. +DEBUG:root:FIELD: environment SLOT: environment +DEBUG:root: VAL: Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂). +DEBUG:root:FIELD: host SLOT: host +DEBUG:root: VAL: Brachypodium distachyon +INFO:ontogpt.engines.knowledge_engine:GROUNDING Brachypodium distachyon using Host +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Brachypodium distachyon to Host +DEBUG:root: VAL: Pseudomonas fluorescens-group strains +INFO:ontogpt.engines.knowledge_engine:GROUNDING Pseudomonas fluorescens-group strains using Host +INFO:ontogpt.engines.knowledge_engine:Singularized Pseudomonas fluorescens-group strains to pseudomonas fluorescens-group strain +INFO:ontogpt.engines.knowledge_engine:GROUNDING pseudomonas fluorescens-group strain using Host +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Pseudomonas fluorescens-group strains to Host +DEBUG:root:FIELD: host_type SLOT: host_type +DEBUG:root: VAL: Plant; Bacteria +DEBUG:root:FIELD: target_microbes SLOT: target_microbes +DEBUG:root: VAL: Pseudomonas fluorescens +INFO:ontogpt.engines.knowledge_engine:GROUNDING Pseudomonas fluorescens using Microbe +INFO:ontogpt.engines.knowledge_engine:Singularized Pseudomonas fluorescens to pseudomonas fluorescen +INFO:ontogpt.engines.knowledge_engine:GROUNDING pseudomonas fluorescen using Microbe +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Pseudomonas fluorescens to Microbe +DEBUG:root:FIELD: biological_system SLOT: biological_system +DEBUG:root: VAL: Rhizosphere +DEBUG:root:FIELD: conditions SLOT: conditions +DEBUG:root: VAL: 20-hour light +DEBUG:root: VAL: 4-hour dark +DEBUG:root: VAL: 24°C daytime +DEBUG:root: VAL: 18°C nighttime +DEBUG:root: VAL: microaerobic (15% O₂) +DEBUG:root:FIELD: experimental_factors SLOT: experimental_factors +DEBUG:root: VAL: Presence of Brachypodium root exudates +INFO:ontogpt.engines.knowledge_engine:GROUNDING Presence of Brachypodium root exudates using ExperimentalFactor +INFO:ontogpt.engines.knowledge_engine:Singularized Presence of Brachypodium root exudates to presence of brachypodium root exudate +INFO:ontogpt.engines.knowledge_engine:GROUNDING presence of brachypodium root exudate using ExperimentalFactor +INFO:ontogpt.engines.knowledge_engine:Could not ground and normalize Presence of Brachypodium root exudates to ExperimentalFactor +DEBUG:root:FIELD: is_biolog_experiment SLOT: is_biolog_experiment +DEBUG:root: VAL: yes +INFO:root:Looking for yes in FlexibleBooleanEnum +DEBUG:root:FIELD: type_of_biolog_experiment SLOT: type_of_biolog_experiment +DEBUG:root: VAL: Phenotype MicroArray +DEBUG:root:FIELD: plates SLOT: plates +DEBUG:root: VAL: PM01 +DEBUG:root: VAL: PM2A +DEBUG:root:FIELD: replicates SLOT: replicates +DEBUG:root: VAL: 3 +DEBUG:root:FIELD: protocol_steps SLOT: protocol_steps +DEBUG:root: VAL: Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C) +DEBUG:root: VAL: root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C +DEBUG:root: VAL: root exudates analyzed by LC-MS and GC-MS +DEBUG:root: VAL: LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection +DEBUG:root: VAL: GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC +DEBUG:root: VAL: cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions +DEBUG:root: VAL: RNA extracted and depleted of rRNA +DEBUG:root: VAL: stranded library preparation and sequencing via Illumina HiSeq 2500 +DEBUG:root: VAL: Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours +DEBUG:root: VAL: collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes +DEBUG:root: VAL: Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02 +DEBUG:root: VAL: Transcriptome data analyzed using DESeq2 and comparative genomic tools. +DEBUG:root:FIELD: plate_reader_types_of_data_collected SLOT: plate_reader_types_of_data_collected +DEBUG:root: VAL: Optical density +DEBUG:root: VAL: formazan (absorbance at 590 nm) +DEBUG:root:FIELD: od_protocol SLOT: od_protocol +DEBUG:root: VAL: Not provided +DEBUG:root:FIELD: respiration_protocol SLOT: respiration_protocol +DEBUG:root: VAL: Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C. +DEBUG:root:FIELD: instrument_used SLOT: instrument_used +DEBUG:root: VAL: Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer +DEBUG:root: VAL: Ultra Performance Liquid Chromatograph (UPLC) +DEBUG:root: VAL: Pegasus 4D TOF Mass Spectrometer +DEBUG:root: VAL: Agilent 7890A GC +DEBUG:root: VAL: Illumina HiSeq 2500 +DEBUG:root: VAL: OmniLog Phenotype MicroArray System +DEBUG:root:FIELD: analysis_software SLOT: analysis_software +DEBUG:root: VAL: OmniLog Parametric Analysis software v1.20.02 +DEBUG:root: VAL: DESeq2 +DEBUG:root: VAL: OrthoMCL +DEBUG:root: VAL: COG-based phylogeny +DEBUG:root:FIELD: incubation_temperature SLOT: incubation_temperature +DEBUG:root: VAL: 25°C +DEBUG:root:FIELD: incubation_duration SLOT: incubation_duration +DEBUG:root: VAL: 48 hours +DEBUG:root:Creating object from dict {'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['AUTO:Pseudomonas%20fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +INFO:root:{'label': 'This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', 'experiment_motivation': 'To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', 'experiment_design': 'Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', 'environment': 'Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', 'host': ['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], 'host_type': 'Plant; Bacteria', 'target_microbes': ['AUTO:Pseudomonas%20fluorescens'], 'biological_system': 'Rhizosphere', 'conditions': ['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], 'experimental_factors': ['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], 'is_biolog_experiment': 'yes', 'type_of_biolog_experiment': 'Phenotype MicroArray', 'plates': ['PM01', 'PM2A'], 'replicates': '3', 'protocol_steps': ['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], 'plate_reader_types_of_data_collected': ['Optical density', 'formazan (absorbance at 590 nm)'], 'od_protocol': 'Not provided', 'respiration_protocol': 'Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', 'instrument_used': ['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], 'analysis_software': ['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], 'incubation_temperature': '25°C', 'incubation_duration': '48 hours'} +DEBUG:root:Creating object from dict {'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [Experiment(label='This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', experiment_motivation='To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', experiment_design='Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', environment='Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', host=['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], host_type='Plant; Bacteria', target_microbes=['AUTO:Pseudomonas%20fluorescens'], biological_system='Rhizosphere', conditions=['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], experimental_factors=['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], is_biolog_experiment='yes', type_of_biolog_experiment='Phenotype MicroArray', plates=['PM01', 'PM2A'], replicates='3', protocol_steps=['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], plate_reader_types_of_data_collected=['Optical density', 'formazan (absorbance at 590 nm)'], od_protocol='Not provided', respiration_protocol='Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', instrument_used=['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], analysis_software=['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], incubation_temperature='25°C', incubation_duration='48 hours')]} +INFO:root:{'study_title': 'Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas', 'authors': ['Olga V. Mavrodi', 'Janiece R. McWilliams', 'Jacob O. Peter', 'Anna Berim', 'Karl A. Hassan', 'Liam D. H. Elbourne', 'Melissa K. LeTourneau', 'David R. Gang', 'Ian T. Paulsen', 'David M. Weller', 'Linda S. Thomashow', 'Alex S. Flynt', 'Dmitri V. Mavrodi'], 'doi': '10.3389/fmicb.2021.651282', 'date': '14 April 2021', 'experiments': [Experiment(label='This experiment investigates the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon by analyzing bacterial transcriptional and metabolic responses to plant root exudates.', experiment_motivation='To examine molecular communication between beneficial rhizosphere bacteria and their host plant, specifically focusing on bacterial responses to plant-derived metabolites present within root exudates.', experiment_design='Surface-sterilized Brachypodium distachyon plants were grown hydroponically to collect root exudates, which were characterized by LC-MS and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with root exudates under controlled microaerobic conditions, followed by transcriptomic analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays.', environment='Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); microaerobic environment (15% O₂).', host=['AUTO:Brachypodium%20distachyon', 'AUTO:Pseudomonas%20fluorescens-group%20strains'], host_type='Plant; Bacteria', target_microbes=['AUTO:Pseudomonas%20fluorescens'], biological_system='Rhizosphere', conditions=['20-hour light', '4-hour dark', '24°C daytime', '18°C nighttime', 'microaerobic (15% O₂)'], experimental_factors=['AUTO:Presence%20of%20Brachypodium%20root%20exudates'], is_biolog_experiment='yes', type_of_biolog_experiment='Phenotype MicroArray', plates=['PM01', 'PM2A'], replicates='3', protocol_steps=['Surface-sterilized plants grown hydroponically in sterile jars filled with glass beads and distilled water under controlled conditions (20-hour light 24°C, 4-hour dark 18°C)', 'root exudates collected, sterilized via 0.22 µm filtration, lyophilized, stored at –80°C', 'root exudates analyzed by LC-MS and GC-MS', 'LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection', 'GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to Agilent 7890A GC', 'cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown on medium with root exudates under microaerobic (15% O₂) conditions', 'RNA extracted and depleted of rRNA', 'stranded library preparation and sequencing via Illumina HiSeq 2500', 'Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation in OmniLog Phenotype MicroArray System at 25°C for 48 hours', 'collection of optical density and colorimetric respiration data (formazan production, absorbance at 590 nm) every 15 minutes', 'Biolog data analysis performed using OmniLog Parametric Analysis software v1.20.02', 'Transcriptome data analyzed using DESeq2 and comparative genomic tools.'], plate_reader_types_of_data_collected=['Optical density', 'formazan (absorbance at 590 nm)'], od_protocol='Not provided', respiration_protocol='Colorimetric respiration measured by formazan production (absorbance detection at 590 nm), collected every 15 minutes over a period of 48 hours at 25°C.', instrument_used=['Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer', 'Ultra Performance Liquid Chromatograph (UPLC)', 'Pegasus 4D TOF Mass Spectrometer', 'Agilent 7890A GC', 'Illumina HiSeq 2500', 'OmniLog Phenotype MicroArray System'], analysis_software=['OmniLog Parametric Analysis software v1.20.02', 'DESeq2', 'OrthoMCL', 'COG-based phylogeny'], incubation_temperature='25°C', incubation_duration='48 hours')]} +INFO:root:Output format: yaml +--- +input_text: |+ + ORIGINAL RESEARCH + published: 14 April 2021 + doi: 10.3389/fmicb.2021.651282 + Edited by: + Barbara Pivato, + Institut National de Recherche pour + l’agriculture, l’alimentation et + l’environnement (INRAE), France + Reviewed by: + Cara Helene Haney, + The University of British Columbia, + Canada + Xingang Zhou, + Northeast Agricultural University, + China + *Correspondence: + Dmitri V. Mavrodi + dmitri.mavrodi@usm.edu + Alex S. Flynt + alex.flynt@usm.edu + †These authors have contributed + equally to this work + Specialty section: + This article was submitted to + Microbe and Virus Interactions with + Plants, + a section of the journal + Frontiers in Microbiology + Received: 09 January 2021 + Accepted: 08 March 2021 + Published: 14 April 2021 + Citation: + Mavrodi OV, McWilliams JR, + Peter JO, Berim A, Hassan KA, + Elbourne LDH, LeTourneau MK, + Gang DR, Paulsen IT, Weller DM, + Thomashow LS, Flynt AS and + Mavrodi DV (2021) Root Exudates + Alter the Expression of Diverse + Metabolic, Transport, Regulatory, + and Stress Response Genes + in Rhizosphere Pseudomonas. + Front. Microbiol. 12:651282. + doi: 10.3389/fmicb.2021.651282 + Root Exudates Alter the Expression + of Diverse Metabolic, Transport, + Regulatory, and Stress Response + Genes in Rhizosphere Pseudomonas + Olga V. Mavrodi1†, Janiece R. McWilliams1†, Jacob O. Peter1, Anna Berim2, + Karl A. Hassan3, Liam D. H. Elbourne4, Melissa K. LeTourneau5, David R. Gang2, + Ian T. Paulsen4, David M. Weller5, Linda S. Thomashow5, Alex S. Flynt1* and + Dmitri V. Mavrodi1* + 1 School of Biological, Environmental, and Earth Sciences, The University of Southern Mississippi, Hattiesburg, MS, + United States, 2 Institute of Biological Chemistry, Washington State University, Pullman, WA, United States, 3 School + of Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW, Australia, 4 Department of Molecular + Sciences, Macquarie University, Sydney, NSW, Australia, 5 USDA Agricultural Research Service, Wheat Health, Genetics + and Quality Research Unit, Pullman, WA, United States + Plants live in association with microorganisms that positively influence plant + development, vigor, and fitness in response to pathogens and abiotic stressors. + The bulk of the plant microbiome is concentrated belowground at the plant rootsoil + interface. Plant roots secrete carbon-rich rhizodeposits containing primary and + secondary low molecular weight metabolites, lysates, and mucilages. These exudates + provide nutrients for soil microorganisms and modulate their affinity to host plants, but + molecular details of this process are largely unresolved. We addressed this gap by + focusing on the molecular dialog between eight well-characterized beneficial strains + of the Pseudomonas fluorescens group and Brachypodium distachyon, a model for + economically important food, feed, forage, and biomass crops of the grass family. We + collected and analyzed root exudates of B. distachyon and demonstrated the presence + of multiple carbohydrates, amino acids, organic acids, and phenolic compounds. The + subsequent screening of bacteria by Biolog Phenotype MicroArrays revealed that many + of these metabolites provide carbon and energy for the Pseudomonas strains. RNAseq + profiling of bacterial cultures amended with root exudates revealed changes in the + expression of genes encoding numerous catabolic and anabolic enzymes, transporters, + transcriptional regulators, stress response, and conserved hypothetical proteins. Almost + half of the differentially expressed genes mapped to the variable part of the strains’ + pangenome, reflecting the importance of the variable gene content in the adaptation of + P. fluorescens to the rhizosphere lifestyle. Our results collectively reveal the diversity + of cellular pathways and physiological responses underlying the establishment of + mutualistic interactions between these beneficial rhizobacteria and their plant hosts. + Keywords: Pseudomonas, Brachypodium, rhizosphere, root exudates, transcriptome + Frontiers in Microbiology | www.frontiersin.org 1 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + INTRODUCTION + Plants are meta-organisms or holobionts that rely in part on + their microbiome for specific functions and traits. The ability + of the plant microbiome to influence plant development, vigor, + health, and fitness in response to abiotic stressors associated + with global climate change is documented by numerous studies + (Lugtenberg and Kamilova, 2009). There is mounting evidence + that plants actively recruit beneficial microbiomes, but many + aspects of this process are still very much a black box (Reinhold- + Hurek et al., 2015). The foundation for the differential affinity + of rhizobacteria toward host plants is built upon complex + chemical cross talk between microorganisms and plant roots. + Up to 40% of photosynthetically fixed carbon is released by + plant roots in the form of exudates and secretions, lysates, and + mucilages (Curl and Truelove, 1986; Lynch, 1990; Whipps, 1990; + Badri and Vivanco, 2009). The release of these compounds is + actively controlled in response to environmental stimuli, and the + composition of root exudates varies greatly according to plant + species and physiological condition (Lynch, 1990; Nguyen, 2003; + Phillips et al., 2004; De-la-Pena et al., 2008). The presence and + composition of exudates strongly impact soil microorganisms, + which is consistent with the idea that plants actively select and + shape their root microbiota (Zolla et al., 2013). + Primary root exudates include simple and complex sugars, + amino acids, polypeptides and proteins, organic, aliphatic and + fatty acids, sterols, and phenolics (Nguyen, 2003; Badri and + Vivanco, 2009; Badri et al., 2009). These compounds serve as + carbon and energy sources for rhizobacteria, and the presence + of the intact corresponding catabolic pathways is essential + for competitive colonization of roots and disease suppression + (Lugtenberg et al., 2001; Kamilova et al., 2005; Lugtenberg and + Kamilova, 2009). Root exudates also contain numerous signal + molecules and secondary metabolites, the significance of which + is only now emerging (Walker et al., 2003; Bais et al., 2005, + 2006). A handful of analyses of plant-induced gene expression + by transcriptional profiling in vitro (Mark et al., 2005) or in + the rhizosphere (Silby and Levy, 2004; Ramos-Gonzalez et al., + 2005; Matilla et al., 2007; Barret et al., 2009) have identified + multiple genes that are differentially regulated by exposure to + roots or root exudates. Bacterial pathways expressed during + rhizosphere colonization control utilization of plant-derived + metabolites (Simons et al., 1996, 1997; Camacho-Carvajal, 2001; + Lugtenberg and Kamilova, 2009), motility and chemotaxis (de + Weert et al., 2002; Lugtenberg and Kamilova, 2009), phase + variation (Dekkers et al., 1998; Sanchez-Contreras et al., 2002; + van den Broek et al., 2005), outer membrane integrity (de Weert + et al., 2006; Lugtenberg and Kamilova, 2009), and the ability + to sequester limiting resources (Raaijmakers et al., 1995) and + resist environmental stresses (Sarniguet et al., 1995; Miller and + Wood, 1996; van Veen et al., 1997; Schnider-Keel et al., 2001). In + its spatial and temporal properties, root colonization resembles + biofilm formation, and biofilm-related pathways also have been + implicated in adhesion to seeds and roots and rhizosphere + colonization (Espinosa-Urgel et al., 2000; Hinsa et al., 2003; + Yousef-Coronado et al., 2008; Fuqua, 2010; Martinez-Gil et al., + 2010; Nielsen et al., 2011; Zboralski and Filion, 2020). Finally, + root exudates strongly affect the expression of diverse plant + growth promotion and biocontrol genes (Vacheron et al., 2013). + Over the past decade, the genomes of numerous rhizosphere + strains have been sequenced and analyzed, but functional + genomics studies of rhizosphere competence lag behind the + availability of sequence data. + This study explored the molecular dialog between the + model host plant Brachypodium distachyon and several wellcharacterized + rhizosphere strains of the Pseudomonas fluorescens + group. Brachypodium is a small annual grass originating in semiarid + regions of the Middle East that has emerged as a prime model + for economically important food, feed, forage, and biomass crops + of the grass family (Bevan et al., 2010; Schwartz et al., 2010; + Brkljacic et al., 2011; Hong et al., 2011; Tyler et al., 2014). The + biology, extensive collection of resources, and research tools + make B. distachyon an attractive model to investigate interactions + between plants and root-associated microbes. Pseudomonads + are ubiquitous Gram-negative g-proteobacteria that colonize + eukaryotic hosts and include both commensals and economically + important pathogens of plants and animals (Moore et al., + 2006; Schroth et al., 2006; Yahr and Parsek, 2006). The genus + Pseudomonas currently comprises > 100 named species that have + been separated based on multilocus sequence analysis into 14 + species groups (Garrido-Sanz et al., 2016; Hesse et al., 2018). + The P. fluorescens group is the most diverse regarding both + the genetic distances within it, the number of species and the + large pangenome that makes up > 50% of the pangenome + of the genus as a whole (Loper et al., 2012). The group also + encompasses an unusually high proportion of strains that inhabit + the plant rhizosphere and possess plant growth promoting and + biocontrol properties. Naylor et al. (2017) profiled bacterial + communities associated with root tissues and rhizosphere of + 18 different plant species of the Poaceae family. That study + identified Pseudomonas among taxa constituting the core grass + root microbiome and demonstrated that these bacteria were + enriched in C3 plants, including wheat, rye, barley, oat, and + Brachypodium. We confirmed the capacity of B. distachyon + Bd21 to serve as a host for rhizobacteria of the P. fluorescens + group in preliminary greenhouse assays with biocontrol strains + P. synxantha 2-79, P. brassicacearum Q8r1-96, and P. protegens + Pf-5. Results of these experiments revealed that all strains + successfully established and colonized the roots of Brachypodium + (Supplementary Table 1). + In this study, we focused on eight well-studied strains + of the P. fluorescens complex that are supported by years + of studies, numerous refereed publications, and high-quality + genome sequences. By profiling transcriptomes of these strains + during growth in root exudates of B. distachyon, we revealed the + diversity of cellular pathways and physiological responses that + underlie the establishment of mutualistic interactions between + beneficial rhizobacteria and the host plant. Our results also + confirmed that root exudates contain carbohydrates, amino acids, + organic acids, and phenolics that serve as carbon and energy + sources for rhizobacteria. The root exudates also contained + osmoprotectants that may help microorganisms to persist in the + rhizosphere of drought-stressed plants. The diversity of microbial + genes perturbed by root exudates reflects the importance of + Frontiers in Microbiology | www.frontiersin.org 2 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + the variable genome in adaptation of individual strains of + Pseudomonas to the rhizosphere lifestyle. + MATERIALS AND METHODS + Bacterial Strains Used in the Study + The eight Pseudomonas strains used for this study are + P. synxantha 2-79 (Thomashow and Weller, 1988), P. fluorescens + SBW25 (Silby et al., 2009), Pseudomonas sp. R1-43-08 (Parejko + et al., 2012), P. brassicacearum Q8r1-96 (Raaijmakers and Weller, + 1998), P. fluorescens Q2-87 (Bangera and Thomashow, 1996), + P. chlororaphis 30-84 (Thomashow et al., 1990), P. fluorescens + Pf0-1 (Silby et al., 2009), and P. protegens Pf-5 (Howell and + Stipanovic, 1980). The selected organisms have been studied + extensively for their role in biological control and plant + growth promotion (Supplementary Table 2). The strains were + maintained in the laboratory as frozen stocks (􀀀80 C) and + routinely cultured in King’s medium B (King et al., 1954) or + 21C medium, which contained (per 1 L): 1.0 g of NH4Cl, 3.5 g + of Na2HPO4 2H2O, 2.8 g of KH2PO4, 3.0 g of glucose, and + 20 ml of a microelement solution (Smibert and Kreig, 1994; + Halverson and Firestone, 2000). + Propagation of Plants and Collection of + Root Exudates + B. distachyon Bd21 was established from seed material obtained + from the USDA-ARS Plant Germplasm Introduction and Testing + Research Unit (Pullman, WA, United States). Brachypodium + seeds were imbibed for 3 days at 4 C and sown in 7   7 cm + pots filled with Sunshine Potting Mix #4 (Sun Gro Horticulture, + Agawam, MA, United States). Plants were grown in an IR- + 89X (Percival Scientific, Perry, IA, United States) controlled + environment chamber retrofitted with 6500K and 3000K T5 54W + grow lights (Spectralux) under a 20-h light, 24 C/4-h dark, 18 C + cycle. Plants were watered and fertilized with Jack’s professional + water-soluble fertilizer (20:20:20) (JR Peters, Allentown, PA, + United States). After 12 weeks and plant senescence, seeds + were collected, processed, and stored under desiccant and dark + conditions at room temperature. + To collect root exudates, seeds of B. distachyon Bd21 were + surface-sterilized, pregerminated, and placed in sterile 1 L widemouth + glass jars containing 113 g of 6-mm glass beads and + 25 ml distilled water. Jars were covered with vented caps + and plants were grown hydroponically in an environmental + controlled growth chamber under conditions described above. + After 6 days, root exudates were extracted from individual + jars and their sterility was confirmed by spotting on nutrient + agar. Multiple batches of root exudates were collected, filtered + (0.22 mm), aliquoted in Falcon tubes (10 ml), lyophilized, and + stored at –80 C. + Metabolomic Profiling of Root Exudates + Exudates were analyzed for primary metabolites at the Murdock + Metabolomics Laboratory at Washington State University + (Pullman, WA, United States). Freeze-dried residues were + suspended in 500 ml 50% aqueous acetonitrile and clarified by + centrifugation for 20 min at 21,000   g and 4 C. The liquid + chromatography mass spectrometry analysis was conducted + with a Synapt G2-S quadrupole-ion mobility spectrometrytime + of flight mass spectrometer system equipped with an + acquity ultra-performance liquid chromatograph (UPLC) and + an acquity photodiode array detector (all from Waters, Milford, + MA, United States). The exudate metabolites were separated + on a SeQuant ZIC-pHILIC HPLC column (2.1   100 mm, + 3 mm) (Milllipore Sigma, Burlington, MA, United States) using + acetonitrile with 0.1% formic acid as solvent B and water + with 0.1% formic acid as solvent A at a flow rate of 400 ml + min􀀀1 and the following linear gradient extending over 14 min: + 0 min, 80% B; 4 min, 80% B, 6 min: 10% B; 7.5 min, 10% + B; 10 min, 80% B; and 14 min, 80% B. Mass spectra were + collected in positive ion mode over a range of m/z 50–1,200 + with a scan time of 0.2 s. The Q-TOF-MS source was at + 3.0 kV and 120 C; the sampling cone at 40 V, desolvation + temperature was 250 C; cone gas and desolvation gas flow were + at 0 and 850 L h􀀀1, respectively. Leucine enkephalin was used + for post-acquisition mass correction. Target compounds were + visualized using selected ion chromatograms at 0.05 Da window + width. The compound identification was based on comparison + of chromatographic behavior and accurate masses to those of + authentic standards. + For gas chromatography, derivatization was carried out using + a modification of the procedure of Lee and Fiehn (2008). + The freeze-dried residues were suspended in 950 ml aqueous + methanol (84%, v/v) and clarified by centrifugation for 15 min + at 21,000   g at 4 C. The supernatants were spiked with 1 + mg of the internal standard salicylic acid-d6 (C/D/N Isotopes, + Quebec, Canada) and dried in vacuo. The dry residues were + suspended in 10 ml of O-methoxylamine hydrochloride (30 mg + ml􀀀1 in anhydrous pyridine, both from Millipore Sigma) and + incubated while mixing (1,000 RPM) for 90 min at 30 C. + Subsequently, samples were derivatized with 90 ml of MSTFA + with 1% TMCS (Thermo Fisher Scientific, Waltham, MA, + United States) for 30 min at 37 C. Gas chromatography-mass + spectroscopy analysis was performed using a Pegasus 4D timeof- + flight mass spectrometer (LECO, Saint Joseph MI) equipped + with a MPS2 autosampler (Gerstel, Linthicum, MD) and a 7890A + oven (Agilent Technologies, Santa Clara, CA, United States). + The derivatization products were separated on a 30-m, 0.25 mm + i.d., 0.25 mm df Rxi-5Sil column (Restek, Bellefonte, PA, + United States) with an IntegraGuard precolumn using ultrapure + He at a constant flow of 0.9 ml min􀀀1 as carrier gas. The linear + thermal gradient started with a 1-min hold at 70 C, followed + by a ramp to 300 C at 10 C min􀀀1. The final temperature was + held for 5 min prior to returning to initial conditions. Mass + spectra were collected at 17 spectra s􀀀1. Peak identification was + conducted using the Fiehn primary metabolite library (Kind + et al., 2009) and an identity score cutoff of 700. Additionally, + authentic standards for a number of primary metabolites were + analyzed under identical conditions and the data used to compare + the chromatographic behavior. Peak alignment and spectrum + comparisons were carried out using the Statistical Compare + feature of ChromaTOF software (LECO). + Frontiers in Microbiology | www.frontiersin.org 3 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + Isolation of RNA From Bacteria Cultured + in Root Exudates and RNA-Seq + The strains were pregrown overnight at 25 C on 21C-glucose + agar and then subcultured into 96-well microplates containing + liquid 21C-glucose medium amended with Brachypodium + exudates. The liquid medium was prepared by dissolving the + lyophilized root exudate material in an appropriate volume of + 21C-glucose medium to concentrate root exudates 20-fold. The + growth medium was sterilized by passing it through a 0.22- + mm membrane filter. The control cultures were grown under + identical conditions in the absence of exudates. All treatments + were inoculated at OD600 of 0.1 and incubated for 20 to + 22 h until cultures entered late-exponential growth phase at + 25 C in an atmosphere of 15% oxygen [created by a ProOx + P110 oxygen controller (BioSpherix, Parish, NY, United States) + with a hypoxia C-chamber]. The cells were stabilized by the + addition RNAprotect reagent (QIAGEN, Germantown, MD, + United States) and total RNA was purified using a RNeasy Protect + Bacteria Mini Kit (QIAGEN) from three biological replicates of + each strain cultured under control conditions and in exudates. + The quality assessment of the extracted RNA samples was + performed with a NanoDrop OneC Spectrophotometer (Thermo + Fisher Scientific) and a 2100 Bioanalyzer (Agilent Technologies) + and revealed A260/A280 and A260/A230 values of > 2.0 and a mean + RNA integrity numbers (RIN) value of > 9.2. + Three biological replicates of RNA samples were shipped on + dry ice to the DOE Joint Genome Institute (Walnut Creek, CA, + United States), where rRNA was depleted and stranded RNASeq + libraries were prepared, quantified by qPCR and sequenced + using a HiSeq 2500 instrument (Illumina). The fastq file reads + were filtered and processed with BBDuk1 to remove reads that + contained 1 or more “N” bases, had an average quality score + across the read less than 10 or had a minimum length < 51 bp + or 33% of the full read length. Reads mapped with BBMap (see + text footnote 2) to masked human, cat, dog, and mouse references + at 93% identity were removed. Another category of removed + sequences matched RNA spike-in, PhiX, common microbial + contaminants, and ribosomal RNAs. The processed reads from + each library were aligned to the reference genome using BBMap + with only unique mappings allowed (BAMs/directory). If a + read mapped to more than one location it was ignored. + featureCounts (Liao et al., 2014) was used to generate raw + gene counts, which were normalized to adjust for the length of + each gene and total number of reads mapped for each library. + The normalization formula used: n = [r/(l/1,000)]/(t/1,000,000), + where n = normalized read count for gene (G) for library (L); + r = raw read count for gene G for library L; l = gene G length; and + t = total reads mapped for library L. Raw gene counts were used to + evaluate the level of correlation between biological samples using + Pearson’s correlation. + Bioinformatic Analysis + Count tables generated by the JGI RNA-Seq pipeline were input + into DESeq2 (Love et al., 2014) to normalize and determine + 1https://sourceforge.net/projects/bbmap/ + differential expression. Statistical significance was established + through DESeq2 by using three biological replicates for control + and root exudate conditions. Scatterplots were generated from + the DESeq2 data table outputs using ggplot2. Genes differentially + expressed between control and root exudate samples (log2 foldchanges + –2   to   2, adjusted p value   0.05) were used in + downstream analysis. The core genome and pangenome for the + Pseudomonas strains used in this study were computed using the + OthoMCL v.2.0, Species Tree Builder v.2.2.0, and Phylogenetic + Pangenome Accumulation v1.4.0 apps implemented in the U.S. + Department of Energy Systems Biology Knowledgebase (KBase) + (Arkin et al., 2018). Additional comparisons were conducted + with the PGAweb pangenome analysis pipeline (Chen et al., + 2018). Differentially expressed genes were assigned to core, + non-core, and singleton parts of each strain’s proteome by + BLASTp with an E value cutoff of e-06, identity of 40% + and coverage of 60%. Functional annotation of differentially + expressed genes was carried out with the Blast2GO (Conesa + and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., + 2018). Additional manual curation was performed using tools + implemented in the Integrated Microbial Genomes (IMG) + database (Markowitz et al., 2012), Pseudomonas Genome + Database (Winsor et al., 2009), Kyoto Encyclopedia of Genes and + Genomes (KEGG) (Kanehisa et al., 2008), and Geneious 10.2.3 + (Biomatters, Auckland, New Zealand). Metabolic functions + encoded by the differentially expressed genes were mapped + using iPath 3.0 (Darzi et al., 2018). Phylogenetic analyses + were carried out by building multiple sequence alignments + with MAFFT v7.222 (Katoh and Standley, 2013) and inferring + neighbor-joining (NJ) phylogenies with Geneious Tree Builder. + The resultant phylogenetic trees were visualized with iTOL + (Letunic and Bork, 2016). Reproducibility of clades within the + inferred NJ trees was assessed by bootstrap resampling with + 1,000 replicates. + Characterization of Carbon Source + Utilization With Biolog Phenotype + Microarrays + The utilization of carbon sources was analyzed using Phenotype + MicroArrays (Biolog, Hayward, CA, United States) as follows. + The bacteria were cultured overnight on Luria-Bertani agar + at 25 C, after which cells were harvested and suspended in + inoculating fluid (IF-0). The transmittance of the suspension + was adjusted to 42% using a Biolog turbidimeter. The cell + suspension was mixed with IF􀀀0 containing Dye Mix A + (Biolog) to achieve a final transmittance of 85%. One hundred + microliter aliquots of the adjusted cell suspension were + inoculated into PM01 and PM02A plates, which were then + incubated in an OmniLog Phenotype MicroArray System + (Biolog) at 25 C for 48 h. The formation of formazan was + recorded at 15 min intervals, and data were analyzed using + OmniLog Parametric Analysis software v1.20.02 (Biolog). + Relative growth of the studied strains was normalized to + growth on D-glucose and visualized using Heatmapper + (Babicki et al., 2016). + Frontiers in Microbiology | www.frontiersin.org 4 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + Data Availability + Sequences generated in this project were deposited + under NCBI BioProject accession numbers PRJNA439743 + through PRJNA439790. + RESULTS + Metabolomic Profiling of Root Exudates + of B. distachyon + Metabolomics analysis of lyophilized root exudates revealed + the presence of numerous plant metabolites, 86 of which were + identified by matching their spectra to the LECO/Fiehn + Metabolomics library (Supplementary Table 3). These + metabolites included (i) carbohydrates and their derivatives + (glucose, fructose, xylose, sucrose, trehalose, maltose, galactose, + and others); (ii) sugar alcohols (b-mannosylglycerate, myoinositol, + galactinol, 2-deoxyerythritol, ribitol, threitol and + cellobitol); (iii) amino acids and derivatives (glutamine, + tyrosine, glutamic acid, asparagine, aspartic acid, valine, + phenylalanine, isoleucine, glycine, serine, proline, leucine, + tryptophan, cysteine, methionine, citrulline, and others); (iv) + organic acids (aconitic, allantoic, g-aminobutyric, azelaic, citric, + fumaric, 2-furoic, D-glyceric, 3-hydroxypropionic, a-ketoadipic, + malic, methylmalonic, nicotinic, quinic, succinic, threonic); + and (v) assorted metabolites including heterocyclic compounds, + phenolics, and biogenic amines, etc (3-hydroxypyridine, + maleimide, noradrenaline, 4-hydroxy-3-methoxybenzoate, + 5-methoxytryptamine, uracil, aminomalonic acid, palmitic acid, + and urea). Results of the analysis also revealed that root exudates + of B. distachyon contain hydroxyectoine and the quaternary + amine (QA) glycine betaine (Supplementary Figure 1). + Phylogenetic and Pangenome Analyses + of Pseudomonas Strains Used in the + Study + We used a set of phylogenetic markers suggested by Mulet et al. + (2010) to investigate the relatedness of the eight strains used in + this study to distinct lineages recognized within the P. fluorescens + species complex. The multilocus sequence analysis based on + the concatenated sequences of the housekeeping genes rrs (16S + rRNA), gyrB, rpoB, and rpoD identified R1-43-08 (along with + strains 2-79 and SBW25) as a member of the P. fluorescens + subgroup (Figure 1). The rest of the strains clustered closely with + four additional subgroups of the P. fluorescens complex, namely + P. corrugata (strains Q2-87 and Q8r1-96), P. koreensis (Pf0-1), + P. protegens (Pf-5), and P. chlororaphis (30–84). The genomes of + the eight rhizosphere Pseudomonas strains varied in size by 1.43 + megabase (ranging from 5.65 to 7.07 Mb) and contained between + 5,166 and 6,363 protein-coding genes (Figure 2A). The shared + gene content was characterized with OrthoMCL, which uses allagainst- + all BLASTp followed by the Markov Cluster algorithm to + identify protein groups shared between the compared genomes, + as well as groups representing species-specific gene expansion + families (Li et al., 2003). The pangenome analysis revealed a + core comprised of approximately 3,179 orthologs that were + shared among all strains and represented 50.0% to 61.5% + of each predicted proteome (Figures 2A,B). The non-core + pangenome contained genes shared by two or more (but not + all) strains and contained between 1,482 and 2,080 orthologs, + which corresponded to 28.7–36.3% of individual proteomes. The + rest of the predicted protein-coding genes were strain-specific + singletons that comprised 7.5% to 15.1% of the strain’s predicted + proteomes. In respect to divergence from the core genome, strain + Pf-5 was found to possess the highest proportion of unique genes + (n = 949) followed by 2-79 (n = 887). The entire pangenome of + the Pseudomonas strains encompassed over 12,000 homolog and + singleton gene families. + Further homolog family-based comparisons identified Q8r1- + 96 and R1-43-08 as the most distantly related strains, with + 3349 shared homologs (Supplementary Table 4A). Q8r1-96 and + Q2-87, which shared 4,489 homologs, were the most closely + related strains. The partitioning of homolog gene families into + the core, non-core, and singleton parts of the pangenome agreed + with phylogenetic relationships of the strains deduced from the + analysis of a selected subset of COGs (Clusters of Orthologous + Groups) (Figure 2C and Supplementary Tables 4B,C). The + COG-based phylogeny supported the multilocus sequence + analysis and revealed that the eight Pseudomonas strains form + three distinct clusters, the first of which contained 2-79, R1-30- + 84, and SBW25. The second cluster included Q8r1-96 and Q2-87, + whereas the third encompassed strains 30-84, Pf-5, and Pf0-1. + Correlating the Composition of Root + Exudates With Metabolic Profiles of + Pseudomonas Strains + We used the Phenotype MicroArray PM1 and PM2 plates to + profile the eight Pseudomonas strains for the utilization of + 190 different carbon sources. Results of the analysis identified + 90 compounds that supported growth and clustered by their + intensities of utilization into three distinct groups (Figure 3). + Group I was comprised of 30 highly metabolized carbon + sources, which included several amino acids and intermediates of + glycolysis, pyruvate metabolism, and citrate cycle. Approximately + half of these compounds were catabolized by all eight strains, + and included several organic acids (fumaric, citric, gluconic, + malic, and pyroglutamic), amino acids (Glu, Asn, Gln, Asp, Pro, + Ala, and g-aminobutyric acid), carbohydrates (glucose, mannose, + and mannitol), and the purine nucleoside inosine. Group II + was composed of 44 chemically diverse carbon sources that + were variably utilized by the strains. These compounds were + carbohydrates, organic acids, amino acids, phenolics, and polyols, + and included known compatible solutes and intermediates of + metabolism of pentoses, galactose, starch, and sucrose. Group + III encompassed the rest of the Phenotype MicroArray test + panel and contained compounds that were not catabolized + by the tested strains. Among several notable exceptions were + a-hydroxyglutamic acid- g-lactone, putrescine, and itaconic, + citramalic, and succinamic acids, which supported the growth + of strains 2-79, 30-84, Pf-5, and SBW25. We further matched + the carbon metabolic profiles of the Pseudomonas strains against + the list of plant-derived metabolites from the root exudates of + Frontiers in Microbiology | www.frontiersin.org 5 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 1 | Neighbor joining phylogeny showing the relationship of the eight strains used in this study (indicated by red triangles) to different species of the + P. fluorescens complex. The phylogeny was established based on the concatenated sequences of the housekeeping genes rrs (16S rRNA), gyrB (subunit B of DNA + gyrase), rpoB (b subunit of RNA polymerase), and rpoD (sigma 70 factor subunit of RNA polymerase). Distance matrices were calculated by the Jukes-Cantor + method. Colored circles on tree nodes indicate bootstrap values (1,000 replicates) that vary between 60% (smallest circle) and 100% (largest circles). + B. distachyon Bd21. Interestingly, many carbon sources from + the Phenotype MicroArray panel were also present in the root + exudates of B. distachyon Bd21, and some of these compounds + (glucose, mannose, galactose, fructose, g-aminobutyric acid, + aspartic acid, citric acid, malic acid, fumaric acid, quinic acid, + alanine, glutamine, and glutamic acid) were catabolized by all + strains used in this study, while others (e.g., xylose, trehalose, + m-inositol) were actively utilized only by certain organisms + (Figure 3). The comparison of catabolic profiles across the + eight studied Pseudomonas strains revealed the presence of three + distinct clusters. The first cluster contained strains Q8r1-96 and + Q2-87, which consumed very similar sets of carbon sources, + as well as strain Pf0-1. The second cluster was composed of + 2-79, R1-43-08, SBW25, and 30-84, whereas the third cluster + was represented by a single strain, Pf-5. The overall similarity + of the catabolic profiles partially agreed with the separation + of the strains into different subgroups of the P. fluorescens + complex (see above). + Analysis of the RNA-seq Results + In order to understand the cellular responses of rhizosphere + Pseudomonas to plant exometabolites, we analyzed the + transcriptome changes in cultures grown in the presence of + root exudates. Under field conditions, rhizobacteria colonize + plant roots in the form of surface-attached microaerobic biofilms + (Hojberg et al., 1999). To mimic these conditions, the eight + Pseudomonas strains were grown statically at 72% air saturation + in 21C-glucose medium amended with root exudates and then + processed to extract total RNA (Supplementary Figure 2). + A total of 995 million raw sequencing reads were generated from + the RNA samples by using the Illumina HiSeq-2500 platform, + averaging 20.7 million reads per sample. The removal of lowquality + and rRNA sequences resulted in a total of 793 million + filtered reads that were mapped onto the eight Pseudomonas + genomes with a mean of 7.48 million mapped fragments per + genome. The differentially abundant transcripts were identified + by setting a p value of 0.05 (adjusted for multiple testing) and + Frontiers in Microbiology | www.frontiersin.org 6 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 2 | Pangenome analysis of the studied Pseudomonas strains. (A) The innermost circle shows the number of orthologous protein families shared among all + eight strains used in this study. The second circle shows orthologs present in two or more (but not all) strains, whereas the outermost circle represents strain-specific + singletons. Values in brackets under strain names correspond to the total number of protein-coding genes predicted in each genome. (B) The gradual expansion of + the pangenome (blue color) and contraction of the core genome (green color) following the sequential addition of genomes from the dataset. Box plots indicate the + 25th and 75th percentiles and medians (horizontal lines) with whiskers corresponding to the 10th and 90th percentiles. The input order was randomized to avoid any + bias due to the sequential addition of new genomes. The pangenome size increases steadily without reaching a plateau even after the addition of 11,939 + non-redundant gene families. At the same time, the core genome converged to 3,179 genes. (C) The pangenome-based phylogenomic analysis of the studied + strains. Here, the pangenome was defined with OrthoMCL, and orthologous gene sets were then partitioned into the core, singleton, and non-core (the remaining + ortholog sets) categories. These categories were calculated for each node in the Maximum Likelihood species tree, using the set of genomes for which that node + represents the ancestor. The results of the ortholog partitioning are shown in pie charts placed at tree nodes. Numbers indicate bootstrap support values. The + analysis was conducted in KBase (Arkin et al., 2018). + the log2 fold-change (FC) threshold    2.0 (Figure 4 and + Supplementary Tables 5–12). When compared with the control + conditions, an average of 204 genes per strain were differentially + expressed in the presence of root exudates, with the highest + (n = 425) and lowest (n = 112) numbers observed, respectively, + in SBW25 and Q2-87 (Figure 4). Overall, more genes were + induced than repressed in response to exudates, but the actual + numbers in each category varied substantially depending on the + identity of the Pseudomonas strain. In most strains, the bulk of + the differentially expressed genes was almost equally distributed + between the core (mean, 48.2%) and non-core (mean, 45.8%) + parts of the genome, whereas the strain-specific singleton genes + constituted on average only 5.9% (Figure 4B). One notable + exception was observed in Q8r1-96, where all differentially + expressed genes belonged to the core (73.8%) and non-core + (26.2%) parts of the genome. Another notable pattern was + observed in R1-43-08, where the majority of genes affected by + the presence of root exudate fell into the non-core category + (56.3%). The highest proportion of differentially expressed + singletons (11.3 and 10.4%, respectively) was identified in + strains SBW25 and Pf-5. + We further explored how the identified differentially + expressed genes were distributed across genomes of the eight + studied rhizosphere strains. The pairwise BLASTp comparisons + identified 2-79 and SBW25 as two strains that shared the highest + number of genes (n = 101) induced or repressed in response + to root exudates (Table 1). The second pair of strains with + a significant number of similar differentially expressed genes + (n = 86) was Q8r1-96 and Pf-5, which was followed by Pf0-1 + and 30-84, which shared 56 differentially expressed genes. These + patterns of shared genes were also observed when the results of + the pairwise BLASTp comparisons were converted into a binary + gene presence/absence matrix, which was then subjected to + cluster analysis using a UPGMA algorithm based on Sorensen’s + dissimilarity index or examined by non-metric multidimensional + scaling (NMDS) (Figure 5). + The differentially expressed Pseudomonas genes were + subjected to Blast2Go analysis and Gene Ontology (GO) + annotation (Figure 6). Metabolic process, catalytic activity, + and membrane were the most common annotation terms + across the three primary GO term categories (i.e., biological + process, molecular function, and cellular component). A total + of 1,694 GO terms was assigned to 805 upregulated genes, with + the majority of the GO terms related to molecular function + (682, 40.3%), followed by biological process (669, 39.5%), and + cellular component (343, 20.2%). In the 539 downregulated gene + category, 1,101 GO terms were assigned to biological process + (420, 38.1%), molecular function (417, 37.9%), and cellular + component (264, 24.0%). Within biological process, metabolic + process, cellular process, localization, response to stimulus, and + regulation were over-represented. Within molecular function, + the largest proportion was assigned to catalytic activity, binding, + and transporter activity categories. Within cellular component, + the majority were assigned to membrane, membrane part, cell, + and cell part categories. Across the eight strains, 37–42% of + differentially expressed genes had no Gene Ontology IDs and + encoded various conserved hypothetical proteins. + Functional Classification of Shared + Differentially Expressed Genes + The interrogation of RNA-seq data revealed multiple cellular + pathways that were differentially regulated in bacterial cultures + incubated with root exudates (Supplementary Figures 3, 4). + Frontiers in Microbiology | www.frontiersin.org 7 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 3 | Biolog Phenotype MicroArray profiling the eight rhizosphere + Pseudomonas strains used in the study. The hierarchical clustering analysis + was carried out using the average linkage method with Euclidean distances. + Carbon sources identified by red arrowheads were also detected in the sterile + root exudates of B. distachyon Bd21. + Although none of these differentially regulated pathways were + shared by all eight strains, the cross-strain comparisons revealed + several types of common and specific transcriptomic responses + that were elicited by the presence of plant exometabolites + (Table 2). The visual representation of core gene expression + patterns is provided in Supplementary Figure 5, which + shows heatmaps of expression profiles and p-adj values for + core genes shared by the studied strains. The figure is + accompanied by Supplementary Table 13 that lists predicted + functions of genes constituting the four distinct clusters + observed after hierarchical clustering of gene expression + values. The first category of shared differentially expressed + pathways functioned in the uptake and catabolism of selected + carbohydrates, quaternary ammonium compounds (QAs), and + phenolics. All strains except for R1-43-08, responded to root + exudates by inducing the fructose-specific phosphoenolpyruvate + (PEP)-dependent phosphotransferase system (PTSFru). The + components of this system are encoded by a conserved + operon and include the cytoplasmic polyprotein EI/HPr/EIIAFru + (FruB), the 1-phosphofructokinase FruK, and the fructosespecific + permease EIIBC (FruA) (Chavarria et al., 2016). The + PTSFru system functions by acquiring high-energy phosphates + from PEP and sequentially passing them, via the EI/HPr/EIIAFru + domains of FruB, to the EIIB component of FruA. The + phosphates are ultimately transferred by the EIIC transporter + to fructose yielding fructose 1-phosphate, which is channeled + into the central metabolic pathways through the action of the + phosphofructokinase FruK. + In all strains except for Q8r1-96 and Pf-5, the exposure to root + exudates resulted in the induction of two genes adjacent to the + fru cluster that encoded a Major Facilitator Superfamily (MFS) + transporter and an L-arabinonate dehydratase (Table 2). These + genes are predicted to participate in the uptake and catabolism + of L-arabinose, where L-arabinonate dehydratase plays an + important role by converting L-arabinonate to 2-dehydro-3- + deoxy-L-arabinonate (Rahman et al., 2017). In SBW25, R1- + 43-08, and Q2-87, we also observed the induction of genes + encoding components of the AraFGH complex, an ATP-Binding + Cassette (ABC) superfamily transporter involved in the import of + arabinose into the cell (Supplementary Tables 6, 7, 9). Finally, all + strains except SBW25 and R1-43-08 responded to the presence + of exudates by upregulating a conserved gene encoding an + aldose epimerase superfamily protein. Such enzymes equilibrate + alpha- and beta-anomers of aldoses and ensure that stereospecific + enzymes involved in the metabolism of free sugars do not act as + metabolic bottlenecks (Abayakoon et al., 2018). Although some + aldose epimerases have been linked to specific pathways, the + Pseudomonas gene identified in this study could not be assigned + to a particular metabolic process based on sequence analysis and + genomic location. + Several Pseudomonas strains responded to the presence + of root exudates by upregulating genes involved in the + uptake and catabolism of myo-inositol and possibly other + stereoisomers of inositol (Table 2). The upregulated catabolic + genes encode the dehydrogenase IolG, which oxidizes myoinositol + to its corresponding ketone, as well as IolE, IolD, IolB, + and IolC that collectively convert the 2-keto-myo-inositol to + acetyl-CoA and the glycolysis intermediate dihydroxyacetone + phosphate (Yoshida et al., 2008; Kohler et al., 2011). In R1- + 43-08, Q8r1-96, Q2-87, and Pf-5, the upregulated functions + Frontiers in Microbiology | www.frontiersin.org 8 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 4 | (A) Log ratio versus abundance plots (MA-plots) showing the changes in gene expression in response to root exudates. The differentially expressed core + and non-core genes are shown in red and blue, respectively. Green color indicates genes with a log2 fold-change and/or adjusted p values below the established + threshold. (B) Circular diagrams depicting the distribution of differentially expressed genes among the core, non-core, and singleton proteomes of individual + Pseudomonas strains. (C) The number of genes per genome that were induced and repressed by B. distachyon root exudates. + also involved components of the putative inositol-specific + ABC transporter. The cross-genome comparisons revealed + that in all studied strains except for Pf0-1, components + of the myo-inositol utilization pathway were encoded + within a well-conserved gene cluster which, in addition to + catabolic and transport functions, also encodes a dedicated + transcriptional repressor. + All studied strains of Pseudomonas carry multiple genes + involved in scavenging the quaternary ammonium compounds + choline, glycine betaine (GB), carnitine, choline-O-sulfate, and + Frontiers in Microbiology | www.frontiersin.org 9 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + TABLE 1 | The number of differentially expressed genes shared among the eight + studied strains of rhizosphere Pseudomonas. + Strain 2-79 SBW25 R1-43-08 Q8r1-96 Q2-87 30-84 Pf0-1 Pf-5 + 2–79 260 + SBW25 101 425 + R1-43-08 30 25 151 + Q8r1-96 32 39 21 145 + Q2-87 27 28 25 31 112 + 30-84 27 23 24 32 28 136 + Pf0-1 38 50 29 29 50 56 205 + Pf-5 36 41 52 86 29 55 40 230 + The pairwise comparisons were conducted by BLASTp with the following cutoff + parameters: E-value < 1e-06, minimum percent identity > 40%, and minimum + percent coverage > 65%. The black diagonal cells show the number of differentially + expressed genes per strain. In other words, these are self comparison values. + sarcosine from the environment. Many of these genes were + differentially expressed, including those encoding parts of the + ABC transporter CbcXWV, which is predicted to function in + the uptake of choline under water-replete conditions (Table 2). + Among enzymes induced in the presence of root exudates were + the choline dehydrogenase BetA, which converts choline to + glycine betaine and a network of enzymes (i.e., the Rieske family + oxygenase GbcAB, the dimethyglycine demethylase DgcAB, and + the sarcosine oxidase SoxBDAG) that sequentially convert GB + to glycine. In 2-79 and SBW25, this group of differentially + regulated genes also included an AraC-family transcriptional + activator GbdR, which perceives intercellular levels of GB and + induces genes involved in the transport and catabolism of + glycine betaine and detoxification of the catabolic byproducts + (Hampel et al., 2014). + The last category of activated catabolic pathways included the + catechol branch of the b-ketoadipate pathway for the degradation + of aromatic compounds. In strains 30-84, Pf0-1, and Pf-5, + growth on root exudates resulted in upregulation of catechol- + 1,2-dioxygenase, muconate cycloisomerase, and muconolactone + isomerase, which collectively cleave the catechol ring and + convert it to b-ketoadipate enol-lactone (Harwood and Parales, + 1996). Finally, analysis of the P. synxantha 2-79 transcriptome + identified an induction of benABC genes encoding subunits + of benzoate 1,2-dioxygenase, an oxidoreductase that generates + catechol from benzoate. + In addition to various catabolic pathways, the exposure to root + exudates also induced several genes involved in the homeostasis + of copper (Table 2). Four of these genes form a conserved cluster + in genomes of the strains and encode the periplasmic coppersensing + two-component system CinRS, the plastocyanin/azurinlike + protein CinA, and the NADPH-dependent pre-Q0 reductase + CinQ. Also, in strains Q2-87, 30-84, Pf0-1, and Pf-5, we observed + upregulation of a conserved operon encoding the multicopper + oxidase CopA, the periplasmic copper-binding protein CopC, the + inner membrane protein CopD, and outer membrane protein + CopB. In several Gram-negative bacteria, these Cop proteins are + thought to have dual functions and participate both in the uptake + of essential copper as well as in the sequestration of excess copper + in the periplasm and outer membrane. + The analysis of shared downregulated pathways revealed that + most of the strains respond to the presence of root exudates by + repressing genes involved in the uptake and catabolism of sulfur + compounds (Table 2). In strains SBW25, R1-43-08, Q8r1-96, + Q2-87, Pf0-1, and Pf-5, this response involved the ssuEADCB + operon responsible for the utilization of alkanesulfonates + as sulfur sources. The ssu operon is highly conserved in + fluorescent pseudomonads and encodes the FMNH2-dependent + monooxygenase SsuD and the NAD (P)H-dependent FMN + reductase SsuE, which together catalyze the desulfonation of + alkanesulfonates. Also, the ssu locus contains genes for the + molybdopterin-binding protein SsuF and the alkanesulfonatespecific + ABC-type transporter consisting of the sulfonate + substrate-binding protein SsuA, sulfonate permease protein + SsuC, and sulfonate transport ATP-binding protein SsuB. Finally, + in R1-43-08, Q2-87, Pf0-1, and Pf-5, growth on root exudates + coincided with repression of the tauABCD operon, which + allows these strains to utilize taurine (2-aminoethanesulfonate) + as a sulfur source. The repressed tau genes encoded the + 2-oxoglutarate-dependent taurine dioxygenase TauD and + substrate-binding, ATP-binding, and permease components of + the taurine-specific ABC transporter TauABC. + Other Differentially Expressed Pathways + In addition to their effect on several shared cellular pathways, + growth on root exudates resulted in the induction or repression of + numerous strain-specific genes. In closely related P. synxantha 2- + 79 and P. fluorescens SBW25, we observed differential expression + of genes involved in energy metabolism, transport of amino + acids, and surface attachment (Supplementary Tables 5, 6). + Other notable differentially expressed pathways included 2– + 79 gene clusters that encode enzymes for the catabolism of + trehalose, a prophage, and toxin/antitoxin system, as well as the + SBW25 operon predicted to control the synthesis of the capsular + exopolysaccharide colonic acid. The response of Pseudomonas sp. + R1-43-08 to root exudates also involved differential expression + of different energy metabolism pathways. In addition, we + observed the upregulation of genes involved in the uptake and + catabolism of xylose (also upregulated in 2–79) and repression + of enzymes for the biosynthesis of phenazine-1-carboxylic acid + and assimilation of inorganic sulfur and L-cysteine biosynthesis + (Supplementary Table 7). + The analysis of the Q8r1-96 transcriptome revealed + perturbation of different metabolic pathways including genes + encoding components of cytochrome C oxidase, transport + and catabolism of sorbitol/mannitol, metabolism of butanoic + acid, and biosynthesis of exopolysaccharides alginate and + poly-b-1-6-N-acetylglucosamine (Supplementary Table 8). In + P. fluorescens Q2-87, we identified differential expression of + genes involved in metabolism of galactose, tryptophan, tyrosine, + glycine, serine, and threonine (Supplementary Table 9), + while in P. chlororaphis 30-84, growth on exudates activated + the biosynthesis of molybdopterin cofactor, catabolism of + galactonate and acetoin, and uptake and catabolism of putrescine + (Supplementary Table 10). The response of P. protegens Pf-5 to + root exudates involved upregulation of acetoin dehydrogenase, + which converts acetoin to acetaldehyde and acetyl-CoA, as + Frontiers in Microbiology | www.frontiersin.org 10 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + FIGURE 5 | Comparison of the eight Pseudomonas strains based on the content (presence/absence) of genes differentially expressed in the presence of root + exudates. (A) UPGMA clustering based on the Sorensen’s dissimilarity index. (B) non-metric multidimensional scaling (NMDS) analysis. + FIGURE 6 | Gene Ontology (GO) classification of Pseudomonas genes that were induced (red bars) or repressed (gray bars) in response to root exudates of + B. distachyon Bd21. The terms were derived from 93 different functional groups (GO subcategories level 4). The GO terms were assigned with Blast2GO (Conesa + and Gotz, 2008) and visualized in WEGO 2.0 (Ye et al., 2018). On a WEGO histogram, the percentage of 100 is defined as the total number of genes assigned a GO + term. However, the subcategories do not add up to 100% because many genes fall into more than one functional class and are therefore annotated by multiple GO + terms. + well as pathways for the utilization of glycolate and putrescine + (Supplementary Table 11). Also induced were genes for + the production of pyrrolnitrin and PhlG hydrolase, which + modulate the metabolic loads attributed to the synthesis of + 2,4-diacetylphloroglucinol. The differentially expressed genes of + P. fluorescens Pf0-1 included, among others, operons encoding + cytochrome C oxidase and enzymes for catabolism of malonic + acid (Supplementary Table 12). Yet another interesting + finding involved the induction of assorted genes acting in the + homeostasis of iron and defense against reactive oxygen species + (ROS). We observed activation of iron dicitrate transporters + (SBW25 and 30-84), genes for the biosynthesis of siderophores + ornicorrugatin (SBW25) and pyochelin (Pf-5), heme-degrading + enzymes (2–79, 30–84), TonB siderophore receptors, and + components of the energy-transducing inner membrane + complex TonB-ExbB-ExbD (2–79 and Pf-5). The differentially + expressed ROS defense pathways were represented by different + catalases in strains 2–79, R1-43-08, Q8r1-96, Q2-87, Pf0-1, and + Pf-5 and organic hydroperoxide resistance proteins in strains + SBW25 and R1-43-08. Finally, in SBW25, Q2-87, 30–84, and + Pf0-1, the addition of exudates resulted in the upregulation of + peroxiredoxins that detoxify H2O2, peroxynitrite, and aliphatic + and aromatic hydroperoxides. + DISCUSSION + Our analysis of B. distachyon root exudates revealed a complex + mix of primary and secondary metabolites, thus supporting + the view of the plant rhizosphere as a carbon-rich niche for + Frontiers in Microbiology | www.frontiersin.org 11 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + TABLE 2 | The distribution and predicted functions of selected differentially expressed genesa. + Predicted function Strainb + 2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 + Uptake and catabolism of fructose + D-fructose PTS system, IIC component 2756598827 (2.7) 649634314 (2.3) 2597873629 (3.6) 2597850083 (2.7) 2597856046 (3.6) 637740645 (2.9) 637318202 (2.8) + 1-phosphofructokinase 2756598828 (2.9) 649634313 (3.1) 2597873628 (3.9) 2597850082 (2.7) 2597856045 (3.4) 637740644 (3.1) 637318201 (2.7) + D-fructose PTS system, IIA component 2756598829 (2.6) 649634312 (3.0) 2597873627 (3.7) 2597850081 (2.6) 2597856044 (3.6) 637740643 (2.9) 637318200 (3.2) + Uptake and catabolism of arabinose + MFS superfamily transporter 2756599521 (2.2) 649635836 (3.6) 2756590067 (4.9) 2597851595 (3.1) 2597859759 (4.2) 637743102 (2.9) + L-arabinonate dehydratase 2756599520 (3.1) 649635835 (4.3) 2756590066 (5.5) 2597851594 (5.4) 2597859760 (4.0) 637743103 (3.5) + Interconversion of alpha- and beta-anomers of aldoses + Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) + Uptake and catabolism of quaternary ammonium compounds + Choline dehydrogenase BetA 2597874908 (2.3) 2597851450 (2.1) + Transcriptional regulator GbdR 2756597125 (􀀀3.7) 649639087 (􀀀3.9) + Membrane dipeptidase, dgc operon 2756597136 (1.9) 2756592046 (2.0) 2597878321 (3.8) 2597849833 (2.9) 2597860696 (2.6) 637323077 (3.3) + Hypothetical protein, dgcAB operon 2756597137 (2.4) 2756592045 (2.3) 2597878320 (3.4) 2597849834 (2.7) 2597860695 (2.3) 637323076 (3.3) + Dimethyl Gly demethylase DgcA 2756597138 (2.7) 2756592044 (2.0) 2597878317 (3.3) 2597849835 (2.7) 2597860694 (2.2) 637323075 (3.2) + Dimethyl Gly demethylase DgcB 2756597139 (2.3) 2597878318 (3.6) 2597849836 (2.6) 2597860693 (2.4) 637323074 (3.0) + Betaine demethylase, GbcA subunit 2756597143 (2.1) 2756592039 (2.3) 2597878312 (4.3) 637323070 (3.7) + Betaine demethylase, GbcB subunit 2597878311 (3.9) 2597860689 (2.0) 637323069 (3.2) + Ser hydroxymethyltransferase, sox operon 2756597149 (2.1) 2756592033 (2.0) 2597878308 (2.9) 2597849846 (2.7) 637323064 (3.1) + Sarcosine oxidase, g subunit, SoxG 2597878304 (2.9) + Sarcosine oxidase, a subunit, SoxA 2756592028 (2.1) 2597878305 (3.0) 637323061 (2.4) + Sarcosine oxidase, d subunit, SoxD 2597878306 (2.5) 637323062 (2.2) + Sarcosine oxidase, b subunit, SoxB 2756597150 (2.4) 2756592032 (1.9) 2597878307 (2.6) 637323063 (2.3) + Betaine substrate-binding protein CbcX 2756590368 (2.3) 2597878336 (2.3) 2597850794 (3.1) 637742655 (-3.5) + ABC transporter, ATP-binding protein CbcV 2597878338 (2.3) 637742656 (-3.5) + Uptake and catabolism of myo-inositol + 5-dehydro-2-deoxygluconokinase, IolC 2756592881 (2.5) 2597876275 (4.3) 2597857598 (2.4) 637319925 (4.2) + 2-keto-myo-inositol dehydratase, IolE 2756592884 (2.5) 2597876273 (4.6) 2597857602 (2.4) 637319928 (4.1) + 5-deoxy-glucuronate isomerase, IolB 2756592883 (2.3) 2597876272 (4.3) 2597857600 (2.4) 637319927 (4.3) + 2-keto-myo-inositol isomerase, IolL 2756592882 (2.7) 2597857599 (2.4) 637319926 (4.4) + 3D-(3,5/4)-trihydroxycyclohexane-1,2-dione + acylhydrolase, IolD + 2756592885 (2.4) 2597876269 (4.8) 2597857603 (2.5) 637319929 (4.3) + Myo-inositol 2-dehydrogenase, IolG 2756595203 (1.9) 2756592886 (2.2) 2597876268 (4.8) 2597857604 (2.0) 637319930 (3.9) + Inositol transport substrate-binding protein 2756592888 (2.0) 2597876265 (3.8) 2597851513 (2.8) 637319932 (3.8) + Inositol transport permease protein 2756592890 (1.9) 2597876263 (3.6) 2597851515 (2.2) 637319934 (3.5) + Inositol transport ATP-binding protein 2756592889 (2.2) 2597876264 (3.8) 2597851514 (2.7) 637319933 (3.6) + (Continued) + Frontiers in Microbiology | www.frontiersin.org 12 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + TABLE 2 | Continued + Predicted function Strainb + 2–79 SBW25 R1-43-08 Q8r1-96 Q2-87 30–84 Pf0-1 Pf-5 + Uptake and catabolism of fructose + Catabolism of phenolics + Muconate cycloisomerase 2597859089 (3.1) 637742838 (3.4) 637321199 (4.8) + Muconolactone delta-isomerase 2597859088 (2.6) 637742837 (3.5) 637321198 (4.4) + Catechol 1,2-dioxygenase 2597859087 (2.0) 637742836 (2.9) 637321197 (3.5) + AraC-type DNA-binding protein 2597859086 (2.0) 637321196 (2.2) + Benzoate 1,2-dioxygenase, a subunit 2756599329 (2.7) 2597859085 (3.4) 637742843 (3.9) 637321195 (3.8) + Benzoate 1,2-dioxygenase, b subunit 2756599330 (2.2) 2597859084 (3.3) 637742842 (4.2) 637321194 (4.8) + Benzoate 1,2-dioxygenase, reductase subunit 2756599331 (2.1) 2597859083 (3.4) 637742841 (3.9) 637321193 (4.3) + Copper homeostasis + pre-Q0 reductase/7-cyano-7-deazaguanine + reductase CinQ + 2756597439 (2.7) 649635068 (1.8) 2597874689 (3.4) 637319306 (2.5) + Cupredoxin-like copper-binding protein CinA 2756597440 (2.8) 649635067 (4.3) 2756590986 (2.03) 2597853017 (7.3) 2597857153 (5.7) 637743734 (3.5) 637319305 (3.4) + Heavy metal response regulator CinR 649635066 (1.8) 2597874687 (2.1) 2597853018 (3.2) 2597857152 (2.4) 637743735 (2.1) 637319304 (2.5) + Heavy metal sensor histidine kinase CinS 649635065 (2.0) 2597857151 (2.2) 637743736 (2.0) 637319303 (2.0) + Copper resistance protein CopA 2597877412 (5.9) 2597850492 (5.6) 2597857966 (6.9) 637743691 (5.2) 637320232 (6.4) + Copper resistance protein CopB 2597850491 (5.3) 2597857965 (6.8) 637743692 (5.5) 637320231 (6.0) + Copper resistance protein CopC 2597850490 (4.6) 2597857964 (7.2) 637743693 (5.4) + Copper resistance protein CopD 2597850489 (4.9) 2597857963 (6.9) 637743694 (5.2) + Conserved hypothetical protein + Aldose epimerase superfamily protein 2756599919 (2.7) 2597878613 (4.2) 2597849545 (3.1) 2597860977 (4.2) 637742166 (3.4) 637323358 (3.5) + Uptake and catabolism of sulfonates + FMN-dependent monooxygenase SsuE 2756592254 (􀀀1.9) 637745334 (􀀀3.5) + Sulfonate substrate-binding protein SsuA 649639261 (􀀀2.0) 2756592253 (􀀀2.2) 2597878518 (􀀀2.6) 2597849636 (􀀀3.9) 637745333 (􀀀7.0) 637323272 (􀀀2.7) + FMN-dependent monooxygenase SsuD 649639260 (􀀀1.9) 2756592252 (􀀀2.3) 2597849637 (􀀀2.8) 637745332 (􀀀5.8) 637323271 (􀀀2.3) + Sulfonate permease protein SsuC 649639259 (􀀀2.4) 2756592251 (􀀀2.1) 2597878516 (􀀀1.9) 2597849638 (􀀀3.2) 637745331 (􀀀5.8) 637323270 (􀀀2.5) + Sulfonate transport ATP-binding protein SsuB 649639258 (􀀀2.5) 2756592250 (􀀀2.3) 2597849639 (􀀀2.5) 637745330 (􀀀5.6) 637323269 (􀀀2.7) + Molybdopterin binding protein SsuF 649639257 (􀀀3.1) 2756592249 (􀀀2.7) 2597878514 (􀀀2.9) 2597849641 (􀀀3.3) 637745339 (􀀀6.2) 637323268 (􀀀2.7) + Uptake and catabolism of taurine + Taurine substrate-binding protein TauA 2597854917 (􀀀2.0) 637740095 (􀀀4.1) + Taurine transport ATP-binding protein TauB 2756592398 (􀀀1.9) 637740094 (􀀀3.7) 637317614 (􀀀1.9) + Taurine permease protein TauC 2756592398 (􀀀1.9) 637740093 (􀀀3.9) 637317613 (􀀀2.0) + Taurine dioxygenase TauD 2597854920 (􀀀1.8) 637740092 (􀀀3.6) + aThe shared differentially expressed genes were identified by BLASTp with the cutoff parameters of E-value < 1e-06, minimum percent identity > 40%, and minimum percent coverage > 65%. + bValues in columns indicate JGI IMG gene IDs followed by the corresponding fold-change (FC) values (shown in brackets). + Frontiers in Microbiology | www.frontiersin.org 13 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + soil microorganisms. Our results were in agreement with a + recent report of 27 different sugars, amino acids, and organic + acids in Brachypodium exudates (Kawasaki et al., 2016). We + confirmed the presence of exometabolites identified in that + study, along with dozens of additional analytes that were + identified by matching their mass-spectra and retention indices + to the LECO/Fiehn Metabolomics library (Supplementary + Table 3). The complementation of the metabolomic analysis + with profiling of the bacteria by Biolog Phenotype MicroArrays + revealed that a substantial proportion of the characterized + exudate constituents were catabolized by a collection of eight + Pseudomonas strains from across the P. fluorescens group that is + known to form associations with plant roots. The amendment + of Pseudomonas cultures with root exudates caused changes in + the expression of multiple genes encoding catabolic and anabolic + enzymes, predicted transporters, transcriptional regulators, stress + response, and conserved hypothetical proteins. In most strains, + these differentially expressed genes were almost equally split + between the core and variable genome regions, mirroring + the substantial strain-to-strain variation in the genome size + and gene content within the P. fluorescens species complex + (Loper et al., 2012). + The analysis of transcriptome responses to root exudates + revealed several types of cellular pathways present in the strains + used in this study. The first category of such pathways was + involved in the catabolism of carbohydrates such as fructose, + arabinose, myo-inositol, xylose, trehalose, and galactose. Among + these catabolic traits, the ability to utilize fructose as a carbon + source is highly conserved among fluorescent pseudomonads. + In contrast, growth on arabinose, myo-inositol, xylose, and + trehalose is variably present and was traditionally used to + differentiate species and biovars within the P. fluorescens + group (Barrett et al., 1986). We speculate that such variably + distributed pathways contribute to the differential affinity of + pseudomonads toward host plants and/or to determine which + strains flourish in response to growing roots and changing + environments. Several independent studies have confirmed + the importance of carbohydrate catabolism pathways for the + biology of rhizosphere pseudomonads. For example, in vivo + expression technology (IVET) profiling of P. fluorescens SBW25 + identified xylose isomerase among genome regions essential + for the colonization of sugar beet seedlings (Liu et al., 2015), + whereas a genome-wide Tn-Seq screen of Pseudomonas simiae + identified genes for the catabolism of myo-inositol among traits + essential for the colonization of Arabidopsis thaliana roots + (Cole et al., 2017). + The response of rhizosphere Pseudomonas to Brachypodium + root exudates also involved pathways for the uptake and + metabolism of amino acids. We observed differential + expression of genes encoding the hydrophobic (HAAT) and + polar (PAAT) amino acid uptake transporters in strains + 2-79, SBW25, Q2-87, Pf0-1, and Pf-5. Other related genes + encoded enzymes for the catabolism of valine and glutamic + acid (2-79); metabolism of tryptophan, glycine, serine, and + threonine (Q2-87); and biosynthesis of methionine (Q8r1- + 96). It is plausible that the abundance of amino acids in + root exudates is also linked to the repression of pathways + involved in the catabolism of sulfonates and taurine that + was observed in several strains (Table 2). Although the + preferred source of sulfur for P. fluorescens is unknown, in the + closely related P. aeruginosa, the sulfur starvation response + is triggered by the growth on any sulfur compound other + than sulfate, thiocyanate, and cysteine (Hummerjohann et al., + 1998). This fact, together with the presence of cysteine and + cystine in the root exudates, suggest that root exudates of + Brachypodium may serve as an important source of sulfur for + rhizosphere Pseudomonas. These findings also agree well with + the reported scarcity of inorganic sulfate in the soil, and the + presence of sulfur mostly in the form of organic compounds, + including amino acids, proteins, sulfate esters, and sulfonates + (Autry and Fitzgerald, 1990). + Another interesting result of this study was the concerted + activation of copper and iron homeostasis pathways observed + in all of the Pseudomonas strains used in this work. In bacteria, + an excess of copper is toxic and triggers oxidative stress due to + the formation of free radicals, as well as disruption of protein + metalation and stability of iron-sulfur clusters (Bondarczuk + and Piotrowska-Seget, 2013). On the other hand, copper is an + essential trace element used as a cofactor in different enzymes. + Similarly, although elevated levels of iron cause redox stress, this + element is also found in active energy metabolism enzymes and is + crucial for bacterial growth (Andrews et al., 2003). The analysis + of metal homeostasis genes identified in this study suggests + that their induction was likely triggered by the deficiency of + copper and iron in bacterial cultures grown in the presence of + root exudates. We attribute this effect to the ability of some + components of root exudates to chelate soil metals. + Despite the abundance of iron in the soil, its bioavailability + is limited due to the low solubility of Fe (III) oxyhydrates + at neutral pH. The non-graminaceous plants circumvent this + problem by acidifying the rhizosphere and secreting flavins, + phenolics, and organic acids that chelate iron. The reduction + of these ferric chelates releases soluble ferrous iron taken up + by root cells (Kobayashi and Nishizawa, 2012). Graminaceous + plants, like Brachypodium, acquire iron by secreting into the + soil non-protein amino acids of the mugineic acid (MA) group, + which act as Fe (III)-chelating phytosiderophores. In addition to + iron, low-molecular-weight organic acids and phytosiderophores + bind other divalent and trivalent metals (including copper) and + contribute to heavy-metal tolerance in plants (Chen et al., 2017). + It is plausible that the presence of these plant exometabolites + is responsible for the deficit of iron and copper observed in + Pseudomonas cultures grown in the presence of root exudates. + These results further underscore the importance of diverse and + redundant metal homeostasis pathways found in genomes of the + P. fluorescens group for the ability of these organisms to colonize + and persist in the plant rhizosphere. + Recently, Klonowska et al. (2018) examined transcriptomic + responses of symbiotic nitrogen-fixing bacteria to root + exudates of the legume plant Mimosa pundica, which has + an unusual ability to support both alpha- (Rhizobium) and + beta-rhizobia (Cupriavidus and Burkholderia). Using RNA-seq, + the authors characterized genes involved in the perception of + root exudates in the nodulating bacteria Burkholderia phymatum + Frontiers in Microbiology | www.frontiersin.org 14 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + STM815, Cupriavidus taiwanensis LMG19424, and Rhizobium + mesoamericanum STM3625. Interestingly, the analysis of + differentially expressed genes revealed induction of pathways + involved in the catabolism of fructose, xylose, myo-inositol, + and protocatechuate/catechol. Also upregulated were some + copper homeostasis, siderophore biosynthesis, and oxidative + stress genes. Finally, the analytical profiling of M. pundica + exudates revealed an overlap with Brachypodium in the types of + carbohydrates, amino acids, and organic acids present. These + findings suggest that differentially expressed genes shared by + multiple strains of the group P. fluorescens are not unique to + the Brachypodium-Pseudomonas system but represent a set + of conserved cellular pathways involved in the perception of + plant exometabolites by different clades of rhizosphere-dwelling + Proteobacteria. + Most strains included in this study were originally selected + based on the ability to colonize the rhizosphere and produce + secondary metabolites that alleviate the plant stress response + and/or inhibit soilborne pathogens. It has been suggested + that plant metabolites released into the rhizosphere affect + the biocontrol activity of plant-beneficial pseudomonads (de + Werra et al., 2011). We provide further support to this + hypothesis by demonstrating that in some strains, root exudates + modulate the expression of genes for the catabolism of + the plant growth-promoting metabolites acetoin and 2,3- + butanediol. The exposure to exudates also affected the expression + of genes for the synthesis of well-characterized antifungal + compounds pyrrolnitrin, phenazine-1-carboxylic acid, and 2,4- + diacetylphloroglucinol. The modulatory effects were strainspecific, + suggesting significant differences in the regulatory + networks involved in the perception of plant signals and + regulation of the production of antibiotics and growthpromoting + metabolites. + The final significant finding of this study was the induction + of catabolism of quaternary amines (QAs) observed in multiple + strains of the P. fluorescens group during growth on root + exudates. This observation was supported by the detection of + glycine betaine in the root secretions of B. distachyon. The + presence of QAs in plant tissues and the capacity of these + metabolites to provide stress protection and nutrients to plant + pathogens and symbionts were reported before (Boncompagni + et al., 1999; Chen et al., 2013; Kabbadj et al., 2017), but our + study is among the first to highlight the potential importance of + these metabolites for rhizosphere interactions. Pseudomonads do + not synthesize QAs de novo but have evolved many pathways to + scavenge them from eukaryotic hosts, where these metabolites + are abundant due to the prominence of phosphatidylcholine + in cellular membranes. Strains of P. fluorescens carry genes + for the conversion of choline, carnitine, and glycine betaine to + glycine, as well as quaternary amine transporters of the BCCT + and ABC families that are also conserved in the opportunistic + human pathogen P. aeruginosa and the plant pathogen P. syringae + (Galvao et al., 2006; Chen et al., 2013; Wargo, 2013b). + In P. aeruginosa, choline catabolism genes are essential for + the ability of this pathogen to persist during lung infection + (Wargo, 2013a). Similarly, a P. syringae mutant deficient in + BetT, OpuC, and CbcXWV quaternary amine transporters had + reduced fitness during colonization of bean and soybean leaves + under greenhouse and field conditions (Chen et al., 2013). + Depending on water availability, P. aeruginosa and P. syringae + catabolize exogenously supplied QAs as carbon and nitrogen + sources or accumulate them as osmoprotectants (Chen et al., + 2013; Wargo, 2013b). Our ongoing work in P. synxantha 2– + 79 unraveled similar physiological responses and demonstrated + that QA transporters function differentially and redundantly + in the uptake of quaternary amines as nutrients (Pablo and + Mavrodi, unpublished). In contrast, under water stress, the QAs + choline, betaine, and carnitine are accumulated preferentially for + osmoprotection. Under drought stress, a 2–79 mutant devoid + of all known QA transporters was less competitive in the + colonization of the Brachypodium rhizosphere than its wild-type + parental strain. Interestingly, our metabolomic profiling of root + exudates also revealed proline, glutamine, and hydroxyectoine. + These metabolites act as compatible solutes in different groups + of microorganisms (Yancey et al., 1982; Empadinhas and da + Costa, 2008), suggesting an important role of root exudates + in the ability of Pseudomonas to persist in the rhizosphere of + drought-stressed plants. + DATA AVAILABILITY STATEMENT + The original contributions presented in the study are publicly + available. This data can be found here: NCBI BioProject accession + numbers PRJNA439743 through PRJNA439790. + AUTHOR CONTRIBUTIONS + DM, OM, and LT conceived the research project. OM and + JM collected root exudates. OM and DM cultured strains + and extracted total RNA. AB and DG performed metabolomic + analysis of root exudates. DM, JP, and AF analyzed RNA-seq + data. LE, KH, and IP conducted Biolog analyses. DM, AF, OM, + DW, and LT wrote the manuscript. All authors contributed to the + manuscript revision. + FUNDING + This study was funded by NSF grant IOS-1656872 and by an + award from the DOE Joint Genome Institute’s Community + Science Program. The authors also acknowledge support from + Australian Research Council Discovery grant (DP160103746) + and Mississippi INBRE, funded by an Institutional Development + Award (IDeA) from the National Institute of General + Medical Sciences of the National Institutes of Health under + grant P20GM103476. + SUPPLEMENTARY MATERIAL + The Supplementary Material for this article can be found online + at: https://www.frontiersin.org/articles/10.3389/fmicb.2021. + 651282/full#supplementary-material + Frontiers in Microbiology | www.frontiersin.org 15 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + REFERENCES + Abayakoon, P., Lingford, J. P., Jin, Y., Bengt, C., Davies, G. J., Yao, S., et al. (2018). + Discovery and characterization of a sulfoquinovose mutarotase using kinetic + analysis at equilibrium by exchange spectroscopy. Biochem. J. 475, 1371–1383. + doi: 10.1042/bcj20170947 + Andrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003). Bacterial iron + homeostasis. FEMS Microbiol. Rev. 27, 215–237. doi: 10.1016/s0168-6445(03) + 00055-x + Arkin, A. P., Cottingham, R. W., Henry, C. S., Harris, N. L., Stevens, R. L., Maslov, + S., et al. (2018). KBase: the United States department of energy systems biology + knowledgebase. Nat. Biotechnol. 36, 566–569. + Autry, A. R., and Fitzgerald, J. W. (1990). Sulfonate S: a major form of forest soil + organic sulfur. Biol. Fertil. Soils 10, 50–56. + Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J. R., Maciejewski, A., et al. + (2016). Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res. 44, + W147–W153. + Badri, D. V., and Vivanco, J. M. (2009). Regulation and function of root exudates. + Plant Cell Environ. 32, 666–681. doi: 10.1111/j.1365-3040.2009.01926.x + Badri, D. V., Weir, T. L., Van Der Lelie, D., and Vivanco, J. M. (2009). Rhizosphere + chemical dialogues: plant-microbe interactions. Curr. Opin. Biotechnol. 20, + 642–650. doi: 10.1016/j.copbio.2009.09.014 + Bais, H. P., Prithiviraj, B., Jha, A. K., Ausubel, F. M., and Vivanco, J. M. (2005). + Mediation of pathogen resistance by exudation of antimicrobials from roots. + Nature 434, 217–221. doi: 10.1038/nature03356 + Bais, H. P., Weir, T. L., Perry, L. G., Gilroy, S., and Vivanco, J. M. (2006). The role of + root exudates in rhizosphere interations with plants and other organisms. Annu. + Rev. Plant Biol. 57, 233–266. doi: 10.1146/annurev.arplant.57.032905.105159 + Bangera, M. G., and Thomashow, L. S. (1996). Characterization of a genomic + locus required for synthesis of the antibiotic 2,4-diacetylphloroglucinol by the + biological control agent Pseudomonas fluorescens Q2-87. Mol. Plant Microbe + Interact. 9, 83–90. + Barret, M., Frey-Klett, P., Guillerm-Erckelboudt, A. Y., Boutin, M., Guernec, G., + and Sarniguet, A. (2009). Effect of wheat roots infected with the pathogenic + fungus Gaeumannomyces graminis var. tritici on gene expression of the + biocontrol bacterium Pseudomonas fluorescens Pf29Arp. Mol. Plant-Microbe + Interact. 22, 1611–1623. doi: 10.1094/mpmi-22-12-1611 + Barrett, E. L., Solanes, R. E., Tang, J. S., and Palleroni, N. J. (1986). Pseudomonas + fluorescens biovar V: its resolution into distinct component groups and the + relationship of these groups to other P. fluorescens biovars, to P. putida, and to + psychrotrophic pseudomonads associated with food spoilage. J. Gen. Microbiol. + 132, 2709–2721. doi: 10.1099/00221287-132-10-2709 + Bevan, M. W., Garvin, D. F., and Vogel, J. P. (2010). Brachypodium distachyon + genomics for sustainable food and fuel production. Curr. Opin. Biotechnol. 21, + 211–217. doi: 10.1016/j.copbio.2010.03.006 + Boncompagni, E., Osteras, M., Poggi, M. C., and Le Rudulier, D. (1999). + Occurrence of choline and glycine betaine uptake and metabolism in the family + Rhizobiaceae and their roles in osmoprotection. Appl. Environ. Microbiol. 65, + 2072–2077. doi: 10.1128/aem.65.5.2072-2077.1999 + Bondarczuk, K., and Piotrowska-Seget, Z. (2013). Molecular basis of active copper + resistance mechanisms in Gram-negative bacteria. Cell Biol. Toxicol. 29, 397– + 405. doi: 10.1007/s10565-013-9262-1 + Brkljacic, J., Grotewold, E., Scholl, R., Mockler, T., Garvin, D. F., Vain, P., et al. + (2011). Brachypodium as a model for the grasses: today and the future. Plant + Physiol. 157, 3–13. + Camacho-Carvajal, M. M. (2001). Molecular Characterization of the Roles of Type + 4 pili, NDH-I and PyrR in Rhizosphere Colonization of Pseudomonas fluorescens + WCS365. Dissertation, University of Leiden, Leiden. + Chavarria, M., Goni-Moreno, A., De Lorenzo, V., and Nikel, P. I. (2016). A + metabolic widget adjusts the phosphoenolpyruvate-dependent fructose influx + in Pseudomonas putida. mSystems 1:e00154-16. + Chen, C., Li, S., McKeever, D. R., and Beattie, G. A. (2013). The widespread + plant-colonizing bacterial species Pseudomonas syringae detects and exploits an + extracellular pool of choline in hosts. Plant J. 75, 891–902. doi: 10.1111/tpj. + 12262 + Chen, X., Zhang, Y., Zhang, Z., Zhao, Y., Sun, C., Yang, M., et al. (2018). PGAweb: + a web server for bacterial pan-genome analysis. Front. Microbiol. 9:1910. doi: + 10.3389/fmicb.2018.01910 + Chen, Y. T., Wang, Y., and Yeh, K. C. (2017). Role of root exudates in metal + acquisition and tolerance. Curr. Opin. Plant Biol. 39, 66–72. doi: 10.1016/j.pbi. + 2017.06.004 + Cole, B. J., Feltcher, M. E., Waters, R. J., Wetmore, K. M., Mucyn, T. S., Ryan, E. M., + et al. (2017). Genome-wide identification of bacterial plant colonization genes. + PLoS Biol. 15:e2002860. doi: 10.1371/journal.pbio.2002860 + Conesa, A., and Gotz, S. (2008). Blast2GO: a comprehensive suite for functional + analysis in plant genomics. Int. J. Plant Genomics 2008:619832. + Curl, E. A., and Truelove, B. (1986). The Rhizosphere. Berlin: Springer-Verlag. + Darzi, Y., Letunic, I., Bork, P., and Yamada, T. (2018). iPath3.0: interactive + pathways explorer v3. Nucleic Acids Res. 46, W510–W513. + de Weert, S., Dekkers, L., Bitter, W., Tuinman, S., Wijfjes, A., Van Boxtel, R., et al. + (2006). The two-component colR/S system of Pseudomonas fluorescens WCS365 + plays a role in rhizosphere competence through maintaining the structure and + function of the outer membrane. FEMS Microbiol. Ecol. 58, 205–213. doi: + 10.1111/j.1574-6941.2006.00158.x + de Weert, S., Vermeiren, H., Mulders, I. H., Kuiper, I., Hendrickx, N., Bloemberg, + G. V., et al. (2002). Flagella-driven chemotaxis towards exudate components is + an important trait for tomato root colonization by Pseudomonas fluorescens. + Mol. Plant Microbe Interact. 15, 1173–1180. doi: 10.1094/mpmi.2002.15.11. + 1173 + de Werra, P., Huser, A., Tabacchi, R., Keel, C., and Maurhofer, M. (2011). Plantand + microbe-derived compounds affect the expression of genes encoding + antifungal compounds in a pseudomonad with biocontrol activity. Appl. + Environ. Microbiol. 77, 2807–2812. doi: 10.1128/aem.01760-10 + Dekkers, L. C., Phoelich, C. C., Van Der Fits, L., and Lugtenberg, B. J. (1998). + A site-specific recombinase is required for competitive root colonization by + Pseudomonas fluorescens WCS365. Proc. Natl. Acad. Sci. U.S.A. 95, 7051–7056. + doi: 10.1073/pnas.95.12.7051 + De-la-Pena, C., Lei, Z., Watson, B. S., Sumner, L. W., and Vivanco, J. M. (2008). + Root-microbe communication through protein secretion. J. Biol. Chem. 283, + 25247–25255. doi: 10.1074/jbc.m801967200 + Empadinhas, N., and da Costa, M. S. (2008). Osmoadaptation mechanisms + in prokaryotes: distribution of compatible solutes. Int. Microbiol. 11, + 151–161. + Espinosa-Urgel, M., Salido, A., and Ramos, J. L. (2000). Genetic analysis of + functions involved in adhesion of Pseudomonas putida to seeds. J. Bacteriol. 182, + 2363–2369. doi: 10.1128/jb.182.9.2363-2369.2000 + Fuqua, C. (2010). Passing the baton between laps: adhesion and cohesion in + Pseudomonas putida biofilms. Mol.Microbiol. 77, 533–536. doi: 10.1111/j.1365- + 2958.2010.07250.x + Galvao, T. C., De Lorenzo, V., and Canovas, D. (2006). Uncoupling of choline- + O-sulphate utilization from osmoprotection in Pseudomonas putida. Mol. + Microbiol. 62, 1643–1654. doi: 10.1111/j.1365-2958.2006.05488.x + Garrido-Sanz, D., Meier-Kolthoff, J. P., Göker, M., Martín, M., Rivilla, R., + and Redondo-Nieto, M. (2016). Genomic and genetic diversity within the + Pseudomonas fluorescens complex. PLoS One 11:e0150183. doi: 10.1371/journal. + pone.0150183 + Halverson, L. J., and Firestone, M. K. (2000). Differential effects of permeating and + nonpermeating solutes on the fatty acid composition of Pseudomonas putida. + Appl. Environ. Microbiol. 66, 2414–2421. doi: 10.1128/aem.66.6.2414-2421. + 2000 + Hampel, K. J., Labauve, A. E., Meadows, J. A., Fitzsimmons, L. F., Nock, A. M., and + Wargo, M. J. (2014). Characterization of the GbdR regulon in Pseudomonas + aeruginosa. J. Bacteriol. 196, 7–15. doi: 10.1128/jb.01055-13 + Harwood, C. S., and Parales, R. E. (1996). The beta-ketoadipate pathway and + the biology of self-identity. Annu. Rev. Microbiol. 50, 553–590. doi: 10.1146/ + annurev.micro.50.1.553 + Hesse, C., Schulz, F., Bull, C. T., Shaffer, B. T., Yan, Q., Shapiro, N., et al. (2018). + Genome-based evolutionary history of Pseudomonas spp. Environ. Microbiol. + 20, 2142–2159. + Hinsa, S. M., Espinosa-Urgel, M., Ramos, J. L., and O’Toole, G. A. (2003). + Transition from reversible to irreversible attachment during biofilm formation + by Pseudomonas fluorescens WCS365 requires an ABC transporter and a large + secreted protein. Mol. Microbiol. 49, 905–918. doi: 10.1046/j.1365-2958.2003. + 03615.x + Hojberg, O., Schnider, U., Winteler, H. V., Sorensen, J., and Haas, D. (1999). + Oxygen-sensing reporter strain of Pseudomonas fluorescens for monitoring + Frontiers in Microbiology | www.frontiersin.org 16 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + the distribution of low-oxygen habitats in soil. Appl. Environ. Microbiol. 65, + 4085–4093. doi: 10.1128/aem.65.9.4085-4093.1999 + Hong, S. Y., Park, J. H., Cho, S. H., Yang, M. S., and Park, C. M. (2011). Phenological + growth stages of Brachypodium distachyon: codification and description. Weed + Res. 51, 612–620. doi: 10.1111/j.1365-3180.2011.00877.x + Howell, C. R., and Stipanovic, R. D. (1980). Suppression of Pythium ultimum + induced damping-off of cotton seedlings by Pseudomonas fluorescens and + its antibiotic pyoluteorin. Phytopathology 70, 712–715. doi: 10.1094/phyto- + 70-712 + Hummerjohann, J., Kuttel, E., Quadroni, M., Ragaller, J., Leisinger, T., and Kertesz, + M. A. (1998). Regulation of the sulfate starvation response in Pseudomonas + aeruginosa: role of cysteine biosynthetic intermediates.Microbiology 144, 1375– + 1386. doi: 10.1099/00221287-144-5-1375 + Kabbadj, A., Makoudi, B., Mouradi, M., Pauly, N., Frendo, P., and Ghoulam, + C. (2017). Physiological and biochemical responses involved in water deficit + tolerance of nitrogen-fixing Vicia faba. PLoS One 12:e0190284. doi: 10.1371/ + journal.pone.0190284 + Kamilova, F., Validov, S., Azarova, T., Mulders, I., and Lugtenberg, B. (2005). + Enrichment for enhanced competitive plant root tip colonizers selects for a new + class of biocontrol bacteria. Environ. Microbiol. 7, 1809–1817. doi: 10.1111/j. + 1462-2920.2005.00889.x + Kanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., et al. (2008). + KEGG for linking genomes to life and the environment. Nucleic Acids Res. 36, + D480–D484. + Katoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment + software version 7: improvements in performance and usability. Mol. Biol. Evol. + 30, 772–780. doi: 10.1093/molbev/mst010 + Kawasaki, A., Donn, S., Ryan, P. R., Mathesius, U., Devilla, R., Jones, A., et al. + (2016). Microbiome and exudates of the root and rhizosphere of Brachypodium + distachyon, a model for wheat. PLoS One 11:e0164533. doi: 10.1371/journal. + pone.0164533 + Kind, T., Wohlgemuth, G., Lee, D. Y., Lu, Y., Palazoglu, M., Shahbaz, S., et al. + (2009). FiehnLib: mass spectral and retention index libraries for metabolomics + based on quadrupole and time-of-flight gas chromatography/mass + spectrometry. Anal. Chem. 81, 10038–10048. doi: 10.1021/ac9019522 + King, E. O., Ward, M. K., and Raney, D. E. (1954). Two simple media for the + demonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44, 301–307. + Klonowska, A., Melkonian, R., Miche, L., Tisseyre, P., and Moulin, L. (2018). + Transcriptomic profiling of Burkholderia phymatum STM815, Cupriavidus + taiwanensis LMG19424 and Rhizobium mesoamericanum STM3625 in response + to Mimosa pudica root exudates illuminates the molecular basis of their + nodulation competitiveness and symbiotic evolutionary history.BMC Genomics + 19:105. doi: 10.1186/s12864-018-4487-2 + Kobayashi, T., and Nishizawa, N. K. (2012). Iron uptake, translocation, and + regulation in higher plants. Annu. Rev. Plant Biol. 63, 131–152. doi: 10.1146/ + annurev-arplant-042811-105522 + Kohler, P. R., Choong, E. L., and Rossbach, S. (2011). The RpiR-like repressor + IolR regulates inositol catabolism in Sinorhizobium meliloti. J. Bacteriol. 193, + 5155–5163. doi: 10.1128/jb.05371-11 + Lee, D., and Fiehn, O. (2008). High quality metabolomic data for Chlamydomonas + reinhardtii. Plant Meth. 4:7. doi: 10.1186/1746-4811-4-7 + Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for + the display and annotation of phylogenetic and other trees. Nucleic Acids Res. + 44, W242–W245. + Li, L., Stoeckert, C. J. Jr., and Roos, D. S. (2003). OrthoMCL: identification of + ortholog groups for eukaryotic genomes. Genome Res. 13, 2178–2189. doi: + 10.1101/gr.1224503 + Liao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient + general purpose program for assigning sequence reads to genomic features. + Bioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656 + Liu, Y., Rainey, P. B., and Zhang, X. X. (2015). Molecular mechanisms of xylose + utilization by Pseudomonas fluorescens: overlapping genetic responses to xylose, + xylulose, ribose and mannitol. Mol. Microbiol. 98, 553–570. doi: 10.1111/mmi. + 13142 + Loper, J. E., Hassan, K. A., Mavrodi, D. V., Davis, E. W., Lim, C. K., Shaffer, + B. T., et al. (2012). Comparative genomics of plant-associated Pseudomonas + spp.: insights into diversity and inheritance of traits involved in multitrophic + interactions. PLoS Genet. 8:e1002784. doi: 10.1371/journal.pgen.1002784 + Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change + and dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550. + Lugtenberg, B., and Kamilova, F. (2009). Plant-growth-promoting rhizobacteria. + Annu. Rev. Microbiol. 63, 541–556. + Lugtenberg, B. J., Dekkers, L., and Bloemberg, G. V. (2001). Molecular + determinants of rhizosphere colonization by Pseudomonas. Annu. Rev. + Phytopathol. 39, 461–490. + Lynch, J. M. (1990). “Microbial metabolites,” in The Rhizosphere, ed. J. M. Lynch + (Chichester: JohnWiley & Sons), 177–206. + Mark, G. L., Dow, J. M., Kiely, P. D., Higgins, H.,Haynes, J., Baysse, C., et al. (2005). + Transcriptome profiling of bacterial responses to root exudates identifies genes + involved in microbe-plant interactions. Proc. Natl. Acad. Sci. U.S.A. 102, 17454– + 17459. doi: 10.1073/pnas.0506407102 + Markowitz, V. M., Chen, I. M., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y., + et al. (2012). IMG: the integrated microbial genomes database and comparative + analysis system. Nucleic Acids Res. 40, D115–D122. + Martinez-Gil, M., Yousef-Coronado, F., and Espinosa-Urgel, M. (2010). LapF, + the second largest Pseudomonas putida protein, contributes to plant root + colonization and determines biofilm architecture. Mol. Microbiol. 77, 549–561. + doi: 10.1111/j.1365-2958.2010.07249.x + Matilla, M. A., Espinosa-Urgel, M., Rodriguez-Herva, J. J., Ramos, J. L., and + Ramos-Gonzalez, M. I. (2007). Genomic analysis reveals the major driving + forces of bacterial life in the rhizosphere. Genome Biol. 8:R179. + Miller, K. J., and Wood, J. M. (1996). Osmoadaptation by rhizosphere bacteria. + Annu. Rev. Microbiol. 50, 101–136. doi: 10.1146/annurev.micro.50.1.101 + Moore, E. R. B., Tindall, B. J.,Martins Dos Santos, V. A. P., Pieper, D. H., Ramos, J.- + L., and Palleroni, N. J. (2006). “Nonmedical Pseudomonas,” in The Prokaryotes, + eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stackebrandt + (New York, NY: Springer), 646–703. + Mulet, M., Lalucat, J., and Garcia-Valdes, E. (2010). DNA sequence-based analysis + of the Pseudomonas species. Environ. Microbiol. 12, 1513–1530. + Naylor, D., DeGraaf, S., Purdom, E., and Coleman-Derr, D. (2017). Drought + and host selection influence bacterial community dynamics in the grass root + microbiome. ISME J. 11, 2691–2704. doi: 10.1038/ismej.2017.118 + Nguyen, C. (2003). Rhizodeposition of organic C by plants: mechanisms and + controls. Agronomie 23, 375–396. doi: 10.1051/agro:2003011 + Nielsen, L., Li, X., and Halverson, L. J. (2011). Cell-cell and cell-surface + interactions mediated by cellulose and a novel exopolysaccharide contribute + to Pseudomonas putida biofilm formation and fitness under water-limiting + conditions. Environ. Microbiol. 13, 1342–1356. doi: 10.1111/j.1462-2920.2011. + 02432.x + Parejko, J. A., Mavrodi, D. V., Mavrodi, O. V., Weller, D. M., and Thomashow, + L. S. (2012). Population structure and diversity of phenazine-1-carboxylic acid + producing fluorescent Pseudomonas spp. from dryland cereal fields of central + Washington state (USA) Microb. Ecol. 63, 226–241. doi: 10.1007/s00248-012- + 0015-0 + Phillips, D. A., Fox, T. C., King, M. D., Bhuvaneswari, T. V., and Teuber, L. R. + (2004). Microbial products trigger amino acid exudation from plant roots. Plant + Physiol. 136, 2887–2894. doi: 10.1104/pp.104.044222 + Raaijmakers, J. M., Vandersluis, I., Koster, M., Bakker, P. A. H. M., Weisbeek, + P. J., and Schippers, B. (1995). Utilization of heterologous siderophores and + rhizosphere competence of fluorescent Pseudomonas spp. Can. J. Microbiol. 41, + 126–135. doi: 10.1139/m95-017 + Raaijmakers, J. M., and Weller, D. M. (1998). Natural plant protection by 2,4- + diacetylphloroglucinol-producing Pseudomonas spp. in take-all decline soils. + Mol. Plant Microbe Interact. 11, 144–152. doi: 10.1094/mpmi.1998.11.2.144 + Rahman, M. M., Andberg, M., Thangaraj, S. K., Parkkinen, T., Penttila, M., Janis, + J., et al. (2017). The crystal structure of a bacterial L-arabinonate dehydratase + contains a [2Fe-2S] cluster. ACS Chem. Biol. 12, 1919–1927. doi: 10.1021/ + acschembio.7b00304 + Ramos-Gonzalez, M. I., Campos, M. J., and Ramos, J. L. (2005). Analysis of + Pseudomonas putida KT2440 gene expression in the maize rhizosphere: in vivo + expression technology capture and identification of root-activated promoters. + J. Bacteriol. 187, 4033–4041. doi: 10.1128/jb.187.12.4033-4041.2005 + Reinhold-Hurek, B., Bunger, W., Burbano, C. S., Sabale, M., and Hurek, T. + (2015). Roots shaping their microbiome: global hotspots for microbial activity. + Annu. Rev. Phytopathol. 53, 403–424. doi: 10.1146/annurev-phyto-082712-10 + 2342 + Frontiers in Microbiology | www.frontiersin.org 17 April 2021 | Volume 12 | Article 651282 + Mavrodi et al. The Effect of Root Exudates on the Pseudomonas Transcriptome + Sanchez-Contreras, M., Martin, M., Villacieros, M., O’Gara, F., Bonilla, I., and + Rivilla, R. (2002). Phenotypic selection and phase variation occur during + alfalfa root colonization by Pseudomonas fluorescens F113. J. Bacteriol. 184, + 1587–1596. doi: 10.1128/jb.184.6.1587-1596.2002 + Sarniguet, A., Kraus, J., Henkels, M. D., Muehlchen, A. M., and Loper, J. E. (1995). + The sigma factor ss affects antibiotic production and biological control activity + of Pseudomonas fluorescens Pf-5. Proc. Natl. Acad. Sci. U.S.A. 92, 12255–12259. + doi: 10.1073/pnas.92.26.12255 + Schnider-Keel, U., Lejbolle, K. B., Baehler, E., Haas, D., and Keel, C. (2001). The + sigma factor AlgU (AlgT) controls exopolysaccharide production and tolerance + towards desiccation and osmotic stress in the biocontrol agent Pseudomonas + fluorescens CHA0. Appl. Environ. Microbiol. 67, 5683–5693. doi: 10.1128/aem. + 67.12.5683-5693.2001 + Schroth, M. N., Hildebrand, D. C., and Panopoulos, N. J. (2006). + “Phytopathogenic pseudomonads and related plant-associated pseudomonads,” + in The Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.- + H. Schleifer, and E. Stackebrandt (New York, NY: Springer), 714–740. + doi: 10.1007/0-387-30746-x_23 + Schwartz, C. J., Doyle, M. R., Manzaneda, A. J., Rey, P. J., Mitchell-Olds, T., and + Amasino, R. M. (2010). Natural variation of flowering time and vernalization + responsiveness in Brachypodium distachyon. Bioenergy Res. 3, 38–46. doi: 10. + 1007/s12155-009-9069-3 + Silby, M. W., Cerdeno-Tarraga, A. M., Vernikos, G. S., Giddens, S. R., Jackson, + R. W., Preston, G. M., et al. (2009). Genomic and genetic analyses of diversity + and plant interactions of Pseudomonas fluorescens. Genome Biol. 10:R51. + Silby, M. W., and Levy, S. B. (2004). Use of in vivo expression technology to + identify genes important in growth and survival of Pseudomonas fluorescens + Pf0-1 in soil: Discovery of expressed sequences with novel genetic organization. + J. Bacteriol. 186, 7411–7419. doi: 10.1128/jb.186.21.7411-7419.2004 + Simons, M., Van Der Bij, A. J., Brand, I., De Weger, L. A., Wijffelman, C. A., + and Lugtenberg, B. J. (1996). Gnotobiotic system for studying rhizosphere + colonization by plant growth-promoting Pseudomonas bacteria. Mol. Plant + Microbe Interact. 9, 600–607. doi: 10.1094/mpmi-9-0600 + Simons, M., Van Der Bij, A. J., Brand, J., De Weger, L. A., Wijffelman, D. A., and + Lugtenberg, B. J. J. (1997). Amino acid synthesis is necessary for tomato root + colonization by Pseudomonas fluorescens strain WCS365. Mol. Plant Microbe + Interact. 10, 102–106. doi: 10.1094/mpmi.1997.10.1.102 + Smibert, R. M., and Kreig, N. R. (1994). “Phenotypic characterization,” in Methods + for General and Molecular Bacteriology, eds P. Gerhardt, R. G. E. Murray, W. A. + Wood, and N. R. Kreig (Washington, DC: American Society of Microbiology), + 607–654. + Thomashow, L. S., and Weller, D. M. (1988). Role of a phenazine antibiotic from + Pseudomonas fluorescens in biological control of Gaeumannomyces graminis + var. tritici. J. Bacteriol. 170, 3499–3508. doi: 10.1128/jb.170.8.3499-3508.1988 + Thomashow, L. S., Weller, D. M., Bonsall, R. F., and Pierson, L. S. (1990). + Production of the antibiotic phenazine-1-carboxylic acid by fluorescent + Pseudomonas species in the rhizosphere of wheat. Appl. Environ. Microbiol. 56, + 908–912. doi: 10.1128/aem.56.4.908-912.1990 + Tyler, L., Fangel, J. U., Fagerstrom, A. D., Steinwand, M. A., Raab, T. K., Willats, + W. G., et al. (2014). Selection and phenotypic characterization of a core + collection of Brachypodium distachyon inbred lines. BMC Plant Biol. 14:25. + doi: 10.1186/1471-2229-14-25 + Vacheron, J., Desbrosses, G., Bouffaud, M. L., Touraine, B., Moenne-Loccoz, Y., + Muller, D., et al. (2013). Plant growth-promoting rhizobacteria and root system + functioning. Front. Plant Sci. 4:356. doi: 10.3389/fpls.2013.00356 + van den Broek, D., Bloemberg, G. V., and Lugtenberg, B. (2005). The role of + phenotypic variation in rhizosphere Pseudomonas bacteria. Environ. Microbiol. + 7, 1686–1697. doi: 10.1111/j.1462-2920.2005.00912.x + van Veen, J. A., Van Overbeek, L. S., and Van Elsas, J. D. (1997). Fate and activity + of microorganisms introduced into soil. Microbiol. Mol. Biol. Rev. 61, 121–135. + doi: 10.1128/.61.2.121-135.1997 + Walker, T. S., Bais, H. P., Halligan, K. M., Stermitz, F. R., and Vivanco, J. M. + (2003). Metabolic profiling of root exudates of Arabidopsis thaliana. J. Agric. + Food Chem. 51, 2548–2554. + Wargo, M. J. (2013a). Choline catabolism to glycine betaine contributes to + Pseudomonas aeruginosa survival during murine lung infection. PLoS One + 8:e56850. doi: 10.1371/journal.pone.0056850 + Wargo, M. J. (2013b). Homeostasis and catabolism of choline and glycine betaine: + lessons from Pseudomonas aeruginosa. Appl. Environ. Microbiol. 79, 2112– + 2120. + Whipps, J. M. (1990). “Carbon economy,” in The Rhizosphere, ed. J. M. Lynch + (Chichester: JohnWiley & Sons), 59–97. + Winsor, G. L., Van Rossum, T., Lo, R., Khaira, B., Whiteside, M. D., Hancock, + R. E., et al. (2009). Pseudomonas Genome Database: facilitating user-friendly, + comprehensive comparisons of microbial genomes. Nucleic Acids Res. 37, + D483–D488. + Yahr, T. L., and Parsek, M. R. (2006). “Pseudomonas aeruginosa,” in The + Prokaryotes, eds M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. + Stackebrandt (New York, NY: Springer), 704–713. + Yancey, P. H., Clark, M. E., Hand, S. C., Bowlus, R. D., and Somero, G. N. (1982). + Living with water stress: evolution of osmolyte systems. Science 217, 1214–1222. + Ye, J., Zhang, Y., Cui, H., Liu, J., Wu, Y., Cheng, Y., et al. (2018). WEGO 2.0: a web + tool for analyzing and plotting GO annotations, 2018 update. Nucleic Acids Res. + 46, W71–W75. + Yoshida, K., Yamaguchi, M., Morinaga, T., Kinehara, M., Ikeuchi, M., Ashida, H., + et al. (2008). myo-Inositol catabolism in Bacillus subtilis. J. Biol. Chem. 283, + 10415–10424. + Yousef-Coronado, F., Travieso, M. L., and Espinosa-Urgel, M. (2008). Different, + overlapping mechanisms for colonization of abiotic and plant surfaces by + Pseudomonas putida. FEMS Microbiol. Lett. 288, 118–124. + Zboralski, A., and Filion, M. (2020). Genetic factors involved in rhizosphere + colonization by phytobeneficial Pseudomonas spp. Comput. Struct. Biotechnol. + J. 18, 3539–3554. + Zolla, G., Bakker, M. G., Badri, D. V., Chaparro, J. M., Sheflin, A. M.,Manter, D. K., + et al. (2013). “Understanding root-microbiome interactions,” in Molecular + Microbial Ecology of the Rhizosphere, ed. F. J. De Bruijn (Hoboken, NJ: John + Wiley & Sons), 745–754. + Conflict of Interest: The authors declare that the research was conducted in the + absence of any commercial or financial relationships that could be construed as a + potential conflict of interest. + Copyright © 2021 Mavrodi, McWilliams, Peter, Berim, Hassan, Elbourne, + LeTourneau, Gang, Paulsen, Weller, Thomashow, Flynt and Mavrodi. This is an + open-access article distributed under the terms of the Creative Commons Attribution + License (CC BY). The use, distribution or reproduction in other forums is permitted, + provided the original author(s) and the copyright owner(s) are credited and that the + original publication in this journal is cited, in accordance with accepted academic + practice. No use, distribution or reproduction is permitted which does not comply + with these terms. + Frontiers in Microbiology | www.frontiersin.org 18 April 2021 | Volume 12 | Article 651282 + +raw_completion_output: |- + study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas + authors: Olga V. Mavrodi; Janiece R. McWilliams; Jacob O. Peter; Anna Berim; Karl A. Hassan; Liam D. H. Elbourne; Melissa K. LeTourneau; David R. Gang; Ian T. Paulsen; David M. Weller; Linda S. Thomashow; Alex S. Flynt; Dmitri V. Mavrodi + doi: 10.3389/fmicb.2021.651282 + date: 14 April 2021 + experiments: The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. +prompt: |+ + From the text below, extract the following entities in the following format: + + label: + experiment_motivation: + experiment_design: + environment: + host: + host_type: + target_microbes: + biological_system: + conditions: + experimental_factors: + is_biolog_experiment: + type_of_biolog_experiment: + plates: + replicates: + protocol_steps: + plate_reader_types_of_data_collected: + od_protocol: + respiration_protocol: + instrument_used: + analysis_software: + incubation_temperature: + incubation_duration: + + + Text: + The experiment aimed to examine the molecular communication between beneficial rhizosphere bacteria (Pseudomonas spp.) and their host plant (Brachypodium distachyon). Researchers collected root exudates from B. distachyon by growing surface-sterilized plants hydroponically in sterile jars filled with glass beads and distilled water in a controlled environment (20-hour light at 24°C, 4-hour dark at 18°C). Root exudates were extracted, sterilized with 0.22 µm filtration, lyophilized, and stored at –80°C. The composition of the collected root exudates was analyzed via liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS). For LC-MS, a Synapt G2-S quadrupole-ion mobility time-of-flight mass spectrometer and an Ultra Performance Liquid Chromatograph (UPLC) with photodiode array detection were used, while GC-MS analysis was performed with a Pegasus 4D time-of-flight mass spectrometer coupled to an Agilent 7890A GC. Sixty-day-old cultures of eight Pseudomonas fluorescens-group strains were grown on medium amended with Brachypodium root exudates under controlled microaerobic (15% O2) conditions. Differential gene expression in response to root exudates was assessed using transcriptomics (RNA-seq), with RNA samples depleted of rRNA prior to stranded library preparation and sequencing on the Illumina HiSeq 2500 platform. Metabolic capabilities of these Pseudomonas strains were further characterized using Biolog Phenotype MicroArrays (PM01 and PM2A plates), incubated in an OmniLog Phenotype MicroArray System at 25°C for 48 hours to profile the catabolism of diverse carbon sources. Optical density and colorimetric respiration data (formazan production, absorbance detection at 590 nm) were collected every 15 minutes. Each experiment had three biological replicates. Analysis of Biolog data was performed using OmniLog Parametric Analysis software v1.20.02. For transcriptomic data, DESeq2 and various comparative genomic tools (e.g., OrthoMCL, COG-based phylogeny) were used to identify differentially expressed genes and functional annotations. Data collection and incubations lasted for 20-22 hours at 25°C. The experimental analyses revealed changes in gene expression associated with plant metabolite catabolism, transport, stress responses (e.g., oxidative and metal stresses), biocontrol activities, and suggested the importance of root-secreted osmoprotectants for bacterial rhizosphere colonization. + + === + +extracted_object: + study_title: Root Exudates Alter the Expression of Diverse Metabolic, Transport, + Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas + authors: + - Olga V. Mavrodi + - Janiece R. McWilliams + - Jacob O. Peter + - Anna Berim + - Karl A. Hassan + - Liam D. H. Elbourne + - Melissa K. LeTourneau + - David R. Gang + - Ian T. Paulsen + - David M. Weller + - Linda S. Thomashow + - Alex S. Flynt + - Dmitri V. Mavrodi + doi: 10.3389/fmicb.2021.651282 + date: 14 April 2021 + experiments: + - label: This experiment investigates the molecular communication between beneficial + rhizosphere bacteria (Pseudomonas spp.) and the host plant Brachypodium distachyon + by analyzing bacterial transcriptional and metabolic responses to plant root + exudates. + experiment_motivation: To examine molecular communication between beneficial + rhizosphere bacteria and their host plant, specifically focusing on bacterial + responses to plant-derived metabolites present within root exudates. + experiment_design: Surface-sterilized Brachypodium distachyon plants were grown + hydroponically to collect root exudates, which were characterized by LC-MS + and GC-MS. Eight Pseudomonas fluorescens-group strains were cultured with + root exudates under controlled microaerobic conditions, followed by transcriptomic + analysis (RNA-seq) and metabolic profiling using Biolog Phenotype MicroArrays. + environment: Controlled environment (20-hour light at 24°C, 4-hour dark at 18°C); + microaerobic environment (15% O₂). + host: + - AUTO:Brachypodium%20distachyon + - AUTO:Pseudomonas%20fluorescens-group%20strains + host_type: Plant; Bacteria + target_microbes: + - AUTO:Pseudomonas%20fluorescens + biological_system: Rhizosphere + conditions: + - 20-hour light + - 4-hour dark + - 24°C daytime + - 18°C nighttime + - microaerobic (15% O₂) + experimental_factors: + - AUTO:Presence%20of%20Brachypodium%20root%20exudates + is_biolog_experiment: yes + type_of_biolog_experiment: Phenotype MicroArray + plates: + - PM01 + - PM2A + replicates: '3' + protocol_steps: + - Surface-sterilized plants grown hydroponically in sterile jars filled with + glass beads and distilled water under controlled conditions (20-hour light + 24°C, 4-hour dark 18°C) + - root exudates collected, sterilized via 0.22 µm filtration, lyophilized, + stored at –80°C + - root exudates analyzed by LC-MS and GC-MS + - LC-MS analysis performed with Synapt G2-S quadrupole-ion mobility TOF Mass + Spectrometer and Ultra Performance Liquid Chromatograph (UPLC) with photodiode + array detection + - GC-MS analysis performed with Pegasus 4D TOF Mass Spectrometer coupled to + Agilent 7890A GC + - cultures (60-day-old) of eight Pseudomonas fluorescens-group strains grown + on medium with root exudates under microaerobic (15% O₂) conditions + - RNA extracted and depleted of rRNA + - stranded library preparation and sequencing via Illumina HiSeq 2500 + - Biolog Phenotype MicroArray (PM01 and PM2A plates) inoculation and incubation + in OmniLog Phenotype MicroArray System at 25°C for 48 hours + - collection of optical density and colorimetric respiration data (formazan + production, absorbance at 590 nm) every 15 minutes + - Biolog data analysis performed using OmniLog Parametric Analysis software + v1.20.02 + - Transcriptome data analyzed using DESeq2 and comparative genomic tools. + plate_reader_types_of_data_collected: + - Optical density + - formazan (absorbance at 590 nm) + od_protocol: Not provided + respiration_protocol: Colorimetric respiration measured by formazan production + (absorbance detection at 590 nm), collected every 15 minutes over a period + of 48 hours at 25°C. + instrument_used: + - Synapt G2-S quadrupole-ion mobility TOF Mass Spectrometer + - Ultra Performance Liquid Chromatograph (UPLC) + - Pegasus 4D TOF Mass Spectrometer + - Agilent 7890A GC + - Illumina HiSeq 2500 + - OmniLog Phenotype MicroArray System + analysis_software: + - OmniLog Parametric Analysis software v1.20.02 + - DESeq2 + - OrthoMCL + - COG-based phylogeny + incubation_temperature: 25°C + incubation_duration: 48 hours +named_entities: + - id: AUTO:Brachypodium%20distachyon + label: Brachypodium distachyon + original_spans: + - 2671:2693 + - 7905:7927 + - 75924:75946 + - 82382:82404 + - 84035:84057 + - 92737:92759 + - 94848:94870 + - id: AUTO:Pseudomonas%20fluorescens-group%20strains + label: Pseudomonas fluorescens-group strains + - id: AUTO:Pseudomonas%20fluorescens + label: Pseudomonas fluorescens + original_spans: + - 2637:2659 + - 7986:8008 + - 75071:75093 + - 75391:75413 + - 75571:75593 + - 76860:76882 + - 78450:78472 + - 78876:78898 + - 79419:79441 + - 80612:80634 + - 81674:81696 + - 81944:81966 + - 82617:82639 + - 86158:86180 + - 91547:91569 + - 91814:91836 + - 92145:92167 + - 93005:93027 + - 93180:93202 + - 93811:93833 + - 94250:94272 + - id: AUTO:Presence%20of%20Brachypodium%20root%20exudates + label: Presence of Brachypodium root exudates From 89d76f4cbe9a4000b6f256380f77e167903ef1ef Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Tue, 18 Mar 2025 22:04:37 -0700 Subject: [PATCH 12/25] adding METPO annotator --- src/ontogpt/templates/biolog.py | 3 +- src/ontogpt/templates/biolog.py_last | 352 ++++++++++++++++++++ src/ontogpt/templates/biolog_simple.py_last | 123 +++++++ 3 files changed, 476 insertions(+), 2 deletions(-) create mode 100644 src/ontogpt/templates/biolog.py_last create mode 100644 src/ontogpt/templates/biolog_simple.py_last diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 805fd3b53..018bfea89 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -77,7 +77,7 @@ def __contains__(self, key:str) -> bool: 'prefix_reference': 'https://example.org/PaperExtractionSchema/'}, 'rdf': {'prefix_prefix': 'rdf', 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, - 'source_file': 'src/ontogpt/templates/biolog.yaml', + 'source_file': '/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml', 'title': 'Paper Extraction Schema'} ) class NullDataOptions(str, Enum): @@ -673,4 +673,3 @@ def pattern_original_spans(cls, v): Host.model_rebuild() Microbe.model_rebuild() ExperimentalFactor.model_rebuild() - diff --git a/src/ontogpt/templates/biolog.py_last b/src/ontogpt/templates/biolog.py_last new file mode 100644 index 000000000..baedb1889 --- /dev/null +++ b/src/ontogpt/templates/biolog.py_last @@ -0,0 +1,352 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "0.1.0" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'paperex', + 'default_range': 'string', + 'description': 'Schema for extracting structured data from papers, including ' + 'Biolog Phenotype MicroArray experiments.', + 'id': 'https://example.org/PaperExtractionSchema', + 'imports': ['linkml:types'], + 'license': 'https://creativecommons.org/publicdomain/zero/1.0/', + 'name': 'PaperExtractionSchema', + 'prefixes': {'linkml': {'prefix_prefix': 'linkml', + 'prefix_reference': 'https://w3id.org/linkml/'}, + 'paperex': {'prefix_prefix': 'paperex', + 'prefix_reference': 'https://example.org/PaperExtractionSchema/'}, + 'rdf': {'prefix_prefix': 'rdf', + 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, + 'source_file': '/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml', + 'title': 'Paper Extraction Schema'} ) + +class FlexibleBooleanEnum(str, Enum): + """ + Minimal enumeration for capturing yes/no/true/false in lowercase, plus 'not provided' if data are missing or ambiguous + + """ + yes = "yes" + no = "no" + true = "true" + false = "false" + not_provided = "not provided" + + + +class NamedEntity(ConfiguredBaseModel): + """ + A generic named entity for all non-root classes. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + pass + + +class Paper(ConfiguredBaseModel): + """ + A single paper or study. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema', 'tree_root': True}) + + study_title: str = Field(default=..., description="""The paper's title.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Extract the paper's title from the " + 'text. \n' + 'If no title is found, return "Not ' + 'provided".\n'}}, + 'domain_of': ['Paper']} }) + authors: List[Author] = Field(default=..., description="""Authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide a YAML list of authors under the ' + 'key `authors:`.\n' + 'Example:\n' + '\n' + ' authors:\n' + ' - name: "Olga V. Mavrodi"\n' + ' - name: "Janiece R. McWilliams"\n' + '\n' + 'If no authors are found, return an ' + '*empty* list:\n' + '\n' + ' authors: []\n' + '\n' + '(Avoid leaving `authors:` blank with no ' + 'list or value.)\n'}}, + 'domain_of': ['Paper']} }) + doi: Optional[str] = Field(default=None, description="""DOI of the publication.""", json_schema_extra = { "linkml_meta": {'alias': 'doi', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the DOI if available. If not ' + "found, use 'Not provided'.\n"}}, + 'domain_of': ['Paper']} }) + date: Optional[str] = Field(default=None, description="""Publication date.""", json_schema_extra = { "linkml_meta": {'alias': 'date', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the publication date. If not ' + "specified, 'Not provided'.\n"}}, + 'domain_of': ['Paper']} }) + experiments: List[Experiment] = Field(default=..., description="""Experiments (e.g. Biolog Phenotype MicroArray) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Summarize each distinct experiment under ' + '`experiments:`\n' + 'with a YAML list. For example:\n' + '\n' + ' experiments:\n' + ' - experiment_motivation: ' + '"Screening with Biolog..."\n' + ' ...\n' + ' - experiment_motivation: "RNA-seq ' + 'profiling..."\n' + ' ...\n' + '\n' + 'If none are found, return an empty ' + 'list:\n' + '\n' + ' experiments: []\n' + '\n' + '(Do not leave it blank.)\n'}}, + 'domain_of': ['Paper']} }) + + +class Author(NamedEntity): + """ + An author of the paper. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the full name of this author. If ' + "none, 'Not provided'."}}, + 'domain_of': ['Author']} }) + + +class Experiment(NamedEntity): + """ + A single experiment from the paper. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + experiment_motivation: Optional[str] = Field(default=None, description="""Rationale for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_motivation', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe the purpose or reason for doing ' + 'this experiment.'}}, + 'domain_of': ['Experiment']} }) + experiment_design: Optional[str] = Field(default=None, description="""Summary of how the experiment was designed.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_design', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Summarize the experimental design or ' + 'approach.'}}, + 'domain_of': ['Experiment']} }) + environment: Optional[str] = Field(default=None, description="""Location or conditions in which the experiment occurred (lab, greenhouse, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'environment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract any mention of environment or ' + "setting. If none, 'Not provided'."}}, + 'domain_of': ['Experiment']} }) + host: Optional[List[Host]] = Field(default=None, description="""One or more hosts in the experiment (e.g., a plant).""", json_schema_extra = { "linkml_meta": {'alias': 'host', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Identify any host organisms used (like ' + 'Brachypodium distachyon).'}}, + 'domain_of': ['Experiment']} }) + host_type: Optional[str] = Field(default=None, description="""Type of host (e.g., plant, animal).""", json_schema_extra = { "linkml_meta": {'alias': 'host_type', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Specify the host type (plant, animal, ' + "etc.). If not mentioned, 'Not " + "provided'."}}, + 'domain_of': ['Experiment']} }) + target_microbes: Optional[List[Microbe]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "List the microbe(s). If none, 'Not " + "provided'."}}, + 'domain_of': ['Experiment']} }) + biological_system: Optional[str] = Field(default=None, description="""Biological system under study.""", json_schema_extra = { "linkml_meta": {'alias': 'biological_system', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Extract e.g. 'rhizosphere' or 'root " + "exudates system'. If missing, 'Not " + "provided'."}}, + 'domain_of': ['Experiment']} }) + conditions: Optional[str] = Field(default=None, description="""Experimental conditions (temperature, pH, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'conditions', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Pull out relevant conditions. If absent, ' + "'Not provided'."}}, + 'domain_of': ['Experiment']} }) + experimental_factors: Optional[List[ExperimentalFactor]] = Field(default=None, description="""Factors tested or measured (e.g., presence of root exudates, carbon sources).""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'List factors or treatments. If none, ' + "'Not provided'."}}, + 'domain_of': ['Experiment']} }) + is_biolog_experiment: Optional[FlexibleBooleanEnum] = Field(default=None, description="""Whether the experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Indicate 'yes' or 'no' if explicitly " + 'stated. \n' + "If not clear, use 'not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + type_of_biolog_experiment: Optional[str] = Field(default=None, description="""Type/subtype of the Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'type_of_biolog_experiment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'If it is a Biolog experiment, specify ' + "(e.g. 'Phenotype MicroArray').\n" + "If absent, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + plates: Optional[List[str]] = Field(default=None, description="""Biolog plates used (e.g., PM01, PM02A).""", json_schema_extra = { "linkml_meta": {'alias': 'plates', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Identify plate types like PM01, PM02A, ' + 'EcoPlate, GEN III, etc.\n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + replicates: Optional[str] = Field(default=None, description="""Number of replicates.""", json_schema_extra = { "linkml_meta": {'alias': 'replicates', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract any mention of replicates. If ' + "none, 'Not provided'."}}, + 'domain_of': ['Experiment']} }) + protocol_steps: Optional[List[str]] = Field(default=None, description="""Key steps in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'protocol_steps', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'List the main protocol steps. \n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + plate_reader_types_of_data_collected: Optional[List[str]] = Field(default=None, description="""Data types collected (OD, formazan color, respiration, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'plate_reader_types_of_data_collected', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract data types measured by plate ' + 'reader or OmniLog. \n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + od_protocol: Optional[str] = Field(default=None, description="""Protocol for measuring optical density.""", json_schema_extra = { "linkml_meta": {'alias': 'od_protocol', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'If relevant, describe OD measurement ' + "steps. If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + respiration_protocol: Optional[str] = Field(default=None, description="""Protocol for measuring respiration or metabolic activity.""", json_schema_extra = { "linkml_meta": {'alias': 'respiration_protocol', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'If formazan or respiration is mentioned, ' + "describe. Otherwise, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + instrument_used: Optional[List[str]] = Field(default=None, description="""Instrument or equipment used (e.g. OmniLog Phenotype MicroArray System).""", json_schema_extra = { "linkml_meta": {'alias': 'instrument_used', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "If there's mention of an instrument " + '(OmniLog, etc.), list it. \n' + 'If multiple are referenced, list them. \n' + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + analysis_software: Optional[List[str]] = Field(default=None, description="""Software used to analyze or interpret data.""", json_schema_extra = { "linkml_meta": {'alias': 'analysis_software', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': "Extract any software used (e.g. 'OmniLog " + "Parametric Analysis v1.20.02'). \n" + "If none, 'Not provided'.\n"}}, + 'domain_of': ['Experiment']} }) + incubation_temperature: Optional[str] = Field(default=None, description="""Temperature at which cultures/plates were incubated.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_temperature', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'If a temperature is mentioned, capture ' + "it; else, 'Not provided'."}}, + 'domain_of': ['Experiment']} }) + incubation_duration: Optional[str] = Field(default=None, description="""Total incubation time for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_duration', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'If a time is given (48h, etc.), record ' + "it; else, 'Not provided'."}}, + 'domain_of': ['Experiment']} }) + + +class Host(NamedEntity): + """ + A host organism or system. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Host.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract details about this host. If ' + "none, 'Not provided'."}}, + 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor']} }) + + +class Microbe(NamedEntity): + """ + Information about a microbial species or strain. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Microbe.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract details about this microbe. If ' + "none, 'Not provided'."}}, + 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor']} }) + + +class ExperimentalFactor(NamedEntity): + """ + A factor manipulated or measured in an experiment. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for ExperimentalFactor.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract details about any factor (e.g. ' + 'carbon source, exudate presence).\n' + "If absent, 'Not provided'.\n"}}, + 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor']} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +NamedEntity.model_rebuild() +Paper.model_rebuild() +Author.model_rebuild() +Experiment.model_rebuild() +Host.model_rebuild() +Microbe.model_rebuild() +ExperimentalFactor.model_rebuild() diff --git a/src/ontogpt/templates/biolog_simple.py_last b/src/ontogpt/templates/biolog_simple.py_last new file mode 100644 index 000000000..58cd303d2 --- /dev/null +++ b/src/ontogpt/templates/biolog_simple.py_last @@ -0,0 +1,123 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "0.1.0" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'https://example.org/PaperExtractionSchema/', + 'description': 'A simplified schema describing the fields to extract from a ' + 'paper that includes study metadata and a minimal notion of ' + 'experiments.\n', + 'id': 'https://example.org/PaperExtractionSchema', + 'imports': ['linkml:types'], + 'name': 'PaperExtractionSchema', + 'prefixes': {'linkml': {'prefix_prefix': 'linkml', + 'prefix_reference': 'https://w3id.org/linkml/'}, + 'xsd': {'prefix_prefix': 'xsd', + 'prefix_reference': 'http://www.w3.org/2001/XMLSchema#'}}, + 'source_file': '/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog_simple.yaml', + 'title': 'Paper Extraction Schema', + 'types': {'boolean': {'base': 'bool', + 'from_schema': 'https://example.org/PaperExtractionSchema', + 'name': 'boolean'}, + 'string': {'base': 'str', + 'from_schema': 'https://example.org/PaperExtractionSchema', + 'name': 'string'}}} ) + + +class Paper(ConfiguredBaseModel): + """ + Top-level class representing a single paper/study. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema', 'tree_root': True}) + + study_title: str = Field(default=..., description="""Title of the study.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', 'domain_of': ['Paper']} }) + authors: List[Author] = Field(default=..., description="""List of authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', 'domain_of': ['Paper']} }) + doi: Optional[str] = Field(default=None, description="""DOI of the publication.""", json_schema_extra = { "linkml_meta": {'alias': 'doi', 'domain_of': ['Paper']} }) + date: Optional[str] = Field(default=None, description="""Date of the publication or the study.""", json_schema_extra = { "linkml_meta": {'alias': 'date', 'domain_of': ['Paper']} }) + experiments: Optional[List[Experiment]] = Field(default=None, description="""List of experiments mentioned in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'domain_of': ['Paper']} }) + + +class Author(ConfiguredBaseModel): + """ + Represents an author of the paper. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', 'domain_of': ['Author']} }) + + +class Experiment(ConfiguredBaseModel): + """ + A minimal experiment class (could be Biolog or otherwise). + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + experiment_title: Optional[str] = Field(default=None, description="""Title or short description of the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_title', 'domain_of': ['Experiment']} }) + is_biolog_experiment: Optional[bool] = Field(default=None, description="""Flag indicating if this experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', 'domain_of': ['Experiment']} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +Paper.model_rebuild() +Author.model_rebuild() +Experiment.model_rebuild() From faeaa6b1a6afa846fbfbf5726417da637f49ec0e Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 10:40:56 -0400 Subject: [PATCH 13/25] Fixes for the table extractions --- src/ontogpt/templates/table_arrays.py | 281 +++++++++++++++++++++++ src/ontogpt/templates/table_arrays.yaml | 4 +- src/ontogpt/templates/table_values.py | 285 ++++++++++++++++++++++++ src/ontogpt/templates/table_values.yaml | 4 +- 4 files changed, 570 insertions(+), 4 deletions(-) create mode 100644 src/ontogpt/templates/table_arrays.py create mode 100644 src/ontogpt/templates/table_values.py diff --git a/src/ontogpt/templates/table_arrays.py b/src/ontogpt/templates/table_arrays.py new file mode 100644 index 000000000..d1b4804a5 --- /dev/null +++ b/src/ontogpt/templates/table_arrays.py @@ -0,0 +1,281 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "None" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'STE', + 'description': 'An OntoGPT schema for extracting simple numerical tables with ' + 'a separate header row, label row, and data arrays', + 'id': 'https://w3id.org/ontogpt/table_arrays', + 'imports': ['core'], + 'name': 'SimpleTableExtraction', + 'source_file': 'src/ontogpt/templates/table_arrays.yaml'} ) + +class NullDataOptions(str, Enum): + UNSPECIFIED_METHOD_OF_ADMINISTRATION = "UNSPECIFIED_METHOD_OF_ADMINISTRATION" + NOT_APPLICABLE = "NOT_APPLICABLE" + NOT_MENTIONED = "NOT_MENTIONED" + + + +class ExtractionResult(ConfiguredBaseModel): + """ + A result of extracting knowledge on text + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + input_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_id', 'domain_of': ['ExtractionResult']} }) + input_title: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_title', 'domain_of': ['ExtractionResult']} }) + input_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_text', 'domain_of': ['ExtractionResult']} }) + raw_completion_output: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'raw_completion_output', 'domain_of': ['ExtractionResult']} }) + prompt: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'prompt', 'domain_of': ['ExtractionResult']} }) + extracted_object: Optional[Any] = Field(default=None, description="""The complex objects extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'extracted_object', 'domain_of': ['ExtractionResult']} }) + named_entities: Optional[List[Any]] = Field(default=None, description="""Named entities extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'named_entities', 'domain_of': ['ExtractionResult']} }) + + +class NamedEntity(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class CompoundExpression(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + pass + + +class Triple(CompoundExpression): + """ + Abstract parent for Relation Extraction tasks + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject', 'domain_of': ['Triple']} }) + predicate: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'predicate', 'domain_of': ['Triple']} }) + object: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object', 'domain_of': ['Triple']} }) + qualifier: Optional[str] = Field(default=None, description="""A qualifier for the statements, e.g. \"NOT\" for negation""", json_schema_extra = { "linkml_meta": {'alias': 'qualifier', 'domain_of': ['Triple']} }) + subject_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the subject of the statement, e.g. \"high dose\" or \"intravenously administered\"""", json_schema_extra = { "linkml_meta": {'alias': 'subject_qualifier', 'domain_of': ['Triple']} }) + object_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the object of the statement, e.g. \"severe\" or \"with additional complications\"""", json_schema_extra = { "linkml_meta": {'alias': 'object_qualifier', 'domain_of': ['Triple']} }) + + +class TextWithTriples(ConfiguredBaseModel): + """ + A text containing one or more relations of the Triple type. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + triples: Optional[List[Triple]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'triples', 'domain_of': ['TextWithTriples']} }) + + +class TextWithEntity(ConfiguredBaseModel): + """ + A text containing one or more instances of a single type of entity. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + entities: Optional[List[str]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'entities', 'domain_of': ['TextWithEntity']} }) + + +class RelationshipType(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core', + 'id_prefixes': ['RO', 'biolink']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Publication(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: Optional[str] = Field(default=None, description="""The publication identifier""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'domain_of': ['NamedEntity', 'Publication']} }) + title: Optional[str] = Field(default=None, description="""The title of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'title', 'domain_of': ['Publication']} }) + abstract: Optional[str] = Field(default=None, description="""The abstract of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'abstract', 'domain_of': ['Publication']} }) + combined_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'combined_text', 'domain_of': ['Publication']} }) + full_text: Optional[str] = Field(default=None, description="""The full text of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'full_text', 'domain_of': ['Publication']} }) + + +class AnnotatorResult(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject_text', 'domain_of': ['AnnotatorResult']} }) + object_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_id', 'domain_of': ['AnnotatorResult']} }) + object_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_text', 'domain_of': ['AnnotatorResult']} }) + + +class Table(ConfiguredBaseModel): + """ + “A table extracted from a publication, containing rows and columns in array form.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_arrays'}) + + table_id: Optional[str] = Field(default=None, description="""“Identifier for the table, e.g. ‘Table 1’, ‘Table 2’, etc.”""", json_schema_extra = { "linkml_meta": {'alias': 'table_id', 'domain_of': ['Table']} }) + source_document: Optional[str] = Field(default=None, description="""“Identifier or reference to the PDF or publication source.”""", json_schema_extra = { "linkml_meta": {'alias': 'source_document', 'domain_of': ['Table']} }) + caption: Optional[str] = Field(default=None, description="""“Text of the table caption or title.”""", json_schema_extra = { "linkml_meta": {'alias': 'caption', 'domain_of': ['Table']} }) + header_row: Optional[TableHeaderRow] = Field(default=None, description="""“The row containing the column headers.”""", json_schema_extra = { "linkml_meta": {'alias': 'header_row', 'domain_of': ['Table']} }) + label_row: Optional[TableLabelRow] = Field(default=None, description="""“The row containing row labels.”""", json_schema_extra = { "linkml_meta": {'alias': 'label_row', 'domain_of': ['Table']} }) + data_rows: Optional[List[TableDataRow]] = Field(default=None, description="""“A list of data rows, each representing a row in the table body.”""", json_schema_extra = { "linkml_meta": {'alias': 'data_rows', 'domain_of': ['Table']} }) + + +class TableHeaderRow(ConfiguredBaseModel): + """ + “A single row containing all column headers.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_arrays'}) + + header_values: Optional[List[str]] = Field(default=None, description="""“An array of strings corresponding to each column header.”""", json_schema_extra = { "linkml_meta": {'alias': 'header_values', 'domain_of': ['TableHeaderRow']} }) + + +class TableLabelRow(ConfiguredBaseModel): + """ + “A single row containing the labels for each of the data rows.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_arrays'}) + + label_values: Optional[List[str]] = Field(default=None, description="""“An array of strings corresponding to each row label, aligning with data_rows.”""", json_schema_extra = { "linkml_meta": {'alias': 'label_values', 'domain_of': ['TableLabelRow']} }) + + +class TableDataRow(ConfiguredBaseModel): + """ + “A row of data in the table body, indexed and containing an array of numeric values.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_arrays'}) + + row_index: Optional[int] = Field(default=None, description="""“Numeric index of this row (starting from 0 or 1).”""", json_schema_extra = { "linkml_meta": {'alias': 'row_index', 'domain_of': ['TableDataRow']} }) + values: Optional[List[float]] = Field(default=None, description="""“An array of numeric values in this data row.”""", json_schema_extra = { "linkml_meta": {'alias': 'values', 'domain_of': ['TableDataRow']} }) + note: Optional[str] = Field(default=None, description="""“Optional note or comment about this particular row.”""", json_schema_extra = { "linkml_meta": {'alias': 'note', 'domain_of': ['TableDataRow']} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +ExtractionResult.model_rebuild() +NamedEntity.model_rebuild() +CompoundExpression.model_rebuild() +Triple.model_rebuild() +TextWithTriples.model_rebuild() +TextWithEntity.model_rebuild() +RelationshipType.model_rebuild() +Publication.model_rebuild() +AnnotatorResult.model_rebuild() +Table.model_rebuild() +TableHeaderRow.model_rebuild() +TableLabelRow.model_rebuild() +TableDataRow.model_rebuild() + diff --git a/src/ontogpt/templates/table_arrays.yaml b/src/ontogpt/templates/table_arrays.yaml index 61c111b40..201f49d16 100644 --- a/src/ontogpt/templates/table_arrays.yaml +++ b/src/ontogpt/templates/table_arrays.yaml @@ -1,8 +1,9 @@ +id: https://w3id.org/ontogpt/table_arrays name: SimpleTableExtraction description: An OntoGPT schema for extracting simple numerical tables with a separate header row, label row, and data arrays default_prefix: STE imports: - - ontogpt:core + - core classes: Table: description: “A table extracted from a publication, containing rows and columns in array form.” @@ -65,4 +66,3 @@ slots: note: description: “Optional note or comment about this particular row.” range: string -templates: {} diff --git a/src/ontogpt/templates/table_values.py b/src/ontogpt/templates/table_values.py new file mode 100644 index 000000000..e13771b37 --- /dev/null +++ b/src/ontogpt/templates/table_values.py @@ -0,0 +1,285 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "None" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'STE', + 'description': 'An OntoGPT schema for extracting simple numerical tables with ' + 'row/column labels from PDF publications', + 'id': 'https://w3id.org/ontogpt/table_values', + 'imports': ['core'], + 'name': 'SimpleTableExtraction', + 'source_file': 'src/ontogpt/templates/table_values.yaml'} ) + +class NullDataOptions(str, Enum): + UNSPECIFIED_METHOD_OF_ADMINISTRATION = "UNSPECIFIED_METHOD_OF_ADMINISTRATION" + NOT_APPLICABLE = "NOT_APPLICABLE" + NOT_MENTIONED = "NOT_MENTIONED" + + + +class ExtractionResult(ConfiguredBaseModel): + """ + A result of extracting knowledge on text + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + input_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_id', 'domain_of': ['ExtractionResult']} }) + input_title: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_title', 'domain_of': ['ExtractionResult']} }) + input_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_text', 'domain_of': ['ExtractionResult']} }) + raw_completion_output: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'raw_completion_output', 'domain_of': ['ExtractionResult']} }) + prompt: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'prompt', 'domain_of': ['ExtractionResult']} }) + extracted_object: Optional[Any] = Field(default=None, description="""The complex objects extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'extracted_object', 'domain_of': ['ExtractionResult']} }) + named_entities: Optional[List[Any]] = Field(default=None, description="""Named entities extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'named_entities', 'domain_of': ['ExtractionResult']} }) + + +class NamedEntity(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class CompoundExpression(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + pass + + +class Triple(CompoundExpression): + """ + Abstract parent for Relation Extraction tasks + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject', 'domain_of': ['Triple']} }) + predicate: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'predicate', 'domain_of': ['Triple']} }) + object: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object', 'domain_of': ['Triple']} }) + qualifier: Optional[str] = Field(default=None, description="""A qualifier for the statements, e.g. \"NOT\" for negation""", json_schema_extra = { "linkml_meta": {'alias': 'qualifier', 'domain_of': ['Triple']} }) + subject_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the subject of the statement, e.g. \"high dose\" or \"intravenously administered\"""", json_schema_extra = { "linkml_meta": {'alias': 'subject_qualifier', 'domain_of': ['Triple']} }) + object_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the object of the statement, e.g. \"severe\" or \"with additional complications\"""", json_schema_extra = { "linkml_meta": {'alias': 'object_qualifier', 'domain_of': ['Triple']} }) + + +class TextWithTriples(ConfiguredBaseModel): + """ + A text containing one or more relations of the Triple type. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + triples: Optional[List[Triple]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'triples', 'domain_of': ['TextWithTriples']} }) + + +class TextWithEntity(ConfiguredBaseModel): + """ + A text containing one or more instances of a single type of entity. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + entities: Optional[List[str]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'entities', 'domain_of': ['TextWithEntity']} }) + + +class RelationshipType(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core', + 'id_prefixes': ['RO', 'biolink']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Publication(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: Optional[str] = Field(default=None, description="""The publication identifier""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'domain_of': ['NamedEntity', 'Publication']} }) + title: Optional[str] = Field(default=None, description="""The title of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'title', 'domain_of': ['Publication']} }) + abstract: Optional[str] = Field(default=None, description="""The abstract of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'abstract', 'domain_of': ['Publication']} }) + combined_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'combined_text', 'domain_of': ['Publication']} }) + full_text: Optional[str] = Field(default=None, description="""The full text of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'full_text', 'domain_of': ['Publication']} }) + + +class AnnotatorResult(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject_text', 'domain_of': ['AnnotatorResult']} }) + object_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_id', 'domain_of': ['AnnotatorResult']} }) + object_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_text', 'domain_of': ['AnnotatorResult']} }) + + +class Table(ConfiguredBaseModel): + """ + “A table extracted from a publication, containing rows and columns of numerical data.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_values'}) + + table_id: Optional[str] = Field(default=None, description="""“Identifier for the table, e.g. ‘Table 1’, ‘Table 2’, etc.”""", json_schema_extra = { "linkml_meta": {'alias': 'table_id', 'domain_of': ['Table']} }) + source_document: Optional[str] = Field(default=None, description="""“Identifier or reference to the PDF or publication source.”""", json_schema_extra = { "linkml_meta": {'alias': 'source_document', 'domain_of': ['Table']} }) + caption: Optional[str] = Field(default=None, description="""“Text of the table caption or title.”""", json_schema_extra = { "linkml_meta": {'alias': 'caption', 'domain_of': ['Table']} }) + rows: Optional[List[TableRow]] = Field(default=None, description="""“List of row objects in the table.”""", json_schema_extra = { "linkml_meta": {'alias': 'rows', 'domain_of': ['Table']} }) + columns: Optional[List[TableColumn]] = Field(default=None, description="""“List of column objects in the table.”""", json_schema_extra = { "linkml_meta": {'alias': 'columns', 'domain_of': ['Table']} }) + cells: Optional[List[TableCell]] = Field(default=None, description="""“List of cell objects representing the intersection of each row and column.”""", json_schema_extra = { "linkml_meta": {'alias': 'cells', 'domain_of': ['Table']} }) + + +class TableRow(ConfiguredBaseModel): + """ + “A single row in the table, with a label identifying the row.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_values'}) + + row_label: Optional[str] = Field(default=None, description="""“Label describing the row (e.g., a variable, category, subject).”""", json_schema_extra = { "linkml_meta": {'alias': 'row_label', 'domain_of': ['TableRow']} }) + row_index: Optional[int] = Field(default=None, description="""“Numeric index of this row (starting from 0 or 1).”""", json_schema_extra = { "linkml_meta": {'alias': 'row_index', 'domain_of': ['TableRow', 'TableCell']} }) + + +class TableColumn(ConfiguredBaseModel): + """ + “A single column in the table, with a label identifying the column.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_values'}) + + column_label: Optional[str] = Field(default=None, description="""“Label describing the column (e.g., timepoint, parameter).”""", json_schema_extra = { "linkml_meta": {'alias': 'column_label', 'domain_of': ['TableColumn']} }) + column_index: Optional[int] = Field(default=None, description="""“Numeric index of this column (starting from 0 or 1).”""", json_schema_extra = { "linkml_meta": {'alias': 'column_index', 'domain_of': ['TableColumn', 'TableCell']} }) + + +class TableCell(ConfiguredBaseModel): + """ + “An individual cell in the table, containing a numerical value and optional qualifiers.” + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://w3id.org/ontogpt/table_values'}) + + row_index: Optional[int] = Field(default=None, description="""“Numeric index of this row (starting from 0 or 1).”""", json_schema_extra = { "linkml_meta": {'alias': 'row_index', 'domain_of': ['TableRow', 'TableCell']} }) + column_index: Optional[int] = Field(default=None, description="""“Numeric index of this column (starting from 0 or 1).”""", json_schema_extra = { "linkml_meta": {'alias': 'column_index', 'domain_of': ['TableColumn', 'TableCell']} }) + value: Optional[float] = Field(default=None, description="""“Numeric value in this cell.”""", json_schema_extra = { "linkml_meta": {'alias': 'value', 'domain_of': ['TableCell']} }) + unit: Optional[str] = Field(default=None, description="""“Optional unit (e.g., mg, kg, mL).”""", json_schema_extra = { "linkml_meta": {'alias': 'unit', 'domain_of': ['TableCell']} }) + note: Optional[str] = Field(default=None, description="""“Optional note or comment about this particular cell.”""", json_schema_extra = { "linkml_meta": {'alias': 'note', 'domain_of': ['TableCell']} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +ExtractionResult.model_rebuild() +NamedEntity.model_rebuild() +CompoundExpression.model_rebuild() +Triple.model_rebuild() +TextWithTriples.model_rebuild() +TextWithEntity.model_rebuild() +RelationshipType.model_rebuild() +Publication.model_rebuild() +AnnotatorResult.model_rebuild() +Table.model_rebuild() +TableRow.model_rebuild() +TableColumn.model_rebuild() +TableCell.model_rebuild() + diff --git a/src/ontogpt/templates/table_values.yaml b/src/ontogpt/templates/table_values.yaml index 74e6a3751..5448ce30f 100644 --- a/src/ontogpt/templates/table_values.yaml +++ b/src/ontogpt/templates/table_values.yaml @@ -1,3 +1,4 @@ +id: https://w3id.org/ontogpt/table_values name: SimpleTableExtraction description: An OntoGPT schema for extracting simple numerical tables with row/column labels from PDF publications default_prefix: STE @@ -6,7 +7,7 @@ default_prefix: STE # STE: https://w3id.org/your-namespace/ # RO: http://purl.obolibrary.org/obo/RO_ imports: - - ontogpt:core + - core classes: Table: description: “A table extracted from a publication, containing rows and columns of numerical data.” @@ -79,4 +80,3 @@ slots: note: description: “Optional note or comment about this particular cell.” range: string -templates: {} From 0b33b523e2fe86dbe79aba193ef1f2c430d80572 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 10:54:18 -0400 Subject: [PATCH 14/25] Refine organism extraction; add slot for group name --- src/ontogpt/templates/biolog.yaml | 43 +++++++++++++++++++++++++++---- 1 file changed, 38 insertions(+), 5 deletions(-) diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 7ded619a3..0dfa52d99 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -180,7 +180,7 @@ classes: range: string required: false - host: + host_organism: description: "One or more hosts in the experiment." examples: - value: "Brachypodium distachyon" @@ -204,18 +204,36 @@ classes: range: string required: false - target_microbes: - description: "Microbes targeted in the experiment." + target_organisms: + description: "Organisms, including microbes, targeted in the experiment." examples: - value: "Pseudomonas fluorescens" annotations: prompt: | - Extract a list of target microbes studied in - a semicolon-delimited list. + Extract a semicolon-delimited list of target organisms, + including microbes, targeted in the study. + prompt.examples: + Pseudomonas fluorescens; Bacillus subtilis; Pseudomonas simiae WCS 417 range: Microbe multivalued: true required: false + target_organism_group: + description: >- + "A high-level description of organisms, including microbes, + targeted in the experiment." + examples: + - value: "Pseudomonas strains" + annotations: + prompt: | + Extract a semicolon-delimited list of groups of target organisms, + including microbes, targeted in the study. + prompt.examples: + Pseudomonas strains; Pseudomonas group; Roseobacter clade + range: MicrobeGroup + multivalued: true + required: false + biological_system: description: "Biological system under study." examples: @@ -435,6 +453,21 @@ classes: range: string required: false + MicrobeGroup: + is_a: NamedEntity + description: "Information about a microbial group or taxon." + attributes: + name: + description: "Name of a microbial group or taxon." + examples: + - value: "Pseudomonas strains" + annotations: + prompt: | + The name of the microbial group. + prompt.examples: Pseudomonas strains + range: string + required: false + ExperimentalFactor: is_a: NamedEntity description: "A factor manipulated or measured in an experiment." From 815a9760637948588af528320a068873de1a1781 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 11:09:01 -0400 Subject: [PATCH 15/25] Add trait extraction --- src/ontogpt/templates/biolog.py | 131 ++++++++++++++++++++++++++++-- src/ontogpt/templates/biolog.yaml | 28 +++++++ 2 files changed, 151 insertions(+), 8 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 018bfea89..4b161aaef 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -77,7 +77,7 @@ def __contains__(self, key:str) -> bool: 'prefix_reference': 'https://example.org/PaperExtractionSchema/'}, 'rdf': {'prefix_prefix': 'rdf', 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, - 'source_file': '/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml', + 'source_file': 'src/ontogpt/templates/biolog.yaml', 'title': 'Paper Extraction Schema'} ) class NullDataOptions(str, Enum): @@ -374,7 +374,7 @@ class Experiment(ConfiguredBaseModel): 'value': 'Extract the environment or setting.'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Lab-based assay under 25°C'}]} }) - host: Optional[List[str]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host', + host_organism: Optional[List[str]] = Field(default=None, description="""One or more hosts in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'host_organism', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Identify the host organisms or systems ' 'used\n' @@ -390,13 +390,30 @@ class Experiment(ConfiguredBaseModel): "(e.g. 'plant')."}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Plant'}]} }) - target_microbes: Optional[List[str]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', + target_organisms: Optional[List[str]] = Field(default=None, description="""Organisms, including microbes, targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_organisms', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract a list of target microbes ' - 'studied in\n' - 'a semicolon-delimited list.\n'}}, + 'value': 'Extract a semicolon-delimited list of ' + 'target organisms,\n' + 'including microbes, targeted in the ' + 'study. \n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Pseudomonas fluorescens; ' + 'Bacillus subtilis; Pseudomonas ' + 'simiae WCS 417'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) + target_organism_group: Optional[List[str]] = Field(default=None, description="""\"A high-level description of organisms, including microbes, targeted in the experiment.\"""", json_schema_extra = { "linkml_meta": {'alias': 'target_organism_group', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract a semicolon-delimited list of ' + 'groups of target organisms,\n' + 'including microbes, targeted in the ' + 'study.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Pseudomonas strains; ' + 'Pseudomonas group; Roseobacter ' + 'clade'}}, + 'domain_of': ['Experiment'], + 'examples': [{'value': 'Pseudomonas strains'}]} }) biological_system: Optional[str] = Field(default=None, description="""Biological system under study.""", json_schema_extra = { "linkml_meta": {'alias': 'biological_system', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract the focal biological system ' @@ -411,6 +428,15 @@ class Experiment(ConfiguredBaseModel): '7).\n'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'pH 7, 25°C'}]} }) + traits: Optional[List[str]] = Field(default=None, description="""Organismal traits measured or observed.""", json_schema_extra = { "linkml_meta": {'alias': 'traits', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract a semicolon-delimited list of ' + 'traits\n' + 'measured or observed in the study.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Growth rate; Lactose ' + 'production; motility'}}, + 'domain_of': ['Experiment']} }) experimental_factors: Optional[List[str]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'List experimental factors or treatments\n' @@ -543,7 +569,7 @@ class Host(NamedEntity): 'annotations': {'prompt': {'tag': 'prompt', 'value': 'The name of the host organism or ' 'system.\n'}}, - 'domain_of': ['Host', 'Microbe'], + 'domain_of': ['Host', 'Microbe', 'MicrobeGroup'], 'examples': [{'value': 'B. distachyon'}]} }) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -584,7 +610,7 @@ class Microbe(NamedEntity): 'annotations': {'prompt': {'tag': 'prompt', 'value': 'The name of the microbial species or ' 'strain.\n'}}, - 'domain_of': ['Host', 'Microbe'], + 'domain_of': ['Host', 'Microbe', 'MicrobeGroup'], 'examples': [{'value': 'Pseudomonas fluorescens'}]} }) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, @@ -615,6 +641,48 @@ def pattern_original_spans(cls, v): return v +class MicrobeGroup(NamedEntity): + """ + Information about a microbial group or taxon. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + name: Optional[str] = Field(default=None, description="""Name of a microbial group or taxon.""", json_schema_extra = { "linkml_meta": {'alias': 'name', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'The name of the microbial group.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Pseudomonas strains'}}, + 'domain_of': ['Host', 'Microbe', 'MicrobeGroup'], + 'examples': [{'value': 'Pseudomonas strains'}]} }) + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity', 'Experiment'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + class ExperimentalFactor(NamedEntity): """ A factor manipulated or measured in an experiment. @@ -657,6 +725,50 @@ def pattern_original_spans(cls, v): return v +class Trait(NamedEntity): + """ + An organism's physical trait measured or observed in an experiment. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', + 'value': 'sqlite:obo:ecocore, sqlite:obo:pato, ' + 'sqlite:obo:go, sqlite:obo:oba, ' + 'bioportal:biodivthes, ' + 'bioportal:metpo'}, + 'prompt': {'tag': 'prompt', + 'value': 'A physical trait of an organism measured ' + 'or observed in the study.\n'}}, + 'from_schema': 'https://example.org/PaperExtractionSchema', + 'id_prefixes': ['ECOCORE', 'PATO', 'GO', 'OBA', 'BIODIVTHES', 'METPO']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity', 'Experiment'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + # Model rebuild # see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model ExtractionResult.model_rebuild() @@ -672,4 +784,7 @@ def pattern_original_spans(cls, v): Experiment.model_rebuild() Host.model_rebuild() Microbe.model_rebuild() +MicrobeGroup.model_rebuild() ExperimentalFactor.model_rebuild() +Trait.model_rebuild() + diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 0dfa52d99..9ea504b4c 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -256,6 +256,18 @@ classes: range: string required: false + traits: + description: "Organismal traits measured or observed." + annotations: + prompt: | + Extract a semicolon-delimited list of traits + measured or observed in the study. + prompt.examples: + Growth rate; Lactose production; motility + multivalued: true + range: Trait + required: false + experimental_factors: description: "Factors tested or measured." examples: @@ -482,3 +494,19 @@ classes: If absent, 'Not provided'. range: string required: false + + Trait: + is_a: NamedEntity + description: "An organism's physical trait measured or observed in an experiment." + id_prefixes: + - ECOCORE + - PATO + - GO + #- ENVTHES + - OBA + - BIODIVTHES + - METPO + annotations: + prompt: | + A physical trait of an organism measured or observed in the study. + annotators: sqlite:obo:ecocore, sqlite:obo:pato, sqlite:obo:go, sqlite:obo:oba, bioportal:biodivthes, bioportal:metpo \ No newline at end of file From 4f2674eaa0979cb1d7fc29a9cf9004375c89df26 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 11:42:01 -0400 Subject: [PATCH 16/25] Add traits extraction to study and experiment --- src/ontogpt/templates/biolog.py | 14 ++++++++++++-- src/ontogpt/templates/biolog.yaml | 14 +++++++++++++- 2 files changed, 25 insertions(+), 3 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 4b161aaef..f296e9db0 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -284,6 +284,15 @@ class Paper(ConfiguredBaseModel): "found, use 'Not provided'."}}, 'domain_of': ['Paper'], 'examples': [{'value': '14 April 2021'}]} }) + traits: Optional[List[str]] = Field(default=None, description="""Organismal traits measured or observed.""", json_schema_extra = { "linkml_meta": {'alias': 'traits', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract a semicolon-delimited list of ' + 'traits\n' + 'focused on in the study.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Growth rate; Lactose ' + 'production; motility'}}, + 'domain_of': ['Paper', 'Experiment']} }) experiments: List[Experiment] = Field(default=..., description="""Experiments (e.g., Biolog) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a list of brief descriptions of ' @@ -432,11 +441,12 @@ class Experiment(ConfiguredBaseModel): 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract a semicolon-delimited list of ' 'traits\n' - 'measured or observed in the study.\n'}, + 'measured or observed in the ' + 'experiment.\n'}, 'prompt.examples': {'tag': 'prompt.examples', 'value': 'Growth rate; Lactose ' 'production; motility'}}, - 'domain_of': ['Experiment']} }) + 'domain_of': ['Paper', 'Experiment']} }) experimental_factors: Optional[List[str]] = Field(default=None, description="""Factors tested or measured.""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'List experimental factors or treatments\n' diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 9ea504b4c..110bd0fbe 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -88,6 +88,18 @@ classes: range: string required: false + traits: + description: "Organismal traits measured or observed." + annotations: + prompt: | + Extract a semicolon-delimited list of traits + focused on in the study. + prompt.examples: + Growth rate; Lactose production; motility + multivalued: true + range: Trait + required: false + experiments: description: "Experiments (e.g., Biolog) included in the paper." examples: @@ -261,7 +273,7 @@ classes: annotations: prompt: | Extract a semicolon-delimited list of traits - measured or observed in the study. + measured or observed in the experiment. prompt.examples: Growth rate; Lactose production; motility multivalued: true From 014c1d5b832aba1a739db84bba0c0405ec5f582b Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 11:49:54 -0400 Subject: [PATCH 17/25] Adjust how METPO is used to annotate traits --- src/ontogpt/templates/biolog.py | 7 ++----- src/ontogpt/templates/biolog.yaml | 17 +++++++++-------- 2 files changed, 11 insertions(+), 13 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index f296e9db0..f8237faad 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -740,15 +740,12 @@ class Trait(NamedEntity): An organism's physical trait measured or observed in an experiment. """ linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', - 'value': 'sqlite:obo:ecocore, sqlite:obo:pato, ' - 'sqlite:obo:go, sqlite:obo:oba, ' - 'bioportal:biodivthes, ' - 'bioportal:metpo'}, + 'value': 'bioportal:metpo'}, 'prompt': {'tag': 'prompt', 'value': 'A physical trait of an organism measured ' 'or observed in the study.\n'}}, 'from_schema': 'https://example.org/PaperExtractionSchema', - 'id_prefixes': ['ECOCORE', 'PATO', 'GO', 'OBA', 'BIODIVTHES', 'METPO']}) + 'id_prefixes': ['https']}) id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index 110bd0fbe..a7b2c26de 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -511,14 +511,15 @@ classes: is_a: NamedEntity description: "An organism's physical trait measured or observed in an experiment." id_prefixes: - - ECOCORE - - PATO - - GO - #- ENVTHES - - OBA - - BIODIVTHES - - METPO + # - ECOCORE + # - PATO + # - GO + # - ENVTHES + # - OBA + # - BIODIVTHES + - https annotations: prompt: | A physical trait of an organism measured or observed in the study. - annotators: sqlite:obo:ecocore, sqlite:obo:pato, sqlite:obo:go, sqlite:obo:oba, bioportal:biodivthes, bioportal:metpo \ No newline at end of file + # annotators: sqlite:obo:ecocore, sqlite:obo:pato, sqlite:obo:go, sqlite:obo:oba, bioportal:biodivthes, bioportal:metpo + annotators: bioportal:metpo \ No newline at end of file From 5d8ee7d30ad165e7fd9ca2019fd5790c7da65a64 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 11:53:43 -0400 Subject: [PATCH 18/25] Prompt tuning in plate name extraction --- src/ontogpt/templates/biolog.py | 20 ++++++++++++++------ src/ontogpt/templates/biolog.yaml | 11 +++++++---- 2 files changed, 21 insertions(+), 10 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index f8237faad..4e74e8292 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -470,13 +470,18 @@ class Experiment(ConfiguredBaseModel): 'examples': [{'value': 'Phenotype MicroArray'}]} }) plates: Optional[List[str]] = Field(default=None, description="""Biolog plates used (e.g., EcoPlate, GEN III, PM01, PM02A).""", json_schema_extra = { "linkml_meta": {'alias': 'plates', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the name(s) of any Biolog ' - 'plates\n' - '(e.g., PM01, PM02A) in a ' + 'value': 'Extract the name(s) of any phenotype ' + 'microarray plates\n' + '(also known as Biolog plates, 96 well ' + 'plates, microarrays,\n' + 'microarray plates, or microplates) in a ' 'semicolon-delimited list.\n' + 'Names may be in the format PM01, PM02A, ' + 'etc.\n' "If not mentioned, 'Not provided'.\n"}, 'prompt.examples': {'tag': 'prompt.examples', - 'value': 'PM01; PM02A'}}, + 'value': 'PM01; PM02A; EcoPlate, GEN ' + 'III'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'PM01'}, {'value': 'PM02A'}]} }) replicates: Optional[str] = Field(default=None, description="""Number of replicates.""", json_schema_extra = { "linkml_meta": {'alias': 'replicates', @@ -535,13 +540,16 @@ class Experiment(ConfiguredBaseModel): 'domain_of': ['Experiment'], 'examples': [{'value': 'Colorimetric change at 590 nm for formazan ' 'detection'}]} }) - instrument_used: Optional[List[str]] = Field(default=None, description="""Instrument or equipment used for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'instrument_used', + instrument_used: Optional[List[str]] = Field(default=None, description="""Instruments or equipment used for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'instrument_used', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract the name(s) of the ' 'instrument/equipment used\n' '(e.g., OmniLog system) in a ' 'semicolon-delimited list.\n' - "If none, 'Not provided'.\n"}}, + "If none, 'Not provided'.\n"}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'OmniLog Phenotype MicroArray ' + 'System; OmniLog Turbidimeter'}}, 'domain_of': ['Experiment'], 'examples': [{'value': 'OmniLog Phenotype MicroArray System (Biolog)'}]} }) analysis_software: Optional[List[str]] = Field(default=None, description="""Software used to analyze or interpret data.""", json_schema_extra = { "linkml_meta": {'alias': 'analysis_software', diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index a7b2c26de..e1fc1c508 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -322,10 +322,12 @@ classes: - value: "PM02A" annotations: prompt: | - Extract the name(s) of any Biolog plates - (e.g., PM01, PM02A) in a semicolon-delimited list. + Extract the name(s) of any phenotype microarray plates + (also known as Biolog plates, 96 well plates, microarrays, + microarray plates, or microplates) in a semicolon-delimited list. + Names may be in the format PM01, PM02A, etc. If not mentioned, 'Not provided'. - prompt.examples: PM01; PM02A + prompt.examples: PM01; PM02A; EcoPlate, GEN III range: string multivalued: true required: false @@ -402,7 +404,7 @@ classes: required: false instrument_used: - description: "Instrument or equipment used for the experiment." + description: "Instruments or equipment used for the experiment." examples: - value: "OmniLog Phenotype MicroArray System (Biolog)" annotations: @@ -410,6 +412,7 @@ classes: Extract the name(s) of the instrument/equipment used (e.g., OmniLog system) in a semicolon-delimited list. If none, 'Not provided'. + prompt.examples: OmniLog Phenotype MicroArray System; OmniLog Turbidimeter range: string multivalued: true required: false From b115fb38c68945179fd9e59b2a1e1dca04b6f362 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Wed, 19 Mar 2025 12:03:23 -0400 Subject: [PATCH 19/25] Expand protocol step extraction; move culture conditions into step --- src/ontogpt/templates/biolog.py | 66 +++++++++++++++++++++------ src/ontogpt/templates/biolog.yaml | 76 ++++++++++++++++++++++--------- 2 files changed, 107 insertions(+), 35 deletions(-) diff --git a/src/ontogpt/templates/biolog.py b/src/ontogpt/templates/biolog.py index 4e74e8292..4a26403a2 100644 --- a/src/ontogpt/templates/biolog.py +++ b/src/ontogpt/templates/biolog.py @@ -491,7 +491,7 @@ class Experiment(ConfiguredBaseModel): "If not mentioned, 'Not provided'.\n"}}, 'domain_of': ['Experiment'], 'examples': [{'value': '3'}]} }) - protocol_steps: Optional[List[str]] = Field(default=None, description="""Key steps in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'protocol_steps', + protocol_steps: Optional[List[ProtocolStep]] = Field(default=None, description="""Key steps in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'protocol_steps', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'List the full protocol steps or key ' 'experimental procedures,\n' @@ -563,18 +563,6 @@ class Experiment(ConfiguredBaseModel): 'domain_of': ['Experiment'], 'examples': [{'value': 'OmniLog Parametric Analysis software v1.20.02 ' '(Biolog)'}]} }) - incubation_temperature: Optional[str] = Field(default=None, description="""Temperature used for incubating the plates or cultures.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_temperature', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the incubation temperature if ' - 'provided (e.g. 25°C).\n'}}, - 'domain_of': ['Experiment'], - 'examples': [{'value': '25°C'}]} }) - incubation_duration: Optional[str] = Field(default=None, description="""Total duration of incubation (e.g., 48 h).""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_duration', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract how long the incubation lasted ' - '(e.g. 48 hours).\n'}}, - 'domain_of': ['Experiment'], - 'examples': [{'value': '48 h'}]} }) class Host(NamedEntity): @@ -784,6 +772,57 @@ def pattern_original_spans(cls, v): return v +class ProtocolStep(ConfiguredBaseModel): + """ + A step in an experimental protocol. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) + + step_description: Optional[str] = Field(default=None, description="""A step in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'step_description', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Description of a step in the ' + 'experimental protocol.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '1) Grow cells overnight on LB ' + 'agar'}}, + 'domain_of': ['ProtocolStep'], + 'examples': [{'value': 'Grow cells overnight on LB agar'}]} }) + culture_media: Optional[str] = Field(default=None, description="""Culture media used in the step.""", json_schema_extra = { "linkml_meta": {'alias': 'culture_media', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'The culture media used in the step,\n' + 'if applicable. Do not include a value\n' + 'for this field if not relevant.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'LB agar'}}, + 'domain_of': ['ProtocolStep'], + 'examples': [{'value': 'LB agar'}]} }) + culture_temperature: Optional[str] = Field(default=None, description="""Culture incubation temperature used in the step.""", json_schema_extra = { "linkml_meta": {'alias': 'culture_temperature', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'The culture conditions used in the ' + 'step,\n' + 'if applicable, including temperature.\n' + 'Do not include a value for this field ' + 'if\n' + 'not relevant.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '25°C'}}, + 'domain_of': ['ProtocolStep'], + 'examples': [{'value': '25°C'}]} }) + culture_duration: Optional[str] = Field(default=None, description="""Culture incubation duration used in the step.""", json_schema_extra = { "linkml_meta": {'alias': 'culture_duration', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'The culture incubation duration used in ' + 'the step,\n' + 'if applicable, expressed as value and ' + 'unit.\n' + 'Do not include a value for this field ' + 'if\n' + 'not relevant.\n'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '48 h'}}, + 'domain_of': ['ProtocolStep'], + 'examples': [{'value': '48 h'}]} }) + + # Model rebuild # see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model ExtractionResult.model_rebuild() @@ -802,4 +841,5 @@ def pattern_original_spans(cls, v): MicrobeGroup.model_rebuild() ExperimentalFactor.model_rebuild() Trait.model_rebuild() +ProtocolStep.model_rebuild() diff --git a/src/ontogpt/templates/biolog.yaml b/src/ontogpt/templates/biolog.yaml index e1fc1c508..85d04f988 100644 --- a/src/ontogpt/templates/biolog.yaml +++ b/src/ontogpt/templates/biolog.yaml @@ -364,7 +364,7 @@ classes: Do not include newlines. prompt.examples: 1) Grow cells overnight on LB agar; 2) Harvest and suspend in IF-0 - range: string + range: ProtocolStep multivalued: true required: false @@ -430,26 +430,6 @@ classes: multivalued: true required: false - incubation_temperature: - description: "Temperature used for incubating the plates or cultures." - examples: - - value: "25°C" - annotations: - prompt: | - Extract the incubation temperature if provided (e.g. 25°C). - range: string - required: false - - incubation_duration: - description: "Total duration of incubation (e.g., 48 h)." - examples: - - value: "48 h" - annotations: - prompt: | - Extract how long the incubation lasted (e.g. 48 hours). - range: string - required: false - # Placeholder, to be expanded Host: is_a: NamedEntity @@ -525,4 +505,56 @@ classes: prompt: | A physical trait of an organism measured or observed in the study. # annotators: sqlite:obo:ecocore, sqlite:obo:pato, sqlite:obo:go, sqlite:obo:oba, bioportal:biodivthes, bioportal:metpo - annotators: bioportal:metpo \ No newline at end of file + annotators: bioportal:metpo + + ProtocolStep: + description: "A step in an experimental protocol." + attributes: + step_description: + description: "A step in the experimental protocol." + examples: + - value: "Grow cells overnight on LB agar" + annotations: + prompt: | + Description of a step in the experimental protocol. + prompt.examples: 1) Grow cells overnight on LB agar + range: string + required: false + culture_media: + description: "Culture media used in the step." + examples: + - value: "LB agar" + annotations: + prompt: | + The culture media used in the step, + if applicable. Do not include a value + for this field if not relevant. + prompt.examples: LB agar + range: string + required: false + culture_temperature: + description: "Culture incubation temperature used in the step." + examples: + - value: "25°C" + annotations: + prompt: | + The culture conditions used in the step, + if applicable, including temperature. + Do not include a value for this field if + not relevant. + prompt.examples: 25°C + range: string + required: false + culture_duration: + description: "Culture incubation duration used in the step." + examples: + - value: "48 h" + annotations: + prompt: | + The culture incubation duration used in the step, + if applicable, expressed as value and unit. + Do not include a value for this field if + not relevant. + prompt.examples: 48 h + range: string + required: false \ No newline at end of file From 9c685e137dd5c79d57275301b0258aa147a3c486 Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Mon, 24 Mar 2025 18:26:27 -0700 Subject: [PATCH 20/25] growth experiment --- src/ontogpt/templates/growth_experiment.yaml | 244 +++++++++++++++++++ 1 file changed, 244 insertions(+) create mode 100644 src/ontogpt/templates/growth_experiment.yaml diff --git a/src/ontogpt/templates/growth_experiment.yaml b/src/ontogpt/templates/growth_experiment.yaml new file mode 100644 index 000000000..9d2595e0e --- /dev/null +++ b/src/ontogpt/templates/growth_experiment.yaml @@ -0,0 +1,244 @@ +id: https://example.org/MicrobialCulturingExtraction +name: MicrobialCulturingExtraction +title: "Microbial Culturing Extraction Schema" +description: > + A more sophisticated schema for extracting structured data about microbial + liquid culturing media and growth conditions from research articles. +license: https://creativecommons.org/publicdomain/zero/1.0/ +version: 0.1.0 + +prefixes: + rdf: "http://www.w3.org/1999/02/22-rdf-syntax-ns#" + linkml: "https://w3id.org/linkml/" + microex: "https://example.org/MicrobialCulturingExtraction/" + +default_prefix: microex +default_range: string + +imports: + - linkml:types + - core + +######################## +# ENUMS +######################## + +enums: + # If needed, reuse this flexible boolean enumeration + FlexibleBooleanEnum: + description: > + A minimal enumeration for capturing yes/no/true/false in lowercase, + plus 'not provided' if data are missing or ambiguous. + permissible_values: + 'yes': {} + 'no': {} + 'true': {} + 'false': {} + 'not provided': {} + +######################## +# CLASSES +######################## + +classes: + + MicrobialCulturing: + tree_root: true + description: "Container for microbial culturing information extracted from text." + attributes: + media: + description: "One or more culture media used in the study." + range: Medium + multivalued: true + required: false + annotations: + prompt: >- + List each distinct culture medium described in the text, extracting + details into the Medium class. Separate multiple media with semicolons + if they are clearly distinct. + prompt.examples: "LB medium; M9 minimal medium" + + growth_experiments: + description: "One or more growth experiments or conditions described in the study." + range: GrowthExperiment + multivalued: true + required: false + annotations: + prompt: >- + Summarize each distinct growth experiment or set of conditions + into the GrowthExperiment class. If multiple growth experiments + are described, separate them accordingly. + prompt.examples: "Aerobic growth at 30°C; Anaerobic growth in sealed flasks at 37°C" + + + Medium: + description: "Details about a liquid culturing medium." + attributes: + identifier: + description: "A name or code used to identify the medium (e.g., LB, M9, etc.)." + examples: + - value: "LB" + annotations: + prompt: "Extract the main name or code for this medium." + prompt.examples: "LB; M9" + range: string + required: false + + label: + description: "A human-readable label or short name for the medium." + examples: + - value: "Lysogeny Broth" + annotations: + prompt: "Provide the full or commonly used label for the medium." + prompt.examples: "Lysogeny Broth" + range: string + required: false + + ingredients: + description: "Ingredients of the medium, including chemical compounds and concentrations." + examples: + - value: "Tryptone 10 g/L, Yeast Extract 5 g/L, NaCl 10 g/L" + annotations: + prompt: >- + Provide all ingredients used in the medium, including concentrations + where specified. If multiple ingredients, separate them with semicolons. + prompt.examples: "Tryptone 10 g/L; Yeast Extract 5 g/L; NaCl 10 g/L" + range: string + required: false + + modifications: + description: "Any modifications or special additives to the standard medium recipe." + examples: + - value: "Adjusted NaCl to 5 g/L; Added 2 g/L glucose" + annotations: + prompt: "Describe any deviations from the standard recipe." + prompt.examples: "Extra glucose; Reduced NaCl" + range: string + required: false + + preparation_steps: + description: "Description of how the medium is prepared." + examples: + - value: "Sterilized by autoclaving at 121°C for 15 minutes; pH adjusted to 7.0" + annotations: + prompt: >- + Outline the preparation steps for the medium, including sterilization, + pH adjustment, or other relevant details. + prompt.examples: "Autoclave at 121°C for 15 minutes; Adjust pH to 7.0" + range: string + required: false + + + GrowthExperiment: + description: "Growth conditions and experimental methods for culturing." + attributes: + temperature: + description: "Temperature(s) at which the culture is grown (e.g., in Celsius)." + examples: + - value: "30°C" + annotations: + prompt: "Specify the growth temperature(s). If multiple, separate with semicolons." + prompt.examples: "30°C; 37°C" + range: string + required: false + + salinity: + description: "Salinity or salt concentration conditions (e.g., % NaCl)." + examples: + - value: "3% NaCl" + annotations: + prompt: >- + Provide salinity or salt concentration if given. If multiple or a range, + separate with semicolons. + prompt.examples: "3% NaCl; 2 M NaCl" + range: string + required: false + + pH: + description: "pH level at which the culture is maintained." + examples: + - value: "pH 7.0" + annotations: + prompt: "Provide the pH level for the culture (e.g., pH 7.0)." + prompt.examples: "pH 7.2" + range: string + required: false + + oxygen_status: + description: "Oxygen conditions (e.g., aerobic, anaerobic, microaerophilic)." + examples: + - value: "Aerobic" + - value: "Anaerobic" + annotations: + prompt: >- + Indicate whether the culture is aerobic, anaerobic, microaerophilic, + or specify an oxygen concentration or partial pressure. + prompt.examples: "Aerobic; Microaerophilic; ~5% O2" + range: string + required: false + + sparging_equipment: + description: "Equipment or method used for sparging or aeration (if any)." + examples: + - value: "Air sparger with 0.2 μm filter at 1 L/min" + annotations: + prompt: "Provide details on aeration equipment or methods." + prompt.examples: "Air sparger at 1 L/min" + range: string + required: false + + inoculation_method: + description: "Method used to inoculate the culture (e.g., inoculum source or process)." + examples: + - value: "Inoculated with a single colony from a streak plate" + annotations: + prompt: "Describe how the culture was inoculated, including inoculum origin and technique." + prompt.examples: "Single colony from LB agar plate; 1% v/v seed culture" + range: string + required: false + + mixing_method: + description: "Method of agitation or mixing (e.g., stirring speed, shaker, etc.)." + examples: + - value: "Shaken at 200 rpm on an orbital shaker" + annotations: + prompt: "Provide details on how the culture was mixed, if at all." + prompt.examples: "200 rpm on orbital shaker" + range: string + required: false + + culturing_duration: + description: "Duration of the culture growth or experiment." + examples: + - value: "24 hours" + - value: "48 h" + annotations: + prompt: >- + State how long the culture was grown or observed. Include time units + (e.g., 24 h, 48 h). + prompt.examples: "24 h; 48 h" + range: string + required: false + + light_method: + description: "Type or intensity of light provided (if relevant) for phototrophic organisms." + examples: + - value: "Continuous light at 50 μmol photons m⁻² s⁻¹" + annotations: + prompt: >- + Provide the type of lighting (e.g., wavelength, intensity) or indicate + dark conditions if relevant. + prompt.examples: "Continuous white light at 50 μmol photons m⁻² s⁻¹" + range: string + required: false + + other_methods: + description: "Any other relevant growth or handling methods not captured above." + examples: + - value: "Maintained under 1 atm CO2; Pressure of 10 MPa" + annotations: + prompt: "Describe any additional methods or conditions (e.g., special additives, pressure)." + prompt.examples: "10% CO2 atmosphere; High pressure environment" + range: string + required: false + From 53513e0c9fe596a9da7df160aa45aa465b0a6304 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 25 Mar 2025 11:09:39 -0400 Subject: [PATCH 21/25] Schema tuning --- src/ontogpt/templates/growth_experiment.py | 457 +++++++++++++++++++ src/ontogpt/templates/growth_experiment.yaml | 78 +++- 2 files changed, 512 insertions(+), 23 deletions(-) create mode 100644 src/ontogpt/templates/growth_experiment.py diff --git a/src/ontogpt/templates/growth_experiment.py b/src/ontogpt/templates/growth_experiment.py new file mode 100644 index 000000000..d904874b8 --- /dev/null +++ b/src/ontogpt/templates/growth_experiment.py @@ -0,0 +1,457 @@ +from __future__ import annotations + +import re +import sys +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) + + +metamodel_version = "None" +version = "0.1.0" + + +class ConfiguredBaseModel(BaseModel): + model_config = ConfigDict( + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) + pass + + + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'microex', + 'default_range': 'string', + 'description': 'A more sophisticated schema for extracting structured data ' + 'about microbial liquid culturing media and growth conditions ' + 'from research articles.\n', + 'id': 'https://example.org/MicrobialCulturingExtraction', + 'imports': ['linkml:types', 'core'], + 'license': 'https://creativecommons.org/publicdomain/zero/1.0/', + 'name': 'MicrobialCulturingExtraction', + 'prefixes': {'linkml': {'prefix_prefix': 'linkml', + 'prefix_reference': 'https://w3id.org/linkml/'}, + 'microex': {'prefix_prefix': 'microex', + 'prefix_reference': 'https://example.org/MicrobialCulturingExtraction/'}, + 'rdf': {'prefix_prefix': 'rdf', + 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, + 'source_file': 'src/ontogpt/templates/growth_experiment.yaml', + 'title': 'Microbial Culturing Extraction Schema'} ) + +class NullDataOptions(str, Enum): + UNSPECIFIED_METHOD_OF_ADMINISTRATION = "UNSPECIFIED_METHOD_OF_ADMINISTRATION" + NOT_APPLICABLE = "NOT_APPLICABLE" + NOT_MENTIONED = "NOT_MENTIONED" + + +class FlexibleBooleanEnum(str, Enum): + """ + A minimal enumeration for capturing yes/no/true/false in lowercase, plus 'not provided' if data are missing or ambiguous. + + """ + yes = "yes" + no = "no" + true = "true" + false = "false" + not_provided = "not provided" + + + +class ExtractionResult(ConfiguredBaseModel): + """ + A result of extracting knowledge on text + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + input_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_id', 'domain_of': ['ExtractionResult']} }) + input_title: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_title', 'domain_of': ['ExtractionResult']} }) + input_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_text', 'domain_of': ['ExtractionResult']} }) + raw_completion_output: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'raw_completion_output', 'domain_of': ['ExtractionResult']} }) + prompt: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'prompt', 'domain_of': ['ExtractionResult']} }) + extracted_object: Optional[Any] = Field(default=None, description="""The complex objects extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'extracted_object', 'domain_of': ['ExtractionResult']} }) + named_entities: Optional[List[Any]] = Field(default=None, description="""Named entities extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'named_entities', 'domain_of': ['ExtractionResult']} }) + + +class NamedEntity(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity', 'Medium'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class CompoundExpression(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + pass + + +class Triple(CompoundExpression): + """ + Abstract parent for Relation Extraction tasks + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject', 'domain_of': ['Triple']} }) + predicate: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'predicate', 'domain_of': ['Triple']} }) + object: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object', 'domain_of': ['Triple']} }) + qualifier: Optional[str] = Field(default=None, description="""A qualifier for the statements, e.g. \"NOT\" for negation""", json_schema_extra = { "linkml_meta": {'alias': 'qualifier', 'domain_of': ['Triple']} }) + subject_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the subject of the statement, e.g. \"high dose\" or \"intravenously administered\"""", json_schema_extra = { "linkml_meta": {'alias': 'subject_qualifier', 'domain_of': ['Triple']} }) + object_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the object of the statement, e.g. \"severe\" or \"with additional complications\"""", json_schema_extra = { "linkml_meta": {'alias': 'object_qualifier', 'domain_of': ['Triple']} }) + + +class TextWithTriples(ConfiguredBaseModel): + """ + A text containing one or more relations of the Triple type. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + triples: Optional[List[Triple]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'triples', 'domain_of': ['TextWithTriples']} }) + + +class TextWithEntity(ConfiguredBaseModel): + """ + A text containing one or more instances of a single type of entity. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + entities: Optional[List[str]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'entities', 'domain_of': ['TextWithEntity']} }) + + +class RelationshipType(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core', + 'id_prefixes': ['RO', 'biolink']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity', 'Medium'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Publication(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: Optional[str] = Field(default=None, description="""The publication identifier""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'domain_of': ['NamedEntity', 'Publication']} }) + title: Optional[str] = Field(default=None, description="""The title of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'title', 'domain_of': ['Publication']} }) + abstract: Optional[str] = Field(default=None, description="""The abstract of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'abstract', 'domain_of': ['Publication']} }) + combined_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'combined_text', 'domain_of': ['Publication']} }) + full_text: Optional[str] = Field(default=None, description="""The full text of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'full_text', 'domain_of': ['Publication']} }) + + +class AnnotatorResult(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject_text', 'domain_of': ['AnnotatorResult']} }) + object_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_id', 'domain_of': ['AnnotatorResult']} }) + object_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_text', 'domain_of': ['AnnotatorResult']} }) + + +class MicrobialCulturing(ConfiguredBaseModel): + """ + Container for microbial culturing information extracted from text. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/MicrobialCulturingExtraction', + 'tree_root': True}) + + media: Optional[List[Medium]] = Field(default=None, description="""One or more culture media used in the study.""", json_schema_extra = { "linkml_meta": {'alias': 'media', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'List each distinct culture medium ' + 'described in the text. Separate multiple ' + 'media with semicolons if they are ' + 'clearly distinct. For each, include the ' + 'main name or code, ingredients, any ' + 'modifications, and preparation steps.'}}, + 'domain_of': ['MicrobialCulturing']} }) + growth_experiments: Optional[List[GrowthExperiment]] = Field(default=None, description="""One or more growth experiments or conditions described in the study.""", json_schema_extra = { "linkml_meta": {'alias': 'growth_experiments', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Summarize each distinct growth ' + 'experiment or set of conditions. If ' + 'multiple growth experiments are ' + 'described, separate them accordingly in ' + 'a semicolon-delimited list with no ' + 'numbering and no newlines. For each, ' + 'include temperature, salinity, pH, ' + 'oxygen status, sparging equipment, ' + 'inoculation method, mixing method, ' + 'culturing duration, light method (if ' + 'relevant), and any other methods or ' + 'conditions.'}}, + 'domain_of': ['MicrobialCulturing']} }) + + +class Medium(ConfiguredBaseModel): + """ + Details about a liquid culturing medium. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/MicrobialCulturingExtraction'}) + + identifier: Optional[str] = Field(default=None, description="""A name or code used to identify the medium (e.g., LB, M9, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'identifier', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Extract the main name or code for this ' + 'medium.'}}, + 'domain_of': ['Medium'], + 'examples': [{'value': 'LB'}]} }) + label: Optional[str] = Field(default=None, description="""A human-readable label or short name for the medium.""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide the full or commonly used label ' + 'for the medium.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Lysogeny Broth'}}, + 'domain_of': ['NamedEntity', 'Medium'], + 'examples': [{'value': 'Lysogeny Broth'}]} }) + ingredients: Optional[List[str]] = Field(default=None, description="""Ingredients of the medium, including chemical compounds and concentrations.""", json_schema_extra = { "linkml_meta": {'alias': 'ingredients', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide all ingredients used in the ' + 'medium, including concentrations where ' + 'specified. If multiple ingredients, ' + 'separate them with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Tryptone 10 g/L; Yeast Extract ' + '5 g/L; NaCl 10 g/L'}}, + 'domain_of': ['Medium'], + 'examples': [{'value': 'Tryptone 10 g/L, Yeast Extract 5 g/L, NaCl 10 g/L'}]} }) + modifications: Optional[List[str]] = Field(default=None, description="""Any modifications or special additives to the standard medium recipe.""", json_schema_extra = { "linkml_meta": {'alias': 'modifications', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe any deviations from the ' + 'standard recipe. If multiple ' + 'modifications, separate them with ' + 'semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Extra glucose; Reduced NaCl'}}, + 'domain_of': ['Medium'], + 'examples': [{'value': 'Adjusted NaCl to 5 g/L; Added 2 g/L glucose'}]} }) + preparation_steps: Optional[List[str]] = Field(default=None, description="""Description of how the medium is prepared.""", json_schema_extra = { "linkml_meta": {'alias': 'preparation_steps', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Outline the preparation steps for the ' + 'medium, including sterilization, pH ' + 'adjustment, or other relevant details. ' + 'If multiple steps, separate them with ' + 'semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Autoclave at 121°C for 15 ' + 'minutes; Adjust pH to 7.0'}}, + 'domain_of': ['Medium'], + 'examples': [{'value': 'Sterilized by autoclaving at 121°C for 15 minutes; pH ' + 'adjusted to 7.0'}]} }) + + +class GrowthExperiment(ConfiguredBaseModel): + """ + Growth conditions and experimental methods for culturing. + """ + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/MicrobialCulturingExtraction'}) + + temperature: Optional[List[str]] = Field(default=None, description="""Temperature(s) at which the culture is grown (e.g., in Celsius).""", json_schema_extra = { "linkml_meta": {'alias': 'temperature', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Specify the growth temperature(s). If ' + 'multiple, separate with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '30°C; 37°C'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': '30°C'}]} }) + salinity: Optional[List[str]] = Field(default=None, description="""Salinity or salt concentration conditions (e.g., % NaCl).""", json_schema_extra = { "linkml_meta": {'alias': 'salinity', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide salinity or salt concentration ' + 'if given. If multiple or a range, ' + 'separate with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '3% NaCl; 2 M NaCl'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': '3% NaCl'}]} }) + pH: Optional[List[str]] = Field(default=None, description="""pH level at which the culture is maintained.""", json_schema_extra = { "linkml_meta": {'alias': 'pH', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide the pH level for the culture ' + '(e.g., pH 7.0). If multiple or a range, ' + 'separate with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'pH 7.2'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'pH 7.0'}]} }) + oxygen_status: Optional[List[str]] = Field(default=None, description="""Oxygen conditions (e.g., aerobic, anaerobic, microaerophilic).""", json_schema_extra = { "linkml_meta": {'alias': 'oxygen_status', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Indicate whether the culture is aerobic, ' + 'anaerobic, microaerophilic, or specify ' + 'an oxygen concentration or partial ' + 'pressure. If multiple, separate with ' + 'semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Aerobic; Microaerophilic; ~5% ' + 'O2'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'Aerobic'}, {'value': 'Anaerobic'}]} }) + sparging_equipment: Optional[List[str]] = Field(default=None, description="""Equipment or method used for sparging or aeration (if any).""", json_schema_extra = { "linkml_meta": {'alias': 'sparging_equipment', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide details on aeration equipment or ' + 'methods. Include flow rate, filter size, ' + 'or other relevant information. If ' + 'multiple, separate with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Air sparger at 1 L/min'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'Air sparger with 0.2 μm filter at 1 L/min'}]} }) + inoculation_method: Optional[List[str]] = Field(default=None, description="""Method(s) used to inoculate the culture (e.g., inoculum source or process).""", json_schema_extra = { "linkml_meta": {'alias': 'inoculation_method', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe how the culture was inoculated, ' + 'including inoculum origin and technique. ' + 'If multiple methods, separate with ' + 'semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Single colony from LB agar ' + 'plate; 1% v/v seed culture'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'Inoculated with a single colony from a streak plate'}]} }) + mixing_method: Optional[List[str]] = Field(default=None, description="""Method of agitation or mixing (e.g., stirring speed, shaker, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'mixing_method', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide details on how the culture was ' + 'mixed, if at all. If multiple methods, ' + 'separate with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '200 rpm on orbital shaker'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'Shaken at 200 rpm on an orbital shaker'}]} }) + culturing_duration: Optional[List[str]] = Field(default=None, description="""Duration of the culture growth or experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'culturing_duration', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'State how long the culture was grown or ' + 'observed. Include time units (e.g., 24 ' + 'h, 48 h). If multiple durations, ' + 'separate with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '24 h; 48 h'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': '24 hours'}, {'value': '48 h'}]} }) + light_method: Optional[List[str]] = Field(default=None, description="""Type or intensity of light provided (if relevant) for phototrophic organisms.""", json_schema_extra = { "linkml_meta": {'alias': 'light_method', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide the type of lighting (e.g., ' + 'wavelength, intensity) or indicate dark ' + 'conditions if relevant. If multiple ' + 'light conditions, separate with ' + 'semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Continuous white light at 50 ' + 'μmol photons m⁻² s⁻¹'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'Continuous light at 50 μmol photons m⁻² s⁻¹'}]} }) + other_methods: Optional[List[str]] = Field(default=None, description="""Any other relevant growth or handling methods not captured above.""", json_schema_extra = { "linkml_meta": {'alias': 'other_methods', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Describe any additional methods or ' + 'conditions (e.g., special additives, ' + 'pressure). If multiple methods, separate ' + 'with semicolons.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': '10% CO2 atmosphere; High ' + 'pressure environment'}}, + 'domain_of': ['GrowthExperiment'], + 'examples': [{'value': 'Maintained under 1 atm CO2; Pressure of 10 MPa'}]} }) + + +# Model rebuild +# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model +ExtractionResult.model_rebuild() +NamedEntity.model_rebuild() +CompoundExpression.model_rebuild() +Triple.model_rebuild() +TextWithTriples.model_rebuild() +TextWithEntity.model_rebuild() +RelationshipType.model_rebuild() +Publication.model_rebuild() +AnnotatorResult.model_rebuild() +MicrobialCulturing.model_rebuild() +Medium.model_rebuild() +GrowthExperiment.model_rebuild() + diff --git a/src/ontogpt/templates/growth_experiment.yaml b/src/ontogpt/templates/growth_experiment.yaml index 9d2595e0e..f541a90eb 100644 --- a/src/ontogpt/templates/growth_experiment.yaml +++ b/src/ontogpt/templates/growth_experiment.yaml @@ -53,10 +53,11 @@ classes: required: false annotations: prompt: >- - List each distinct culture medium described in the text, extracting - details into the Medium class. Separate multiple media with semicolons - if they are clearly distinct. - prompt.examples: "LB medium; M9 minimal medium" + List each distinct culture medium described in the text. + Separate multiple media with semicolons if they are clearly + distinct. For each, include the main name or code, + ingredients, any modifications, and preparation steps. + # prompt.examples: "LB medium; M9 minimal medium" growth_experiments: description: "One or more growth experiments or conditions described in the study." @@ -65,10 +66,13 @@ classes: required: false annotations: prompt: >- - Summarize each distinct growth experiment or set of conditions - into the GrowthExperiment class. If multiple growth experiments - are described, separate them accordingly. - prompt.examples: "Aerobic growth at 30°C; Anaerobic growth in sealed flasks at 37°C" + Summarize each distinct growth experiment or set of conditions. + If multiple growth experiments are described, separate them accordingly + in a semicolon-delimited list with no numbering and no newlines. + For each, include temperature, salinity, pH, oxygen status, sparging + equipment, inoculation method, mixing method, culturing duration, + light method (if relevant), and any other methods or conditions. + # prompt.examples: "Aerobic growth at 30°C; Anaerobic growth in sealed flasks at 37°C" Medium: @@ -80,7 +84,7 @@ classes: - value: "LB" annotations: prompt: "Extract the main name or code for this medium." - prompt.examples: "LB; M9" + # prompt.examples: "LB; M9" range: string required: false @@ -105,16 +109,20 @@ classes: prompt.examples: "Tryptone 10 g/L; Yeast Extract 5 g/L; NaCl 10 g/L" range: string required: false + multivalued: true modifications: description: "Any modifications or special additives to the standard medium recipe." examples: - value: "Adjusted NaCl to 5 g/L; Added 2 g/L glucose" annotations: - prompt: "Describe any deviations from the standard recipe." + prompt: >- + Describe any deviations from the standard recipe. + If multiple modifications, separate them with semicolons. prompt.examples: "Extra glucose; Reduced NaCl" range: string required: false + multivalued: true preparation_steps: description: "Description of how the medium is prepared." @@ -123,11 +131,12 @@ classes: annotations: prompt: >- Outline the preparation steps for the medium, including sterilization, - pH adjustment, or other relevant details. + pH adjustment, or other relevant details. If multiple steps, separate + them with semicolons. prompt.examples: "Autoclave at 121°C for 15 minutes; Adjust pH to 7.0" range: string required: false - + multivalued: true GrowthExperiment: description: "Growth conditions and experimental methods for culturing." @@ -137,10 +146,13 @@ classes: examples: - value: "30°C" annotations: - prompt: "Specify the growth temperature(s). If multiple, separate with semicolons." + prompt: >- + Specify the growth temperature(s). + If multiple, separate with semicolons. prompt.examples: "30°C; 37°C" range: string required: false + multivalued: true salinity: description: "Salinity or salt concentration conditions (e.g., % NaCl)." @@ -153,16 +165,20 @@ classes: prompt.examples: "3% NaCl; 2 M NaCl" range: string required: false + multivalued: true pH: description: "pH level at which the culture is maintained." examples: - value: "pH 7.0" annotations: - prompt: "Provide the pH level for the culture (e.g., pH 7.0)." + prompt: >- + Provide the pH level for the culture (e.g., pH 7.0). + If multiple or a range, separate with semicolons. prompt.examples: "pH 7.2" range: string required: false + multivalued: true oxygen_status: description: "Oxygen conditions (e.g., aerobic, anaerobic, microaerophilic)." @@ -172,40 +188,52 @@ classes: annotations: prompt: >- Indicate whether the culture is aerobic, anaerobic, microaerophilic, - or specify an oxygen concentration or partial pressure. + or specify an oxygen concentration or partial pressure. If multiple, + separate with semicolons. prompt.examples: "Aerobic; Microaerophilic; ~5% O2" range: string required: false + multivalued: true sparging_equipment: description: "Equipment or method used for sparging or aeration (if any)." examples: - value: "Air sparger with 0.2 μm filter at 1 L/min" annotations: - prompt: "Provide details on aeration equipment or methods." + prompt: >- + Provide details on aeration equipment or methods. + Include flow rate, filter size, or other relevant information. + If multiple, separate with semicolons. prompt.examples: "Air sparger at 1 L/min" range: string required: false + multivalued: true inoculation_method: - description: "Method used to inoculate the culture (e.g., inoculum source or process)." + description: "Method(s) used to inoculate the culture (e.g., inoculum source or process)." examples: - value: "Inoculated with a single colony from a streak plate" annotations: - prompt: "Describe how the culture was inoculated, including inoculum origin and technique." + prompt: >- + Describe how the culture was inoculated, including inoculum origin and technique. + If multiple methods, separate with semicolons. prompt.examples: "Single colony from LB agar plate; 1% v/v seed culture" range: string required: false + multivalued: true mixing_method: description: "Method of agitation or mixing (e.g., stirring speed, shaker, etc.)." examples: - value: "Shaken at 200 rpm on an orbital shaker" annotations: - prompt: "Provide details on how the culture was mixed, if at all." + prompt: >- + Provide details on how the culture was mixed, if at all. + If multiple methods, separate with semicolons. prompt.examples: "200 rpm on orbital shaker" range: string required: false + multivalued: true culturing_duration: description: "Duration of the culture growth or experiment." @@ -215,10 +243,11 @@ classes: annotations: prompt: >- State how long the culture was grown or observed. Include time units - (e.g., 24 h, 48 h). + (e.g., 24 h, 48 h). If multiple durations, separate with semicolons. prompt.examples: "24 h; 48 h" range: string required: false + multivalued: true light_method: description: "Type or intensity of light provided (if relevant) for phototrophic organisms." @@ -227,18 +256,21 @@ classes: annotations: prompt: >- Provide the type of lighting (e.g., wavelength, intensity) or indicate - dark conditions if relevant. + dark conditions if relevant. If multiple light conditions, separate with semicolons. prompt.examples: "Continuous white light at 50 μmol photons m⁻² s⁻¹" range: string required: false + multivalued: true other_methods: description: "Any other relevant growth or handling methods not captured above." examples: - value: "Maintained under 1 atm CO2; Pressure of 10 MPa" annotations: - prompt: "Describe any additional methods or conditions (e.g., special additives, pressure)." + prompt: >- + Describe any additional methods or conditions (e.g., special additives, pressure). + If multiple methods, separate with semicolons. prompt.examples: "10% CO2 atmosphere; High pressure environment" range: string required: false - + multivalued: true From 81eeadfba8a18c8a234e0faa40fd8d52ce570b41 Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 25 Mar 2025 11:28:03 -0400 Subject: [PATCH 22/25] Further prompt tuning --- src/ontogpt/templates/biolog.py_last | 352 ------------ src/ontogpt/templates/biolog_simple.py_last | 123 ----- .../templates/environmental_metadata.py | 519 ++++++++++++++---- src/ontogpt/templates/growth_experiment.py | 11 +- src/ontogpt/templates/growth_experiment.yaml | 10 +- 5 files changed, 418 insertions(+), 597 deletions(-) delete mode 100644 src/ontogpt/templates/biolog.py_last delete mode 100644 src/ontogpt/templates/biolog_simple.py_last diff --git a/src/ontogpt/templates/biolog.py_last b/src/ontogpt/templates/biolog.py_last deleted file mode 100644 index baedb1889..000000000 --- a/src/ontogpt/templates/biolog.py_last +++ /dev/null @@ -1,352 +0,0 @@ -from __future__ import annotations - -import re -import sys -from datetime import ( - date, - datetime, - time -) -from decimal import Decimal -from enum import Enum -from typing import ( - Any, - ClassVar, - Dict, - List, - Literal, - Optional, - Union -) - -from pydantic import ( - BaseModel, - ConfigDict, - Field, - RootModel, - field_validator -) - - -metamodel_version = "None" -version = "0.1.0" - - -class ConfiguredBaseModel(BaseModel): - model_config = ConfigDict( - validate_assignment = True, - validate_default = True, - extra = "forbid", - arbitrary_types_allowed = True, - use_enum_values = True, - strict = False, - ) - pass - - - - -class LinkMLMeta(RootModel): - root: Dict[str, Any] = {} - model_config = ConfigDict(frozen=True) - - def __getattr__(self, key:str): - return getattr(self.root, key) - - def __getitem__(self, key:str): - return self.root[key] - - def __setitem__(self, key:str, value): - self.root[key] = value - - def __contains__(self, key:str) -> bool: - return key in self.root - - -linkml_meta = LinkMLMeta({'default_prefix': 'paperex', - 'default_range': 'string', - 'description': 'Schema for extracting structured data from papers, including ' - 'Biolog Phenotype MicroArray experiments.', - 'id': 'https://example.org/PaperExtractionSchema', - 'imports': ['linkml:types'], - 'license': 'https://creativecommons.org/publicdomain/zero/1.0/', - 'name': 'PaperExtractionSchema', - 'prefixes': {'linkml': {'prefix_prefix': 'linkml', - 'prefix_reference': 'https://w3id.org/linkml/'}, - 'paperex': {'prefix_prefix': 'paperex', - 'prefix_reference': 'https://example.org/PaperExtractionSchema/'}, - 'rdf': {'prefix_prefix': 'rdf', - 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, - 'source_file': '/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog.yaml', - 'title': 'Paper Extraction Schema'} ) - -class FlexibleBooleanEnum(str, Enum): - """ - Minimal enumeration for capturing yes/no/true/false in lowercase, plus 'not provided' if data are missing or ambiguous - - """ - yes = "yes" - no = "no" - true = "true" - false = "false" - not_provided = "not provided" - - - -class NamedEntity(ConfiguredBaseModel): - """ - A generic named entity for all non-root classes. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - pass - - -class Paper(ConfiguredBaseModel): - """ - A single paper or study. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema', 'tree_root': True}) - - study_title: str = Field(default=..., description="""The paper's title.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': "Extract the paper's title from the " - 'text. \n' - 'If no title is found, return "Not ' - 'provided".\n'}}, - 'domain_of': ['Paper']} }) - authors: List[Author] = Field(default=..., description="""Authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Provide a YAML list of authors under the ' - 'key `authors:`.\n' - 'Example:\n' - '\n' - ' authors:\n' - ' - name: "Olga V. Mavrodi"\n' - ' - name: "Janiece R. McWilliams"\n' - '\n' - 'If no authors are found, return an ' - '*empty* list:\n' - '\n' - ' authors: []\n' - '\n' - '(Avoid leaving `authors:` blank with no ' - 'list or value.)\n'}}, - 'domain_of': ['Paper']} }) - doi: Optional[str] = Field(default=None, description="""DOI of the publication.""", json_schema_extra = { "linkml_meta": {'alias': 'doi', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the DOI if available. If not ' - "found, use 'Not provided'.\n"}}, - 'domain_of': ['Paper']} }) - date: Optional[str] = Field(default=None, description="""Publication date.""", json_schema_extra = { "linkml_meta": {'alias': 'date', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the publication date. If not ' - "specified, 'Not provided'.\n"}}, - 'domain_of': ['Paper']} }) - experiments: List[Experiment] = Field(default=..., description="""Experiments (e.g. Biolog Phenotype MicroArray) included in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Summarize each distinct experiment under ' - '`experiments:`\n' - 'with a YAML list. For example:\n' - '\n' - ' experiments:\n' - ' - experiment_motivation: ' - '"Screening with Biolog..."\n' - ' ...\n' - ' - experiment_motivation: "RNA-seq ' - 'profiling..."\n' - ' ...\n' - '\n' - 'If none are found, return an empty ' - 'list:\n' - '\n' - ' experiments: []\n' - '\n' - '(Do not leave it blank.)\n'}}, - 'domain_of': ['Paper']} }) - - -class Author(NamedEntity): - """ - An author of the paper. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract the full name of this author. If ' - "none, 'Not provided'."}}, - 'domain_of': ['Author']} }) - - -class Experiment(NamedEntity): - """ - A single experiment from the paper. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - experiment_motivation: Optional[str] = Field(default=None, description="""Rationale for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_motivation', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Describe the purpose or reason for doing ' - 'this experiment.'}}, - 'domain_of': ['Experiment']} }) - experiment_design: Optional[str] = Field(default=None, description="""Summary of how the experiment was designed.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_design', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Summarize the experimental design or ' - 'approach.'}}, - 'domain_of': ['Experiment']} }) - environment: Optional[str] = Field(default=None, description="""Location or conditions in which the experiment occurred (lab, greenhouse, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'environment', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract any mention of environment or ' - "setting. If none, 'Not provided'."}}, - 'domain_of': ['Experiment']} }) - host: Optional[List[Host]] = Field(default=None, description="""One or more hosts in the experiment (e.g., a plant).""", json_schema_extra = { "linkml_meta": {'alias': 'host', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Identify any host organisms used (like ' - 'Brachypodium distachyon).'}}, - 'domain_of': ['Experiment']} }) - host_type: Optional[str] = Field(default=None, description="""Type of host (e.g., plant, animal).""", json_schema_extra = { "linkml_meta": {'alias': 'host_type', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Specify the host type (plant, animal, ' - "etc.). If not mentioned, 'Not " - "provided'."}}, - 'domain_of': ['Experiment']} }) - target_microbes: Optional[List[Microbe]] = Field(default=None, description="""Microbes targeted in the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'target_microbes', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': "List the microbe(s). If none, 'Not " - "provided'."}}, - 'domain_of': ['Experiment']} }) - biological_system: Optional[str] = Field(default=None, description="""Biological system under study.""", json_schema_extra = { "linkml_meta": {'alias': 'biological_system', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': "Extract e.g. 'rhizosphere' or 'root " - "exudates system'. If missing, 'Not " - "provided'."}}, - 'domain_of': ['Experiment']} }) - conditions: Optional[str] = Field(default=None, description="""Experimental conditions (temperature, pH, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'conditions', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Pull out relevant conditions. If absent, ' - "'Not provided'."}}, - 'domain_of': ['Experiment']} }) - experimental_factors: Optional[List[ExperimentalFactor]] = Field(default=None, description="""Factors tested or measured (e.g., presence of root exudates, carbon sources).""", json_schema_extra = { "linkml_meta": {'alias': 'experimental_factors', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'List factors or treatments. If none, ' - "'Not provided'."}}, - 'domain_of': ['Experiment']} }) - is_biolog_experiment: Optional[FlexibleBooleanEnum] = Field(default=None, description="""Whether the experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': "Indicate 'yes' or 'no' if explicitly " - 'stated. \n' - "If not clear, use 'not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - type_of_biolog_experiment: Optional[str] = Field(default=None, description="""Type/subtype of the Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'type_of_biolog_experiment', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'If it is a Biolog experiment, specify ' - "(e.g. 'Phenotype MicroArray').\n" - "If absent, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - plates: Optional[List[str]] = Field(default=None, description="""Biolog plates used (e.g., PM01, PM02A).""", json_schema_extra = { "linkml_meta": {'alias': 'plates', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Identify plate types like PM01, PM02A, ' - 'EcoPlate, GEN III, etc.\n' - "If none, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - replicates: Optional[str] = Field(default=None, description="""Number of replicates.""", json_schema_extra = { "linkml_meta": {'alias': 'replicates', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract any mention of replicates. If ' - "none, 'Not provided'."}}, - 'domain_of': ['Experiment']} }) - protocol_steps: Optional[List[str]] = Field(default=None, description="""Key steps in the experimental protocol.""", json_schema_extra = { "linkml_meta": {'alias': 'protocol_steps', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'List the main protocol steps. \n' - "If none, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - plate_reader_types_of_data_collected: Optional[List[str]] = Field(default=None, description="""Data types collected (OD, formazan color, respiration, etc.).""", json_schema_extra = { "linkml_meta": {'alias': 'plate_reader_types_of_data_collected', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract data types measured by plate ' - 'reader or OmniLog. \n' - "If none, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - od_protocol: Optional[str] = Field(default=None, description="""Protocol for measuring optical density.""", json_schema_extra = { "linkml_meta": {'alias': 'od_protocol', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'If relevant, describe OD measurement ' - "steps. If none, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - respiration_protocol: Optional[str] = Field(default=None, description="""Protocol for measuring respiration or metabolic activity.""", json_schema_extra = { "linkml_meta": {'alias': 'respiration_protocol', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'If formazan or respiration is mentioned, ' - "describe. Otherwise, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - instrument_used: Optional[List[str]] = Field(default=None, description="""Instrument or equipment used (e.g. OmniLog Phenotype MicroArray System).""", json_schema_extra = { "linkml_meta": {'alias': 'instrument_used', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': "If there's mention of an instrument " - '(OmniLog, etc.), list it. \n' - 'If multiple are referenced, list them. \n' - "If none, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - analysis_software: Optional[List[str]] = Field(default=None, description="""Software used to analyze or interpret data.""", json_schema_extra = { "linkml_meta": {'alias': 'analysis_software', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': "Extract any software used (e.g. 'OmniLog " - "Parametric Analysis v1.20.02'). \n" - "If none, 'Not provided'.\n"}}, - 'domain_of': ['Experiment']} }) - incubation_temperature: Optional[str] = Field(default=None, description="""Temperature at which cultures/plates were incubated.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_temperature', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'If a temperature is mentioned, capture ' - "it; else, 'Not provided'."}}, - 'domain_of': ['Experiment']} }) - incubation_duration: Optional[str] = Field(default=None, description="""Total incubation time for the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'incubation_duration', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'If a time is given (48h, etc.), record ' - "it; else, 'Not provided'."}}, - 'domain_of': ['Experiment']} }) - - -class Host(NamedEntity): - """ - A host organism or system. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Host.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract details about this host. If ' - "none, 'Not provided'."}}, - 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor']} }) - - -class Microbe(NamedEntity): - """ - Information about a microbial species or strain. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for Microbe.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract details about this microbe. If ' - "none, 'Not provided'."}}, - 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor']} }) - - -class ExperimentalFactor(NamedEntity): - """ - A factor manipulated or measured in an experiment. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - placeholder_attribute: Optional[str] = Field(default=None, description="""Placeholder attribute for ExperimentalFactor.""", json_schema_extra = { "linkml_meta": {'alias': 'placeholder_attribute', - 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'Extract details about any factor (e.g. ' - 'carbon source, exudate presence).\n' - "If absent, 'Not provided'.\n"}}, - 'domain_of': ['Host', 'Microbe', 'ExperimentalFactor']} }) - - -# Model rebuild -# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model -NamedEntity.model_rebuild() -Paper.model_rebuild() -Author.model_rebuild() -Experiment.model_rebuild() -Host.model_rebuild() -Microbe.model_rebuild() -ExperimentalFactor.model_rebuild() diff --git a/src/ontogpt/templates/biolog_simple.py_last b/src/ontogpt/templates/biolog_simple.py_last deleted file mode 100644 index 58cd303d2..000000000 --- a/src/ontogpt/templates/biolog_simple.py_last +++ /dev/null @@ -1,123 +0,0 @@ -from __future__ import annotations - -import re -import sys -from datetime import ( - date, - datetime, - time -) -from decimal import Decimal -from enum import Enum -from typing import ( - Any, - ClassVar, - Dict, - List, - Literal, - Optional, - Union -) - -from pydantic import ( - BaseModel, - ConfigDict, - Field, - RootModel, - field_validator -) - - -metamodel_version = "None" -version = "0.1.0" - - -class ConfiguredBaseModel(BaseModel): - model_config = ConfigDict( - validate_assignment = True, - validate_default = True, - extra = "forbid", - arbitrary_types_allowed = True, - use_enum_values = True, - strict = False, - ) - pass - - - - -class LinkMLMeta(RootModel): - root: Dict[str, Any] = {} - model_config = ConfigDict(frozen=True) - - def __getattr__(self, key:str): - return getattr(self.root, key) - - def __getitem__(self, key:str): - return self.root[key] - - def __setitem__(self, key:str, value): - self.root[key] = value - - def __contains__(self, key:str) -> bool: - return key in self.root - - -linkml_meta = LinkMLMeta({'default_prefix': 'https://example.org/PaperExtractionSchema/', - 'description': 'A simplified schema describing the fields to extract from a ' - 'paper that includes study metadata and a minimal notion of ' - 'experiments.\n', - 'id': 'https://example.org/PaperExtractionSchema', - 'imports': ['linkml:types'], - 'name': 'PaperExtractionSchema', - 'prefixes': {'linkml': {'prefix_prefix': 'linkml', - 'prefix_reference': 'https://w3id.org/linkml/'}, - 'xsd': {'prefix_prefix': 'xsd', - 'prefix_reference': 'http://www.w3.org/2001/XMLSchema#'}}, - 'source_file': '/Users/marcin/Documents/VIMSS/ontology/LLMs/ontogpt/src/ontogpt/templates/biolog_simple.yaml', - 'title': 'Paper Extraction Schema', - 'types': {'boolean': {'base': 'bool', - 'from_schema': 'https://example.org/PaperExtractionSchema', - 'name': 'boolean'}, - 'string': {'base': 'str', - 'from_schema': 'https://example.org/PaperExtractionSchema', - 'name': 'string'}}} ) - - -class Paper(ConfiguredBaseModel): - """ - Top-level class representing a single paper/study. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema', 'tree_root': True}) - - study_title: str = Field(default=..., description="""Title of the study.""", json_schema_extra = { "linkml_meta": {'alias': 'study_title', 'domain_of': ['Paper']} }) - authors: List[Author] = Field(default=..., description="""List of authors of the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'authors', 'domain_of': ['Paper']} }) - doi: Optional[str] = Field(default=None, description="""DOI of the publication.""", json_schema_extra = { "linkml_meta": {'alias': 'doi', 'domain_of': ['Paper']} }) - date: Optional[str] = Field(default=None, description="""Date of the publication or the study.""", json_schema_extra = { "linkml_meta": {'alias': 'date', 'domain_of': ['Paper']} }) - experiments: Optional[List[Experiment]] = Field(default=None, description="""List of experiments mentioned in the paper.""", json_schema_extra = { "linkml_meta": {'alias': 'experiments', 'domain_of': ['Paper']} }) - - -class Author(ConfiguredBaseModel): - """ - Represents an author of the paper. - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - name: str = Field(default=..., description="""Name of the author.""", json_schema_extra = { "linkml_meta": {'alias': 'name', 'domain_of': ['Author']} }) - - -class Experiment(ConfiguredBaseModel): - """ - A minimal experiment class (could be Biolog or otherwise). - """ - linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/PaperExtractionSchema'}) - - experiment_title: Optional[str] = Field(default=None, description="""Title or short description of the experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'experiment_title', 'domain_of': ['Experiment']} }) - is_biolog_experiment: Optional[bool] = Field(default=None, description="""Flag indicating if this experiment is a Biolog experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'is_biolog_experiment', 'domain_of': ['Experiment']} }) - - -# Model rebuild -# see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model -Paper.model_rebuild() -Author.model_rebuild() -Experiment.model_rebuild() diff --git a/src/ontogpt/templates/environmental_metadata.py b/src/ontogpt/templates/environmental_metadata.py index 32c1b77eb..436e833c2 100644 --- a/src/ontogpt/templates/environmental_metadata.py +++ b/src/ontogpt/templates/environmental_metadata.py @@ -1,181 +1,454 @@ -from __future__ import annotations -from datetime import datetime, date -from enum import Enum +from __future__ import annotations -from typing import List, Dict, Optional, Any, Union -from pydantic import BaseModel as BaseModel, ConfigDict, Field, field_validator import re import sys -if sys.version_info >= (3, 8): - from typing import Literal -else: - from typing_extensions import Literal +from datetime import ( + date, + datetime, + time +) +from decimal import Decimal +from enum import Enum +from typing import ( + Any, + ClassVar, + Dict, + List, + Literal, + Optional, + Union +) + +from pydantic import ( + BaseModel, + ConfigDict, + Field, + RootModel, + field_validator +) metamodel_version = "None" version = "None" + class ConfiguredBaseModel(BaseModel): model_config = ConfigDict( - validate_assignment=True, - validate_default=True, - extra = 'forbid', - arbitrary_types_allowed=True, - use_enum_values = True) + validate_assignment = True, + validate_default = True, + extra = "forbid", + arbitrary_types_allowed = True, + use_enum_values = True, + strict = False, + ) pass + + +class LinkMLMeta(RootModel): + root: Dict[str, Any] = {} + model_config = ConfigDict(frozen=True) + + def __getattr__(self, key:str): + return getattr(self.root, key) + + def __getitem__(self, key:str): + return self.root[key] + + def __setitem__(self, key:str, value): + self.root[key] = value + + def __contains__(self, key:str) -> bool: + return key in self.root + + +linkml_meta = LinkMLMeta({'default_prefix': 'envmd', + 'default_range': 'string', + 'description': 'A template for categorizing Environmental Data Initiative ' + 'data entries. See ' + 'https://github.com/EDIorg/EDIorg-repository-index', + 'id': 'http://w3id.org/ontogpt/environmental-metadata', + 'imports': ['linkml:types', 'core'], + 'license': 'https://creativecommons.org/publicdomain/zero/1.0/', + 'name': 'environmental-metadata', + 'prefixes': {'ENVO': {'prefix_prefix': 'ENVO', + 'prefix_reference': 'http://purl.obolibrary.org/obo/ENVO_'}, + 'envmd': {'prefix_prefix': 'envmd', + 'prefix_reference': 'http://w3id.org/ontogpt/environmental-metadata'}, + 'linkml': {'prefix_prefix': 'linkml', + 'prefix_reference': 'https://w3id.org/linkml/'}, + 'rdf': {'prefix_prefix': 'rdf', + 'prefix_reference': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'}}, + 'source_file': 'src/ontogpt/templates/environmental_metadata.yaml', + 'title': 'Environmental sample metadata'} ) + class NullDataOptions(str, Enum): - - UNSPECIFIED_METHOD_OF_ADMINISTRATION = "UNSPECIFIED_METHOD_OF_ADMINISTRATION" - NOT_APPLICABLE = "NOT_APPLICABLE" - NOT_MENTIONED = "NOT_MENTIONED" - - -class Dataset(ConfiguredBaseModel): - - packageid: Optional[str] = Field(None, description="""The internal identifier for the dataset""") - topic: Optional[List[str]] = Field(default_factory=list, description="""the general scientific area of study concerning the sample(s)""") - location: Optional[List[str]] = Field(default_factory=list, description="""the geographic location where the sample was isolated""") - environmental_material: Optional[List[str]] = Field(default_factory=list, description="""the environmental material that was sampled""") - environments: Optional[List[str]] = Field(default_factory=list, description="""the environmental context in which the study was conducted""") - methods: Optional[List[str]] = Field(default_factory=list) - - + class ExtractionResult(ConfiguredBaseModel): """ A result of extracting knowledge on text """ - input_id: Optional[str] = Field(None) - input_title: Optional[str] = Field(None) - input_text: Optional[str] = Field(None) - raw_completion_output: Optional[str] = Field(None) - prompt: Optional[str] = Field(None) - extracted_object: Optional[Any] = Field(None, description="""The complex objects extracted from the text""") - named_entities: Optional[List[Any]] = Field(default_factory=list, description="""Named entities extracted from the text""") - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) -class NamedEntity(ConfiguredBaseModel): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - + input_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_id', 'domain_of': ['ExtractionResult']} }) + input_title: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_title', 'domain_of': ['ExtractionResult']} }) + input_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'input_text', 'domain_of': ['ExtractionResult']} }) + raw_completion_output: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'raw_completion_output', 'domain_of': ['ExtractionResult']} }) + prompt: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'prompt', 'domain_of': ['ExtractionResult']} }) + extracted_object: Optional[Any] = Field(default=None, description="""The complex objects extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'extracted_object', 'domain_of': ['ExtractionResult']} }) + named_entities: Optional[List[Any]] = Field(default=None, description="""Named entities extracted from the text""", json_schema_extra = { "linkml_meta": {'alias': 'named_entities', 'domain_of': ['ExtractionResult']} }) -class Topic(NamedEntity): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - -class Location(NamedEntity): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - +class NamedEntity(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) -class EnvironmentalMaterial(NamedEntity): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) -class Environment(NamedEntity): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v -class Method(NamedEntity): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - class CompoundExpression(ConfiguredBaseModel): - - None - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + pass + class Triple(CompoundExpression): """ Abstract parent for Relation Extraction tasks """ - subject: Optional[str] = Field(None) - predicate: Optional[str] = Field(None) - object: Optional[str] = Field(None) - qualifier: Optional[str] = Field(None, description="""A qualifier for the statements, e.g. \"NOT\" for negation""") - subject_qualifier: Optional[str] = Field(None, description="""An optional qualifier or modifier for the subject of the statement, e.g. \"high dose\" or \"intravenously administered\"""") - object_qualifier: Optional[str] = Field(None, description="""An optional qualifier or modifier for the object of the statement, e.g. \"severe\" or \"with additional complications\"""") - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'abstract': True, 'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject', 'domain_of': ['Triple']} }) + predicate: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'predicate', 'domain_of': ['Triple']} }) + object: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object', 'domain_of': ['Triple']} }) + qualifier: Optional[str] = Field(default=None, description="""A qualifier for the statements, e.g. \"NOT\" for negation""", json_schema_extra = { "linkml_meta": {'alias': 'qualifier', 'domain_of': ['Triple']} }) + subject_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the subject of the statement, e.g. \"high dose\" or \"intravenously administered\"""", json_schema_extra = { "linkml_meta": {'alias': 'subject_qualifier', 'domain_of': ['Triple']} }) + object_qualifier: Optional[str] = Field(default=None, description="""An optional qualifier or modifier for the object of the statement, e.g. \"severe\" or \"with additional complications\"""", json_schema_extra = { "linkml_meta": {'alias': 'object_qualifier', 'domain_of': ['Triple']} }) + class TextWithTriples(ConfiguredBaseModel): """ A text containing one or more relations of the Triple type. """ - publication: Optional[Publication] = Field(None) - triples: Optional[List[Triple]] = Field(default_factory=list) - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + triples: Optional[List[Triple]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'triples', 'domain_of': ['TextWithTriples']} }) + class TextWithEntity(ConfiguredBaseModel): """ A text containing one or more instances of a single type of entity. """ - publication: Optional[Publication] = Field(None) - entities: Optional[List[str]] = Field(default_factory=list) - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + publication: Optional[Publication] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'publication', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'domain_of': ['TextWithTriples', 'TextWithEntity']} }) + entities: Optional[List[str]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'entities', 'domain_of': ['TextWithEntity']} }) + class RelationshipType(NamedEntity): - - id: str = Field(..., description="""A unique identifier for the named entity""") - label: Optional[str] = Field(None, description="""The label (name) of the named thing""") - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core', + 'id_prefixes': ['RO', 'biolink']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + class Publication(ConfiguredBaseModel): - - id: Optional[str] = Field(None, description="""The publication identifier""") - title: Optional[str] = Field(None, description="""The title of the publication""") - abstract: Optional[str] = Field(None, description="""The abstract of the publication""") - combined_text: Optional[str] = Field(None) - full_text: Optional[str] = Field(None, description="""The full text of the publication""") - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + id: Optional[str] = Field(default=None, description="""The publication identifier""", json_schema_extra = { "linkml_meta": {'alias': 'id', 'domain_of': ['NamedEntity', 'Publication']} }) + title: Optional[str] = Field(default=None, description="""The title of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'title', 'domain_of': ['Publication']} }) + abstract: Optional[str] = Field(default=None, description="""The abstract of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'abstract', 'domain_of': ['Publication']} }) + combined_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'combined_text', 'domain_of': ['Publication']} }) + full_text: Optional[str] = Field(default=None, description="""The full text of the publication""", json_schema_extra = { "linkml_meta": {'alias': 'full_text', 'domain_of': ['Publication']} }) + class AnnotatorResult(ConfiguredBaseModel): - - subject_text: Optional[str] = Field(None) - object_id: Optional[str] = Field(None) - object_text: Optional[str] = Field(None) - - + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/core'}) + + subject_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'subject_text', 'domain_of': ['AnnotatorResult']} }) + object_id: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_id', 'domain_of': ['AnnotatorResult']} }) + object_text: Optional[str] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'object_text', 'domain_of': ['AnnotatorResult']} }) + + +class Dataset(ConfiguredBaseModel): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'http://w3id.org/ontogpt/environmental-metadata', + 'tree_root': True}) + + packageid: Optional[str] = Field(default=None, description="""The internal identifier for the dataset""", json_schema_extra = { "linkml_meta": {'alias': 'packageid', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'single unique identifier for the ' + 'dataset'}}, + 'domain_of': ['Dataset']} }) + topic: Optional[List[str]] = Field(default=None, description="""the general scientific area of study concerning the sample(s)""", json_schema_extra = { "linkml_meta": {'alias': 'topic', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'semicolon-separated list of scientific ' + 'areas of study concerning the ' + 'sample(s)'}}, + 'domain_of': ['Dataset']} }) + location: Optional[List[str]] = Field(default=None, description="""the geographic location where the sample was isolated""", json_schema_extra = { "linkml_meta": {'alias': 'location', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'semicolon-separated list of geographic ' + 'locations of sample isolations'}}, + 'domain_of': ['Dataset']} }) + environmental_material: Optional[List[str]] = Field(default=None, description="""the environmental material that was sampled""", json_schema_extra = { "linkml_meta": {'alias': 'environmental_material', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'semicolon-separated list of ' + 'environmental materials'}}, + 'domain_of': ['Dataset']} }) + environments: Optional[List[str]] = Field(default=None, description="""the environmental context in which the study was conducted""", json_schema_extra = { "linkml_meta": {'alias': 'environments', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'semicolon-separated list of ' + 'environmental contexts in which the ' + 'study was conducted'}}, + 'domain_of': ['Dataset']} }) + methods: Optional[List[str]] = Field(default=None, json_schema_extra = { "linkml_meta": {'alias': 'methods', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'semicolon-separated list of methods used ' + 'in measuring sample variables'}}, + 'domain_of': ['Dataset']} }) + + +class Topic(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', + 'value': 'bioportal:ENVTHES'}}, + 'from_schema': 'http://w3id.org/ontogpt/environmental-metadata', + 'id_prefixes': ['ENVTHES']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Location(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', + 'value': 'sqlite:obo:envo, bioportal:GAZ'}}, + 'from_schema': 'http://w3id.org/ontogpt/environmental-metadata', + 'id_prefixes': ['ENVO', 'GAZ']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class EnvironmentalMaterial(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', + 'value': 'sqlite:obo:envo, bioportal:ENVTHES'}}, + 'from_schema': 'http://w3id.org/ontogpt/environmental-metadata', + 'id_prefixes': ['ENVO', 'ENVTHES']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Environment(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', + 'value': 'sqlite:obo:envo, bioportal:ENVTHES'}}, + 'from_schema': 'http://w3id.org/ontogpt/environmental-metadata', + 'id_prefixes': ['ENVO', 'ENVTHES']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v + + +class Method(NamedEntity): + linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'annotations': {'annotators': {'tag': 'annotators', + 'value': 'bioportal:ENVTHES'}}, + 'from_schema': 'http://w3id.org/ontogpt/environmental-metadata', + 'id_prefixes': ['ENVTHES']}) + + id: str = Field(default=..., description="""A unique identifier for the named entity""", json_schema_extra = { "linkml_meta": {'alias': 'id', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['this is populated during the grounding and normalization step'], + 'domain_of': ['NamedEntity', 'Publication']} }) + label: Optional[str] = Field(default=None, description="""The label (name) of the named thing""", json_schema_extra = { "linkml_meta": {'alias': 'label', + 'aliases': ['name'], + 'annotations': {'owl': {'tag': 'owl', + 'value': 'AnnotationProperty, AnnotationAssertion'}}, + 'domain_of': ['NamedEntity'], + 'slot_uri': 'rdfs:label'} }) + original_spans: Optional[List[str]] = Field(default=None, description="""The coordinates of the original text span from which the named entity was extracted, inclusive. For example, \"10:25\" means the span starting from the 10th character and ending with the 25th character. The first character in the text has index 0. Newlines are treated as single characters. Multivalued as there may be multiple spans for a single text.""", json_schema_extra = { "linkml_meta": {'alias': 'original_spans', + 'annotations': {'prompt.skip': {'tag': 'prompt.skip', 'value': 'true'}}, + 'comments': ['This is determined during grounding and normalization', + 'But is based on the full input text'], + 'domain_of': ['NamedEntity']} }) + + @field_validator('original_spans') + def pattern_original_spans(cls, v): + pattern=re.compile(r"^\d+:\d+$") + if isinstance(v,list): + for element in v: + if isinstance(v, str) and not pattern.match(element): + raise ValueError(f"Invalid original_spans format: {element}") + elif isinstance(v,str): + if not pattern.match(v): + raise ValueError(f"Invalid original_spans format: {v}") + return v # Model rebuild # see https://pydantic-docs.helpmanual.io/usage/models/#rebuilding-a-model -Dataset.model_rebuild() ExtractionResult.model_rebuild() NamedEntity.model_rebuild() -Topic.model_rebuild() -Location.model_rebuild() -EnvironmentalMaterial.model_rebuild() -Environment.model_rebuild() -Method.model_rebuild() CompoundExpression.model_rebuild() Triple.model_rebuild() TextWithTriples.model_rebuild() @@ -183,4 +456,10 @@ class AnnotatorResult(ConfiguredBaseModel): RelationshipType.model_rebuild() Publication.model_rebuild() AnnotatorResult.model_rebuild() +Dataset.model_rebuild() +Topic.model_rebuild() +Location.model_rebuild() +EnvironmentalMaterial.model_rebuild() +Environment.model_rebuild() +Method.model_rebuild() diff --git a/src/ontogpt/templates/growth_experiment.py b/src/ontogpt/templates/growth_experiment.py index d904874b8..bd011faee 100644 --- a/src/ontogpt/templates/growth_experiment.py +++ b/src/ontogpt/templates/growth_experiment.py @@ -285,7 +285,7 @@ class Medium(ConfiguredBaseModel): 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Extract the main name or code for this ' 'medium.'}}, - 'domain_of': ['Medium'], + 'domain_of': ['Medium', 'GrowthExperiment'], 'examples': [{'value': 'LB'}]} }) label: Optional[str] = Field(default=None, description="""A human-readable label or short name for the medium.""", json_schema_extra = { "linkml_meta": {'alias': 'label', 'annotations': {'prompt': {'tag': 'prompt', @@ -337,6 +337,15 @@ class GrowthExperiment(ConfiguredBaseModel): """ linkml_meta: ClassVar[LinkMLMeta] = LinkMLMeta({'from_schema': 'https://example.org/MicrobialCulturingExtraction'}) + identifier: Optional[str] = Field(default=None, description="""A unique identifier for the growth experiment.""", json_schema_extra = { "linkml_meta": {'alias': 'identifier', + 'annotations': {'prompt': {'tag': 'prompt', + 'value': 'Provide a unique identifier for this ' + 'growth experiment. This may be a summary ' + 'of the experiment in a few words.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Aerobic growth of ' + 'Staphylococcus aureus'}}, + 'domain_of': ['Medium', 'GrowthExperiment']} }) temperature: Optional[List[str]] = Field(default=None, description="""Temperature(s) at which the culture is grown (e.g., in Celsius).""", json_schema_extra = { "linkml_meta": {'alias': 'temperature', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Specify the growth temperature(s). If ' diff --git a/src/ontogpt/templates/growth_experiment.yaml b/src/ontogpt/templates/growth_experiment.yaml index f541a90eb..4522b1e24 100644 --- a/src/ontogpt/templates/growth_experiment.yaml +++ b/src/ontogpt/templates/growth_experiment.yaml @@ -74,7 +74,6 @@ classes: light method (if relevant), and any other methods or conditions. # prompt.examples: "Aerobic growth at 30°C; Anaerobic growth in sealed flasks at 37°C" - Medium: description: "Details about a liquid culturing medium." attributes: @@ -141,6 +140,15 @@ classes: GrowthExperiment: description: "Growth conditions and experimental methods for culturing." attributes: + identifier: + description: "A unique identifier for the growth experiment." + annotations: + prompt: >- + Provide a unique identifier for this growth experiment. + This may be a summary of the experiment in a few words. + prompt.examples: "Aerobic growth of Staphylococcus aureus" + range: string + required: false temperature: description: "Temperature(s) at which the culture is grown (e.g., in Celsius)." examples: From 2d09094f7a5cb42c00dfdde50f1491b8d23b561a Mon Sep 17 00:00:00 2001 From: caufieldjh Date: Tue, 25 Mar 2025 14:14:29 -0400 Subject: [PATCH 23/25] Further tweaks for more consistent parsing --- src/ontogpt/templates/growth_experiment.py | 30 +++++++++++++------- src/ontogpt/templates/growth_experiment.yaml | 15 +++++----- 2 files changed, 27 insertions(+), 18 deletions(-) diff --git a/src/ontogpt/templates/growth_experiment.py b/src/ontogpt/templates/growth_experiment.py index bd011faee..6fc996336 100644 --- a/src/ontogpt/templates/growth_experiment.py +++ b/src/ontogpt/templates/growth_experiment.py @@ -251,27 +251,37 @@ class MicrobialCulturing(ConfiguredBaseModel): media: Optional[List[Medium]] = Field(default=None, description="""One or more culture media used in the study.""", json_schema_extra = { "linkml_meta": {'alias': 'media', 'annotations': {'prompt': {'tag': 'prompt', - 'value': 'List each distinct culture medium ' + 'value': 'Summarize each distinct culture medium ' 'described in the text. Separate multiple ' - 'media with semicolons if they are ' - 'clearly distinct. For each, include the ' - 'main name or code, ingredients, any ' - 'modifications, and preparation steps.'}}, + 'summaries with semicolons. For each, ' + 'include the main name or code, ' + 'ingredients, any modifications, and ' + 'preparation steps. Do not include ' + 'newlines or numbering.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Blood agar (BA), made from 0.5% ' + 'Peptone, 0.3% beef ' + 'extract/yeast extract, 1.5% ' + 'agar, 0.5% NaCl, and distilled ' + 'water; M9 minimal medium'}}, 'domain_of': ['MicrobialCulturing']} }) growth_experiments: Optional[List[GrowthExperiment]] = Field(default=None, description="""One or more growth experiments or conditions described in the study.""", json_schema_extra = { "linkml_meta": {'alias': 'growth_experiments', 'annotations': {'prompt': {'tag': 'prompt', 'value': 'Summarize each distinct growth ' - 'experiment or set of conditions. If ' - 'multiple growth experiments are ' - 'described, separate them accordingly in ' - 'a semicolon-delimited list with no ' + 'experiment or set of conditions. ' + 'Separate all summaries in a ' + 'semicolon-delimited list with no ' 'numbering and no newlines. For each, ' 'include temperature, salinity, pH, ' 'oxygen status, sparging equipment, ' 'inoculation method, mixing method, ' 'culturing duration, light method (if ' 'relevant), and any other methods or ' - 'conditions.'}}, + 'conditions.'}, + 'prompt.examples': {'tag': 'prompt.examples', + 'value': 'Aerobic growth at 30°C; ' + 'Anaerobic growth in sealed ' + 'flasks at 37°C'}}, 'domain_of': ['MicrobialCulturing']} }) diff --git a/src/ontogpt/templates/growth_experiment.yaml b/src/ontogpt/templates/growth_experiment.yaml index 4522b1e24..62ba6d803 100644 --- a/src/ontogpt/templates/growth_experiment.yaml +++ b/src/ontogpt/templates/growth_experiment.yaml @@ -53,11 +53,11 @@ classes: required: false annotations: prompt: >- - List each distinct culture medium described in the text. - Separate multiple media with semicolons if they are clearly - distinct. For each, include the main name or code, - ingredients, any modifications, and preparation steps. - # prompt.examples: "LB medium; M9 minimal medium" + Summarize each distinct culture medium described in the text. + Separate multiple summaries with semicolons. For each, + include the main name or code, ingredients, any modifications, + and preparation steps. Do not include newlines or numbering. + prompt.examples: "Blood agar (BA), made from 0.5% Peptone, 0.3% beef extract/yeast extract, 1.5% agar, 0.5% NaCl, and distilled water; M9 minimal medium" growth_experiments: description: "One or more growth experiments or conditions described in the study." @@ -67,12 +67,11 @@ classes: annotations: prompt: >- Summarize each distinct growth experiment or set of conditions. - If multiple growth experiments are described, separate them accordingly - in a semicolon-delimited list with no numbering and no newlines. + Separate all summaries in a semicolon-delimited list with no numbering and no newlines. For each, include temperature, salinity, pH, oxygen status, sparging equipment, inoculation method, mixing method, culturing duration, light method (if relevant), and any other methods or conditions. - # prompt.examples: "Aerobic growth at 30°C; Anaerobic growth in sealed flasks at 37°C" + prompt.examples: "Aerobic growth at 30°C; Anaerobic growth in sealed flasks at 37°C" Medium: description: "Details about a liquid culturing medium." From 7a70fb61bf2214887aafaf1b3d77e7e1c4a00c1a Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Tue, 1 Apr 2025 10:38:02 -0700 Subject: [PATCH 24/25] biolog PM01 --- src/ontogpt/templates/biolog_PM01.yaml | 585 +++++++++++++++++++++++++ 1 file changed, 585 insertions(+) create mode 100644 src/ontogpt/templates/biolog_PM01.yaml diff --git a/src/ontogpt/templates/biolog_PM01.yaml b/src/ontogpt/templates/biolog_PM01.yaml new file mode 100644 index 000000000..7a58c81d2 --- /dev/null +++ b/src/ontogpt/templates/biolog_PM01.yaml @@ -0,0 +1,585 @@ +id: https://example.org/PaperExtractionSchema +name: PaperExtractionSchema +title: Paper Extraction Schema +description: "Schema for extracting structured data from papers, including Biolog Phenotype MicroArray experiments. Extended to capture PM01 well-level results." +license: https://creativecommons.org/publicdomain/zero/1.0/ +version: 0.2.0 + +prefixes: + rdf: "http://www.w3.org/1999/02/22-rdf-syntax-ns#" + linkml: "https://w3id.org/linkml/" + paperex: "https://example.org/PaperExtractionSchema/" + +default_prefix: paperex +default_range: string + +imports: + - linkml:types + - core # Adjust or remove if your environment does not have a 'core' schema. + +######################## +# ENUMS +######################## + +enums: + FlexibleBooleanEnum: + description: > + A minimal enumeration for capturing yes/no/true/false in lowercase, + plus 'not provided' if data are missing or ambiguous. + permissible_values: + 'yes': {} + 'no': {} + 'true': {} + 'false': {} + 'not provided': {} + + PM01ResultEnum: + description: > + Captures whether the PM01 assay result was positive, negative, or not provided. + permissible_values: + positive: {} + negative: {} + not provided: {} + + PM01ChemicalEnum: + description: > + Enumeration of all 96 wells/chemicals in Biolog PM01. + permissible_values: + L-Arabinose: {} + N-Acetyl-D-Glucosamine: {} + D-Saccharic\ acid: {} + Succinic\ acid: {} + D-Galactose: {} + L-Aspartic\ acid: {} + L-Proline: {} + D-Alanine: {} + D-Trehalose: {} + D-Mannose: {} + Dulcitol: {} + D-Serine: {} + D-Sorbitol: {} + Glycerol: {} + L-Fucose: {} + D-Glucuronic\ acid: {} + D-Gluconic\ acid: {} + DL-a-Glycerol\ Phosphate: {} + D-Xylose: {} + L-Lactic\ acid: {} + Formic\ acid: {} + D-Mannitol: {} + L-Glutamic\ acid: {} + D-Glucose-6-Phosphate: {} + D-Galactonic\ acid-g-Lactone: {} + DL-Malic\ acid: {} + D-Ribose: {} + Tween\ 20: {} + L-Rhamnose: {} + D-Fructose: {} + Acetic\ acid: {} + a-D-Glucose: {} + Maltose: {} + D-Melibiose: {} + Thymidine: {} + L-Asparagine: {} + D-Aspartic\ acid: {} + D-Glucosaminic\ acid: {} + 1,2-Propanediol: {} + Tween\ 40: {} + a-Ketoglutaric\ acid: {} + a-Ketobutyric\ acid: {} + a-Methyl-D-Galactoside: {} + a-D-Lactose: {} + Lactulose: {} + Sucrose: {} + Uridine: {} + L-Glutamine: {} + m-Tartaric\ acid: {} + D-Glucose-1-Phosphate: {} + D-Fructose-6-Phosphate: {} + Tween\ 80: {} + a-Hydroxyglutaric\ acid-g-Lactone: {} + a-Hydroxybutyric\ acid: {} + b-Methyl-D-Glucoside: {} + Adonitol: {} + Maltotriose: {} + "2`-Deoxyadenosine": {} + Adenosine: {} + Gly-Asp: {} + Citric\ acid: {} + m-Inositol: {} + D-Threonine: {} + Fumaric\ acid: {} + Bromosuccinic\ acid: {} + Propionic\ acid: {} + Mucic\ acid: {} + Glycolic\ acid: {} + Glyoxylic\ acid: {} + D-Cellobiose: {} + Inosine: {} + Gly-Glu: {} + Tricarballylic\ acid: {} + L-Serine: {} + L-Threonine: {} + L-Alanine: {} + Ala-Gly: {} + Acetoacetic\ acid: {} + N-Acetyl-D-Mannosamine: {} + Mono-Methylsuccinate: {} + Methylpyruvate: {} + D-Malic\ acid: {} + L-Malic\ acid: {} + Gly-Pro: {} + p-Hydroxyphenyl\ Acetic\ acid: {} + m-Hydroxyphenyl\ Acetic\ acid: {} + Tyramine: {} + D-Psicose: {} + L-Lyxose: {} + Glucuronamide: {} + Pyruvic\ acid: {} + L-Galactonic\ acid-g-Lactone: {} + D-Galacturonic\ acid: {} + Phenylethylamine: {} + 2-Aminoethanol: {} + +######################## +# CLASSES +######################## + +classes: + + Paper: + tree_root: true + description: "A single paper or study." + attributes: + study_title: + description: "The paper's title." + annotations: + prompt: >- + Extract the paper's title. If not found, return only 'Not provided'. + prompt.examples: + Root Exudates Alter the Expression of Diverse Metabolic, Transport, Regulatory, and Stress Response Genes in Rhizosphere Pseudomonas. + required: true + + authors: + description: "Authors of the paper." + annotations: + prompt: | + Extract a list of authors, semicolon delimited. + If multiple authors appear in one string, please split them carefully. + If not found, use 'Not provided'. + range: string + multivalued: true + required: true + + doi: + description: "DOI of the publication." + annotations: + prompt: "Extract the DOI. If not found, use 'Not provided'." + range: string + required: false + + date: + description: "Publication date." + annotations: + prompt: "Extract the publication date. If not found, use 'Not provided'." + range: string + required: false + + traits: + description: "Organismal traits measured or observed." + annotations: + prompt: | + Extract a semicolon-delimited list of traits + focused on in the study. + prompt.examples: + Growth rate; Lactose production; motility + multivalued: true + range: Trait + required: false + + experiments: + description: "Experiments (e.g., Biolog) included in the paper." + annotations: + prompt: | + Extract a list of brief descriptions of experiments + described in the paper. This list must be semicolon-delimited. + For each description, include all of the following details if + provided: experiment motivation (why it was performed), + experiment design (how it was performed, including methods, + tools, organisms, and chemicals used), environment (the location + and conditions in which the experiment occurred, including + metrics like temperature), all organisms used and their + high-level type (e.g., plant, animal), the biological system + under study (e.g., rhizosphere), experimental conditions + (temperature, pH, etc.), and experimental factors tested or measured. + Also note whether the experiment is a Biolog experiment and if so, + the type or subtype of experiment (e.g., Phenotype MicroArray), + the Biolog plates used (e.g., PM01), the number of replicates, all + key steps in the experimental protocol, the types of data collected + by the plate reader (e.g., OD, respiration), the protocol for + measuring optical density (e.g., OD600 measured every 2 hours), + the protocol for measuring respiration (e.g., Colorimetric change + at 590 nm for formazan detection), the instrument or equipment used + (e.g., OmniLog Phenotype MicroArray System), the software used for + data analysis (e.g., OmniLog Parametric Analysis software), the + incubation temperature (e.g., 25°C), the total duration of incubation + (e.g., 48 h), and any other relevant details. + If nothing is mentioned regarding experiments, use 'Not provided'. + Do not provide any details here not related to experiments. + Do not include newlines. + range: Experiment + multivalued: true + required: true + + Experiment: + description: "A single experiment from the paper." + attributes: + + label: + description: >- + A single sentence description of this experiment. + annotations: + prompt: "A single sentence description of this experiment." + + experiment_motivation: + description: "Rationale for the experiment." + annotations: + prompt: | + Explain the motivation or purpose of the experiment. + If missing, use 'Not provided'. + range: string + required: false + + experiment_design: + description: "Summary of how the experiment was designed." + annotations: + prompt: "Summarize the overall design. If missing, use 'Not provided'." + range: string + required: false + + environment: + description: "Location or conditions in which the experiment occurred." + annotations: + prompt: "Extract the environment or setting." + range: string + required: false + + host_organism: + description: "One or more hosts in the experiment." + annotations: + prompt: | + Identify the host organisms or systems used + in a semicolon-delimited list. + prompt.examples: + Brachypodium distachyon; Arabidopsis thaliana + range: Host + multivalued: true + required: false + + host_type: + description: "Type of host (e.g., plant, animal)." + annotations: + prompt: >- + Describe the host type if applicable (e.g. 'plant'). + range: string + required: false + + target_organisms: + description: "Organisms, including microbes, targeted in the experiment." + annotations: + prompt: | + Extract a semicolon-delimited list of target organisms, + including microbes, targeted in the study. + prompt.examples: + Pseudomonas fluorescens; Bacillus subtilis; Pseudomonas simiae WCS 417 + range: Microbe + multivalued: true + required: false + + target_organism_group: + description: >- + A high-level description of organisms, including microbes, + targeted in the experiment. + annotations: + prompt: | + Extract a semicolon-delimited list of groups of target organisms, + including microbes, targeted in the study. + prompt.examples: + Pseudomonas strains; Pseudomonas group; Roseobacter clade + range: MicrobeGroup + multivalued: true + required: false + + biological_system: + description: "Biological system under study." + annotations: + prompt: >- + Extract the focal biological system (e.g., rhizosphere). + range: string + required: false + + conditions: + description: "Experimental conditions (temperature, pH, etc.)." + annotations: + prompt: | + Identify relevant experimental conditions in a + semicolon-delimited list (e.g., 25°C; pH 7). + multivalued: true + range: string + required: false + + traits: + description: "Organismal traits measured or observed." + annotations: + prompt: | + Extract a semicolon-delimited list of traits + measured or observed in the experiment. + prompt.examples: + Growth rate; Lactose production; motility + multivalued: true + range: Trait + required: false + + experimental_factors: + description: "Factors tested or measured." + annotations: + prompt: | + List experimental factors or treatments + in a semicolon-delimited list. + range: ExperimentalFactor + multivalued: true + required: false + + is_biolog_experiment: + description: "Whether the experiment is a Biolog experiment." + annotations: + prompt: | + Indicate if it's a Biolog experiment ('yes' or 'no'). + If not stated, 'not provided'. + range: FlexibleBooleanEnum + required: false + + type_of_biolog_experiment: + description: "Type or subtype of the Biolog experiment." + annotations: + prompt: | + If it is a Biolog experiment, specify the subtype (e.g., Phenotype MicroArray). + Otherwise, 'Not provided'. + range: string + required: false + + plates: + description: "Biolog plates used (e.g., EcoPlate, GEN III, PM01, PM02A)." + annotations: + prompt: | + Extract the name(s) of any phenotype microarray plates + (also known as Biolog plates, 96 well plates, microarrays, + microarray plates, or microplates) in a semicolon-delimited list. + Names may be in the format PM01, PM02A, etc. + If not mentioned, 'Not provided'. + prompt.examples: PM01; PM02A; EcoPlate; GEN III + range: string + multivalued: true + required: false + + replicates: + description: "Number of replicates." + annotations: + prompt: | + Extract the number of replicates as a single number. + If not mentioned, 'Not provided'. + range: string + required: false + + protocol_steps: + description: "Key steps in the experimental protocol." + annotations: + prompt: | + List the full protocol steps or key experimental procedures, + including inoculation details, incubation times/temperatures, + data recording intervals, and data analysis steps. + This should be a semicolon-delimited list. + Do not include newlines. + prompt.examples: + 1) Grow cells overnight on LB agar; 2) Harvest and suspend in IF-0 + range: ProtocolStep + multivalued: true + required: false + + plate_reader_types_of_data_collected: + description: "Data types collected (OD, respiration, formazan formation, etc.)." + annotations: + prompt: | + Identify data types measured by the plate reader + (e.g. formazan, OD) in a semi-colon delimited list. + If none, 'Not provided'. + range: string + multivalued: true + required: false + + od_protocol: + description: "Protocol for measuring optical density." + annotations: + prompt: | + Describe OD measurement protocol (if relevant). + If not described, 'Not provided'. + range: string + required: false + + respiration_protocol: + description: "Protocol for measuring respiration or related metabolic activity." + annotations: + prompt: | + Describe the respiration or metabolic activity measurement protocol (e.g. formazan production). + If none, 'Not provided'. + range: string + required: false + + instrument_used: + description: "Instruments or equipment used for the experiment." + annotations: + prompt: | + Extract the name(s) of the instrument/equipment used + (e.g., OmniLog system) in a semicolon-delimited list. + If none, 'Not provided'. + prompt.examples: OmniLog Phenotype MicroArray System; OmniLog Turbidimeter + range: string + multivalued: true + required: false + + analysis_software: + description: "Software used to analyze or interpret data." + annotations: + prompt: | + Extract any software or platforms used for data analysis + in a semicolon-delimited list. + prompt.examples: OmniLog Parametric Analysis software v1.20.02 + range: string + multivalued: true + required: false + + pm01_results: + description: > + If this experiment utilized the PM01 Biolog plate, captures + each well's result (positive, negative, or not provided). + annotations: + prompt: > + If PM01 results are reported, list them as ` = `. + For example, `2-Aminoethanol = positive`. + Use `positive`, `negative`, or `not provided` to indicate the result. + One entry per well. If no PM01 data are mentioned, omit. + range: BiologPM01WellResult + multivalued: true + required: false + + BiologPM01WellResult: + description: "A single well result from the PM01 Biolog plate." + attributes: + chemical: + description: "The chemical compound tested in the PM01 well." + range: PM01ChemicalEnum + required: true + result: + description: "Whether the strain/organism grew or reacted positively in this well." + range: PM01ResultEnum + required: false + + # Placeholder, to be expanded + Host: + is_a: NamedEntity + description: "Detail of a host organism or system." + attributes: + name: + description: "Name of the host organism or system." + annotations: + prompt: "The name of the host organism or system." + range: string + required: false + + # Placeholder, to be expanded + Microbe: + is_a: NamedEntity + description: "Information about a microbial species or strain." + attributes: + name: + description: "Name of a microbial taxon." + annotations: + prompt: "The name of the microbial species or strain." + range: string + required: false + + MicrobeGroup: + is_a: NamedEntity + description: "Information about a microbial group or taxon." + attributes: + name: + description: "Name of a microbial group or taxon." + annotations: + prompt: "The name of the microbial group." + range: string + required: false + + ExperimentalFactor: + is_a: NamedEntity + description: "A factor manipulated or measured in an experiment." + attributes: + placeholder_attribute: + description: "Placeholder attribute for ExperimentalFactor." + annotations: + prompt: | + Extract details about this experimental factor (e.g. concentration). + If absent, 'Not provided'. + range: string + required: false + + Trait: + is_a: NamedEntity + description: "An organism's physical trait measured or observed in an experiment." + id_prefixes: + - https + annotations: + prompt: | + A physical trait of an organism measured or observed in the study. + annotators: bioportal:metpo + + ProtocolStep: + description: "A step in an experimental protocol." + attributes: + step_description: + description: "A step in the experimental protocol." + annotations: + prompt: "Description of a step in the experimental protocol." + range: string + required: false + + culture_media: + description: "Culture media used in the step." + annotations: + prompt: | + The culture media used in the step, + if applicable. Do not include a value + for this field if not relevant. + range: string + required: false + + culture_temperature: + description: "Culture incubation temperature used in the step." + annotations: + prompt: | + The culture conditions used in the step, + if applicable, including temperature. + Do not include a value for this field if + not relevant. + range: string + required: false + + culture_duration: + description: "Culture incubation duration used in the step." + annotations: + prompt: | + The culture incubation duration used in the step, + if applicable, expressed as value and unit. + Do not include a value for this field if + not relevant. + range: string + required: false From 468d0d4a86e29b1413da9f6ebbddbf02809877c5 Mon Sep 17 00:00:00 2001 From: "marcin p. joachimiak" <4625870+realmarcin@users.noreply.github.com> Date: Tue, 15 Apr 2025 23:08:52 -0700 Subject: [PATCH 25/25] adding ethics and data governance module --- src/ontogpt/templates/data_sheets_schema.yaml | 1374 +++-------------- 1 file changed, 223 insertions(+), 1151 deletions(-) diff --git a/src/ontogpt/templates/data_sheets_schema.yaml b/src/ontogpt/templates/data_sheets_schema.yaml index 406e4905a..4d71730d9 100644 --- a/src/ontogpt/templates/data_sheets_schema.yaml +++ b/src/ontogpt/templates/data_sheets_schema.yaml @@ -1,1475 +1,547 @@ ---- -id: https://w3id.org/bridge2ai/data-sheets-schema -name: data-sheets-schema -title: data-sheets-schema -description: |- - A LinkML schema for Datasheets for Datasets. -license: MIT -see_also: - - https://bridge2ai.github.io/data-sheets-schema +id: http://w3id.org/ontogpt/datasheet +name: datasheet +title: Datasheet Extraction Template +description: >- + A template for extracting metadata about a dataset, + as defined by the Datasheets for Datasets model + (see https://github.com/bridge2ai/data-sheets-schema), + itself based on Gebru et al. (2021) + (https://cacm.acm.org/magazines/2021/12/256932-datasheets-for-datasets/abstract) +license: MIT prefixes: rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns# - biolink: https://w3id.org/biolink/ - csvw: http://www.w3.org/ns/csvw# - data_sheets_schema: https://w3id.org/bridge2ai/data-sheets-schema/ - datasets: https://w3id.org/linkml/report - dcat: http://www.w3.org/ns/dcat# - example: https://example.org/ - formats: http://www.w3.org/ns/formats/ - frictionless: https://specs.frictionlessdata.io/ linkml: https://w3id.org/linkml/ - mediatypes: https://www.iana.org/assignments/media-types/ - pav: http://purl.org/pav/ - schema: http://schema.org/ - sh: https://w3id.org/shacl/ - skos: http://www.w3.org/2004/02/skos/core# - void: http://rdfs.org/ns/void# -default_prefix: data_sheets_schema + d4d: https://w3id.org/ontogpt/datasheet + data_sheets_schema: https://w3id.org/bridge2ai/data-sheets-schema/ + B2AI_TOPIC: https://w3id.org/bridge2ai/b2ai-standards-registry/ + B2AI_SUBSTRATE: https://w3id.org/bridge2ai/b2ai-standards-registry/ + +default_prefix: d4d default_range: string imports: - linkml:types + - core -## TYPES ## - -types: - ror_identifier: - description: Identifier from Research Organization Registry. - typeof: uriorcurie - id_prefixes: - - ror - wikidata_identifier: - description: Identifier from Wikidata open knowledge base. - typeof: uriorcurie - id_prefixes: - - wikidata - -## SUBSETS ## -subsets: - Motivation: - description: >- - The questions in this section are primarily intended to encourage dataset - creators to clearly articulate their reasons for creating the dataset and - to promote transparency about funding interests. The latter may be - particularly relevant for datasets created for research purposes. - Composition: - description: >- - The questions in this section are intended to provide dataset consumers - with the information they need to make informed decisions about using the - dataset for their chosen tasks. Some of the questions are designed to - elicit information about compliance with the EU’s General Data Protection - Regulation (GDPR) or comparable regulations in other jurisdictions. - Collection: - description: >- - The questions in this section are designed to elicit information that may - help researchers and practitioners to create alternative datasets with - similar characteristics. - Preprocessing-Cleaning-Labeling: - description: >- - The questions in this section are intended to provide dataset consumers - with the information they need to determine whether the “raw” data has - been processed in ways that are compatible with their chosen tasks. - Uses: - description: >- - The questions in this section are intended to encourage dataset creators - to reflect on the tasks for which the dataset should and should not be - used. - Distribution: - description: >- - The questions in this section pertain to dataset distribution. - Maintenance: - description: >- - The questions in this section are intended to encourage dataset creators - to plan for dataset maintenance and communicate this plan to dataset - consumers. - -## CLASSES ## classes: - NamedThing: - description: >- - A generic grouping for any identifiable entity - slots: - - id - - name - - description - close_mappings: - - schema:Thing - - # Adapted from linkml Datasets schema - see - # https://github.com/linkml/linkml-model/blob/main/linkml_model/model/schema/datasets.yaml - Information: - abstract: true - description: Grouping for datasets and data files - close_mappings: - - schema:CreativeWork - slots: - - compression - - conforms_to - - conforms_to_class - - conforms_to_schema - - created_by - - created_on - - description - - doi - - download_url - - id - - issued - - keywords - - language - - last_updated_on - - license - - modified_by - - page - - publisher - - status - - title - - version - - was_derived_from - - # From linkml Datasets schema - see - # https://github.com/linkml/linkml-model/blob/main/linkml_model/model/schema/datasets.yaml - FormatDialect: - description: Additional format information for a file - attributes: - comment_prefix: - delimiter: - double_quote: - header: - quote_char: - - Person: - description: An individual human being. - is_a: NamedThing - attributes: - affiliation: - description: >- - The organization(s) to which the person belongs. - range: Organization - multivalued: true - email: - description: >- - The email address of the person. - range: string - - # TODO: align with the Bridge2AI Standards Schema equivalent class - Organization: - description: A collection of people acting in common interests. - is_a: NamedThing - attributes: - email: - description: >- - The email address of the organization. - range: string - ror_id: - description: Unique ROR identifier. - examples: - - value: ROR:02mp31p96 - range: ror_identifier - values_from: - - ROR - wikidata_id: - description: Unique Wikidata identifier. - examples: - - value: WIKIDATA:Q282186 - range: wikidata_identifier - values_from: - - WIKIDATA - - DatasetProperty: - # Individual properties have their own attributes - description: >- - Represents a single property of a dataset, or a set of related properties. - is_a: NamedThing - attributes: - used_software: - description: >- - What software was used as part of this dataset property? - range: Software - multivalued: true - - DatasetCollection: - aliases: - - file collection - - dataset collection - - data resource collection - tree_root: true - description: >- - A collection of related datasets, likely containing multiple files - of multiple potential purposes and properties. - exact_mappings: - - dcat:Dataset - close_mappings: - - dcat:Catalog - is_a: Information - attributes: - resources: - range: Dataset - multivalued: true - Dataset: - aliases: - - data resource - - data file - - data package - class_uri: dcat:Distribution - exact_mappings: - - schema:DataDownload - see_also: - - https://specs.frictionlessdata.io/data-resource description: >- A single component of related observations and/or information that can be read, manipulated, transformed, and otherwise interpreted. - is_a: Information - slots: - - bytes - - dialect - - encoding - - format - - hash - - md5 - - media_type - - path - - sha256 + tree_root: true + is_a: NamedEntity attributes: + title: + description: The title of the dataset. + description: + description: A brief description of the dataset. purposes: - range: Purpose + description: >- + For what purpose was the dataset created? multivalued: true tasks: - range: Task + description: >- + Was there a specific task in mind for the dataset's application? multivalued: true addressing_gaps: - range: AddressingGap + description: >- + Was there a specific gap that needed to be filled by creation of the dataset? multivalued: true creators: - range: Creator + description: >- + Who created the dataset (e.g., which team, research group) and on behalf + of which entity (e.g., company, institution, organization)? multivalued: true + range: Creator funders: - range: FundingMechanism + description: >- + Who funded the creation of the dataset? If there is an associated grant, + please provide the name of the grantor and the grant name and number. multivalued: true + range: FundingMechanism subsets: - range: DataSubset + description: >- + Are there recommended data splits (e.g., training, development/validation, + testing)? If so, please provide a description of these splits, explaining + the rationale behind them. multivalued: true - slot_uri: dcat:distribution - exact_mappings: - - schema:distribution + range: DataSubset instances: + description: >- + What do the instances that comprise the dataset represent (e.g., + documents, photos, people, countries)? range: Instance - multivalued: true anomalies: - range: DataAnomaly + description: >- + Are there any errors, sources of noise, or redundancies in the dataset? + If so, please provide a description. multivalued: true external_resources: - range: ExternalResource + description: >- + Is the dataset self-contained, or does it link to or otherwise rely on + external resources (e.g., websites, tweets, other datasets)? multivalued: true + range: ExternalResource confidential_elements: + description: >- + Does the dataset contain data that might be considered confidential (e.g., + data that is protected by legal privilege or by doctor patient + confidentiality, data that includes the content of individuals’ + non-public communications)? range: Confidentiality - multivalued: true content_warnings: + description: >- + Does the dataset contain data that, if viewed directly, might be + offensive, insulting, threatening, or might otherwise cause anxiety? range: ContentWarning - multivalued: true subpopulations: + description: >- + Does the dataset identify any subpopulations (e.g., by age, gender)? range: Subpopulation - multivalued: true sensitive_elements: + description: >- + Does the dataset contain data that might be considered sensitive in any + way (e.g., data that reveals race or ethnic origins, sexual orientations, + religious beliefs, political opinions or union memberships, or locations; + financial or health data; biometric or genetic data; forms of government + identification, such as social security numbers; criminal history)? range: SensitiveElement - multivalued: true acquisition_methods: + description: >- + How was the data associated with each instance acquired? range: InstanceAcquisition - multivalued: true collection_mechanisms: + description: >- + What mechanisms or procedures were used to collect the data? range: CollectionMechanism - multivalued: true sampling_strategies: - range: SamplingStrategy + description: >- + Does the dataset contain all possible instances or is it a sample? multivalued: true + range: SamplingStrategy data_collectors: + description: >- + Who was involved in the data collection process? range: DataCollector - multivalued: true collection_timeframes: + description: >- + Over what timeframe was the data collected? range: CollectionTimeframe - multivalued: true ethical_reviews: + description: >- + Were any ethical review processes conducted? range: EthicalReview - multivalued: true data_protection_impacts: + description: >- + Has an analysis of the potential impact of the dataset and its use on data subjects been conducted? range: DataProtectionImpact - multivalued: true preprocessing_strategies: + description: >- + Was any preprocessing of the data done? range: PreprocessingStrategy - multivalued: true cleaning_strategies: + description: >- + Was any cleaning of the data done? range: CleaningStrategy - multivalued: true labeling_strategies: + description: >- + Was any preprocessing/cleaning/labeling of the data done? range: LabelingStrategy - multivalued: true raw_sources: + description: >- + Was the “raw” data saved in addition to the preprocessed/cleaned/labeled data? range: RawData - multivalued: true existing_uses: - range: ExistingUse + description: >- + Has the dataset been used for any tasks already? multivalued: true + range: ExistingUse use_repository: + description: >- + Is there a repository that links to any or all papers or systems that use the dataset? range: UseRepository - multivalued: true other_tasks: - range: OtherTask + description: >- + What (other) tasks could the dataset be used for? multivalued: true + range: OtherTask future_use_impacts: + description: >- + Is there anything about the composition of the dataset or the way it was + collected and preprocessed/cleaned/labeled that might impact future uses? range: FutureUseImpact - multivalued: true discouraged_uses: + description: >- + Are there tasks for which the dataset should not be used? range: DiscouragedUse - multivalued: true distribution_formats: - range: DistributionFormat + description: >- + How will the dataset be distributed? multivalued: true + range: DistributionFormat distribution_dates: - range: DistributionDate + description: >- + When will the dataset be distributed? multivalued: true + range: DistributionDate license_and_use_terms: + description: >- + Will the dataset be distributed under a copyright or other IP license, + and/or under applicable terms of use? range: LicenseAndUseTerms ip_restrictions: + description: >- + Have any third parties imposed IP-based or other restrictions on the data + associated with the instances? range: IPRestrictions regulatory_restrictions: + description: >- + Do any export controls or other regulatory restrictions apply to the dataset + or to individual instances? range: ExportControlRegulatoryRestrictions maintainers: - range: Maintainer + description: >- + Who will be supporting/hosting/maintaining the dataset? multivalued: true + range: Maintainer errata: + description: >- + Is there an erratum? If so, please provide a link or other access point. range: Erratum - multivalued: true updates: + description: >- + Will the dataset be updated? range: UpdatePlan retention_limit: + description: >- + If the dataset relates to people, are there applicable limits on the retention + of the data? range: RetentionLimits version_access: + description: >- + Will older versions of the dataset continue to be supported/hosted/maintained? range: VersionAccess extension_mechanism: + description: >- + If others want to extend/augment/build on/contribute to the dataset, is there + a mechanism for them to do so? range: ExtensionMechanism is_deidentified: + description: >- + Is it possible to identify individuals from the dataset? range: Deidentification is_tabular: - range: string + description: Is the dataset in tabular format? + range: boolean DataSubset: description: >- A subset of a dataset, likely containing multiple files of multiple potential purposes and properties. - is_a: Dataset + is_a: NamedEntity attributes: is_data_split: description: >- Is this subset a split of the larger dataset, e.g., is it a set for model training, testing, or validation? - range: string + range: boolean is_subpopulation: description: >- Is this subset a subpopulation of the larger dataset, e.g., is it a set of data for a specific demographic? - range: string - - Software: - description: >- - A software program or library. - is_a: NamedThing - attributes: - version: - range: string - license: - range: string - url: - range: string - - Purpose: - description: >- - For what purpose was the dataset created? - is_a: DatasetProperty - attributes: - response: - range: string - in_subset: - - Motivation - - Task: - description: >- - Was there a specific task in mind for the dataset's application? - is_a: DatasetProperty - attributes: - response: - range: string - in_subset: - - Motivation - - AddressingGap: - description: >- - Was there a specific gap that needed to be filled by creation of the - dataset? - is_a: DatasetProperty - attributes: - response: - range: string - in_subset: - - Motivation + range: boolean Creator: - description: >- - Who created the dataset (e.g., which team, research group) and on behalf - of which entity (e.g., company, institution, organization)? - This may also be considered a team. - is_a: DatasetProperty + is_a: NamedEntity attributes: - principal_investigator: - range: Person affiliation: - # Note that creators have affiliations which may differ - # from that of their members + description: The affiliation of the creator. range: Organization - in_subset: - - Motivation FundingMechanism: - description: >- - Who funded the creation of the dataset? If there is an associated grant, - please provide the name of the grantor and the grant name and number. - is_a: DatasetProperty + is_a: NamedEntity attributes: grantor: - range: Grantor + description: The organization providing funding. + range: Organization grant: + description: The specific grant mechanism. range: Grant - in_subset: - - Motivation - - Grantor: - # Note this will have a name slot due to being a NamedThing - description: >- - What is the name and/or identifier of the organization providing - monetary support or other resources supporting creation of the dataset? - is_a: Organization Grant: - # Note this will have a name slot due to being a NamedThing - description: >- - What is the name and/or identifier of the specific mechanism providing - monetary support or other resources supporting creation of the dataset? - is_a: NamedThing + is_a: NamedEntity attributes: - grant_number: # TODO: formalize this more - probably not a URI though + grant_number: description: The alphanumeric identifier for the grant. range: string + Organization: + is_a: NamedEntity + Instance: - # Note this does not model the instance itself, but rather the - # description of each instance type or class. - description: >- - What do the instances that comprise the dataset represent (e.g., - documents, photos, people, countries)? - is_a: DatasetProperty + is_a: NamedEntity attributes: - representation: # TODO: align with Bridge2AI Standards Schema topics - range: string + data_topic: + description: What general topic does the data represent? + range: uriorcurie + values_from: + - B2AI_TOPIC instance_type: - description: >- - Are there multiple types of instances (e.g., movies, users, and - ratings; people and interactions between them; nodes and edges)? - range: string - data_type: # TODO: align with Bridge2AI Standards Schema types - description: >- - What data does each instance consist of? “Raw” data (e.g., unprocessed - text or images) or features? In either case, please provide a - description. + description: Are there multiple types of instances? range: string + data_substrate: + description: What data does each instance consist of? + range: uriorcurie + values_from: + - B2AI_SUBSTRATE counts: - description: >- - How many instances are there in total (of each type, if appropriate)? + description: How many instances are there in total? range: integer - label: # TODO: clarify what this means - it's fuzzy, and maybe redundant - description: >- - Is there a label or target associated with each instance? + label: + description: Is there a label or target associated with each instance? + range: boolean + label_description: + description: If there is a label, what pattern or format does it follow? range: string - sampling_strategies: - range: SamplingStrategy - multivalued: true - missing_information: - range: MissingInfo - multivalued: true - in_subset: - - Composition SamplingStrategy: - description: >- - Does the dataset contain all possible instances or is it a sample (not - necessarily random) of instances from a larger set? If the dataset is a - sample, then what is the larger set? Is the sample representative of the - larger set (e.g., geographic coverage)? If so, please describe how this - representativeness was validated/verified. If it is not representative of - the larger set, please describe why not (e.g., to cover a more diverse - range of instances, because instances were withheld or unavailable). - is_a: DatasetProperty + is_a: NamedEntity attributes: is_sample: - range: string - multivalued: true + range: boolean is_random: - range: string - multivalued: true + range: boolean source_data: range: string - multivalued: true is_representative: - range: string - multivalued: true + range: boolean representative_verification: range: string - multivalued: true why_not_representative: range: string - multivalued: true strategies: - description: >- - If the dataset is a sample from a larger set, what was the sampling - strategy (e.g., deterministic, probabilistic with specific sampling - probabilities)? + description: What was the sampling strategy? range: string - multivalued: true - in_subset: - - Composition - - Collection MissingInfo: - description: >- - Is any information missing from individual instances? If so, please - provide a description, explaining why this information is missing (e.g., - because it was unavailable). This does not include intentionally removed - information, but might include, e.g., redacted text. - is_a: DatasetProperty + is_a: NamedEntity attributes: missing: range: string - multivalued: true why_missing: range: string - multivalued: true - in_subset: - - Composition - - Relationships: - description: >- - Are relationships between individual instances made explicit (e.g., users’ - movie ratings, social network links)? If so, please describe how these - relationships are made explicit. - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Composition - - Splits: - description: >- - Are there recommended data splits (e.g., training, development/validation, - testing)? If so, please provide a description of these splits, explaining - the rationale behind them. - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Composition - - DataAnomaly: - description: >- - Are there any errors, sources of noise, or redundancies in the dataset? - If so, please provide a description. - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Composition ExternalResource: - description: >- - Is the dataset self-contained, or does it link to or otherwise rely on - external resources (e.g., websites, tweets, other datasets)? If it links - to or relies on external resources, a) are there guarantees that they - will exist, and remain constant, over time; b) are there official - archival versions of the complete dataset (i.e., including the external - resources as they existed at the time the dataset was created); c) are - there any restrictions (e.g., licenses, fees) associated with any of the - external resources that might apply to a dataset consumer? Please provide - descriptions of all external resources and any restrictions associated - with them, as well as links or other access points, as appropriate. - is_a: DatasetProperty + is_a: NamedEntity attributes: external_resources: range: string - multivalued: true future_guarantees: range: string - multivalued: true archival: - range: string - multivalued: true + range: boolean restrictions: range: string - multivalued: true - in_subset: - - Composition Confidentiality: - description: >- - Does the dataset contain data that might be considered confidential (e.g., - data that is protected by legal privilege or by doctor patient - confidentiality, data that includes the content of individuals’ - non-public communications)? - is_a: DatasetProperty + is_a: NamedEntity attributes: + confidential_elements_present: + range: boolean description: range: string - multivalued: true - in_subset: - - Composition ContentWarning: - description: >- - Does the dataset contain data that, if viewed directly, might be - offensive, insulting, threatening, or might otherwise cause anxiety? If - so, please describe why. - is_a: DatasetProperty + is_a: NamedEntity attributes: + content_warnings_present: + range: boolean warnings: range: string - multivalued: true - in_subset: - - Composition Subpopulation: - description: >- - Does the dataset identify any subpopulations (e.g., by age, gender)? If - so, please describe how these subpopulations are identified and provide a - description of their respective distributions within the dataset. - is_a: DatasetProperty + is_a: NamedEntity attributes: + subpopulation_elements_present: + range: boolean identification: range: string - multivalued: true distribution: range: string - multivalued: true - in_subset: - - Composition - # TODO: consider other ways to model this - is it just a boolean? - # or are we concerned about specific PHI subtypes? - # In practice, if data is Safe Harbor de-identified, that's one thing, - # but it could also be anonymized, and the standard for that isn't as - # explicit. See https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5977668/ Deidentification: - description: >- - Is it possible to identify individuals (i.e., one or more natural - persons), either directly or indirectly (i.e., in combination with other - data) from the dataset? - is_a: DatasetProperty + is_a: NamedEntity attributes: + identifiable_elements_present: + range: boolean description: range: string - multivalued: true - in_subset: - - Composition SensitiveElement: - description: >- - Does the dataset contain data that might be considered sensitive in any - way (e.g., data that reveals race or ethnic origins, sexual orientations, - religious beliefs, political opinions or union memberships, or locations; - financial or health data; biometric or genetic data; forms of government - identification, such as social security numbers; criminal history)? - is_a: DatasetProperty + is_a: NamedEntity attributes: + sensitive_elements_present: + range: boolean description: range: string - multivalued: true - in_subset: - - Composition InstanceAcquisition: - description: >- - How was the data associated with each instance acquired? Was the data - directly observable (e.g., raw text, movie ratings), reported by subjects - (e.g., survey responses), or indirectly inferred/derived from other data - (e.g., part-of-speech tags, model-based guesses for age or language)? If - the data was reported by subjects or indirectly inferred/derived from - other data, was the data validated/verified? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true was_directly_observed: - description: >- - Was the data directly observable (e.g., raw text, movie ratings)? - range: string + range: boolean was_reported_by_subjects: - description: >- - Was the data reported by subjects (e.g., survey responses)? - range: string + range: boolean was_inferred_derived: - description: >- - Was the data indirectly inferred/derived from other data (e.g., - part-of-speech tags, model-based guesses for age or language)? - range: string + range: boolean was_validated_verified: - description: >- - Was the data validated/verified? - range: string - in_subset: - - Collection + range: boolean CollectionMechanism: - description: >- - What mechanisms or procedures were used to collect the data (e.g., - hardware apparatuses or sensors, manual human curation, software programs, - software APIs)? How were these mechanisms or procedures validated? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Collection DataCollector: - description: >- - Who was involved in the data collection process (e.g., students, - crowdworkers, contractors) and how were they compensated (e.g., how much - were crowdworkers paid)? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Collection CollectionTimeframe: - description: >- - Over what timeframe was the data collected? Does this timeframe match the - creation timeframe of the data associated with the instances (e.g., - recent crawl of old news articles)? If not, please describe the timeframe - in which the data associated with the instances was created. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Collection EthicalReview: - description: >- - Were any ethical review processes conducted (e.g., by an institutional - review board)? If so, please provide a description of these review - processes, including the outcomes, as well as a link or other access - point to any supporting documentation. - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Collection - - DirectCollection: - description: >- - Did you collect the data from the individuals in question directly, or - obtain it via third parties or other sources (e.g., websites)? - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Collection - - CollectionNotification: - description: >- - Were the individuals in question notified about the data collection? If - so, please describe (or show with screenshots or other information) how - notice was provided, and provide a link or other access point to, or - otherwise reproduce, the exact language of the notification itself. - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Collection - - CollectionConsent: - description: >- - Did the individuals in question consent to the collection and use of - their data? If so, please describe (or show with screenshots or other - information) how consent was requested and provided, and provide a link - or other access point to, or otherwise reproduce, the exact language to - which the individuals consented. - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Collection - - ConsentRevocation: - description: >- - If consent was obtained, were the consenting individuals provided with a - mechanism to revoke their consent in the future or 8 for certain uses? If - so, please provide a description, as well as a link or other access point - to the mechanism (if appropriate). - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Collection - - DataProtectionImpact: - description: >- - Has an analysis of the potential impact of the dataset and its use on - data subjects (e.g., a data protection impact analysis) been conducted? - If so, please provide a description of this analysis, including the - outcomes, as well as a link or other access point to any supporting - documentation. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Collection PreprocessingStrategy: - description: >- - Was any preprocessing of the data done (e.g., - discretization or bucketing, tokenization, SIFT - feature extraction)? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Preprocessing-Cleaning-Labeling CleaningStrategy: - description: >- - Was any cleaning of the data done (e.g., - removal of instances, processing of missing values)? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Preprocessing-Cleaning-Labeling LabelingStrategy: - description: >- - Was any preprocessing/cleaning/labeling of the data done (e.g., - part-of-speech tagging)? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Preprocessing-Cleaning-Labeling RawData: - description: >- - Was the “raw” data saved in addition to the preprocessed/cleaned/labeled - data (e.g., to support unanticipated future uses)? If so, please provide - a link or other access point to the “raw” data. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Preprocessing-Cleaning-Labeling ExistingUse: - description: >- - Has the dataset been used for any tasks already? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Uses - - Maintenance UseRepository: - description: >- - Is there a repository that links to any or all papers or systems that use - the dataset? If so, please provide a link or other access point. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Uses OtherTask: - description: >- - What (other) tasks could the dataset be used for? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Uses FutureUseImpact: - description: >- - Is there anything about the composition of the dataset or the way it was - collected and preprocessed/cleaned/labeled that might impact future uses? - For example, is there anything that a dataset consumer might need to know - to avoid uses that could result in unfair treatment of individuals or - groups (e.g., stereotyping, quality of service issues) or other risks or - harms (e.g., legal risks, financial harms)? If so, please provide a - description. Is there anything a dataset consumer could do to mitigate - these risks or harms? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Uses DiscouragedUse: - description: >- - Are there tasks for which the dataset should not be used? - is_a: DatasetProperty - attributes: - description: - range: string - multivalued: true - in_subset: - - Uses - - ThirdPartySharing: - description: >- - Will the dataset be distributed to third parties outside of the entity - (e.g., company, institution, organization) on behalf of which the dataset - was created? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - in_subset: - - Distribution DistributionFormat: - description: >- - How will the dataset will be distributed (e.g., tarball on website, API, - GitHub)? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Distribution DistributionDate: - description: >- - When will the dataset be distributed? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Distribution LicenseAndUseTerms: - description: >- - Will the dataset be distributed under a copyright or other intellectual - property (IP) license, and/or under applicable terms of use (ToU)? If so, - please describe this license and/or ToU, and provide a link or other - access point to, or otherwise reproduce, any relevant licensing terms or - ToU, as well as any fees associated with these restrictions. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Distribution IPRestrictions: - description: >- - Have any third parties imposed IP-based or other restrictions on the data - associated with the instances? If so, please describe these restrictions, - and provide a link or other access point to, or otherwise reproduce, any - relevant licensing terms, as well as any fees associated with these - restrictions. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Distribution ExportControlRegulatoryRestrictions: - description: >- - Do any export controls or other regulatory restrictions apply to the - dataset or to individual instances? If so, please describe these - restrictions, and provide a link or other access point to, or otherwise - reproduce, any supporting documentation. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Distribution Maintainer: - description: >- - Who will be supporting/hosting/maintaining the dataset? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: - # range: CreatorOrMaintainerEnum range: string - multivalued: true - in_subset: - - Maintenance Erratum: - description: >- - Is there an erratum? If so, please provide a link or other access point. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Maintenance UpdatePlan: - description: >- - Will the dataset be updated (e.g., to correct labeling errors, add new - instances, delete instances)? If so, please describe how often, by whom, - and how updates will be communicated to dataset consumers (e.g., mailing - list, GitHub)? - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Maintenance RetentionLimits: - description: >- - If the dataset relates to people, are there applicable limits on the - retention of the data associated with the instances (e.g., were the - individuals in question told that their data would be retained for a - fixed period of time and then deleted)? If so, please describe these - limits and explain how they will be enforced. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Maintenance VersionAccess: - description: >- - Will older versions of the dataset continue to be - supported/hosted/maintained? If so, please describe how. If not, please - describe how its obsolescence will be communicated to dataset consumers. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: range: string - multivalued: true - in_subset: - - Maintenance ExtensionMechanism: - description: >- - If others want to extend/augment/build on/contribute to the dataset, is - there a mechanism for them to do so? If so, please provide a description. - Will these contributions be validated/verified? If so, please describe - how. If not, why not? Is there a process for communicating/distributing - these contributions to dataset consumers? If so, please provide a - description. - is_a: DatasetProperty + is_a: NamedEntity attributes: description: - range: string - multivalued: true - in_subset: - - Maintenance - -## SLOTS ## -slots: - - # The majority of these slots are adapted from - # the linkml Datasets schema - see - # https://github.com/linkml/linkml-model/blob/main/linkml_model/model/schema/datasets.yaml - - id: - identifier: true - description: the unique name of the dataset - slot_uri: dcterms:identifier - exact_mappings: - - schema:name - - name: - slot_uri: schema:name - - title: - description: the official title of the element - slot_uri: dcterms:title - - description: - description: human readable description of the information - slot_uri: dcterms:description - - language: - description: language in which the information is expressed - - publisher: - slot_uri: dcterms:publisher - range: uriorcurie - - issued: - slot_uri: dcterms:issued - range: string - - page: - slot_uri: dcat:landingPage - - dialect: - slot_uri: csvw:dialect - - bytes: - description: Size of the data in bytes. - range: integer - slot_uri: dcat:byteSize - - path: - close_mappings: - - frictionless:path - - download_url: - description: >- - URL from which the data can be downloaded. This is not the same as the - landing page, which is a page that describes the dataset. Rather, this - URL points directly to the data itself. - range: uri - slot_uri: dcat:downloadURL - exact_mappings: - - schema:url - close_mappings: - - frictionless:path - - format: - description: >- - The format of the data. This is not the same as the media type. - Rather, this is the format of the data in a more specific sense, - e.g., CSV, JSON, etc. - range: FormatEnum - slot_uri: dcterms:format - - compression: - description: >- - The compression format of the data. This is not the same as the media - type. Rather, this is the compression format of the data in a more - specific sense, e.g., zip, gzip, etc. - range: CompressionEnum - - encoding: - description: >- - The encoding of the data. This is not the same as the media type. - Rather, this is the encoding of the data in a more specific sense, - e.g., UTF-8, ASCII, etc. - range: EncodingEnum - - hash: - description: >- - The hash representation of the data, e.g., sha256, md5, etc. - Subtypes have their own slots. - range: string - sha256: - description: >- - The sha256 hash representation of the data. - is_a: hash - md5: - description: >- - The md5 hash representation of the data. - is_a: hash - - media_type: - description: >- - The media type of the data. This is not the same as the format. - Rather, this is the media type of the data in a more general sense, - e.g., text/csv, application/json, etc., though as it is defined here - the media type can be any string. - range: string - examples: - - value: text/csv - - value: application/json - slot_uri: dcat:mediaType - exact_mappings: - - frictionless:mediatype - - schema:encodingFormat - - conforms_to: - description: >- - The standard to which the data conforms. This is not the same as the - media type. Rather, this is the standard to which the data conforms - in a more specific sense, e.g., frictionless, schema.org, etc. - slot_uri: dcterms:conformsTo - range: uriorcurie - - # TODO: Align with Bridge2AI Standards Schema - conforms_to_schema: - description: >- - The schema to which the data conforms. This is not the same as the - media type. Rather, this is the schema to which the data conforms - in a more specific sense, and even more specific than the general - set of standards it conforms to. - is_a: conforms_to - exact_mappings: - - frictionless:schema - - conforms_to_class: - description: >- - The class in the schema to which the data object instantiates. - is_a: conforms_to - - doi: - description: >- - The Digital Object Identifier of the data, with the doi prefix. - range: uriorcurie - examples: - - value: "doi:10.48550/arXiv.2310.03666" - - profile: - description: >- - The frictionless data profile to which the data conforms. - range: uriorcurie - exact_mappings: - - frictionless:profiles - - keywords: - description: >- - Keywords associated with the data. These may be provided by - the data creator or assigned later in a manual or automated - manner. - singular_name: keyword - multivalued: true - range: string - slot_uri: dcat:keyword - exact_mappings: - - schema:keywords - - themes: - description: >- - Themes associated with the data. These may be provided by - the data creator or assigned later in a manual or automated - manner. - singular_name: theme - multivalued: true - range: uriorcurie - slot_uri: dcat:theme - - created_by: - # range: CreatorOrMaintainerEnum - range: string - description: Agent that created the element - slot_uri: pav:createdBy - multivalued: true - - created_on: - range: string - description: Date and Time at which the element was created - slot_uri: pav:createdOn - - last_updated_on: - range: string - description: Date and Time at which the element was last updated - slot_uri: pav:lastUpdatedOn - - modified_by: - # range: CreatorOrMaintainerEnum - range: string - description: agent that modified the element - slot_uri: oslc:modifiedBy - - status: - range: uriorcurie - description: Status of the element in terms of its maturity or life cycle - slot_uri: bibo:status - examples: - - value: "bibo:draft" - - license: - description: license for the data - slot_uri: dcterms:license - exact_mappings: - - frictionless:licenses - - version: - description: particular version of schema - slot_uri: pav:version - exact_mappings: - - schema:version - - dcterms:hasVersion - - was_derived_from: - slot_uri: prov:wasDerivedFrom - description: >- - A derivation is a transformation of an entity into another, an update - of an entity resulting in a new one, or the construction of a new entity - based on a pre-existing entity.@en - -## ENUMS ## -enums: - - CreatorOrMaintainerEnum: - description: >- - The entity responsible for maintaining a dataset. - permissible_values: - Person: - Organization: - - - # These enums are adapted from - # the linkml Datasets schema - see - # https://github.com/linkml/linkml-model/blob/main/linkml_model/model/schema/datasets.yaml - - MediaTypeEnum: - exact_mappings: - - dcterms:MediaType - permissible_values: - csv: - meaning: mediatypes:text/csv - rdf-xml: - meaning: mediatypes:application/rdf+xml - - FormatEnum: - permissible_values: - JSON-LD: - meaning: formats:JSON-LD - N3: - meaning: formats:N3 - N-Triples: - meaning: formats:N-Triples - N-Quads: - meaning: formats:N-Quads - LD Patch: - meaning: formats:LD_Patch - Microdata: - meaning: formats:microdata - OWL XML Serialization: - meaning: formats:OWL_XML - OWL Functional Syntax: - meaning: formats:OWL_Functional - OWL Manchester Syntax: - meaning: formats:OWL_Manchester - POWDER: - meaning: formats:POWDER - POWDER-S: - meaning: formats:POWDER-S - PROV-N: - meaning: formats:PROV-N - PROV-XML: - meaning: formats:PROV-XML - RDFa: - meaning: formats:RDFa - RDF/JSON: - meaning: formats:RDF_JSON - RDF/XML: - meaning: formats:RDF_XML - RIF XML Syntax: - meaning: formats:RIF_XML - SPARQL Results in XML: - meaning: formats:SPARQL_Results_XML - SPARQL Results in JSON: - meaning: formats:SPARQL_Results_JSON - SPARQL Results in CSV: - meaning: formats:SPARQL_Results_CSV - SPARQL Results in TSV: - meaning: formats:SPARQL_Results_TSV - Turtle: - meaning: formats:Turtle - TriG: - meaning: formats:TriG - YAML: - JSON: - - CompressionEnum: - permissible_values: - GZIP: - TAR: - TARGZIP: - ZIP: - - EncodingEnum: - permissible_values: - ASCII: - Big5: - EUC-JP: - EUC-KR: - EUC-TW: - GB2312: - HZ-GB-2312: - ISO-2022-CN-EXT: - ISO-2022-CN: - ISO-2022-JP-2: - ISO-2022-JP: - ISO-2022-KR: - ISO-8859-10: - ISO-8859-11: - ISO-8859-13: - ISO-8859-14: - ISO-8859-15: - ISO-8859-16: - ISO-8859-1: - ISO-8859-2: - ISO-8859-3: - ISO-8859-4: - ISO-8859-5: - ISO-8859-6: - ISO-8859-7: - ISO-8859-8: - ISO-8859-9: - KOI8-R: - KOI8-U: - Shift_JIS: - UTF-16: - UTF-32: - UTF-7: - UTF-8: + range: string \ No newline at end of file