Paraburkholderia phymatum STM815 σ 54 Controls Utilization of Dicarboxylates, Motility, and T6SS-b Expression

: Rhizobia have two major life styles, one as free-living bacteria in the soil, and the other as bacteroids within the root / stem nodules of host legumes where they convert atmospheric nitrogen into ammonia. In the soil, rhizobia have to cope with changing and sometimes stressful environmental conditions, such as nitrogen limitation. In the beta-rhizobial strain Paraburkholderia phymatum STM815, the alternative sigma factor σ 54 (or RpoN) has recently been shown to control nitrogenase activity during symbiosis with Phaseolus vulgaris . In this study, we determined P. phymatum ’s σ 54 regulon under nitrogen-limited free-living conditions. Among the genes signiﬁcantly downregulated in the absence of σ 54 , we found a C 4 -dicarboxylate carrier protein (Bphy_0225), a ﬂagellar biosynthesis cluster (Bphy_2926-64), and one of the two type VI secretion systems (T6SS-b) present in the P. phymatum STM815 genome (Bphy_5978-97). A deﬁned σ 54 mutant was unable to grow on C 4 dicarboxylates as sole carbon source and was less motile compared to the wild-type strain. Both defects could be complemented by introducing rpoN in trans . Using promoter reporter gene fusions, we also conﬁrmed that the expression of the T6SS-b cluster is regulated by σ 54 . Accordingly, we show that σ 54 a ﬀ ects in vitro competitiveness of P. phymatum STM815 against Paraburkholderia diazotrophica .


Introduction
Symbioses between legumes and rhizobia increase soil fertility and crop yield by means of biological nitrogen fixation [1]. Rhizobia are soil bacteria, which adapt to different environmental stresses and eventually nodulate the roots or the stems of compatible legume plants. In the symbiotic organ-the nodule-bacteria live inside the host plant cells and eventually differentiate into nitrogen-fixing bacteroids. Bacteroids convert atmospheric nitrogen (N 2 ) into ammonia that is assimilated and used as a nitrogen source by the plant [2,3]. To support this energetically expensive reaction (16 molecules of ATP and 8 low potential electrons per N 2 reduced), the legume provides the bacteroids with energy in form of reduced carbon compounds. C 4 -dicarboxylates such as succinate, fumarate, and malate have been shown to be the primary carbon source used by bacteroids and are oxidized to CO 2 in the tricarboxylic acid cycle [4].
Rhizobia are phylogenetically diverse and belong to the alpha-proteobacterial (alpha-rhizobia) and the beta-proteobacterial group (beta-rhizobia). Beta-rhizobial strains such as Burkholderia and Cupriavidus were first isolated from nodules in 2001 [5,6], and most of the legume nodulating Burkholderia for one hour under nitrogen-limited conditions. For each strain, three independent biological replicates were prepared and processed for RNA-Seq analysis.
To test the induction of the promoter fusions, the bacteria were grown at 30 • C in AB medium containing 10 mM glucose as C-source (ABG) in a 96-well plate (Falcon, Corning, USA) to exponential phase. The optical density and GFP fluorescence (excitation/emission wavelength equal to 488 nm/520 nm) were measured using a TECAN plate reader (TECAN Infinite M200 PRO, Tecan Trading AG, Switzerland).
The growth of the following P. phymatum STM815 strains-wild type (wt), σ 54 mutant, and σ 54 complemented-was tested in AB minimal medium with 15 mM of three different C 4 -carbon sources: Fumaric acid (Sigma-Aldrich, ABF), malic acid (Sigma-Aldrich, ABM), and succinic acid (Sigma-Aldrich, ABS). For each strain and C-source tested, the growth of three independent biological replicates was measured.

Promoter Fusion Construction
To construct promoter fusions, the two promoters of interest were introduced into the vector pPROBE-NT (Table S1) [29]. To construct pPROBE-T6SS-b (p5978), the promotor of T6SS-b was amplified from P. phymatum STM815 genomic DNA (gDNA) with the primers Bphy_5978_Promotor_F_HindIII and Bphy_5978_Promotor_R_BamH1 (Table S1). The 657 bp long PCR product was restricted with the enzymes HindIII and BamHI and cloned into pPROBE-NT. The correct sequence was confirmed by sequencing at Microsynth (Balgach, Switzerland) using pCO13-R3 primer [30]. The plasmid was then conjugated into P. phymatum STM815 wild type and σ 54 mutant, and the transconjugants were selected on ABS plates. To obtain the second promoter fusion, the promotor region of the T6SS-3 cluster (p6115) was amplified using primers p6115_SalI_For and p6115_EcoRI_rev (Table S1). The obtained PCR fragment (480 bp) was digested and cloned into pPROBE-NT between the SalI and EcoRI sites. Once the sequence of the construct had been confirmed through sequencing at Microsynth using pCO13-R3 primer [30], the plasmid was mobilized into P. phymatum STM815 wild type and σ 54 mutant.

RNA-Sequencing and Data Processing
Total RNA was isolated from flash-frozen, pelleted cells of the wild type and the σ 54 mutant using a modified hot acid phenol protocol [31]. Afterwards, gDNA was digested by DNase treatment and the quality of the total RNA as well as the complete removal of gDNA were checked by PCR [25,26]. The cDNA synthesis was initiated with 150 ng of total RNA and the library was prepared and purified with the Ovation ® Complete Prokaryotic RNA-Seq DR Multiplex Systems (NuGEN, San Carlos, CA, USA) [25,26]. Next, the quality and quantity of the cDNA libraries were analyzed with a TapeStation System (Agilent Technologies, Santa Clara, CA, USA). The prepared cDNA libraries were single-end sequenced with a HiSeq2500 instrument (Illumina, San Diego, CA, USA) at the Functional Genomic Center Zurich (FGCZ). Using the CLC Genomics Workbench v7.0 (QIAGEN CLC bio, Aarhus, Denmark) program, the obtained reads were trimmed to 70 bp and mapped to the P. phymatum STM815 genome, allowing up to two mismatches per read [32]. Afterwards the unique reads were analyzed statistically with the DESeq R-package (version 1.30.0) [33]. The top 200 significantly RpoN-regulated genes with a log 2 (FC) ≥ 1 and ≤ −1 were considered and then ranked by ascending p-value. To get additional information, the top 200 differentially regulated genes were assigned to functional categories (EggNOG v3.0) [34]. Table S2 lists all P. phymatum STM815 genes and their expression profile in the wild type and the σ 54 mutant. The RNA-Seq raw and processed data files of the P. phymatum STM815 wild-type and σ 54 mutant strains are available with the GSE156048 accession number.

Phenotypic Analysis
Plates for the swimming motility test were prepared by using LB salt-free medium containing 1% tryptone (Difco), 0.5% yeast extract (Difco), and 0.2% agar. The bacterial cells were washed twice with 10 mM MgSO 4 and normalized to an OD 600 of 0.5. The plates were inoculated with the bacterial suspension using a 20 µL pipette tip. Subsequently, the plates were incubated at 30 • C inside a box containing wet paper towels. The diameter of the swimming zone was measured after 40 h incubation.
To perform competition experiments on plates, a target strain and an attacker strain were chosen, plated on salt-free LB without antibiotics and incubated overnight at 30 • C. The strains were washed twice with 10 mM MgSO 4 and the OD 600 was adjusted (target strain OD 600 = 0.2, attacker strain OD 600 = 2.0). The strains were mixed in a 1:1 volume ratio. Twenty µl aliquots of the mixture were put on cellulose nitrate membrane filters (GE Healthcare Life Sciences, Marlborough, MA, USA) on ABG plates, and incubated at 30 • C. After 24 h, the cells on the filters were resuspended in one ml 10 mM MgSO 4 . The colony forming units (CFU) were determined on salt-free LB plates, both with and without antibiotics, to recover the attacker and target strains, respectively.

Statistical Analysis
For category distribution, the percentages of upregulated and downregulated genes in each category were calculated as described previously [25]. One-way ANOVA (performed with Prism 7.0), in which the mean of each strain was compared to the mean of the wild type, was used to analyze the motility data. The one unpaired t-test corrected for multiple comparisons using the Holm-Sidak method (Prism 7.0) was employed to statistically analyze the results of the competition on plate. The induction of the promoter fusions was tested with two-way ANOVA corrected for multiple comparisons using the Sidak test (Prism 7.0).

The P. phymatum STM815 σ 54 Regulon in Free-Living Nitrogen-Limiting Conditions
RNA-Seq was employed to determine the P. phymatum STM815 σ 54 regulon under nitrogen-limited free-living conditions. For each strain, three independent biological replicates were prepared, processed, and analyzed as described previously [25]. The unique read counts obtained per sample ranged from 6.9 to 13.2 million reads.
After a DESeq analysis comparing gene expression in the P. phymatum STM815 σ 54 mutant versus wild type, the top 200 differentially regulated genes (DESeq analysis p-value < 2.2 × 10 −8 , with log 2 [FC] ≥ 1 and ≤ −1) were identified and classified in eggNOG functional categories [34] (Figure 1). Among these 200 top-regulated genes, 81 showed increased expression, while 119 were downregulated in the σ 54 mutant (and were therefore positively regulated by σ 54 ). While the two categories "signal transduction mechanisms" and "cell motility" were over-represented among the downregulated genes (Table 1), one category ("cell wall, membrane, envelope biogenesis") was found to be over-represented in the upregulated genes.
In the category "cell motility", we found several genes involved in flagella biosynthesis (in the Bphy_2926-64 cluster) and a gene involved in chemotaxis (Bphy_5592). In the category "signal transduction mechanisms", several transcriptional regulatory genes (e.g., Bphy_3958, Bphy_5314, and Bphy_5669) and genes coding for histidine kinases (e.g., Bphy_0226, Bphy_3957, Bphy_3963, Bphy_4749, Bphy_5668, and Bphy_5975) were identified. Among these top differentially regulated genes, six might be part of a two-component regulatory system (TCRS): Bphy_3957-58, Bphy_3962-63, and Bphy_5668-69. Upstream of Bphy_5668-69, we found Bphy_5667, which codes for a citrate carrier and has an RpoN-binding sequence in the promoter region ( Figure 2A). Moreover, downstream of Bphy_5668-69 we found a gene coding for a diguanylate phosphodiesterase potentially involved in c-di-GMP metabolism (Bphy_5670, Figure 2A). The histidine kinase Bphy_0226 (dctB) is located downstream of a gene coding for a C 4 -dicarboxylate transporter (dctA, Bphy_0225), which was also on the list of top-regulated genes and contained an RpoN box in the promoter region ( Figure 2B). In several bacteria, the DctA transporter has been shown to be involved in the assimilation of C 4 dicarboxylates, i.e., fumarate, malate, and succinate [4,35,36], which are important energy sources for symbiotic rhizobia [37]. Interestingly, most of the genes (13 of 20) belonging to one of the two T6SS clusters present in P. phymatum STM815 (Bphy_5978-97, T6SS-b, Figure 2C) grouped in the category "function unknown" were downregulated in the σ 54 mutant grown under nitrogen starvation. By contrast, the expression of the second T6SS cluster (Bphy_6107-6129, T6SS-3) was not dependent on σ 54 . T6SSs are contact-dependent secretion systems that have been shown to be important for the transport of effectors to other prokaryotic or eukaryotic cells [38]. Interestingly, several genes located in a cluster coding for components of a type IV secretion system (T4SS, Bphy_7524-37) also showed significant downregulation in the σ 54 mutant. Similar T4SSs are usually involved in conjugation [39].  [34]. The asterisks (*) indicate statistical significance for over-represented genes in a particular category (p-value < 0.02).
In the category "cell motility", we found several genes involved in flagella biosynthesis (in the Bphy_2926-64 cluster) and a gene involved in chemotaxis (Bphy_5592). In the category "signal transduction mechanisms", several transcriptional regulatory genes (e.g., Bphy_3958, Bphy_5314, and Bphy_5669) and genes coding for histidine kinases (e.g., Bphy_0226, Bphy_3957, Bphy_3963, Bphy_4749, Bphy_5668, and Bphy_5975) were identified. Among these top differentially regulated genes, six might be part of a two-component regulatory system (TCRS): Bphy_3957-58, Bphy_3962-63, and Bphy_5668-69. Upstream of Bphy_5668-69, we found Bphy_5667, which codes for a citrate carrier and has an RpoN-binding sequence in the promoter region ( Figure 2A). Moreover, downstream of Bphy_5668-69 we found a gene coding for a diguanylate phosphodiesterase potentially involved in c-di-GMP metabolism (Bphy_5670, Figure 2A). The histidine kinase Bphy_0226 (dctB) is located downstream of a gene coding for a C4-dicarboxylate transporter (dctA, Bphy_0225), which was also on the list of top-regulated genes and contained an RpoN box in the promoter region ( Figure 2B). In several bacteria, the DctA transporter has been shown to be involved in the assimilation of C4 dicarboxylates, i.e., fumarate, malate, and succinate [4,35,36], which are important energy sources for symbiotic rhizobia [37]. Interestingly, most of the genes (13 of 20) belonging to one of the two T6SS clusters present in P. phymatum STM815 (Bphy_5978-97, T6SS-b, Figure 2C) grouped in the category "function unknown" were downregulated in the σ 54 mutant grown under nitrogen starvation. By contrast, the expression of the second T6SS cluster (Bphy_6107-6129, T6SS-3) was not dependent on σ 54 . T6SSs are contact-dependent secretion systems that have been shown to be important for the transport of effectors to other prokaryotic or eukaryotic cells [38]. Interestingly, several genes located in a cluster coding for components of a type IV secretion system (T4SS, Bphy_7524-37) also showed significant downregulation in the σ 54 mutant. Similar T4SSs are usually involved in conjugation [39].
In the category "carbohydrate transport and metabolism", we found several genes whose expression was downregulated in the σ 54 mutant: Bphy_2145, coding for a trehalose-6-phosphate synthase (otsA), as well as two genes involved in glycogen metabolism (Bphy_5335 and Bphy_5336,  [34]. The asterisks (*) indicate statistical significance for over-represented genes in a particular category (p-value < 0.02).
In the category "carbohydrate transport and metabolism", we found several genes whose expression was downregulated in the σ 54 mutant: Bphy_2145, coding for a trehalose-6-phosphate synthase (otsA), as well as two genes involved in glycogen metabolism (Bphy_5335 and Bphy_5336, glgX and glgA, Figure 2D). All three genes display an RpoN-binding sequence in their promoter region (Table 1). In the operon Bphy_3959-61, which encodes a potential nitrate/sulfonate/bicarbonate transporter, we found the most highly downregulated genes (log 2 (FC) = −7.1, −7.4, −6.0, for Bphy_3959 to Bphy_3961, respectively). Additionally, the expression of three genes in a cluster potentially coding for a polysaccharide (Bphy_3730-1 and Bphy_3733) was downregulated in the σ 54 mutant compared to the wild type.
As previously mentioned, among the 81 genes upregulated in the mutant, the category "cell wall/membrane/envelope biogenesis" was over-represented, with several genes coding for porins (Bphy_0154, Bphy_1082 and Bphy_2684), as well as a cluster coding for the polysaccharide cepacian (Bphy_1056-77) [40]. Out of the 200 genes significantly regulated by RpoN, ten were located on the symbiotic plasmid including the T4SS and an aminotransferase class-III (Bphy_7651), which were all positively regulated by σ 54 .
By applying a less stringent p-value cut-off than that selected for the top 200 differentially regulated genes (DESeq analysis p-value ≤ 2 × 10 −5 ), additional genes associated with nitrogen metabolism were found to be positively regulated by σ 54 , such as the TCRS ntrBC (Bphy_1479-80), urtB (Bphy_2252) coding for a urea ABC transporter permease and the glnB1-amtB operon, coding for the nitrogen regulatory protein P-II and the ammonium transporter (Bphy_0256-57). As mentioned above, one of the genes significantly downregulated under nitrogen-limited conditions in our transcriptomic analysis was dctA (Bphy_0225), which codes for a C 4 dicarboxylate transporter. DctA is located upstream of a gene cluster encoding the sensor kinase DctB and the response regulator DctD, which belongs to the σ 54 -interacting protein family. While dctB expression was downregulated in absence of σ 54 , dctD expression did not change, suggesting that dctD is in a separate transcriptional unit ( Figure 2B). In the promoter region of dctA, an RpoN-binding box was identified. conditions in our transcriptomic analysis was dctA (Bphy_0225), which codes for a C4 dicarboxylate transporter. DctA is located upstream of a gene cluster encoding the sensor kinase DctB and the response regulator DctD, which belongs to the σ 54 -interacting protein family. While dctB expression was downregulated in absence of σ 54 , dctD expression did not change, suggesting that dctD is in a separate transcriptional unit ( Figure 2B). In the promoter region of dctA, an RpoN-binding box was identified. A similar genomic constellation to dctA-dctB-dctD is found for dicarboxylate transport systems in several bacteria, however, dctB and dctD are usually transcribed divergently compared to dctA [4]. To verify that expression of the DctA transporter gene was regulated by σ 54 , we examined the ability of the wild-type, the σ 54 mutant, and the complemented σ 54 strains to grow in minimal medium in the presence of C4-carbon sources. The growth of these P. phymatum STM815 strains was tested in defined buffered AB minimal medium supplemented with fumaric acid, malic acid, and succinic acid. We found that the σ 54 mutant strain was impaired in the utilization of all tested C4 organic acids and that this defect could be restored in the σ 54 complemented strain ( Table 2). The results confirmed the validity of our transcriptomic analysis and suggested that σ 54 is involved in the assimilation of C4 compounds in P. phymatum STM815. A similar genomic constellation to dctA-dctB-dctD is found for dicarboxylate transport systems in several bacteria, however, dctB and dctD are usually transcribed divergently compared to dctA [4]. To verify that expression of the DctA transporter gene was regulated by σ 54 , we examined the ability of the wild-type, the σ 54 mutant, and the complemented σ 54 strains to grow in minimal medium in the presence of C 4 -carbon sources. The growth of these P. phymatum STM815 strains was tested in defined buffered AB minimal medium supplemented with fumaric acid, malic acid, and succinic acid. We found that the σ 54 mutant strain was impaired in the utilization of all tested C 4 organic acids and that this defect could be restored in the σ 54 complemented strain ( Table 2). The results confirmed the validity of our transcriptomic analysis and suggested that σ 54 is involved in the assimilation of C 4 compounds in P. phymatum STM815.

σ 54 Positively Controls Swimming Motility
In the list of top genes positively regulated by σ 54 at the transcript level (Table 1), the category "cell motility" was over-represented. The expression of Bphy_2938-39 (fliMN), Bphy_2956 (flgJ), and Bphy_2962 (flhF), all of which encode flagellar biosynthesis proteins, was significantly downregulated in a σ 54 mutant (Table 1). Since flagella are used for motility, the swimming ability of P. phymatum STM815 wild type, the σ 54 mutant, and the complemented strain was tested on salt-free LB plates, and the diameter of the swimming cells was measured after 40 h of incubation at 30 • C. In agreement with our RNA-Seq data, the σ 54 mutant was 40% less motile, and the complemented strain showed increased motility relative to the wild-type strain (Figure 3).
Succinic acid + -+ 1 Growth was assessed in three independent replicates by measuring the optical density at 600 nm after incubation in AB minimal medium supplemented with 15 mM of different carbon sources for 30 h at 30 °C and 220 rpm. The "-" sign corresponds to OD600 ≤ 0.08.

σ 54 Positively Controls Swimming Motility
In the list of top genes positively regulated by σ 54 at the transcript level (Table 1), the category "cell motility" was over-represented. The expression of Bphy_2938-39 (fliMN), Bphy_2956 (flgJ), and Bphy_2962 (flhF), all of which encode flagellar biosynthesis proteins, was significantly downregulated in a σ 54 mutant (Table 1). Since flagella are used for motility, the swimming ability of P. phymatum STM815 wild type, the σ 54 mutant, and the complemented strain was tested on salt-free LB plates, and the diameter of the swimming cells was measured after 40 h of incubation at 30 °C. In agreement with our RNA-Seq data, the σ 54 mutant was 40% less motile, and the complemented strain showed increased motility relative to the wild-type strain (Figure 3).

The Presence of σ 54 Influences Interbacterial Competition
Our RNA-Seq data showed that the expression of one P. phymatum STM815 cluster coding for a T6SS (Bphy_5978-97, T6SS-b cluster) was dependent on σ 54 under nitrogen-limiting growth conditions ( Figure 2C). Both P. phymatum STM815 T6SS systems were previously shown to be important for competition with other related Paraburkholderia strains such as P. diazotrophica [22]. In order to validate the RNA-Seq data, the promoters of Bphy_5978 (T6SS-b cluster) or Bphy_6115 (T6SS-3 cluster) were fused with gfp, and expression was measured in the P. phymatum STM815 wildtype and the σ 54 mutant background. In accordance with the transcriptome data, only the expression of the T6SS-b cluster (p5978), and not of the T6SS-3 cluster (p6115), was decreased in a P. phymatum STM815 σ 54 mutant compared to the wild-type strain (Figure 4).

The Presence of σ 54 Influences Interbacterial Competition
Our RNA-Seq data showed that the expression of one P. phymatum STM815 cluster coding for a T6SS (Bphy_5978-97, T6SS-b cluster) was dependent on σ 54 under nitrogen-limiting growth conditions ( Figure 2C). Both P. phymatum STM815 T6SS systems were previously shown to be important for competition with other related Paraburkholderia strains such as P. diazotrophica [22]. In order to validate the RNA-Seq data, the promoters of Bphy_5978 (T6SS-b cluster) or Bphy_6115 (T6SS-3 cluster) were fused with gfp, and expression was measured in the P. phymatum STM815 wild-type and the σ 54 mutant background. In accordance with the transcriptome data, only the expression of the T6SS-b cluster (p5978), and not of the T6SS-3 cluster (p6115), was decreased in a P. phymatum STM815 σ 54 mutant compared to the wild-type strain (Figure 4).  We next tested the ability of the P. phymatum STM815 wild-type, the σ 54 mutant, and the complemented strain to compete against P. diazotrophica for 24 h on ABG minimal medium ( Figure  5). The σ 54 mutant turned out to be slightly, but significantly, less competitive as compared to the P. phymatum STM815 wild type, indicating a role for this alternative sigma factor in controlling interbacterial competition. In line with this observation, a σ 54 complemented strain, which expressed rpoN from the the lac promoter on the pBBR1MCS plasmid, out-competed P. diazotrophica more effectively than the wild-type strain. Competition assay with P. phymatum STM815 wild-type (wt), σ 54 mutant (σ 54 mt), and σ 54 complemented (σ 54 comp) strains using P. diazotrophica as target. The assay was performed on ABG plates for 24 h at 30 °C. The histogram shows the mean colony forming units (CFU) per ml of the recovered P. diazotrophica strain (target). Three independent biological replicates were carried out, and error bars indicate the standard deviation. The results were analyzed using an unpaired t-test corrected for multiple comparisons using the Holm-Sidak test. *, p < 0.01.

Discussion
The alternative sigma factor σ 54 was originally discovered in Salmonella as the sigma factor required for the synthesis of glutamine synthetase [41]. Subsequently, σ 54 was identified in a wide range of Gram-negative and Gram-positive bacteria, and found to play a general role in the control Figure 4. Expression of the T6SS-b cluster (p5978) and the T6SS-3 cluster (p6115) in P. phymatum STM815 wild-type (wt) and σ 54 mutant (σ 54 mt) strains. The histograms show gfp activity from the promoter (measured at a wavelength of 488 nm) normalized by the OD 600 . Three independent biological replicates were performed. The statistical analysis was performed using two-way ANOVA corrected with Sidak's multiple comparisons test. ****, p < 0.0001. We next tested the ability of the P. phymatum STM815 wild-type, the σ 54 mutant, and the complemented strain to compete against P. diazotrophica for 24 h on ABG minimal medium ( Figure 5). The σ 54 mutant turned out to be slightly, but significantly, less competitive as compared to the P. phymatum STM815 wild type, indicating a role for this alternative sigma factor in controlling interbacterial competition. In line with this observation, a σ 54 complemented strain, which expressed rpoN from the the lac promoter on the pBBR1MCS plasmid, out-competed P. diazotrophica more effectively than the wild-type strain.
We next tested the ability of the P. phymatum STM815 wild-type, the σ 54 mutant, and the complemented strain to compete against P. diazotrophica for 24 h on ABG minimal medium ( Figure  5). The σ 54 mutant turned out to be slightly, but significantly, less competitive as compared to the P. phymatum STM815 wild type, indicating a role for this alternative sigma factor in controlling interbacterial competition. In line with this observation, a σ 54 complemented strain, which expressed rpoN from the the lac promoter on the pBBR1MCS plasmid, out-competed P. diazotrophica more effectively than the wild-type strain. Competition assay with P. phymatum STM815 wild-type (wt), σ 54 mutant (σ 54 mt), and σ 54 complemented (σ 54 comp) strains using P. diazotrophica as target. The assay was performed on ABG plates for 24 h at 30 °C. The histogram shows the mean colony forming units (CFU) per ml of the recovered P. diazotrophica strain (target). Three independent biological replicates were carried out, and error bars indicate the standard deviation. The results were analyzed using an unpaired t-test corrected for multiple comparisons using the Holm-Sidak test. *, p < 0.01.

Discussion
The alternative sigma factor σ 54 was originally discovered in Salmonella as the sigma factor required for the synthesis of glutamine synthetase [41]. Subsequently, σ 54 was identified in a wide range of Gram-negative and Gram-positive bacteria, and found to play a general role in the control of nitrogen metabolism and to affect various other cellular functions such as motility, dicarboxylate transport, degradation of xenobiotics, and virulence in plant and human pathogens [42][43][44][45][46][47][48]. Competition assay with P. phymatum STM815 wild-type (wt), σ 54 mutant (σ 54 mt), and σ 54 complemented (σ 54 comp) strains using P. diazotrophica as target. The assay was performed on ABG plates for 24 h at 30 • C. The histogram shows the mean colony forming units (CFU) per ml of the recovered P. diazotrophica strain (target). Three independent biological replicates were carried out, and error bars indicate the standard deviation. The results were analyzed using an unpaired t-test corrected for multiple comparisons using the Holm-Sidak test. *, p < 0.01.

Discussion
The alternative sigma factor σ 54 was originally discovered in Salmonella as the sigma factor required for the synthesis of glutamine synthetase [41]. Subsequently, σ 54 was identified in a wide range of Gram-negative and Gram-positive bacteria, and found to play a general role in the control of nitrogen metabolism and to affect various other cellular functions such as motility, dicarboxylate transport, degradation of xenobiotics, and virulence in plant and human pathogens [42][43][44][45][46][47][48].
In the rhizobia, including P. phymatum STM815, σ 54 is a master regulator during symbiosis since it controls the expression of the genes encoding nitrogenase, i.e., the enzyme that converts N 2 into a form that can be readily assimilated by the plant [49,50].
P. phymatum STM815 is a beta-rhizobium that previously belonged to the genus Burkholderia, which comprises versatile bacteria able to adapt and colonize different environmental niches. Our group has shown that in the closely related opportunistic pathogen Burkholderia cenocepacia H111, σ 54 is an important regulator of several phenotypic traits, such as the utilization of nitrogen sources, motility, EPS and biofilm formation, biosynthesis of poly-hydroxybutyrate (PHB), and virulence towards Caenorhabditis elegans [48]. To control biofilm formation in H111, σ 54 interacts with the EBP BerB, which binds c-di-GMP and thereby regulates expression of the exopolysaccharide Bep [51].
We report here that σ 54 also plays an important role in P. phymatum STM815 grown under free-living conditions, where it regulates motility and the expression of T6SS-b, and controls the uptake of tricarbocylic acid cycle TCA cycle intermediates such as succinate, fumarate, and malate ( Figure 6). Depending on the presence or absence of oxygen in the environment, dicarboxylates are taken up by bacteria using transporters of different protein families. The dicarboxylate transporter DctA, which belongs to the dicarboxylate amino acid-cation symporter (DAACS) family, has been shown to facilitate dicarboxylate uptake under aerobic conditions in several bacteria, including E. coli and Rhizobium leguminosarum [4,52,53]. We demonstrate here that the P. phymatum STM815 σ 54 mutant is unable to take up succinate, fumarate, and malate under aerobic, free-living conditions, and that the expression of dctA (Bphy_0225), which is located upstream of the two component regulatory system genes dctBD, is activated by σ 54 . DctB is a sensor kinase located in the membrane, which is autophosphorylated in the presence of C 4 dicarboxylates and transfers a phosphate group to its cognate response regulator DctD. DctD is a σ 54 -dependent EBP that activates transcription of target genes including dctA [35]. DctA has been shown to be essential for symbiotic nitrogen fixation in Sinorhizobium meliloti and R. leguminosarum [54][55][56]. We are currently evaluating whether dctA is the main dicarboxylate transporter in P. phymatum STM815 and its role during symbiosis. Interestingly, the expression of the B. cenocepacia H111 dctA ortholog (I35_RS14715) is not regulated by σ 54 , suggesting that C 4 uptake may not be dependent on this alternative sigma factor in opportunistic pathogens of the genus Burkholderia.
Among the most highly downregulated P. phymatum STM815 genes (Table 1), we found Bphy_3959-61, which encodes a potential nitrate/sulfonate/bicarbonate ABC transporter. Out of the five genes annotated as citrate transporter in the P. phymatum STM815 genome (Bphy_0810, Bphy_3035, Bphy_4278, Bphy_5667, and Bphy_5880), only Bphy_5667 was significantly regulated by σ 54 (Figure 2A), suggesting that this transporter plays a key role in importing citrate under free-living conditions. Several genes potentially involved in glycogen and trehalose metabolism were found among those that were statistically most significantly regulated by σ 54 . In fact, the expression of genes encoding a glycogen synthase (Bphy_5336, glgA) ( Figure 2D), a glycogen debranching enzyme (Bphy_5335, glgX) which degrades glycogen, and a trehalose-6-phosphate synthase (Bphy_2145, otsA) was activated by σ 54 and RpoN-binding motifs were found in the promoter regions of these genes ( Table 2).
Glycogen is a soluble polysaccharide composed of glucose in an α-1,4-linked linear arrangement with α-1,6-branches that serves as a storage molecule in many organisms, including eukaryotes and prokaryotes. During starvation periods, glycogen provides a source of stored energy and carbon. The non-reducing disaccharide trehalose (composed of two molecules of α -D-glucose) also acts as an energy reserve compound as well as an osmoprotectant. In Bradyrhizobium diazoefficiens, trehalose is produced to allow survival during desiccation, oxidative stress, and during nodule senescence [57][58][59][60][61][62]. The precursor of trehalose, trehalose 6-phosphate, has been proposed to be a signal for plant growth and stress tolerance in Rhizobium etli [63]. In plants, trehalose-6-phosphate is an important signal metabolite that coordinates carbon and nitrogen metabolism [64][65][66]. The three genes (glgA, glgX and otsA) are regulated by σ 54 not only in free-living conditions, but also during symbiotic growth in root nodules [26]. Additionally, the gene encoding a trehalose synthase (Bphy_7407) was regulated by σ 54 during symbiosis, but not under free-living conditions. To the best of our knowledge, the regulation of glycogen and trehalose synthesis via σ 54 has not been reported previously. Trehalose biosynthesis was previously shown to be regulated by sigma factor σ 38 (RpoS) in E. coli [67] and by the extra-cytoplasmic function (ECF) sigma factor RpoE2 of S. meliloti [68]. Interestingly and in contrast to H111, the synthesis of the storage compound PHB is not controlled by σ 54 in P. phymatum STM815.
To date, not much is known about the regulation of T6SS expression in rhizobia. In this study, we show that σ 54 is involved in the regulation of one of the two P. phymatum STM815 T6SS gene clusters (Figures 2 and 4), and that a σ 54 mutant strain is slightly less competitive with P. diazotrophica than wild type P. phymatum STM815 ( Figure 5). RpoN regulation of T6SS-b is probably indirect, since we did not find a RpoN-binding sequence in the promoter region of the first gene in the operon. However, the presence of two dicarboxylate transporter genes in T6SS-b (Bphy_5990 and Bphy_5992) could be the reason for the indirect control of this cluster by σ 54 . Our results also suggest that the two P. phymatum STM815 T6SSs are used under different conditions and are therefore subject to different regulatory mechanisms, with T6SS-3 expression being σ 54 -independent. Bernard and colleagues [69] showed that σ 54 controls expression of T6SSs in several environmental strains, including Vibrio cholerae, Aeromonas hydrophila, Marinomonas MWYL1 and the plant pathogenic bacteria Pectobacterium atrosepticum, and Pseudomonas syringae pv. tomato [69]. Interestingly, several T6SS clusters investigated to date contain an EBP encoding gene (vasH), which has been demonstrated to be required for maximal activation of T6SS expression [69,70]. No vasH homolog was apparent in the two T6SS clusters present in P. phymatum STM815. Additionally, chromatin immunoprecipitation coupled with next-generation sequencing (ChIP-Seq) and RNA-Seq experiments performed in V. cholerae identified two genes coding for two T6SS hallmark proteins, Hcp and VgrG3, as direct σ 54 targets [71,72].
In summary, in the beta-rhizobial strain P. phymatum STM815 σ 54 controls important phenotypic traits, such as motility, competition, and transport of C4-dicarboxylates, which all contribute to the high competitiveness of this strain in infecting the roots of various legumes. The identity of the different EBPs that work together with σ 54 to control expression of these different traits will be an interesting area for future investigations.
Nitrogen 2020, 1 93 regulated by σ 54 during symbiosis, but not under free-living conditions. To the best of our knowledge, the regulation of glycogen and trehalose synthesis via σ 54 has not been reported previously. Trehalose biosynthesis was previously shown to be regulated by sigma factor σ 38 (RpoS) in E. coli [67] and by the extra-cytoplasmic function (ECF) sigma factor RpoE2 of S. meliloti [68]. Interestingly and in contrast to H111, the synthesis of the storage compound PHB is not controlled by σ 54 in P. phymatum STM815.
To date, not much is known about the regulation of T6SS expression in rhizobia. In this study, we show that σ 54 is involved in the regulation of one of the two P. phymatum STM815 T6SS gene clusters (Figure 2 and Figure 4), and that a σ 54 mutant strain is slightly less competitive with P. diazotrophica than wild type P. phymatum STM815 ( Figure 5). RpoN regulation of T6SS-b is probably indirect, since we did not find a RpoN-binding sequence in the promoter region of the first gene in the operon. However, the presence of two dicarboxylate transporter genes in T6SS-b (Bphy_5990 and Bphy_5992) could be the reason for the indirect control of this cluster by σ 54 . Our results also suggest that the two P. phymatum STM815 T6SSs are used under different conditions and are therefore subject to different regulatory mechanisms, with T6SS-3 expression being σ 54 -independent. Bernard and colleagues [69] showed that σ 54 controls expression of T6SSs in several environmental strains, including Vibrio cholerae, Aeromonas hydrophila, Marinomonas MWYL1 and the plant pathogenic bacteria Pectobacterium atrosepticum, and Pseudomonas syringae pv. tomato [69]. Interestingly, several T6SS clusters investigated to date contain an EBP encoding gene (vasH), which has been demonstrated to be required for maximal activation of T6SS expression [69,70]. No vasH homolog was apparent in the two T6SS clusters present in P. phymatum STM815. Additionally, chromatin immunoprecipitation coupled with next-generation sequencing (ChIP-Seq) and RNA-Seq experiments performed in V. cholerae identified two genes coding for two T6SS hallmark proteins, Hcp and VgrG3, as direct σ 54 targets [71,72].
In summary, in the beta-rhizobial strain P. phymatum STM815 σ 54 controls important phenotypic traits, such as motility, competition, and transport of C4-dicarboxylates, which all contribute to the high competitiveness of this strain in infecting the roots of various legumes. The identity of the different EBPs that work together with σ 54 to control expression of these different traits will be an interesting area for future investigations. Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1: Bacterial strains and plasmids used in this study, Table S2: List of all P. phymatum STM815 genes, log2 of the fold changes