Arabidopsis thaliana exudates induce growth and proteomic changes in Gluconacetobacter diazotrophicus

Background Plants interact with a variety of microorganisms during their life cycle, among which beneficial bacteria deserve special attention. Gluconacetobacter diazotrophicus is a beneficial bacterium able to fix nitrogen and promote plant growth. Despite its biotechnological potential, the mechanisms regulating the interaction between G. diazotrophicus and host plants remain unclear. Methods We analyzed the response of G. diazotrophicus to cocultivation with Arabidopsis thaliana seedlings. Bacterial growth in response to cocultivation and plant exudates was analyzed. Through comparative proteomic analysis, G. diazotrophicus proteins regulated during cocultivation were investigated. Finally, the role of some up-accumulated proteins in the response G. diazotrophicus to cocultivation was analyzed by reverse genetics, using insertion mutants. Results Our results revealed the induction of bacterial growth in response to cocultivation. Comparative proteomic analysis identified 450 bacterial proteins, with 39 up-accumulated, and 12 down-accumulated in response to cocultivation. Among the up-accumulated pathways, the metabolism of pentoses and protein synthesis were highlighted. Proteins potentially relevant to bacterial growth response such as ABC-F-Etta, ClpX, Zwf, MetE, AcnA, IlvC, and AccC were also increased. Reverse genetics analysis, using insertion mutants, revealed that the lack of ABC-F-Etta and AccC proteins severely affects G. diazotrophicus response to cocultivation. Our data demonstrated that specific mechanisms are activated in the bacterial response to plant exudates, indicating the essential role of “ribosomal activity” and “fatty acid biosynthesis” in such a process. This is the first study to demonstrate the participation of EttA and AccC proteins in plant-bacteria interactions, and open new perspectives for understanding the initial steps of such associations.


INTRODUCTION
The interaction between bacteria and plants involves complex recognition and signaling mechanisms that activate specific responses (Dennis, Miller & Hirsch, 2010; Plant-Bacteria cocultivation Ten A. thaliana seedlings were grown in vitro for 10 days in Petri dishes containing 18 mL of half-strength MS medium plus 0.5% sucrose without the addition of hormones. After reach O.D. 600 1.0, G. diazotrophicus wild-type and mutant strains cells were washed three times with 0.85% (v/v) NaCl to prevent carryover of spent growth medium. Then, ten 10-day-old A. thaliana seedlings were inoculated with G. diazotrophicus wild-type and mutant strains at a final concentration of 10 7 cells mL −1 , as described in the previous item (nine repetitions of each treatment, including control). Control treatment was performed by inoculation of G. diazotrophicus wild-type and mutant strains at a concentration of 10 7 cells mL −1 in MS medium without A. thaliana seedlings. Petri dishes were maintained at 23 C under an irradiance of 120 mol/photons m −2 s −1 for 24 h. Bacterial growth was quantified by optical density (O.D. 600 ) analysis.

Protein extraction
Three biological replicates of G. diazotrophicus cells were collected from cultures in MS medium (control) and from G. diazotrophicus/A. thaliana cocultivation, and were used for proteomic analysis. For this purpose, 6.5 mL of each cultivation medium sample was centrifuged to collect the bacteria. Protein extraction of all samples was performed as previously described by Passamani et al. (2017). Specifically, after discarding the supernatant, the pellets were resuspended in 300 mL of the TCA/acetone precipitation buffer (10% trichloroacetic acid in acetone with 20 mM dithiothreitol-DTT) and maintained under constant stirring for 60 min at 4 C. Samples were, then, maintained at −20 C for 60 min and centrifuged (30 min; 4 C; 12.000 g). The supernatants were discarded, and the pellets were rinsed three times with ice-cold acetone containing 20 mM DTT, followed by centrifugation (5 min; 4 C; 12,000 g). The supernatants were discarded, and the pellets were maintained at room temperature to dry. The pellets were resuspended in 300 mL of urea/thiourea extraction buffer (7M urea, 2M thiourea, 1% DTT, 2% Triton X-100, 5 mM pepstatin, 1 mM phenylmethanesulfonylfluoride-PMSF) and maintained under stirring until complete homogenization. The samples were centrifuged (15 min; 4 C; 12.000 g), and the supernatants were collected.
The total protein concentration of the samples was estimated using the 2-D Quant Kit (GE Healthcare, Amersham Place, Little Chalfont, UK) following the manufacturer's recommendations. The absorbance of each sample and a standard curve with bovine albumin serum (BSA, GE Healthcare, Amersham Place, Little Chalfont, UK) was determined in a Synergy 2 Multimode Reader (Biotek Instruments, Winooski, VT, USA) at 485 nm. Protein samples were stored in an ultra-freezer at −80 C.

Protein digestion
Protein extracts (100 mM) were precipitated in methanol/chloroform as described by Nanjo et al. (2012). Then, protein extracts were digested as previously described by Passamani et al. (2017). Specifically, pellets were resuspended in 25 mL of 0.2% (v/v) RapiGest surfactant (Waters, Milford, CT, USA). Samples were, then, vortexed rapidly and incubated in a heated mixer at 80 C for 15 min, and 2.5 µL of 100 mM DTT (Bio-Rad Laboratories, Hercules, CA, USA) was added. Subsequently, the samples were vortexed and incubate for 30 min at 60 C under constant shaking at 350 rpm. Then, 2.5 mL of 300 mM iodoacetamide (GE Healthcare, Piscataway, NJ, USA) was added in each sample. The samples were, then, vortexed rapidly and incubated for 30 min in the dark at room temperature. Subsequently, digestion was performed by adding 20 mL of 50 ng/mL trypsin solution (Promega, Madison, WI, USA) prepared in 50 mM ammonium bicarbonate buffer. The samples were, then, incubated overnight at 37 C. The precipitation of RapiGest was performed by adding 10 mL of 5% (v/v) trifluoroacetic acid, and the tubes were incubated for 30 min at 37 C, followed by centrifugation (30 min; 8 C; 15.000 g). Subsequently, tryptic digestions were desalted using Pierce C18 spin columns (Thermo Scientific, Waltham, MA, USA) according to the manufacturer's instructions. The eluted peptides were then vacuum dried and reconstituted in 50 mM ammonium bicarbonate plus 0.1% formic acid and adjusted to a final concentration of 2 mg mL −1 . The digested samples were then transferred to total recovery flasks (Waters, Milford, CT, USA).

LC-MS/MS analyses
LC-MS/MS analyses were performed as previously described by Passamani et al. (2017). Specifically, a liquid nanoACQUITY ultraperformance UPLC connected to a Q-TOF SYNAPT G2-Si HDMS (Waters, Milford, CT, USA) mass spectrometer was used for LC-MS-HDMS E analysis. The chromatographic step was performed by injecting two mL of the digested samples for sample normalization before the relative quantification of proteins. In the separation step, the samples (digested protein, two mg) were loaded into the nanoACQUITY UPLC 5 mm C18 (180 mm × 20 mm) column at five ml min −1 for 3 min and then into the nanoACQUITY HSS T3 1.8 mm analytical reversed-phase column (75 mm × 150 mm) at 400 nL min −1 . The temperature of the column was set to 45 C. For the elution of the peptides; a binary gradient was used: mobile phase A consisted of water (Tedia, Fairfield, OH, USA) and 0.1% formic acid (Sigma-Aldrich, St. Louis, MO, USA), and mobile phase B consisted of acetonitrile (Sigma-Aldrich, St. Louis, MO, USA) and 0.1% formic acid. Gradient elution was performed as follows: 7% B over 3 min, increased from 7 to 40% B at 90.09 min; then 40 to 85% B at 94.09 min; held constant at 85% until 98.09 min; decreased to 7% B at 100.09 min, and finally held steady at 7% B until the end of the run at 108.09 min. The mass spectrometer was operated in resolution mode (V mode) and positive mode with ionic mobility; with collision energy transfer from 19 to 45 V in high-energy mode; a voltage and capillary cone of 30 V and 2,800 V, respectively; and a source temperature of 70 C. In the TOF parameters, the scan time was set to 0.5 s in continuous mode, with a mass range of 50 to 2,000 Da. Human [Glu1]fibrinopeptide B (Sigma-Aldrich , St. Louis, MO, USA) at 100 fmol.mL −1 was used as an external calibrant, with acquisition performed every 30 s. Mass spectrum acquisition was performed by using MassLynx v4.0 software.

Proteomic data analysis
Spectral processing and database searching were made as previously described by Passamani et al. (2018), through ProteinLynx Global Server (PLGS; version 3.0.2) (Waters, Milford, CT, USA) and ISOQuant workflow software (Distler et al., 2014(Distler et al., , 2016. Specifically, the PLGS analysis was performed with the follows processing parameters: a low energy threshold of 150 (counts), a high energy threshold of 50, and an intensity threshold of 750. Additionally, the analysis was performed with the following parameters adjustment were made: two missed cleavages, minimum fragment ions per peptide equal to 3, minimum ions per peptide equal to 7, minimum peptides per protein equal to 2, fixed carbamidomethyl (C) modifications and variable modifications of oxidation (M) and phosphorylation (STY). The false discovery rate for the identification of peptides and proteins was established according to a maximum of 1%, with a minimum peptide length of six amino acids. Proteomic data were processed using the G. diazotrophicus RIOGENE proteome database (www.uniprot.org/proteomes/UP000001176).
After data analysis in ISOQuant, only the proteins that were present or absent (for single proteins) in all three biological replicates were considered in the differential abundance analysis. Data were analyzed using Student's t-test (two-tailed) as previously described by Leandro et al. (2019). Specifically, proteins with p-values < 0.05 were considered to be up-accumulated if the fold change (FC) was higher than 1.5 and down-accumulated if the FC was less than 0.667. Protein network analyses was performed using STRING database with confidence as the meaning of network edges and 0.700 of interaction score (high confidence) (www.string-db.org).

Sample preparation of the plant, bacterial and cocultivation exudates
Exudates from G. diazotrophicus and A. thaliana cultivations, and G. diazotrophicus/A. thaliana cocultivations were obtained, as illustrated in Fig. S1. Three biological replicates of each treatment were performed.
Arabidopsis thaliana seedlings were cultivated in Petri dishes containing 18 mL of half-strength MS medium at 23 C for 10 days. Plant exudates were collected by filtration of the cultivation medium with syringe filters (0.22 µm) and stored at −80 C.
In order to obtain G. diazotrophicus exudates, the bacteria were inoculated in Petri dishes containing 18 mL of half-strength MS medium and cultivated for 24 h at 23 C. When the bacterial cultures reached an O.D. 600nm = 1.0, the culture was transferred to centrifuge tubes and centrifuged at 8,000×g for 5 min. The supernatant was filtered with a syringe filter (0.22 µm) and stored at −80 C.
Exudates from the cocultivation of A. thaliana seedlings and G. diazotrophicus were obtained from cocultivation assays performed as previously described in item 2.3. After cocultivation, the medium was collected and centrifuged at 8,000×g for 5 min. The supernatant was filtered with a syringe filter (0.22 µm) and stored at −80 C.
Analysis of the G. diazotrophicus PA15 response to exudates Exudates from G. diazotrophicus, A. thaliana seedlings, or cocultivation, were previously obtained as described in "Sample Preparation of the Plant, Bacterial and Cocultivation Exudates." Such exudates were added to G. diazotrophicus cultivation, as shown in Fig. S1. G. diazotrophicus culture (O.D. 600nm = 1.0) produced as described in "G. diazotrophicus PA15 Culture Conditions" was inoculated in 50 mL Erlenmeyer flasks containing half-strength fresh MS medium and each exudate, as the following final proportion: 10% (G. diazotrophicus culture), 45% (fresh MS medium) and 45% (exudate). Control samples received 10% of G. diazotrophicus and 90% of half-strength fresh MS medium. All cultures had a final volume of 10 mL. The cultures were grown at 30 C, with constant agitation (250 rpm) in an orbital shaker. Bacterial growth was analyzed by optical spectrometry (O.D. 600 ). Three biological replicates were performed for each treatment.
Gene sequences were aligned with MEGA-X software version 10.1.7 (Kumar et al., 2018) using the ClustalW algorithm (default settings). MEGA-X was also used to select the best substitution models for phylogenetic analyses, and to generate a phylogenetic tree using the maximum likelihood (ML) method. The parameters selected for accC and ettA phylogenetic threes were the models of substitution gama time reversible (GTR) with the invariant site (I) and GTR with the gamma-distributed site, respectively, with 1,000 bootstraps.

Statistical analyses
The assays were performed with nine replicates in each condition. The growth rate data of G. diazotrophicus wild-type and mutant strains under the control and cocultivation treatments were subjected to a mean test (Tukey) at the 5% probability level to measure significance between treatments. Data analyses were performed by using GraphPad Prism v. 7.00.159 (GraphPad Software, La Jolla, CA, USA).

G. diazotrophicus growth increases in response to cocultivation with
A. thaliana G. diazotrophicus was co-cultivated with A. thaliana seedlings, and its growth was analyzed during the first 24 h of the association. Figure 1 shows that the presence of A. thaliana seedlings increased bacterial growth by approximately 45% after 24 h of cocultivation. These data suggest that plant-produced metabolites are beneficial to bacterial growth. Microscopy analyses did not show morphological differences in G. diazotrophicus in response to cocultivation with A. thaliana (Fig. S2). A. thaliana seedlings constitutively exude metabolites that increase G. diazotrophicus growth To verify whether the exudation of beneficial metabolites by A. thaliana seedlings is constitutive or is induced by the presence of G. diazotrophicus, exudates from G. diazotrophicus/A. thaliana cocultivation were compared to those produced by non-inoculated seedlings and by G. diazotrophicus cultured alone. Thus, MS media containing exudates from non-inoculated seedlings, from G. diazotrophicus cultures, and from G. diazotrophicus/A. thaliana cocultivation were obtained and added to new G. diazotrophicus cultures as shown in Fig. S1. As a control, G. diazotrophicus was grown in the absence of exudates. Figure 2 shows that exudates from non-inoculated A. thaliana (A. thaliana exudate) and G. diazotrophicus/A. thaliana cocultivation (Co-cultive exudate) resulted in a similar increase in bacterial growth. As expected, G. diazotrophicus cultured alone (G. diazotrophicus exudate) did not improve bacterial growth. These results revealed that A. thaliana seedlings constitutively exude metabolites that are beneficial for bacterial growth.
The proteomic profile of G. diazotrophicus alters in response to cocultivation with A. thaliana The main G. diazotrophicus proteins regulated during cocultivation with A. thaliana seedlings were investigated by comparative proteomic analysis. For this purpose, total protein extracts obtained from co-cultivated bacteria were compared to those of bacteria cultivated in the absence of plants. A total of 450 bacterial proteins were identified  (Table S1). Among these proteins, 51 differentially accumulated proteins (DAPs) were observed, with 39 proteins increased, and 12 proteins decreased (Table 1; Fig. 3A).

Protein networks regulated during cocultivation
The functional association networks of the regulated proteins were analyzed with the software String (version 10.5). The analyses were performed for both the up-accumulated and down-accumulated proteins using a confidence level of 0.7. Up-accumulated protein  analysis revealed two major protein groups: "pentose-phosphate pathways" and "ribosomal proteins" (Fig. 3B). When similar analyses were performed for down-accumulated proteins, no interaction networks between the proteins were identified (Fig. S3).

Main up-accumulated proteins
Proteins that are potentially relevant to plant-bacteria interactions were up-accumulated during the first 24 h of cocultivation. Among these proteins, four showed the greatest up-regulation: 5-methyltetrahydropteroyltriglutamate-homocysteine methyltransferase (MetE), which is involved in the synthesis of methionine; aconitate hydratase (AcnA), which is associated with the metabolism of tricarboxylic acids; energy-dependent translational throttle A (EttA), which is involved in the regulation of ribosomal activity, and may also be necessary for protection against antimicrobial compounds; ketol-acid reductoisomerase (IlvC), which is involved in amino acid metabolism (Table 1).
Other up-accumulated proteins included Zwf (glucose-6-phosphate 1-dehydrogenase) and Gnd (6-phosphogluconate dehydrogenase decarboxylating), which are related to the pentose-phosphate pathway (Table 1). Another protein that was increased, ClpX (ATPbinding subunit of the ClpXP protease), is involved in the degradation of unfolded proteins and is potentially related to bacterial quorum sensing (Table 1). Additionally, a key protein for de novo fatty acid biosynthesis, AccC (acetyl-CoA carboxylase biotin carboxylase subunit), and a protein involved with nutrients uptake (TonB-dependent receptor) were also up-accumulated (Table 1).

Down-accumulated proteins
Only 12 proteins were down-accumulated in response to cocultivation. Among these proteins, we highlight three transporters related to the import of sugars: OprB, a Sugar ABC transporter substrate-binding protein, and the D-xylose ABC transporter substrate-binding protein ( Table 1).

Interruption of ettA and accC impairs G. diazotrophicus growth-response to cocultivation with A. thaliana
Using reverse genetic analysis by mutagenesis approach, we checked whether specific proteins, up-accumulated in our proteomic analysis, are essential for G. diazotrophicus response to cocultivation with A. thaliana. Mutant strains defective for proteins potentially relevant for bacterial growth/multiplication were selected. G. diazotrophicus mutant strains defectives for EttA (ΔettA), AccC (ΔaccC), and Zwf (Δzwf) were co-cultivated with A. thaliana seedlings, and their growth was analyzed in comparison to wild-type. Figure 4 shows that the lack of Zwf did not affect G. diazotrophicus growth induction in response to cocultivation with A. thaliana. On the other hand, the lack of EttA and AccC implies the loss of growth induction response, suggesting a key role of these proteins in this process. Additionally, the growth of ΔaccC in the control treatment (MS) was significantly lower than the wild-type strain, suggesting that AccC is also important to bacterial growth under normal culture conditions.
In silico prediction suggests that ettA and accC are organized in operons in G. diazotrophicus genome Phylogenetic analysis results for both ettA and accC genes showed the formation of two distinct groups (Fig. 5). The sequences of G. diazotrophicus are closer related to G. xylinus and A. brasiliense. Although H. seropedicae is a PGPB as G. diazotrophicus, its gene sequences are closer to those from the phytopathogen P. syringae and the human pathogen E. coli (Fig. 5). Moreover, operon prediction results suggest that both ettA and accC are organized in operons in G. diazotrophicus genome. The operon that contains accC is composed of three genes (aroQ, accB, accC) in almost all the bacterial species analyzed, except E. coli, which presents only two genes in this operon (accB and accC) (Fig. 5A). In all bacterial species analyzed, accC is in the last position of the operon (Fig. 5A). So, the interruption in accC probably does not compromise the transcription of the other components of such operon.
In almost all the bacterial species analyzed, ettA is not located in an operon, except in G. diazotrophicus and G. xylinus (Fig. 5B). In these two species, ettA is located in an operon, upstream of a gene that codifies a GCN5-related N-acetyltransferase (GNAT) family protein (Fig. 5B). So, the interruption of ettA may compromise the production of a GNAT family protein in G. diazotrophicus.

DISCUSSION
The present work aimed to evaluate the response of G. diazotrophicus to cocultivation with A. thaliana seedlings with an emphasis on the main bacterial proteins regulated by the interaction. Our results revealed the induction of bacterial growth during cocultivation, suggesting a beneficial effect of metabolites exuded by the plants. The data also revealed Figure 4 Reverse genetics revealed G. diazotrophicus genes involved with response to A. thaliana cocultivation. Insertional mutants of G. diazotrophicus defectives in production of three proteins up-accumulated in our proteomic analysis were selected to perform cocultivation assay with A. thaliana seedlings. The growth of each strain was analyzed after 24 h of cocultivation. Different letters indicate significant differences from the control by the Tukey test at 5% probability (n = 9).
Full-size that the production of such plant-beneficial exudates is a constitutive process. Comparative proteomic analyses demonstrated the regulation of bacterial protein pathways that were mainly related to sugar metabolism (pentoses) and protein synthesis. Among the regulated proteins, MetE, AcnA, IlvC, EttA, Zwf, ClpX, and AccC deserve special attention. Cocultivation with A. thaliana seedlings resulted in increased growth of G. diazotrophicus (Fig. 1). Studies that support this finding have been carried out with G. diazotrophicus under cocultivation with sugarcane plants, where the induction of bacterial multiplication has been reported (Lery et al., 2011;dos Santos et al., 2010). The substances exuded by plant roots include sugars, amino acids, mucilage, flavonoids, organic acids, and volatile compounds capable of attracting bacteria and favoring their multiplication (Sasse, Martinoia & Northen, 2018). Important functions have been attributed to exuded sugars, which would serve as a source of energy for microorganisms (Dennis, Miller & Hirsch, 2010;Sasse, Martinoia & Northen, 2018;Badri & Vivanco, 2009). However, a sugar-rich culture medium (0.5% sucrose) was used in the present work. Nevertheless, the presence of exudates from A. thaliana seedlings positively modulated the growth of G. diazotrophicus, suggesting contributions of other beneficial metabolites to this process. The analysis of G. diazotrophicus growth in the presence of exudates from plants and from plant-bacteria co-cultivation demonstrated that the production of beneficial metabolites by plants is constitutive (Fig. 2). Jacoby, Martyn & Kopriva (2018) showed that A. thaliana roots secrete a large number of secondary metabolites that are used by bacteria. The amino acids valine, tryptophan, threonine, and glutamine, in addition to pyrimidine and purine derivatives including adenosine, cytidine, guanine, and uridine, are consumed by plant-associated bacteria (Jacoby, Martyn & Kopriva, 2018).
Proteomic analyses revealed changes in the protein profile of G. diazotrophicus in response to cocultivation, with 39 proteins classified as up-accumulated and 12 as down-accumulated. Among the most up-accumulated proteins was 5-methyltetrahydropteroyltriglutamatehomocysteine methyltransferase (MetE), which is involved in the conversion of homocysteine to methionine (Winzer, 2002). Additionally, MetE is related to the production of AI-2 quorum sensing (QS) molecule in different microorganisms (Winzer, 2002). However, similar to the rest of the Alphaproteobacteria, G. diazotrophicus lacks a LuxS homolog required for AI-2 synthesis (Rezzonico & Duffy, 2008), so the up-accumulation of MetE in our proteomic analysis should not be related to QS activity. Another up-accumulated protein was Aconitate hydratase (AcnA), which is associated with the metabolism of tricarboxylic acids and may protect bacterial cells from reactive oxygen species (Doi & Takaya, 2015). Ketol-acid reductoisomerase (IlvC), which was also up-accumulated, plays an essential role in the pathway of amino acid biosynthesis (Li et al., 2017). These three proteins were also detected among the most up-accumulated proteins of G. diazotrophicus during its cocultivation with sugarcane plants (Lery et al., 2011). Our results highlight the MetE, IlvC, and AcnA proteins as potentially relevant for the association of G. diazotrophicus with both monocot and dicot plants.
Approximately 23% of the up-accumulated proteins are ribosomal (RpsA, RpsB, RpsH, RplE, RplF, RplN, and RplW), indicating a positive effect of cocultivation on bacterial protein synthesis. In Escherichia coli, the genes coding for RpsH, RplE, RplF and RplN are in an operon called spc that codes for 11 ribosomal proteins, essentials to protein synthesis machinery (Aseev & Boni, 2011). RpsA interacts with the mRNA leader sequence during the formation of the translation initiation complex, acting directly in the regulation of such a process (Komarova et al., 2002). RpsA is also positively regulated in the endophytic bacterium H. seropedicae when exposed to extracts from sugarcane (Cordeiro et al., 2013). RplW is essential for bacterial multiplication and is associated with the "Trigger Factor" in ribosomes (Kramer et al., 2002;Tischendorf, Zeichhardt & Stöffler, 1974). Our data, therefore, suggest the activation of protein synthesis as a critical response to cocultivation, supporting the increased bacterial growth observed.
The proteomic analysis also revealed the up-accumulation of the ABC-F EttA protein that modulates bacterial ribosomes activity in an energy-dependent manner (Boël et al., 2014). In Escherichia coli cells, the accumulation of EttA increases during the stationary phase, when energetic resources decline (Boël et al., 2014). In such a condition, EttA attaches to the ribosome and inhibits protein synthesis (Boël et al., 2014). In contrast, the increase in energy availability leads to the dissociation of EttA of the ribosome, allowing it to enter the elongation cycle resuming protein synthesis (Boël et al., 2014). Interestingly, even with our results showing that after 24 h of cultivation G. diazotrophicus still has available energy resources, characterized by its exponential growth phase, our proteomic analysis showed the up-accumulation of EttA. This result suggests that in G. diazotrophicus the regulation of EttA can be independent of energy availability.
Additionally, the reverse genetics results showed that ΔettA mutant loses the growth response to cocultivation with A. thaliana. ettA gene seems to belong to an operon in G. diazotrophicus genome, located upstream of a gene that codifies a GCN5-related N-acetyltransferase (GNAT) family protein. The phenotype observed for the ΔettA mutant may be the sum of effects caused by the absence of these two proteins (EttA and GNAT). GNATs are widespread in eukaryotes and prokaryotes and are associated with many bacterial processes, such as drug resistance, stress reaction, and regulation of transcription (Xie et al., 2014;Favrot, Blanchard & Vergnolle, 2016). The result obtained to ΔettA allow us to speculate about the role of EttA in the perception of plant metabolites and the subsequent activation of protein synthesis. However, further analyses are necessary to investigate the participation of the protein GNAT in such a process.
Proteins of the pentose-phosphate pathway were also up-accumulated. This pathway is essential for cell metabolism, including the maintenance of carbon homeostasis and the provision of precursors for nucleotide and amino acid biosynthesis (Pickl & Schönheit, 2015;Stincone et al., 2015). According to our data, four members of this pathway were up-accumulated: Zwf (Glucose-6-phosphate 1-dehydrogenase), Gnd (6-phosphogluconate dehydrogenase, decarboxylating), Gpml (2,3-bisphosphoglycerate-independent phosphoglycerate mutase) and a Zwf homolog (A9H0G0). Among these proteins, Zwf and Gnd function in the oxidative phase of the pentose-phosphate pathway, resulting in the production of NADPH (Lim et al., 2002). Despite the positive regulation of this pathway, our reverse genetics results demonstrate that the induction of G. diazotrophicus growth in response to A. thaliana cocultivation was maintained even with the lack of Zwf (A9H326). We hypothesize that the activity of the Zwf homolog (A9H0G0), also up-accumulated in our proteomic analysis, may compensate the absence of A9H326 in the pentose-phosphate pathway.
Additionally, among the proteins up-accumulated in our proteomic analysis potentially involved with the bacterial response to plant exudates, ClpX, and AccC deserve attention. ClpX plays a role in the unfolding and degradation of other proteins. Additionally, it has been shown in Burkholderia cenocepacia that a mutation in the ClpX coding sequence produces an increase in the production of acyl homoserine lactone, an important group of signaling molecules for quorum sensing in Gram-negative bacteria (Veselova et al., 2016). Proteins responsible for quorum sensing are regulated in the presence of hosts and participate in the control of bacterial growth and the regulation of virulence mechanisms (Sasse, Martinoia & Northen, 2018;Pérez-Montaño et al., 2013;Steindler et al., 2009). Additional analyses are necessary to investigate the role of ClpX in controlling unfolding proteins or in quorum sensing mechanisms in the responses of G. diazotrophicus to metabolites exuded from A. thaliana seedlings.
AccC is a component of bacterial acetyl coenzyme-A carboxylase (ACCase) that catalyzes the first step in fatty acid biosynthesis (Cheng et al., 2009). AccC participates in bacterial lipid biosynthesis and consequently impacts cellular membrane formation, essential processes for cell viability and bacterial growth (Cheng et al., 2009). Once G. diazotrophicus growth increased in response to A. thaliana cocultivation, the increase in accumulation of AccC observed in our proteomic analyses was expected. Our data using a defective mutant for AccC (ΔaccC) revealed a lower growth performance in comparison to the wild-type strain. Bacterial multiplication requires the synthesis of lipids to supply the formation of new cell membranes, which may explain our results. Interestingly, another accC homolog is present in the genome of G. diazotrophicus (GDI0201). However, the corresponding protein was not detected in our proteomic analyzes. The presence of such a homolog gene in the genome of G. diazotrophicus did not compensate for the phenotype of ΔaccC in response to A. thaliana cocultivation observed in our results, suggesting a distinct functional role.
Moreover, four proteins related to nutrients uptake were regulated in our proteomic analysis. Among these, the protein TonB-dependent receptor (TBDRs) was highly up-accumulated. TBDRs were described in early studies as an iron transporter. However, currently, its activity is also associated with the active import of other nutrients, such as vitamins B1 and B12 (Schauer, Rodionov & De Reuse, 2008). Molina et al. (2005), in a study with corn, showed that the efficiency of initial steps of seed and root colonization by P. putida is dependent of TBDR activity, once essential mechanisms to this process, as biofilm formation, are highly iron-dependent. Additionally, the up-accumulation of TBDRs was previously reported in G. diazotrophicus in response to sugarcane cocultivation (Lery et al., 2011). Thus, our result reinforces the importance of TBDRs in plant-bacteria associations.
The presence of sugars and other carbon structures in root exudates has been described as the main factor in the induction of bacterial multiplication during the initial steps of plant-bacteria interaction (Babalola, 2010). Moreover, in our assays, a high concentration of sugar was provided in the culture medium. Our proteomic analysis showed that proteins related to sugars uptake (OprB, Sugar ABC transporter substrate-binding, and D-xylose ABC transporter substrate-binding) were down-accumulated in G. diazotrophicus cells co-cultivated with A. thaliana. The up-accumulation of pentose-phosphate pathway proteins in our proteomic analysis reinforces that G. diazotrophicus cells were under high availability of carbon sources in our assays. Previously studies demonstrate that bacterial cells growing under high availability of carbon source exhibit high rates of carbon uptake and metabolism (Russell & Cook, 1995). In this sense, further analyzes are necessary to understand the role of the down-accumulation of proteins involved with sugar uptake at the beginning of the association between G. diazotrophicus and A. thaliana seedlings.

CONCLUSIONS
Taken together, our data provide physiological and molecular aspects related to the association between G. diazotrophicus and A. thaliana seedlings. As summarized in Fig. 6, A. thaliana seedlings constitutively exude beneficial metabolites that increase bacterial growth. These data reinforce the potential involvement of the MetE, AcnA, IlvC, and TonB-dependent receptor proteins as a conserved mechanism during the interaction of G. diazotrophicus with both monocot and dicot plants. Additionally, our results revealed the essential role of the proteins EttA and AccC in the bacterial growth activated in response to cocultivation with A. thaliana. This is the first study to demonstrate the participation of EttA and AccC proteins in plant-bacteria interactions. Our results open new perspectives for further investigations about the initial steps of such associations in other bacteria and plant species.

ADDITIONAL INFORMATION AND DECLARATIONS Funding
This work was supported by the Coordination for the Improvement of Higher Education Personnel (CAPES); the Brazilian National Council for Scientific and Technological Development (CNPq); the Rio de Janeiro Research Foundation (FAPERJ); the Funding Authority for Research and Projects (FINEP); and the State University of North Fluminense "Darcy Ribeiro" (UENF) for granting scholarships to the students. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Grant Disclosures
The following grant information was disclosed by the authors: This work was supported by the Coordination for the Improvement of Higher Education Personnel (CAPES). Brazilian National Council for Scientific and Technological Development (CNPq). Rio de Janeiro Research Foundation (FAPERJ). Funding Authority for Research and Projects (FINEP).