Bradyrhizobium diazoefficiens USDA110 PhaR functions for pleiotropic regulation of cellular processes besides PHB accumulation

Background Bradyrhizobium diazoefficiens USDA110 nodulates soybeans for nitrogen fixation. It accumulates poly-3-hydroxybutyrate (PHB), which is of physiological importance as a carbon/energy source for survival during starvation, infection, and nitrogen fixation conditions. PHB accumulation is orchestrated by not only the enzymes for PHB synthesis but also PHB-binding phasin proteins (PhaPs) stabilizing the PHB granules. The transcription factor PhaR controls the phaP genes. Results Inactivation of phaR led to decreases in PHB accumulation, less cell yield, increases in exopolysaccharide (EPS) production, some improvement in heat stress tolerance, and slightly better growth under microaerobic conditions. Changes in the transcriptome upon phaR inactivation were analyzed. PhaR appeared to be involved in the repression of various target genes, including some PHB-degrading enzymes and others involved in EPS production. Furthermore, in vitro gel shift analysis demonstrated that PhaR bound to the promoter regions of representative targets. For the phaP1 and phaP4 promoter regions, PhaR-binding sites were determined by DNase I footprinting, allowing us to deduce a consensus sequence for PhaR-binding as TGCRNYGCASMA (R: A or G, Y: C or T, S: C or G, M: A or C). We searched for additional genes associated with a PhaR-binding sequence and found that some genes involved in central carbon metabolism, such as pdhA for pyruvate dehydrogenase and pckA for phosphoenolpyruvate carboxykinase, may be regulated positively and directly by PhaR. Conclusions These results suggest that PhaR could regulate various genes not only negatively but also positively to coordinate metabolism holistically in response to PHB accumulation. Electronic supplementary material The online version of this article (10.1186/s12866-018-1317-2) contains supplementary material, which is available to authorized users.


Background
Poly-3-hydroxybutyrate (PHB) is a type of polyhydroxyalkanoate (PHA), which accumulates within the cells of various microorganisms as an energy storage substance under stress conditions, such as oxygen limitation and nutrient starvation [1]. PHB, a potential substitute for petroleum-based plastics, is produced in an enzymatic process involving reactions requiring mild conditions and is biodegradable even under opportunistic anaerobic conditions [2].
PHB or PHA accumulation is of physiological importance in many bacterial species. Azospirillum brasilense is a freeliving soil bacterium that affects the growth of numerous agricultural crops [3]. Mutant strains of A. brasilense lacking the ability to accumulate PHA exhibited decreased resistance to various stress conditions, and degradation products of PHB could protect microorganisms from hydroxyl radicals [4,5]. In the case of free-living Sinorhizobium meliloti, an alfalfa symbiont, PHB is continually synthesized and degraded, which may serve as an energy source under starvation conditions [6]. In addition, a mutant defective in PHB synthesis demonstrated a decreased symbiotic ability and diminished nitrogen fixing ability with weakened nitrogenase activity [7]. In Bradyrhizobium japonicum (Bradyrhizobium diazoefficiens) USDA110, a soybean symbiont bacterium, under the symbiotic state, it was proposed that utilization of PHB might maintain nitrogen fixation activity [8]. In contrast, a mutant of B. diazoefficiens USDA110 defective in PHB accumulation had a better symbiotic performance than the wild type [9]. These observations suggest that the physiological importance of PHB accumulation varies from species to species [10]. Though it has been proposed that PHB is used as an energy source during the bacterial differentiation process from the free-living form to the symbiotic form [11], there remain many unclear points about the physiological significance of PHB.
Ralstonia eutropha was studied intensively to elucidate the mechanisms controlling PHB granule stabilization [13], which requires PhaR and phasin proteins (PhaPs) [14]. When the PHB level is low, the PhaR repressor binds to the promoter regions of its own gene and phaP genes to minimize their expression. PhaR binds not only to DNA but also to PHB. Once PHB production is commenced, the intracellular concentration of PHB reaches a certain level that antagonizes DNA binding of PhaR, and PhaR dissociates from the promoter regions. At the initial stage of PHB accumulation, PhaR binds to PHB granules instead of DNA, and transcription of the phaP promoters is elevated as PhaR dissociates from DNA to induce PhaP synthesis. Since PhaPs have a higher affinity for PHB than PhaR, the increasing amount of PhaPs induces dissociation of PhaR from PHB. PhaPs are amphipathic and stabilize hydrophobic PHB granules. Finally, PhaR released from PHB binds once again to the promoter regions to repress transcription of the phaR and phaP genes. Repeating this process, PHB accumulates in the cytoplasm as large stabilized granules. In B. diazoefficiens USDA110, PHB accumulation is achieved in a similar way as in R. eutropha [15]. There are at least four PhaP paralogs and PhaR, which bound to PHB in vitro; PhaP4 exhibited the highest affinity to PHB, which could compete with PhaR to induce its dissociation from PHB [12].
Disruption of phaR in R. eutropha led to a drastic reduction in PHB accumulation [13,14,16]. In addition, phaR inactivation in Methylobacterium extorquens resulted in slower proliferation and a decrease in PHB accumulation [17]. Furthermore, under anaerobic conditions, disruption of aniA (phaR) in S. meliloti reduced PHB accumulation but increased the accumulation of exopolysaccharides (EPS) [18]. In addition, aniA (phaR)-disruption in Rhizobium etli [19] and B. diazoefficiens [9] produced similar phenotypes as the S. meliloti mutant, even under aerobic conditions. In the present study, we constructed and investigated a phaR-inactive mutant of B. diazoefficiens USDA110. Our transcriptomic analysis reveals that PhaR plays an important role not only in PHB accumulation but also in pleiotropic regulation. Furthermore, we analyzed DNA binding of a recombinant PhaR with a C-terminal His 6 -tag fusion in vitro, and deduced a consensus sequence for PhaR binding, providing a holistic view of PhaR regulon.

Methods
Bacterial strains, plasmids, culture conditions, and oligonucleotide primers Bacterial strains and plasmids used in this study are listed in Table 1. Strains of Escherichia coli were cultured in lysogeny broth (LB) [20] at 37°C with shaking at 180 rpm. Strains of B. diazoefficiens were maintained on peptone salts yeast extract medium (PSY) [21] at 28°C. A fresh colony was selected from a media plate and was inoculated into test tubes containing 5 mL of PSY and precultured at 28°C with shaking. An aliquot of the culture was diluted in 5 mL of yeast extract mannitol medium (YEM) [22] or tryptone yeast medium (TY) [23] and cultured further at 28°C with shaking. When needed, polymyxin B (Pm) and kanamycin (Km) were added at 50 μg/mL, as noted in the following procedures. For cultivation under microaerobic conditions, plates were sealed inside an AnaeroPouch-Anaero anaerobic gas generator system (Mitsubishi Gas Chemical) and incubated at 28°C. The oligonucleotide primers used in this study are listed in Additional file 1: Table S1.

Mutant construction
Two DNA fragments were amplified from the chromosomal DNA of B. diazoefficiens USDA110 by PCR using the primer pairs FFphaR/FRphaR and RFphaR/RRphaR (Additional file 1: Table S1). The fragments were fused by a recombinant PCR reaction using the primer pairs FFphaR/RRphaR (Additional file 1: Table S1). The fusion PCR product, that corresponds to a mutated phaR with an internal deletion from the 21st to 185th codons, was trimmed by two restriction enzymes, XbaI and HindIII, ligated into plasmid pK18mobsacB (Table 1) cleaved previously with the same enzymes, and transformed into E. coli DH5α. Correct construction of the recombinant plasmid was confirmed by DNA sequencing. The plasmid was transferred into E. coli S17-1, which was then cultured overnight at 37°C as the plasmid donor. USDA110 as the recipient was cultured with Pm at 28°C until the OD 600 = 1.0. Equal volumes of the donor and the recipient cultures were mixed and spun down. The collected cells were suspended in sterilized water and pipetted onto a membrane filter (Merck Millipore) placed on a PSY agar plate without antibiotics and incubated overnight at 30°C. Following this incubation, the bacterial cells on the filter were suspended in sterilized water, spread on PSY agar plates containing both Pm and Km, and incubated at 28°C to form colonies. Chromosomal DNA of some colonies was subjected to PCR using the primers FFphaR and RRphaR to choose a candidate clone having the entire plasmid DNA integrated into the target site of the chromosome. The chosen clone was cultured in PSY containing Pm alone to induce a pop-out event-spontaneous intra-chromosomal recombination of the plasmid DNA encoding Km resistance and sacB genes-and plated on PSY containing both 10% sucrose and Pm. The resulting colonies were duplicated on two plates, one containing Km and the other Pm. A clone losing Km resistance, whose correct construction was confirmed by PCR analysis and DNA sequencing with primers phaR-c-F and phaR-c-R (Additional file 1: Table S1), was designated as strain ΔphaR with deletion of phaR.

Phenotypic analysis
To determine PHB accumulation, bacterial cells (1 OD 600 unit) were collected and suspended in 5% sodium hypochlorite and incubated overnight at 37°C. After a brief centrifugation, the precipitate was suspended in distilled water, and mixed with a twofold volume of 99.5% ethanol and acetone 1:1 mixture. After another spinning down, the precipitant was dissolved in chloroform at 50°C. The solution was transferred to another tube and air-dried. After addition of 95% sulfuric acid, the tube was incubated at 100°C for 20 min to lyse the cells, and OD 235 of the solution was measured to calculate the amount of PHB. A standard curve was prepared by using various amounts of PHB (Sigma-Aldrich) as previously described [24].
To evaluate EPS production, bacterial cells (1 OD 600 unit) were spun down, and the supernatant mixed with a threefold volume of ethanol, chilled at − 20°C for 10 min, and centrifuged again. The precipitated EPS was air-dried, suspended in water, mixed with 0.2% anthrone (Nacalai Tesque) in 95% sulfuric acid, incubated at 100°C for 10 min, and cooled under running water. The OD 620 of the solution was measured to calculate the amount of EPS as previously described [25].
To measure the intracellular glycogen, bacterial cells were collected and suspended in 30% KOH and heated at 100°C for 20 min. After adding threefold volume of ethanol, the tube was centrifuged and the supernatant discarded. The precipitate was suspended in water, mixed with an equal volume of 0.2% anthrone in 95% sulfuric acid, incubated at 100°C for 10 min, and cooled under running water. The OD 620 of the solution was measured, and a calibration curve was prepared using glucose to calculate the amount of glycogen as described [26]. Stress tolerance was tested as follows: bacterial strains were precultured in TY and YEM to stationary phase. For the heat shock stress experiments, diluted cultures were incubated at 50°C for 10 min with shaking. For the osmotic stress experiments, cells were mixed with glucose, NaCl, or sucrose at various concentrations and then incubated at 28°C for 24 h with shaking. Following incubation, the cultures were serially diluted, and aliquots were spotted on PSY agar plates and incubated at 28°C.

RNA preparation and analysis
Strains of B. diazoefficiens were grown in 50 mL of TY at 28°C with shaking until the OD 600 = 0.4. Cells from these cultures were harvested and suspended in 1 mL of LETS buffer (100 mM LiCl, 10 mM EDTA, 10 mM Tris-HCl at pH 7.4), vortex mixed with 0.5 mL of glass beads (φ0.5 mm) and 1 mL of phenol for 6 min, and centrifuged. After centrifugation, 1 mL of the upper aqueous layer was mixed with 1 mL of phenol-chloroform for 30 s. After another centrifugation, 0.75 mL of the upper aqueous layer was mixed with 0.1 mL of 1 M LiCl and 2.5 mL of ethanol and centrifuged. The supernatant was discarded, and the precipitate was dissolved in 0.175 mL of distilled water, mixed with 3.5-fold volume of 4 M sodium acetate (pH 6.0), incubated at − 20°C for a minimum of 1 h, and centrifuged. The RNA precipitate was dissolved in distilled water, and treated with DNase I at 37°C for 1 h. Then DNA-free RNA was purified using RNeasy Mini Kit (Qiagen). Concentrations of the purified RNA samples were determined with Nano Vue Plus (GE Healthcare), and the samples were stored in aliquots at − 80°C prior to the usage.
The mRNA transcription start points were determined by the 5′-RACE method using the 5′-Full RACE Core Set (Takara Bio). 50 ng of RNA was used as the template for reverse transcription, producing the first cDNA strand from the specific primers, RT-primer phaP1-phos and RT-primer phaP4-phos (Additional file 1: Table S1). After treatment with RNase H, the resulting single stranded cDNA was incubated with T4 RNA ligase overnight at 15°C. The ligated cDNA served as the template in the subsequent nested PCRs with the following primer pairs: phaP1S1/phaP1A1 followed by phaP1S2/phaP1A2 for phaP1; and phaP4S1/phaP4A1 followed by phaP4S2/ phaP4A2 for phaP4 (Additional file 1: Table S1). The resulting PCR fragment was cloned into a pMD20 (Takara Bio) and sequenced to identify the transcription initiation point.
For quantitative RT-PCR, 10 ng of RNA was used as a template for reverse transcription using a ReverTraAce qPCR RT Kit (Toyobo) and each of the reverse primers (Additional file 1: Table S1; RT primer pairs were systematically designated beginning with "RT-," followed by the respective gene name, and "-F" or "-R" at the end). Reverse transcription products were subjected to relative quantification PCR using Thunderbird Sybr qPCR Mix (Toyobo) and the specific primer pairs (Additional file 1: Table S1) in a Thermal Cycler Dice Real Time System MRQ (Takara Bio).
For RNA-sequencing transcriptome analysis, the total RNA was treated using a Ribo-Zero rRNA Removal Kit (Bacteria) (Illumina) to eliminate rRNA, treated with an Agilent RNA Pico kit (Agilent Technologies), and analyzed for purity using a 2100 Bioanalyzer (Agilent Technologies). The mRNA was converted into a cDNA library by using a NEBNext Ultra Directional RNA Library Prep Kit for Illumina (New England Biolabs), NEB Next Multiplex Oligos for Illumina Index Primers Set I (New England Biolabs), and an Agencourt AMPure XP PCR purification system (Beckman Coulter). Quality of the cDNA library was evaluated using an Agilent High Sensitivity DNA Kit and 2100 Bioanalyzer (Agilent Technologies). The cDNA library was subjected to sequencing using a MiSeq Reagent kit v 3 (150 cycles) (Illumina) in Illumina MiSeq (Illumina).

Databases and software
RNA sequencing data were analyzed using CLC Genomics Workbench version 6.5.1 (Qiagen) as follows. The complete genome sequence of B. japonicum (diazoefficiens) USDA110 was obtained as a reference from the NCBI database (accession number NC_004463.1). Quality-filtered sequencing reads were aligned to the reference genome sequence. Mapping was based on a minimal length of 100 bp with an allowance of up to two mismatches. Relative transcript abundance was measured in reads per kilobase per megabase of library size (RPKM). Results of the analysis were submitted to the DDBJ Sequence Read Archive (DRA) database under accession number DRA005621. Gene functions were predicted using the Kyoto Encyclopedia of Genes and Genomes (http://www.genome.jp/kegg/) and rhizobase (http://genome.microbedb.jp/rhizobase/). PhaR-binding sequences were predicted by GLAM2 version 1056 (http://meme-suite.org/tools/glam2) in MEME suite version 4.11.2 (http://meme-suite.org/) to detect a motif for consensus sequence under the default settings.
Preparation of PhaR-His 6 B. diazoefficiens phaR was expressed in E. coli BL21(DE3) to prepare PhaR with a C-terminal His 6 -tag (PhaR-His 6 ) as follows: a PCR fragment of the phaR-coding region was amplified from the DNA of USDA110 using the primer pair pET28b-phaR-gib-F/pET28b-phaR-gib-His6-R (Additional file 1: Table S1). Another PCR fragment was made from plasmid pET28b (Table 1, Takara Bio) with the primer pair pET28b-inverse NcoI-F/pET28b-inverse BamHI-R (Additional file 1: Table S1). This fragment was ligated with the phaR fragment using a Gibbson assembly kit (New England Biolabs) to obtain the plasmid pET28PhaR (Table 1), the correct construction of which was confirmed by sequencing. pET28PhaR was introduced into BL21(DE3), which was cultured in LB containing Km until the OD 600 reached 0.6, and phaR was induced by the addition of 1 mM IPTG for 4 h. The cells were harvested and suspended in a buffer (50 mM phosphate at pH 8.0, 20% glycerol, and 0.5 M NaCl), and disrupted by sonication in an ice bath. From the lysate, PhaR-His 6 was purified using TALON Metal Affinity Resin (Takara Bio).
Gel shift assay and DNase I footprinting DNA fragments used for the gel shift assay were PCR fragments amplified from the DNA of USDA110 using the specific primer pairs (Additional file 1: Table S1; They were systematically designated beginning with "EMSA-", followed by the respective gene name, and "-F" or "-R" at the end). The DNA fragments were combined in the binding-reaction mixture containing 285 mM Tris-HCl (pH 8.0), 0.14 mg/mL bovine serum albumin, 1.43 mM dithiothreitol, 5% glycerol, 0.04 mg/mL poly-deoxyinosinic-deoxycytidylic acid (Merck), and 3.06 mM EDTA. Addition of serially diluted PhaR-His 6 yielded a total reaction volume of 10 μL, which was incubated for 30 min at 25°C. The reaction was then mixed with 80% glycerol and applied to 6% PAGE in TAE buffer (pH 8.2) at 100 V and 20 mA for 2 h. The gel was stained with SYBR Green, and the DNA was visualized using the Gel Doc XR+ system (Bio-Rad).
For DNase I footprinting analysis, DNA fragments were differentially labeled at the 5′-teminus on either strand of DNA using PCR with FAM-labeled primers (Additional file 1: Table S1, the primers designated beginning with FAM were labeled at 5′-end). For labeling the upper and lower strands of the phaP1 promoter region fragment, the primer pairs FAM-phaP1-F/EMSA-phaP1-R and EMSA-phaP1-F/FAM-phaP1-R were used, respectively. For labeling the upper and lower strands of the phaP4 promoter region fragment, the pairs of FAM-phaP4-F/EMSA--phaP4-R and EMSA-phaP4-F/FAM-phaP4-R were used, respectively. Various amounts of purified PhaR-His 6 and each DNA fragment were combined in the binding reaction and incubated for 30 min at 25°C. Subsequently, the mixture was treated with 0.025-0.1 U/μl DNase I (Takara Bio) for 5 min. The reaction was terminated by the addition of EDTA, and the DNA in the mixture was purified using a QIAquick Nucleotide Removal Kit (Qiagen). Corresponding sequence ladders were prepared using a Thermo Sequenase Dye Primer Manual Cycle Sequencing Kit (Thermo Fisher Scientific) according to the standard procedure. The fragment analysis was outsourced to Sigma-Aldrich, Japan.

Phenotypic changes caused by inactivation of phaR
Two strains of B. diazoefficiens, the wild-type USDA110 and its mutant ΔphaR, were grown in TY and YEM ( Fig. 1a and b, respectively). TY is rich in nitrogen sources, and USDA110 grown in TY does not accumulate PHB [12]. When grown in TY, there were no obvious differences in the growth of the two strains in 7 days (Fig. 1a). In YEM, which has a higher carbon/nitrogen ratio, USDA110 is known to accumulate PHB [12]. USDA110 was able to grow to an OD 600 > 1.0 in 20 days in YEM (Fig. 1b), and its PHB content gradually accumulated up to 140 μg/OD 600 in 12 days (Fig. 1c). In contrast, ΔphaR grew in YEM with less cell yield with an OD 600 ≤ 0.5 in 12 days (Fig. 1b), and its PHB accumulation was observed at the highest level of 35 μg/OD 600 after 4 days, but decreased later to leave only residual amounts after 12 days (Fig. 1c). USDA110 produces EPS when grown in YEM, and its EPS production was observed up to 12 μg/OD 600 in 4 days, and decreased to about 5 μg/ OD 600 after 12 days. The EPS production of ΔphaR was nearly constant, around 10 μg/OD 600 during the 12 day growth (Fig. 1d). Both USDA110 and ΔphaR produced similar intracellular glycogen, at 11.1 ± 1.14 and 11.7 ± 2.32 μg/OD 600 in 12 days, respectively (data not shown).
The defects in phaR led to changes in stress tolerance. After incubation at 50°C, ΔphaR demonstrated better survival than USDA110 on both TY and YEM plates (Additional file 2: Figure S1). In contrast, under the osmotic stress conditions with elevated concentrations of NaCl, sucrose, or glucose, ΔphaR grew similarly to USDA110 (data not shown). Surprisingly, when cultured on TY agar plates under microaerobic conditions, ΔphaR grew better than USDA110, whereas the two strains exhibited no difference in growth under aerobic conditions (Additional file 3: Figure S2).
Taken together, inactivation of phaR caused various changes in the phenotype of B. diazoefficiens, including reduced growth and decreased PHB accumulation, along with elevated production of EPS when grown in YEM. In addition, it enhanced tolerance to heat-shock and microaerobic stresses. These results imply that phaR may be involved in regulation of not only PHB accumulation but also pleiotropic cellular functions.
Transcriptomic changes caused by inactivation of phaR B. diazoefficiens USDA110 does not accumulate PHB in TY, and thus we can assume that in these growth conditions PhaR exerts its regulatory function to repress phaP genes, as PHB granules are not developing [12]. To test this possibility, USDA110 and ΔphaR were grown in TY, and total RNA was extracted and subjected to analyze transcriptomic changes.
Expression of phbA1, phbB1, phbC1, phbC2, phbC3, and phbC4 remained at relatively low levels, whereas expression of phbA2 remained constitutive without obvious change. Interestingly, ΔphaR demonstrated an approximately twofold decrease in phbB2 expression compared with that in USDA110, whereas expression of both phbC5 and phbC3 was elevated. With respect to the PHB-degrading genes, transcription of phaZ1 was elevated in ΔphaR almost threefold over that of USDA110, whereas expression of phaZ2 was only negligible. The quantitative RT-PCR experiments revealed that PhaR could be involved in regulation of not only phaP1 and phaP4 but also phbB2 and phaZ1 involved in PHB synthesis and degradation, respectively. In addition, inactivation of phaR led to various changes in phenotype, as described above. These results demonstrate that PhaR may regulate a large number of targets.
In order to assess the possibility, a transcriptome analysis as a screening test was performed by means of  Table S2). As shown (Additional file 5: Figure S3), the distribution of RPKM values indicated that some genes exhibited altered expression with phaR inactivation. The genes in ΔphaR found to be induced in quantitative RT-PCR (Fig. 2) were also induced in RNA sequencing (Additional file 5: Figure  S3). Nevertheless, RNA sequencing was more sensitive than quantitative RT-PCR, and it suggested that expression of 77 genes was elevated for more than fourfold in ΔphaR (Additional file 4: Table S2). Expression of phaP1, phaP4, and phaZ1 was greatly increased in ΔphaR as compared with USDA110. We found that the transcription of phaP5 (blr2887), which may encode an additional PhaP, prominently increased in ΔphaR. In addition, there was an elevation in the expression of the cyoABCD operon, which encodes the cytochrome o ubiquinol oxidase complex, one of the terminal oxidase complexes of the electron transport chain, which is usually induced in adaptation to microaerobic conditions [27]. Expression of the operon blr2367-blr2371 containing exoZ, encoding an acetyltransferase that transfers an acetyl group to EPS, was also increased, implying that this operon may be involved in the retained synthesis of EPS in ΔphaR (Fig. 1d). In ΔphaR, there were increases in the expression of fixNOQP (cbb3-type cytochrome c oxidase), fixGHIS (Fe-S-binding protein), hemN2 (coproporphyrinogen III oxidase), ppsA (phosphoenolpyruvate synthase), and nnrR (regulator for nitric oxide metabolism), which are regulated by FixK 2 [28]. In addition, the napEDABC operon (periplasmic nitrate reductase system) and nirK (copper-containing nitrite reductase) were also induced, which are reported to be induced by NnrR during adaptation to anaerobic conditions [29,30]. These results suggest that inactivation of phaR may activate the FixK 2 and NnrR regulons. Furthermore, some stress response genes also experienced increased transcription, including blr2761 and bll6069, whose products contain the Usp (universal stress protein) motif [31].
Unexpectedly, a large number of genes were suggested to decrease transcription in ΔphaR compared with USDA110 (Additional file 4: Table S2), as expression of 119 genes was lowered for more than fourfold in ΔphaR, including a putative operon for ABC transporter (blr0308-blr0312), a large gene cluster for polyketide synthesis including another ABC transporter (bll3369-bll3384), and the other genes related to flagellar formation and movement, such as fliF1, fliR2, flhA2, flhB2, flgB2, flgC2, flgE2, flgG2, flgL2, motB2, and motC. The results further suggest that PhaR may upregulate some active transport systems as well as cellular motility.

DNA binding of PhaR
Gel mobility shift assays using serially concentrated PhaR-His 6 produced and purified in E. coli (Additional file 6: Figure S4) revealed that PhaR-His 6 bound to the promoter region of phaP1, phaP4, phaP5, and phaR itself (Fig. 3a-d).
Of the four DNA fragments, PhaR-His 6 bound to the phaP1 promoter most efficiently, as a decrease in the DNA fragment band was observed at 6.89 nM of PhaR-His 6 , and a distinct DNA-protein complex band appeared at 55 nM (Fig. 3a). However, for the phaP4 promoter fragment, shifting of the band occurred at 13.8 nM and the complex appeared at 110 nM (Fig. 3b). For the fragment containing the phaP5 promoter, which was suggested as an additional PhaR target after the transcriptome analysis described above, shifting of the band occurred at 27.8 nM and the complex was formed at 110 nM (Fig. 3c). These results indicated that PhaR-His 6 bound to the promoter regions of phaP1, phaP4, and phaP5. In addition, PhaR-His 6 was shown to bind to the DNA fragment of the phaR promoter (Fig. 3d), but only half of the DNA formed the complex at the highest PhaR-His 6 concentration of 110 nM, suggesting a weaker affinity for the phaR promoter region. Next, gel mobility shift assays were performed as stated above with DNA fragments from the phaZ1 and phaZ3 (blr0899) promoters for PHB degrading enzymes ( Fig. 3e and f, respectively). The phaZ1 fragment shifted at 27.8 nM and completed the complex formation at 110 nM (Fig. 3e), whereas the phaZ3 fragment shifted at 27.8 nM and completed the complex at 55 nM (Fig. 3f ). These results suggested that these genes for PHB degrading enzymes could be regulated directly by PhaR.
Further, we assessed PhaR-His 6 binding to the promoter regions of the possible PhaR targets suggested by the RNA sequencing transcriptome analysis ( Fig. 3g and h). The DNA fragment from the promoter region of cyoA (cyoABCD operon) shifted at 27.8 nM of PhaR-His 6 and completed the complex formation at 55 nM, suggesting higher affinity (Fig. 3g). The fragment from the promoter region of blr2367 for the exoZ-containing operon, shifted at 27.8 nM and completed the complex at 110 nM (Fig. 3h). In contrast, despite that its transcription was elevated more than fourfold in ΔphaR, no PhaR-His 6 binding to the fragments containing promoter regions of fixN, ppsA, nnrR, and hemN2 was detected, suggesting that PhaR may not directly regulate the expression of these genes (data not shown).
We performed additional gel mobility shift assays on the promoter regions of phaP1 and phaP4 in the presence of PHB (Additional file 7: Figure S5). In both cases, the PhaR-DNA complex formation decreased as the concentration of PHB increased. When 50 ng of PHB was added to the reaction mixture, binding of PhaR-His 6 to the DNA was almost abolished, as the DNA bands appeared at the same position as the control without PHB, suggesting that PHB could inactivate DNA binding of PhaR-His 6 .
In order to determine the PhaR-His 6 binding sites, we performed DNase I footprint analyses on the phaP1 and phaP4 promoter regions ( Fig. 4a and b, respectively). We found two sites occupied by PhaR-His 6 within the phaP1 promoter region and one in the phaP4 promoter region. In addition, the respective transcription start points (+ 1) for the two promoters were determined by the 5′-RACE method using the RNA samples prepared from the ΔphaR cells grown in TY, and the corresponding − 10 and − 35 regions were deduced (Fig. 4). The two PhaRbinding sites with 12-mer sequences TGCGACGCACAA and TGCGTCGCACAA in the phaP1 promoter region are located 6 bp upstream and downstream of the − 35 region, respectively, and the binding site with a sequence TGCG CTGCACAA in the phaP4 promoter region overlaps the − 35 region. The presence of two binding sites within the phaP1 promoter region may be responsible for its higher affinity to PhaR-His 6 in vitro.

Additional genes regulated by PhaR
Since PhaR-His 6 bound to the DNA fragments containing the additional six promoter regions, including those of phaP5, phaR, phaZ1, phaZ3, the cyoABCD operon, and the exoZ-containing operon starting with blr2367 (Fig. 3), we speculated that these promoter regions could share a sequence similarity. Nucleotide sequences of the six promoter regions were aligned with three 12-mer sequences of the PhaR-binding sites found in the phaP1 and phaP4 promoter regions (Fig. 4) by GLAM2 in MEME suit to deduce a motif of consensus sequence for PhaR binding as TGCRNYGCASMA (R: A or G, Y: C or T, S: C or G, M: A or C) (Fig. 5).
We searched for sequences similar to the consensus sequence within the putative promoter regions of the genes whose transcription increased or decreased more than twofold in ΔphaR compared with USDA110 (Additional file 4: Table S2). Their putative promoter regions were selected on the basis of the possible transcription start points reported in recent studies [32]. We extracted 28 and 42 genes associated with the PhaR-binding consensus-like sequence, whose transcription increased ( Table 2) and decreased  (Table 3) in ΔphaR, respectively. Among the genes whose transcription was increased in ΔphaR (i.e., PhaR-repressed), the ppc (encoding phosphoenolpyruvate carboxylase) promoter region demonstrated a weaker affinity for PhaR-His 6 , as approximately half of the DNA band shifted only at the highest PhaR-His 6 concentration of 110 nM (Fig. 6). However, among the genes whose transcription decreased (PhaR-activated), DNA fragments containing the promoter regions of bll5961 (encoding a putative regulatory protein of Crp family), exaA (quinoprotein ethanol dehydrogenase), pckA (phosphoenolpyruvate carboxykinase), pdhA (pyruvate dehydrogenase subunit alpha), and phbB2 demonstrated some mobility shift at 110 nM of PhaR-His 6 (Fig. 6).
The results suggested that these consensus-sequence sites might have lower affinities to PhaR but possibly be involved in the regulation by PhaR.

Discussion
In the present study, phenotypic changes resulting from phaR inactivation were investigated. There were no observable differences between ΔphaR and USDA110 grown in TY, whereas ΔphaR demonstrated decreased cell yield and diminished PHB accumulation in YEM (Fig. 1). These results suggest that altered PHB metabolism compromises the balance in intracellular carbon source usage. In fact, expression of pdhAB for acetyl-CoA synthesis was reduced (See figure on previous page.) Fig. 3 PhaR binding to DNA fragments containing the promoter regions of phaP paralogs (a-c), phaR (d), phaZ paralogs (e and f), and others (g and h). Various amounts of purified PhaR-His 6 were incubated with a fixed amount of DNA fragments containing the promoter regions as indicated; (a) phaP1 (bll5155), (b) phaP4 (bll7395), (c) phaP5 (blr2887), and (d) phaR (blr0227), (e) phaZ1 (blr0908), (f) phaZ3 (blr0899), (g) cyoA (blr0149), and (h) blr2367. The reactions were subjected to gel mobility shift assays as described in the main text. Each of the gels has eight lanes containing serially increased amounts of PhaR-His 6 from left to right to give 0, 1.72, 3.44, 6.88, 13.8, 27.5, 55, and 110 nM, respectively. Positions for DNA-protein complex and free DNA fragments are indicated with arrowheads on the left side of the panels. As a negative control (NC), a constant amount of the DNA fragment corresponding to part of the phaP1 coding region is included, which is another PCR fragment amplified using the primer pairs EMSA-phaP1-ORF-F/EMSA-phaP1-ORF-R (Additional file 1: Table S1) in ΔphaR and transcription of ppc and ppsA increased, whereas those of pckA and exaA decreased. Together, these changes may cause a distortion in carbon metabolism. However, amounts of EPS produced in ΔphaR were approximately double those produced in USDA110 (Fig. 1), indicating that part of the energy and resources for PHB accumulation may be used for EPS production in ΔphaR.
In addition, transcription of the operon containing exoZ increased in ΔphaR. Reduced EPS production was reported in an exoZ mutant strain of S. meliloti [33]. Therefore, the opposite effect may occur upon elevated expression of exoZ in ΔphaR, resulting in an increase in EPS production. Recently, the similar phenotypic changes were reported for another mutant with phaR inactivation [9]. On the other hand, we found ΔphaR improved thermal stress tolerance in both YEM and TY regardless of PHB accumulation (Additional file 2: Figure S1). It is known that PhaP expressed in E. coli exerted a molecular chaperone-like effect [34,35]; in ΔphaR, PhaP1, PhaP4, and PhaP5 are produced in excess, which may restore proteins denatured under heat stress conditions. In addition, transcription of both blr2761 and bll6069 was elevated in ΔphaR, each of which encodes a protein containing the Usp motif, similar to E. coli uspA, which is involved in stress tolerance [31]. Interestingly, ΔphaR grew better under microaerobic conditions (Additional file 3: Figure S2). In ΔphaR, some genes that function under hypoxic conditions were enhanced (Additional file 4: Table S2), such as fixNOQP, fixGHIS, napEDABC, and nirK, and were increased, and these genes could be involved in enhancing growth and nitrogen fixation under anaerobic conditions. Nevertheless, the ΔphaR nodulation in soybean was nearly normal, with no obvious alteration in symbiotic nitrogen fixation under laboratory conditions (data not shown). The cyoABCDE operon of E. coli is induced under higher oxygen conditions [36], whereas in R. etli the cyoABCD operon was induced under hypoxic conditions to support growth [27]. Therefore, if the cyoABCD operon functions similarly in B. diazoefficiens as it does in R. etli, the increased expression of the cyoABCD operon in ΔphaR may explain its better growth under the microaerobic conditions. PhaR could directly repress transcription of various genes; at least 28 of the genes may be negatively regulated and associated with a PhaR-binding site, including the cyoABCD operon, the operon involved in EPS production that contains exoZ, and ppc ( Table 2). Inactivation of phaR also led to downregulation of a large number of genes, suggesting PhaR may function as a positive regulator; at least 42 of the genes may be associated with a PhaRbinding site (Table 3). Some genes involved in carbon metabolism, such as the pdhAB operon, pckA, and exaA could be upregulated through PhaR binding to their promoter regions (Fig. 6). In addition, PhaR binds to the promoter region of phbB2 (Fig. 6), previously shown to be important for PHB synthesis [12], the transcription of which decreased twofold in ΔphaR (Fig. 2). In R. etli, PhaB disappeared in the aniA (a phaR homolog) mutant [18], implying direct involvement of PhaR in the production of PhbB2. In any case, further studies are required to understand the mechanism of PhaR in transcriptional activation.
We found two PhaR-His 6 binding sites in the phaP1 promoter and one in the phaP4 promoter, as determined by DNase I footprint analyses (Fig. 4). The gel mobility shift assays indicated a higher PhaR-His 6 affinity for the phaP1 promoter than for the phaP4 promoter (Fig. 3), suggesting that PhaR could exert tighter control over phaP1 than phaP4. In a previous report, transcriptional activity of the phaR promoter in a phaR mutant doubled over that of the wild-type strain of B. diazoefficiens [9]. In the present study, the gel mobility shift assays revealed that PhaR-His 6 binds to the phaR promoter region inefficiently (Fig. 3), implying possible leakiness in its own repression, allowing a constitutive level of PhaR expression in the cell. We demonstrated that DNA binding of PhaR-His 6 was abolished in the presence of PHB in vitro (Additional file 7: Figure S5). These results were in agreement with the previous finding in Paracoccus denitrificans [37] and our previous observation that PhaR competed for PHB binding with PhaPs and was released from PHB as concentrations of PhaPs increased [12]. The liberated PhaR then binds DNA again to repress the PhaP genes. In our previous studies, PhaP4 showed a higher affinity for PHB than PhaP1 [12]. It is suggested that PhaR could control phaP1 more tightly than phaP4 (Fig. 3). Nucleotide sequences of the promoter regions of phaP5, phaR, phaZ1, phaZ3, the cyoABCD operon, and the exoZ-containing operon starting with blr2367 were aligned with three 12-mer sequences for PhaR-binding found in the phaP1 and phaP4 promoter regions by GLAM2 in MEME suit under the default settings for analysis to deduce a motif for PhaR binding. The graphical summary of GLAM2 analysis is shown with the consensus sequence TGCRNYGCASMA (R: A or G, Y: C or T, S: C or G, M: A or C) at the bottom Therefore, given its higher affinity for PHB, PhaP4 may be readily produced to release PhaR from PHB. It is possible that PhaP5 may also be play an important role, as it was induced almost fivefold over phaP4 in ΔphaR (Table 2). Since PhaP1, PhaP4, and PhaP5 were all overexpressed in ΔphaR, PHB could be covered immediately by these PhaPs, thereby inhibiting the expansion of PHB granules. In addition, increasing expression of phaZ1 and phaZ3 could result in degradation of PHB. Previous in vitro experiments revealed that PhaPs on the surface of PHB granules could contribute to activity of PhaZ enzymes, and the possible interaction between PhaPs and PhaZ enzymes may have an effect on the decrease of accumulated PHB [38,39].
In Rhodobacter sphaeroides, PhaR repressed transcription of phaZ for PHB degradation [40]. In the present study, we found that transcription of phaZ1 in ΔphaR increased about fourfold over that in USDA110 (Table 2) and PhaR-His 6 bound to the phaZ1 promoter region in vitro (Fig. 3). Furthermore, a third gene for PHB degradation, phaZ3, was found; its expression was elevated by inactivation of phaR (Table 2), and PhaR-His 6 binding to its promoter region was demonstrated in vitro (Fig. 3). Therefore, PhaR of B. diazoefficiens could negatively and directly control transcription of the two PHB degrading genes, phaZ1 and phaZ3. PhaZ3 has the esterase PHB depolymerase motif, whereas PhaZ1 and PhaZ2 have the PHB depolymerase C-terminus motif (data not shown). In addition, a blast search against the genome of R. eutropha strain H16 revealed that PhaZ3 shares a homology to PhaZ7 of H16 (E-value = 5e- 16). It was demonstrated that induction of phaZ7 in a recombinant E. coli strain greatly reduced PHB   accumulation [41]. Therefore, PhaZ3 may play an important role in PHB degradation in B. diazoefficiens. The transcriptome analysis revealed an additional PhaP gene, phaP5, the expression of which was prominent in ΔphaR compared with USDA110 (Additional file 4: Tables 2). PhaP5 possesses the phasin 2 motif, similar to that found in the other four PhaPs. In the present study, we demonstrated PhaR-His 6 binding to the promoter regions of these PhaP genes using gel mobility shift assays (Fig. 3). It is likely that PhaR binds to the promoter regions of phaP1, phaP4, and phaP5 and represses their transcription under non-PHB-accumulating conditions. These results suggested that PhaP5 may play an important role in PHB accumulation together with PhaP1 and PhaP4.
It was also revealed that some of the genes regulated by FixK 2 , a global transcriptional activator for adaptation to microaerobic conditions, increased their expression in ΔphaR, PhaR-His 6 may bind to the fixK2 promoter region. However, a DNA fragment of the fixK2 promoter region exhibited no interaction with PhaR-His 6 under the conditions used in the present study (data not shown). When cultured under microaerobic conditions, transcription of fixK2 increased in a mutant strain of B. diazoefficiens lacking phaR [9]. In the present study, however, transcription of fixK2 in ΔphaR was not elevated significantly (Additional file 4: Table S2), and no PhaR-binding site was predicted in the promoter region of fixK2. Nevertheless, FixK 2 is activated in response to a moderate decrease in oxygen concentration [28]. As described above, we found that transcription of both the napEDABC operon and nirK, which function in the uptake and reduction of nitric acid under anaerobic conditions, was elevated in ΔphaR [29,30]. This implies that phaR inactivation may make the intracellular environment less aerobic as the metabolic circuit is modulated to accelerate respiration to consume more oxygen, and thus available oxygen could decrease to activate FixK 2 for induction of its targets. PhbC3 could be active as a homodimer catalyzing PHB polymerization, whereas PhbC5 may bind to PhbC3 to form an inactive heterodimer [9]. In the present study, we found that PhaR may not directly control phbC3 and phbC5. It was previously suggested that FixK 2 may be involved in the induction of phbC5 [9]. Expression of phbC5 in ΔphaR appeared to increase slightly (Fig. 2), which may be one of the additional reasons why PHB did not accumulate in ΔphaR (Fig. 1).