Abstract

   The use of biological inputs is an interesting approach to optimize
   crop production and reduce the use of chemical inputs. Understanding
   the chemical communication between bacteria and plants is critical to
   optimizing this approach. Recently, we have shown that Sphingomonas
   (S.) sediminicola can improve both nitrogen supply and yield in pea.
   Here, we used biochemical methods and untargeted metabolomics to
   investigate the chemical dialog between S. sediminicola and pea. We
   also evaluated the metabolic capacities of S. sediminicola by metabolic
   profiling. Our results showed that peas release a wide range of
   hexoses, organic acids, and amino acids during their development, which
   can generally recruit and select fast-growing organisms. In the
   presence of S. sediminicola, a more specific pattern of these molecules
   took place, gradually adapting to the metabolic capabilities of the
   bacterium, especially for pentoses and flavonoids. In turn, S.
   sediminicola is able to produce several compounds involved in cell
   differentiation, biofilm formation, and quorum sensing to shape its
   environment, as well as several molecules that stimulate pea growth and
   plant defense mechanisms.

   Keywords: Sphingomonas sediminicola, pea, root exudates, plant–bacteria
   interaction, molecular dialogue plant–bacteria

1. Introduction

   Plant growth and productivity are closely associated with rhizospheric
   bacteria, such as plant growth-promoting rhizobacteria (PGPR), in a
   reciprocal relationship that benefits both partners [[40]1,[41]2].
   These bacteria can live freely in the rhizosphere of plants or be
   directly associated with plants in a symbiotic relationship. Free PGPR
   contributes to plant health and development by producing plant growth
   regulators [[42]3], increasing the availability of nutrients and water,
   and reducing the effects of diseases, pests, and environmental
   stressors. They can also increase the amount of soil organic matter,
   improve soil texture and structure, and promote micronutrient uptake
   [[43]4,[44]5,[45]6]. Symbiotic PGPR such as those of the order
   Hyphomicrobiales (=Rhizobiales) are useful for agriculture because they
   are able to form nodules on the roots of legumes in which the
   atmospheric nitrogen (N[2]) fixed by the bacteria is converted into
   ammonia for the plant [[46]7].

   The interaction between plants and bacteria requires complex chemical
   communication, which can be viewed as a conversation between the two
   parties in which plants and bacteria exchange chemical signals to
   influence their growth and development [[47]8,[48]9]. First, plants and
   bacteria can sense each other’s chemical signals through specific
   receptors [[49]10,[50]11], which usually enables the recruitment of
   beneficial bacteria by the plant. Plants and bacteria can respond to
   chemical signals by altering their growth, metabolism, and functions
   that allow bacteria to colonize plant roots. This also affects the
   production of chemical signals, forming a feedback loop that leads to
   an equilibrium beneficial to both organisms [[51]12,[52]13]. PGPRs
   recruitment by plants requires a significant amount of energy to
   attract and select bacteria in the rhizosphere [[53]14,[54]15,[55]16].
   Plant roots release a variable and diverse set of compounds, including
   carbohydrates, amino acids, organic acids, and secondary metabolites
   [[56]17,[57]18,[58]19]. The compounds released by the roots are an
   important nutrient source for soil bacteria and have an attractive
   effect [[59]20]. In turn, PGPR produces compounds that promote plant
   growth and health [[60]19,[61]21]. Various studies have shown that the
   microbial community associated with the root can evolve depending on
   the chemical composition of the root exudates, which also depends on
   the plant species and its stage of development [[62]11,[63]19,[64]22].
   For effective interaction, it is important that the bacteria are
   metabolically adapted to the chemical signals released by the plant via
   its root exudates so that they can be recruited and establish
   themselves in the root system.

   Agriculture based on the interaction between plants and bacteria
   promises to limit the use of chemical fertilizers that have negative
   effects on the environment and health acts [[65]23,[66]24]. Since the
   last decades, bacterial fertilizers containing free N[2]-fixing PGPR,
   such as Pseudomonas stutzeri, P. oryzihabitans, or Azospirillum
   brasilense, have been successfully used in agriculture as a strategy to
   improve plant growth in a sustainable way [[67]6,[68]25,[69]26,[70]27].
   This aspect is even more pronounced in legumes such as pea, which is
   normally thought to interact with Rhizobium species [[71]28].
   Therefore, the molecular dialogs during this interaction have been
   studied in detail [[72]29,[73]30]. These interactions, which depend on
   the specific characteristics of the pea species and the bacterial
   species involved, make it possible to predict the effectiveness of the
   use of this bio-input. Despite the importance of this association, it
   is important to note that in conventionally plowed soils where peas are
   rotated, the bacterial community may also be predominantly influenced
   by other bacterial species, such as Sphingomonas
   [[74]31,[75]32,[76]33,[77]34].

   Recently, we have shown that Sphingomonas sediminicola is also able to
   induce the formation of root nodules in peas and increase the
   production of plant biomass [[78]35]. This new interaction raises the
   question of chemical communication between the two partners and offers
   the possibility of optimizing the effect of bacterial bio-input on a
   pea plant. Therefore, we examined a wide range of compounds, including
   carbohydrates, carboxylic acids, amino acids, polyphenols, and
   flavonoids in hydroponic pea cultures inoculated and non-inoculated
   with S. sediminicola. We compared these results with the metabolic
   phenotyping of S. sediminicola based on its ability to degrade various
   carbon, nitrogen, phosphorus, and sulfur sources. To get an overview of
   the effect of S. sediminicola on peas, we also analyzed the compounds
   released by the bacterium.

2. Materials and Methods

2.1. Plant and Bacteria Culture

   Sphingomonas sediminicola (DSM-18106) was grown in R2A medium (VWR,
   Fontenay-sous-Bois, France) for 72 h at 30 °C under constant shaking at
   150 rpm. Pea (Pisum sativum, cv. Douce Provence, Jardiland, Paris,
   France) seeds were surface-sterilized with a 3.5% (v/v) bleach
   solution, cold-stratified for 48 h, and germinated in the dark on a 1%
   (w/v) agar medium at 21 °C. Five days after germination, etiolated
   seedlings were transferred to hydroponic system.

2.2. Hydroponic Growth Conditions

   Peas were cultivated in a homemade hydroponic system using an 8-L
   plastic pot filled with 6 L of sterile distilled water supplemented
   with 0.23 g L^−1 of Murashige and Skoog’s basal salt mixture [[79]36].
   Each device had an air pump with a flow rate of 78 L h^−1 to maintain
   the oxygen supply. Five plants were arranged per system. Substances
   released via pea root exudates were collected at the emergence of the
   first leave, second internode, and flowering by withdrawing 1 L of
   hydroponic solution at each stage and replacing it with 1 L of
   hydroponic solution. Half of the devices were inoculated with S.
   sediminicola to obtain a concentration of 10^6 CFU L^−1.

2.3. Extraction and Analysis of Hydroponic Compounds

   Samples of the hydroponic solution were filtered with Whatman paper (GE
   Healthcare, Chicago, IL, USA), and the chemical compounds in these
   solutions were selectively absorbed by Amberlite XAD-8 resin
   (Sigma-Aldrich, Taufkirchen, Germany), loaded onto the Chromabond C-18
   SPE columns (Macherey-Nagel, Düren, Germany), dropwise desorbed with
   pure methanol then concentrated using a rotary evaporator at 40 °C.

   An ultra-high performance liquid chromatography (UPLC) system (Waters
   Corp., Milford, MA, USA) was used for global analysis of the
   concentrated samples using MassLynx^TM software (v4.2, Waters Corp.,
   Milford, MA, USA). Global detection of biological information was
   performed using high-throughput G2Si High-definition mass spectrometry
   (Waters Q-TOF SYNAPT™, Waters Corp., Milford, MA, USA). Mobile phases
   consisted of a gradient of 0.1% (v/v) formic acid in Milli-Q water (A)
   and MeOH (B). The flow rate was maintained at 0.4 mL min^−1. Elution
   was performed at a gradient of 10% B for 1 min, then increased to 90%
   over 6 min and held for 3 min. Initial conditions were then restored
   over 1 min and held for 3 min for re-equilibration.

   Total polyphenol content was determined spectrometrically by the
   Folin–Ciocalteu method [[80]37] using tannic acid in the range of 0–100
   μg mL^−1 as a reference standard. The flavonoid assay [[81]38] was
   performed using quercetol (0.05 mg mL^−1) as the reference standard.
   D-glucose, D-fructose, and sucrose were determined enzymatically using
   the sucrose/D-glucose kit (R-Biopharm, Pfungstadt, Germany). D-xylose,
   L-arabinose, D-galactose, xylitol, and D-sorbitol were determined using
   specific enzymatic kits (Megazyme Assay Kits, Megazyme International
   Ireland Limited, Wicklow, Ireland).

   For the determination of amino acid content, hydroponic samples were
   concentrated 3-fold by ultrafiltration using a polyethersulfone
   membrane with a cut-off of 3 kDa (Vivaspsin^® 2 concentrator). The
   identification and quantification of amino acids were performed using a
   Kromasil C18—100 Ǻ-5 μm 4.6 × 250 mm column after derivation with
   o-phthalaldehyde (OPA). For derivatization, 50 μL of sample was added
   to 50 μL of borate buffer containing 0.25% (w/v) OPA and 5% (v/v)
   β-mercaptoethanol. An injection of 5 μL of this solution was performed.
   The proline content was determined using a ninhydrin assay [[82]39].

   Organic acid contents were determined by liquid chromatography with UV
   detector (HPLC-UV) method [[83]40]. After centrifugation at 12,000 rpm
   for 2 min, the supernatant of hydroponic samples was filtered to 0.2
   μm, and 50 μL of the filtrate was injected into the analytical system
   equipped with a C18 column (Luna, 5 μm Phenomenex). Organic acids were
   separated in an isocratic medium with a mobile phase consisting of 25
   mM phosphate buffer and methanol (99/1, v/v; pH 2.4). The detection of
   carboxylic acids was performed at 210 nm.

2.4. Microarray Plates for Phenotypic Characterization of S. sediminicola

   We used pre-configured Biolog Phenotype MicroArrays (Biolog Inc.,
   Harvard, CA, USA) to phenotype the metabolic potential of S.
   sediminicola against 190 carbon (PM1 and PM2), 95 nitrogen (PM3), 59
   phosphorus, and 35 sulfur (PM4) sources. According to the
   manufacturer’s instructions, bacteria were grown on an R2A agar plate,
   and colonies were harvested to achieve the required transmittance and
   mixed with a buffer containing IF-0 concentration 1.2X (Biolog Inc.)
   and the redox dye tetrazolium violet. For PM3 and PM4, 125 μL 2 M
   sodium succinate/200 μM ferric citrate was added to the mixture to
   regulate the pH of the suspension. Each plate was incubated in an
   Omnilog (Biolog Inc.) at 25 °C for 96 h. Three biological replicates
   per plate were performed. The phenotypic response of S. sediminicola to
   PM substrates was monitored by the color change in each well. Reduction
   of tetrazolium salt by bacterial dehydrogenases and reductases results
   in a purple formazan dye. Therefore, a color reaction indicates that
   the bacteria are actively metabolizing a substrate in the well, while
   the absence of a color change means that the bacteria cannot metabolize
   the substrate. The rate of color change in each well was monitored
   every 15 min at 490 nm and stored in Omnilog units generated by Biolog
   Data Analysis Software (v1.7, Biolog Inc.). For each source, a color
   intensity greater than 500 Omnilog units were considered positive in at
   least two of three replicates.

2.5. Untargeted Metabolomics of Sphingomonas sediminicola Culture Medium

   The analysis was performed with the supernatant obtained by
   centrifugation of a culture medium in which Sphingomonas sediminicola
   had been grown, reaching a concentration of 10^6 CFU mL^−1. This
   untargeted metabolomic analysis was performed by Beijing Genomics
   Institute (BGI, Tai Po, Hong Kong). The culture was grown in R2A
   medium, according to the method described by Dunn et al. [[84]41].
   Metabolite extracts were analyzed on a Waters 2D UPLC (Waters, Milford,
   MA, USA), coupled to a Q-Exactive mass spectrometer (Thermo Fisher
   Scientific, Waltham, MA, USA) with a heated electrospray ionization
   (HESI) source and controlled by the Xcalibur 2.3 software program
   (Thermo Fisher Scientific, Waltham, MA, USA). Chromatographic
   separation was performed on a Waters ACQUITY UPLC BEH Amide column (1.7
   μm, 2.1 mm × 100 mm, Waters, Milford, MA, USA) at 30 °C.

   For the untargeted metabolomics data, the mass spectrometry raw data
   collected by LC-MS/MS were imported into Compound Discoverer 3.1
   (Thermo Fisher Scientific, Waltham, MA, USA) for data processing. The
   identification of metabolites is a combined result of BMDB (BGI
   Metabolome Database), mzCloud, and ChemSpider (HMDB, KEGG, LipidMaps)
   databases. Main parameters of metabolite identification: Precursor Mass
   Tolerance < 5 ppm, Fragment Mass Tolerance < 10 ppm, RT Tolerance < 0.2
   min. The results of the Compound Discoverer 3.1 export were imported
   into metaX [[85]42] for data preprocessing. A total of 4563 and 1851
   compounds were detected in the positive (pos) and negative (neg) modes,
   respectively. After preprocessing, 1487 (pos) and 869 (neg) of these
   compounds were identified.

2.6. Data Analysis

   To compare the composition of the compounds present in the hydroponic
   solutions, we used the R package agricolae [[86]43] and a
   Kruskal–Wallis nonparametric one-way analysis of variance followed by a
   Conover–Iman post hoc test if significant and Holm’s p-adjust method
   for multiple comparisons.

   Differential metabolites screening conditions were performed with a
   Fold-Change (FC) ≥ 1.2 or ≤0.83 and a p-value < 0.05 obtained through
   Student’s t-test and corrected for false discovery rate (FDR). Cluster
   analysis of differential metabolites was performed with log[2]
   transformation and z-score normalization (zero-mean normalization). The
   clustering algorithm used hierarchical cluster based on Euclidean
   distance. In the bubble plots for pathway enrichment analysis, the
   enrichment factor (RichFactor) on the X-axis is the number of
   differential metabolites annotated for the pathway divided by all
   identified metabolites annotated for the pathway. The larger the value,
   the greater the proportion of differential metabolites annotated for
   the pathway. The dot size represents the number of differential
   metabolites annotated for that metabolic pathway. Volcano plot based on
   log[2]FC and the Heatmap of the differential metabolites were generated
   using the R packages. All statistical analyzes were performed using R
   software (v4.2.2, The R Foundation, [87]https://www.r-project.org/,
   accessed on 13 February 2023).

3. Results

3.1. The Composition of the Hydroponic Solution Is Influenced by S.
sediminicola and the Stage of Pea Development

   The compounds present in the hydroponic solution samples were
   structured around two axes of PCA ordination, which accounted for 60%
   of the variance between samples ([88]Figure 1). Overall, axis 1 is
   related to bacterial inoculation, while axis 2 is related to the stage
   of pea development.

Figure 1.

   [89]Figure 1
   [90]Open in a new tab

   Principal component analysis (PCA) of polyphenol, flavonoid, amino
   acid, and carbohydrate content in hydroponic solutions of peas
   inoculated or not with Sphingomonas sediminicola at the time of first
   leaf emergence, second internode, and flowering.

   At the first leaf emergence stage, the composition of the hydroponic
   solution was very poor in polyphenols and flavonoids ([91]Table S1).
   Serine, glycine, and alanine were the major amino acids. Citrate and
   especially acetate characterized the organic acid content, while
   fructose, glucose, and sucrose determined the carbohydrate content.
   Between the emergence of the first leaf and the formation of the second
   internode, there was little change in the composition of the hydroponic
   solution. However, a change was observed in carbohydrates, with a sharp
   increase in fructose content and a decrease in sorbitol and xylitol
   content. At the flowering phase, the solutions were richer in
   tryptophan and especially in cysteine, which becomes the most abundant
   amino acid in the solutions. Additionally, citrate and fumarate also
   showed an increase during this phase, although not to the same extent
   as lactate, which reached levels similar to acetate. The content of
   carbohydrates also evolved, with fructose remaining proportionally
   dominant and an increase in the content of galactose ([92]Table S1).

   Inoculation of peas with S. sediminicola resulted in a hydroponic
   solution whose composition was almost the same as that of the
   non-inoculated peas until the emergence of the first leaf ([93]Figure
   1, [94]Table S1). However, this inoculation resulted in higher
   flavonoid, galactose, glucose, and xylitol contents than under the
   non-inoculated conditions. Between the emergence of the first leaf and
   the formation of the second internode, the composition of the
   hydroponic solutions of the inoculated peas evolved with an increase in
   the content of many amino acids. The content of polyphenols and
   flavonoids remained the same between the two stages, as did the content
   of carbohydrates. However, the content of glucose increased, while the
   content of galactose decreased ([95]Table S1). The composition of the
   hydroponic solution remained relatively stable between the formation of
   the second internode and flowering of the inoculated peas, with
   consistently higher levels of amino acids, flavonoids, and polyphenols
   than in the non-inoculated condition. UPLC analysis allowed the
   identification of different terpenoids and flavonoids in the solution,
   including pisatin ([96]Figure S1 and [97]Table S2). A clear difference
   between non-inoculated and inoculated conditions was also observed for
   other compounds. This was true for the lactate and galactose content,
   which were very low in the flowering stage of the inoculated peas.

3.2. S. sediminicola Had a Specialized Carbon Metabolism and a Generalist
Nitrogen Metabolism

   To determine whether the compounds released in root exudates can be
   metabolized by S. sediminicola, we examined the metabolic capacities of
   the bacteria ([98]Figure 2; [99]Table S3). Of the 333 substrates
   tested, 189 were degraded by S. sediminicola (>500 Omnilog Unit, OU),
   especially amino acids (55 of 77), nucleic
   acids/nucleotides/nucleosides (31/40), and some specific nitrogen
   sources such as nitrogen dioxide, nitrate, nitrite, and urea
   ([100]Figure 2, [101]Table S2). S. sediminicola showed a strong ability
   to metabolize certain pentoses such as L-lyxose (12116 OU), D-xylose
   (6753 OU), D-ribose (7269 OU), L-arabinose (8509 OU), in contrast to
   numerous hexoses such as D-glucose (664 OU), D-fructose (652 OU),
   D-mannose (842 OU), or D-galactose (634 OU). Similarly, di- and
   trisaccharides were degraded only very slightly or not at all.
   Carboxylic acids form a group of substrates that are only weakly
   degraded by the bacteria but with a preference for 5-keto-D-gluconic
   acid (7482 OU), L-tartaric acid (3834 OU), oxalomalinic acid (1882 OU),
   and for two aminocarboxylic acids, D,L-α-aminocaprylic acid (7369 OU)
   and ε-amino-N-caproic acid (10123 OU).

Figure 2.

   [102]Figure 2
   [103]Open in a new tab

   Metabolic potential of S. sediminicola against PM substrates. For each
   source, the level of substrate degradation, expressed in Omnilog units,
   was considered only if it reached a minimum value of 500 Omnilog units
   in at least two of the three replicates. For degraded substrates, the
   value given is the average of the three replicates; otherwise, a red
   cross indicates that the substrate was not degraded by S. sediminicola.

3.3. Untargeted Metabolomics

   Untargeted metabolomic analysis revealed that there are many
   metabolites whose levels changed drastically during S. sediminicola
   growth ([104]Figure 3a,b and [105]Figure S1). Indeed, the levels of 174
   and 352 compounds identified in the positive and negative modes,
   respectively, decreased during S. sediminicola growth ([106]Figure 3c,d
   and [107]Table S4). Among them were many amino acids (e.g., arginine,
   asparagine, aspartic acid, glutamic acid, lysine, methionine,
   phenylalanine, serine, and tryptophan), peptides, or carbohydrates such
   as ribose, mannose, and threose ([108]Table S1). The levels of 389
   (pos) and 316 (neg) identified molecules, including iminoquinoline,
   p-coumaroyl-homoserine lactone (pC-HSL) and 2-amino-glucopyranosyl
   mannitol, increased during bacterial growth. Additionally, higher
   contents of indole-3-carboxylic acid, methylimidazoleacetic acid, and
   indole-3-acetic acid were detected after S. sediminicola growth.
   9-(α-D-glucosyl)kinetin and kinetin were also found in this condition.
   Furthermore, N-acetylsphingosine, galactopinitol b, 8-methoxykynurenic
   acid, and monoglycosyl-N-acylsphingosine were present in the S.
   sediminicola medium culture. Overall, the bacteria caused an
   accumulation of compounds involved in amino acids, pentoses, and
   secondary metabolic pathways ([109]Figure 3e,f).

Figure 3.

   [110]Figure 3
   [111]Open in a new tab

   Untargeted metabolomics of Sphingomonas sediminicola in R2A medium.
   (a,b) Heatmap of differential metabolites in positive (a) and negative
   (b) ion modes. Each row corresponds to a differentially expressed
   metabolite, while each column represents a specific sample. The color
   gradient, ranging from green to red, indicates the abundance level of
   the differentially expressed metabolites, with green representing low
   abundance and red representing high abundance. (c,d) Volcano plot of
   differential metabolites in positive (a) and negative (b) ion modes.
   Each point represents a metabolite, horizontal coordinates indicate
   different multiplicities of differential metabolites (log[2] values),
   vertical coordinates indicate p-values (−log[10] values), grey
   indicates metabolites with no significant differences, red indicates
   up-regulated metabolites (Up), and green indicates down-regulated
   metabolites (Down). (e,f) bubble plots of metabolic pathway enrichment
   in positive (e) and negative (f) ion modes. The x-axis in the figure
   represents the ratio of differentiated metabolites in a specific
   metabolic pathway to the total number of identified metabolites within
   that pathway (RichFactor). A higher value indicates a greater
   enrichment of differential metabolites in the pathway. The color of the
   dots corresponds to the p-value obtained from the hypergeometric test.
   A smaller p-value indicates a more reliable test. The size of the dots
   represents the number of differential metabolites present in the
   respective metabolic pathway.

4. Discussion

   Plants release root exudates [[112]44,[113]45] to influence soil
   bacterial communities and attract certain bacteria to the rhizosphere
   [[114]46,[115]47]. However, the changes in root exudate composition
   induced by bacteria involved in plant growth remain poorly understood.

   During pea development, carbohydrates, particularly hexoses and
   sucrose, accumulate in the hydroponic solution. Fructose levels notably
   increase during the transition from leaf emergence to flowering, while
   sorbitol and xylitol become undetectable. These compounds, being simple
   sugars, can serve as a carbon and energy source for many bacteria
   [[116]48]. Their presence in the hydroponic solution may serve as a
   chemical communication strategy to recruit bacterial partners capable
   of utilizing labile carbon. Notably, these molecules have been
   described to effectively attract bacterial species such as Bacillus,
   Methylobacterium, or Pseudomonas [[117]22]. In addition, the abrupt
   increase in galactose content in pea hydroponic solution at flowering
   is consistent with Knee et al. [[118]49] but also with the chemotaxic
   character of the molecule. Indeed, this compound induces a chemotaxis
   effect on some PGPR, such as Pseudomonas with legume roots or Bacillus
   velezensis with cucumber roots [[119]49,[120]50].

   In the hydroponic solution of peas inoculated with S. sediminicola,
   hexose contents followed similar trends until the formation of the
   second internode. Compared with the non-inoculated peas, higher
   pentose, and lower galactose contents were observed under the
   conditions with S. sediminicola. This indicates carbohydrate
   communication adapted to S. sediminicola, which fits with the more
   pronounced metabolic preference of the bacterium for pentoses compared
   to hexoses and the predominant use of these compounds in its culture
   medium. Pentoses, like hexoses, are labile carbon sources, but their
   degradation may be more complex than that of hexoses because the
   enzymes required for their degradation may be less abundant in soil
   microorganisms [[121]51]. In addition, pentoses may be incorporated
   into complex polymers such as cellulose and hemicellulose, which are
   important components of fresh soil organic matter [[122]52]. In this
   case, their degradation may be slower and require the activity of
   specialized microorganisms such as S. sediminicola [[123]32].

   The hydroponic solution of non-inoculated peas contains high
   concentrations of organic acids, particularly acetic, citric, lactic,
   and furamic acids. Acetic and citric acids are generally released by
   plants to increase nutrient uptake, such as phosphorus [[124]53],
   manganese [[125]54], iron, and zinc [[126]55]; stimulate biofilm
   production and motility of some PGPR [[127]56,[128]57,[129]58]; induce
   nitrogen-fixing bacteria [[130]59], Rhizobium IC3109 [[131]58] and
   acid-forming bacteria (Acetobacter and Gluconacetobacter [[132]60]); or
   stimulate bacteria such as Pseudomonas, which is capable of producing
   plant growth hormones and protecting plants from disease [[133]61].
   Citric acid has also been shown to recruit phosphate-solubilizing
   bacteria (Pseudomonas putida [[134]62]) or symbiotic bacteria
   (Burkholderia cepacia and Rhizobium leguminosarum
   [[135]58,[136]63,[137]64,[138]65]). Similarly, lactic acid released
   from the roots of legumes has been shown to attract bacteria such as
   Lactobacillus and Pediococcus [[139]66,[140]67]. This organic acid is
   also known for its antimicrobial properties, which can help the plant
   against pathogens in the rhizosphere [[141]68,[142]69]. Furamic acid
   was also present in the hydroponic solution. This acid plays an
   important role in recruiting plant-friendly rhizobacteria, symbiotic
   nitrogen-fixing bacteria, phosphate-solubilizing bacteria, or bacteria
   that produce antimicrobial compounds that protect plants from soil
   pathogens [[143]61,[144]70,[145]71]. Therefore, pea releases a wide
   range of organic acids that serve as a broad-spectrum attractant for
   rhizosphere bacteria to entice any bacteria capable of perceiving and
   metabolizing these compounds, potentially including beneficial bacteria
   that promote plant growth. In the presence of S. sediminicola, the
   hydroponic solution of peas also contains the same organic acids,
   except for lactic acid. The presence of these organic acids in the
   hydroponic solution of peas inoculated with S. sediminicola is
   consistent with the metabolism of the bacteria, which can utilize
   acetic and furamic acids but not lactic acid. ([146]Table S3). Other
   organic acids, such as tartaric acid and butyric acid, which are highly
   metabolized by S. sediminicola and involved in regulating
   plant–microorganism communication, could be further quantified
   [[147]63,[148]72].

   In the later growth stages, plants secrete amino acids and polyphenols,
   which shape bacterial communities [[149]18,[150]73]. Amino acids serve
   as a nitrogen source for bacteria [[151]74], leading to competition for
   nutrient sources among them [[152]75]. Polyphenols and flavonoids also
   act as nutrient sources and chemoattractants for bacteria involved in
   infections or symbiotic relationships, such as Agrobacterium
   tumefaciens or rhizobia [[153]73,[154]76]. Previous studies have shown
   that Pseudomonas fluorescens and P. aeruginosa induce the secretion of
   specific phenolic acids and increase the total content of polyphenols
   at different plant growth stages, especially in advanced stages of
   chickpea plants [[155]77]. Our study indicates that the hydroponic
   solution of peas inoculated with S. sediminicola is enriched with amino
   acids and polyphenols from the formation of the second internode to
   flowering. This suggests that these compounds are a response of the
   peas to the presence of the bacteria and may affect the bacterial
   community structure. In addition, bacteria capable of metabolizing a
   wide range of amino acids from root exudates have a selective advantage
   in the plant rhizosphere [[156]61,[157]78]. The metabolism of S.
   sediminicola exhibits a high affinity for amino acid substrates and
   degrades nearly 80% of all such substrates, which also explains the
   large number of amino acids used by the bacteria in its culture medium.

   Flavonoids are plant secondary metabolites released into the
   rhizosphere, playing a crucial role in chemical communication with
   rhizosphere bacteria [[158]79]. They act as chemoattractants for
   bacteria involved in symbiotic interactions with the plant, such as
   rhizobia [[159]80]. Additionally, flavonoids stimulate biofilm
   production, promoting bacterial colonization of roots, which can lead
   to better plant growth and health [[160]81]. Interestingly, the
   hydroponic solution of peas inoculated with S. sediminicola contained
   flavonoids, which were not detected in the non-inoculated condition.
   Moreover, flavonoid content increased with pea development. Some
   identified flavonoids in the solution corresponded to pisatin, a plant
   phytoalexin with an antimicrobial activity produced by plants in
   response to infection or stress [[161]82,[162]83,[163]84]. Thus, the
   induction of pisatin production in the presence of S. sediminicola can
   be considered a pea defense response. It is noteworthy that S.
   sediminicola has the ability to form nodules on pea roots during
   nitrogen stress, aligning with observations in pea-Rhizobium
   interaction and the regulatory role of isoflavonoids in nodulation
   [[164]35,[165]85,[166]86]. Fragmentation analysis of other peaks from
   biochemical analysis would help in the identification of other
   flavonoid compounds.

   During their growth, rhizospheric bacteria produce various molecules
   for their own growth, including galactoside [[167]87]. Galactopinitol b
   was among the galactosides detected in the culture medium after the
   growth of S. sediminicola. These galactosides serve as carbon and
   energy sources for the bacteria, facilitating their growth in the
   rhizosphere. In addition, certain rhizobia can synthesize
   galactopinitol b, which may play a role in establishing symbiotic
   relationships with plants [[168]88,[169]89,[170]90]. In the culture
   medium, we also detected the presence of various components involved in
   cell differentiation, biofilm formation, and quorum sensing. Thus, the
   bacterial culture medium was enriched with
   methyl-4-hydroxy-6-methyl-2-pyrone as well as p-coumaroyl-homoserine
   lactone (pC-HSL). These compounds are quorum-sensing molecules,
   facilitating communication and coordination among bacteria to regulate
   their collective behavior [[171]91,[172]92]. In the same manner,
   monoglycosyl-N-acylsphingosine, which is involved in biofilm formation,
   may also regulate plant growth and modulate the plant immune system
   [[173]93,[174]94]. Interestingly, enrichment of the R2A medium with
   auxin-associated molecules (indole-3-carboxylic acid,
   methylimidazoleacetic acid, and indole-3-acetic acid) was also
   observed. Therefore, S. sediminicola might affect plant growth and
   development and auxin homeostasis in roots via these molecules
   [[175]95]. Thus, it would be interesting to investigate whether the
   presence of S. sediminicola at the root level affects plant
   development, root system architecture, plant defense mechanisms, but
   also the organization of microbial communities in the rhizosphere.

   Overall, we provide new information on how pea plants modulate the
   composition of their root exudates to recruit bacteria from which they,
   and Sphingomonas sediminicola in particular, can benefit. We also
   describe for the first time the full metabolic and catabolic potential
   of this bacterium. In future experiments, the inclusion of other
   rhizobia would provide valuable comparative data on the effects of
   different bacterial species. Such information would not only improve
   our understanding of the specific interactions between these bacteria
   and pea plants but also shed light on the broader context of
   plant–microbe associations in agricultural systems. It is important to
   note that our study was conducted under controlled hydroponic
   conditions that allowed precise control of nutrient availability and
   environmental factors. However, it would be equally interesting to
   study the chemical dialog between S. sediminicola and pea plants in an
   agricultural soil context. Soil conditions have a significant impact on
   microbial community structure and nutrient dynamics, which could
   influence the nature and extent of plant–microbe interactions. Future
   studies in agricultural soil systems would provide a better
   understanding of the practical implications and transferability of our
   results to real agricultural conditions.

5. Conclusions

   The chemical dialog underlying plant–microbe interactions are essential
   to understanding the use of microorganisms in agriculture. We have
   shown that pea plants employ a strategic mechanism to selectively
   recruit specific bacteria to their rhizosphere that have the potential
   to confer benefits to the plants. During its development, the plant
   modulates its root exudation by first releasing various generalist
   molecules that allow recruitment focused on fast-growing bacteria.
   Then, the exudation evolves further by releasing more specific
   compounds that are tuned to the metabolic potential of the targeted
   bacterial microbiota, such as Sphingomonas sediminicola. In return, the
   bacteria provide various services to the plant that allow it to develop
   better and cope with environmental conditions [[176]35].

   This beneficial interaction between the plant and a bacterium is a step
   towards understanding the interaction between plants and bacteria.
   Indeed, the plant in agricultural soil is in relationship with a large
   variety of organisms, some of which are beneficial, others less so.
   Therefore, it is important to understand how the plant adapts its
   recruitment strategies to select the most important partners. It is
   also important to identify the mechanisms that enable molecular and
   chemical dialog between partners. Understanding all these strategies is
   potentially an important lever for implementing more sustainable
   agriculture.

Acknowledgments