Evidence that Altered Cis Element Spacing Affects PpsR Mediated Redox Control of Photosynthesis Gene Expression in Rubrivivax gelatinosus

PpsR is a major regulator of photosynthesis gene expression among all characterized purple photosynthetic bacteria. This transcription regulator has been extensively characterized in Rhodobacter (Rba.) capsulatus and Rba. sphaeroides which are members of the α-proteobacteria lineage. In this study, we have investigated the biochemical properties and mutational effects of a ppsR deletion strain in the β-proteobacterium Rubrivivax (Rvi.) gelatinosus in order to reveal phylogenetically conserved mechanisms and species-specific characteristics. A deletion of the ppsR gene resulted in de-repression of photosystem synthesis showing that PpsR functions as a repressor of photosynthesis genes in this species. We also constructed a Rvi. gelatinosus PpsR mutant in which a conserved cysteine at position 436 was changed to an alanine to examine whether or not this residue is important for sensing redox, as reported in Rhodobacter species. Surprisingly, the Cys436 Ala mutant retained the ability to repress photosynthesis gene expression under aerobic conditions, suggesting that PpsR from Rvi. gelatinosus has different redox-responding characteristics. Furthermore, biochemical analyses demonstrated that Rvi. gelatinosus PpsR only shows redox-dependent binding to promoters with 9-bp spacing, but not 8-bp spacing, between two PpsR-recognition sequences. These results indicate that redox-dependent binding of PpsR requires appropriate cis configuration of PpsR target sequences in Rvi. gelatinosus. These results also indicate that PpsR homologs from different species regulate photosynthesis genes with altered biochemical properties.


Introduction
Many purple photosynthetic bacteria can grow either using aerobic respiration or anaerobic photosynthesis. These cells tightly control synthesis of their photosynthetic apparatus in response to oxygen tension and light intensity [1,2]. Under aerobic conditions, light-excited bacteriochlorophylls are capable of generating reactive oxygen species (ROS) such as singlet oxygen as a side reaction. ROS are very damaging to cells as they react with many biomolecules including DNA [3]. In order to avoid the production of ROS, most purple photosynthetic bacteria repress synthesis of their photosynthetic apparatus immediately after being exposed to oxygen. One of the main factors responsible for this regulation is the transcription factor PpsR which is universally present among purple photosynthetic bacteria (in some species PpsR is also called CrtJ) [4,5]. However, PpsR/CrtJ homologes from different purple photosynthetic bacterial species only exhibit 26-54% sequence identity indicating that their properties may exhibit variations [6].
Previous studies have demonstrated that PpsR binds to the consensus sequence (TGT-N 12 -ACA) that is present in many photosynthesis gene promoters including those involved in controlling synthesis of carotenoids, bacterichlorophyll and structural light harvesting proteins. PpsR is known to be a redox-responding repressor as it has a higher affinity for target DNA sequences under oxidizing conditions than under reducing conditions in Rba. capsulatus and Rba. sphaeroides [7,8]. This change of binding affinity is thought to be caused by formation of an intramolecular disulfide bond between two cysteine residues (Cys249 and Cys420 in Rba. capsulatus) and/or oxidative modification of the latter cysteine (Cys420) to a sulfenic acid derivative [8,9]. Site-directed mutation of Cys420 in Rba. capsulatus CrtJ or Cys424 in Rba. sphaeroides PpsR resulted in a significant decrease of DNA binding activity of the repressor in vitro [9,10]. Additional expression studies showed that a Cys420 to alanine mutation (C420A) caused elevated aerobic expression of photosynthesis genes to the same level as did a crtJ/ppsR disrupted strain of Rba. capsulatus [9]. This is in contrast to C249A mutation that only partially de-repressed aerobic photosystem expression indicating that the oxidation state of Cys420 is key to redox regulation of PpsR/CrtJ in the tested Rhodobacter species [9].
An interesting divergence occurs in a photosynthetic Bradyrhizobium strain that has two ppsR genes, PpsR1 and PpsR2. Unlike the Rba. capsulatus and Rba. sphaeroides homologs, the DNA binding affinity of Bradyrhizobium PpsR1 is reported to be higher under reducing conditions than under oxidizing conditions. Furthermore PpsR1 contains only one cysteine (Cys429) and this cysteine alone is sufficient for sensing redox [11]. Interestingly, PpsR2 does not contain a similar cysteine in the HTH motif and is thought to be dedicated to light control of photosynthesis gene expression rather than redox control [11]. Finally, PpsR2 from Rhodopseudomonas palustris also lacks a cysteine residue in the HTH motif [12]. Collectively, these observations suggest that there are species-specific alterations in redox-sensing mechanisms employed by PpsR.
To date, most research on PpsR/CrtJ has been confined with members of the α-proteobacteria lineage and, thus the generality of PpsR functionality among purple photosynthetic bacteria is not clear [6,12]. In this study, we examined regulatory characteristics of PpsR from the βproteobacterium Rvi. gelatinosus which exhibits only 33% sequence identity to Rba. capsulatus CrtJ. As is the case with Rhodobacter PpsR homologs, Rvi. gelatinosus PpsR has three cysteine residues with one Cys436 located in the HTH motif. In this study we investigated biochemical and physiological properties of wild type and the C436A mutant PpsR in vitro and in vivo and have highlight dissimilarities that exist from properties of that have been reported for PpsR homologs from other species.

Materials and Methods
Bacterial strains, media, and growth conditions Escherichia (E.) coli strains JM109/λpir, S17-1/λpir and BL21 (DE3) were used for cloning, conjugal transfer of plasmids, and protein overexpression, respectively. E. coli cells were routinely grown in Luria Bertani (LB) medium at 37°C. Kanamycin was used at 50 μg/ml in LB medium.

Cloning and mutagenesis of ppsR
Two 500-bp DNA fragments consisted of N-terminal and C-terminal regions of PpsR were amplified by polymerase chain amplifying (PCR) using KOD-plus-polymerase (TOYOBO). Two sets of primers were used for amplification. One set contained forward primer ppsRxbaIfor (5'-GGTCTAGAGCACTGCTGCTGCCGGC-3') and the reverse primer ppsRsmaIrev (5'-TTTTTCCCGGGTCCAAACGGTCTCATTCGGA-3'). The other set contained forward primer ppsRsmaIfor (5'-TTTTTCCCGGGGAAGCCGAGAAGTAGAGCG-3') and the reverse primer ppsRsacIrev (5'-GGGAGCTCCCAGCGACTTCAGCATCTAC-3'). XbaI, SmaI, SacI and restriction sites were designed at additional polynucleotide tails present in the ppsRxbaIfor, ppsRsmaIrev and ppsRsmaIfor, and ppsRsacIrev primers, respectively (underlined). The first PCR fragment was digested with XbaI and SmaI and the second PCR fragment was digested with SacI and SmaI. After digestion, these two fragments were mixed and ligated together into XbaI-SacI-cut pJPCm [15]. A DNA cassette containing selectable Km r -marker as well as sacB-sacR genes of Bacillus subtilis was then inserted into the plasmid at the outside of the cloned fragment. The plasmid obtained called pJPΔppsR-KmSac was introduced into Rvi. gelatinosus cells by conjugation with the mobilizing E. coli strain S17-1/λpir. Cells undergoing single crossover events were selected by plating exconjugants on PYS plates containing kanamycin and tetracycline. After sequential cultivation in the absence of drug selection, sucrose-resistant but kanamycin-sensitive cells were selected on PYS-agar plates containing 15% sucrose to generate ΔppsR mutant. A deletion of ppsR was confirmed by PCR amplification followed by sequence analysis. For Cys mutant construction, a ppsR clone containing 500-bp of DNA up-and down-stream of the gene was cloned into pJPCm using ppsRxbaIfor and ppsRsacIrev primers. A Cys436 to Ala mutation was then constructed using a QuickChange mutagenesis kit. The resulting plasmid pJPΔppsRC436A was then introduced into Rvi. gelatinosus ΔppsR mutant cells with subsequent double crossover recombination event generating a C436A mutant variant of ppsR.

Membrane preparation and spectrophotometric measurements
Wild-type and mutant strains of Rvi. gelatinosus were grown aerobically and photosynthetically to mid-log phase in PYS medium. Membrane fractions were prepared by French press cell disruption in 5 mM MES buffer (pH 7.0) containing 1 mM MgCl 2 , followed by ultracentrifugation for 20 min at 250,000 xg. The pelleted membranes were suspended and diluted with 5 mM MES buffer (pH 7.0) to a protein concentration of 100 μg protein/ml as measured by a Protein Assay kit (BioRad). Absorption spectra of membranes were recorded from 300 nm to 1000 nm using a UV-1800 spectrophotometer (SHIMADZU).

RNA isolation, quantitative real-time PCR (QRT-PCR)
Wild-type and mutant strains of Rvi. gelatinosus were grown aerobically and photosynthetically to mid-log phase in PYS medium. 1 ml of cells were harvested with total RNA of each sample extracted using SV Total RNA Isolation System (Promega). A typical OD 260 to OD 280 ratio of RNA sample was approximately 2.0.

Overexpression and purification of PpsR and PpsR mutant
A PpsR overexpression vector was constructed by PCR amplification of the PpsR coding sequence using a forward primer containing an NdeI restriction site (5'-GGGGGGCATATGA GACCGTTTGGAGCACC-3'), and a reverse primer containing an HindIII restriction site (5'-TTTAAGCTTCTACTTCTCGGCTTCGGCC-3'). The PCR-amplified DNA segment was cloned into NdeI and HindIII sites in the pSUMO vector (LifeSensors Inc.) to generate the plasmid, pSUMO::PpsR. A Cys mutation in the pSUMO::PpsR construct was subsequently constructed using a QuickChange kit. pSUMO::PpsR was transformed into E. coli strain BL21 (DE3) and the recombinant protein SUMO-PpsR was overexpressed by induction of a 500 ml culture with 0.2 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 16°C overnight (12-16 h). Cells were harvested and resuspended in 20 ml nickel column loading buffer composed of 20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM imidazole and 10% glycerol and lysed by two passes through a French pressure cell operated at 18,000 psi. The lysate was clarified by centrifugation at 30,000 xg for 30 min at 4°C. The resultant supernatant was passed through a 45 μm membrane filter (Millipore) and loaded onto a 1-ml HisTrap column (GE) with an ÄKTApurifier (GE), washed with 30-column-volume of wash buffer containing 20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 20 mM imidazole, and 10% glycerol. SUMO-PpsR was eluted with a gradient of 20 mM imidazole to 500 mM imidazole in the loading buffer over a 15-column-volume total. The SUMO tag was then proteolytically cleaved at room temperature for 1 hour at a 100:1 molar ratio of UlP1 protease. Tag-less PpsR was then isolated by Superose 12 size exclusion chromatography in a PpsR standard buffer containing 20 mM Tris-HCl (pH 8.0), 500 mM NaCl, and 6% glycerol at a rate of 0.5 ml/min [8]. Peak fractions were analyzed by SDS-PAGE and PpsR containing fractions were concentrated. For generating the molecular weight standard curve by the size exclusion chromatography, Gel Filtration Standard (BioRad) was used as protein standard.
Individual footprint reactions were initiated with a 22 μl binding reaction mixture that contained 12.5 mM HEPES (pH 7.8), 5 mM K-acetate (pH 8.0), 2.5 mM Mg-acetate, 1 mM CaCl 2 , 12.5 μg/ml bovine serum albumin [18], 0.3 mg/ml heparin, 200 nM of fluorescence-labeled DNA probe and various amounts of purified protein. Initial binding reaction mixtures were incubated for 30 min at 22°C followed by a 15 min DNase I digestion that was initiated by adding 3 μl of DNase I (New England Biolabs) at an approximately 1:100 dilution (0.02 units/μl) in a footprint binding buffer, which gave partial probe digestion. The digestion reactions were then stopped by adding 25 μl of 0.5 M EDTA (pH 8.0). DNA segments were purified using a MinElute PCR purification kit (QIAGEN), with sample eluted with 15 μl of elution buffer containing 0.5 μl of 500 LIZ Size Standard (Applied Biosystems). Samples were then transferred to a 96-well plate [19,20]. The plate was then sealed using a septum, heated to 95°C for 5 min, and cooled quickly on ice. The samples were separated and detected with a 3730 DNA Analyzer (Applied Biosystems) and analyzed with Peak Scanner Software v1.0 (Applied Biosystems). When reducing conditions was required, fresh DTT was added into all of the buffers and the reaction mixture (10 mM final concentration). Fragment analysis was performed on an Applied Biosystems 3730 automated DNA sequencing machine with fragments detected using the GeneMapper-Generic protocol. Samples were analyzed using SoftGenetics GeneMarker 1.4 analysis software.

PpsR-binding sites in photosynthesis gene promoters in Rvi. gelatinosus
Analysis of the Rvi. gelatinosus genome sequence (accession number AP012320) indicates the presence of putative PpsR-binding motifs were found in pucB and crtI promoter regions in Rvi. gelatinosus and another Rubrivivax species Rvi. benzoatilyticus (Fig 1A). The typical PpsRbinding motif in Rhodobacter species is TGT-N 12 -ACA [13], but it is not perfectly conserved in the Rvi. gelatinosus and Rvi. benzoatilyticus pucB and crtI promoters. Specifically, one of the two PpsR-binding sites, found in the pucB promoter, is TGT-N 12 -ACG in the two Rubrivivax species. The crtI promoter also contains nucleotide variations in the PpsR-binding sites in Rubrivivax species as well as in Rba. capsulatus (Fig 1A). It is also of note that spaces between the two PpsR-binding sites in these promoters are different in these species. Specifically, two PpsR-binding sites in the pucB promoter are separated by 9-bp and 240-bp in Rvi. gelatinosus and Rba. capsulatus, respectively. In the crtI promoters the recognition palindromes are separated by 8-bp and 76-bp in Rvi. gelatinosus and Rba. capsulatus, respectively. In this study, we characterized both pucB and crtI promoters for possible PpsR-binding targets in Rvi. gelatinosus using in vivo and in vitro analyses.

Characterization of ΔppsR and Cys436 to Ala mutants
A computer-based similarity search indicated that Rvi. gelatinosus retains only one ppsR homolog (RGE_33430) on the genome. We tested if PpsR has an in vivo role in controlling photosystem synthesis in Rvi. gelatinosus by constructing a chromosomal deletion of ppsR. We also constructed a strain that had the putative reactive Cys436 present in the HTH domain mutated to Ala. The ΔppsR strain was constructed by a site-specific double recombination event. To avoid the complicating effects on expression of downstream or upstream genes, the clean chromosomal deletion was constructed where the start and stop codon of PpsR were fused (see Materials and methods). The Cys436 to Ala mutation was also constructed using double recombination. Fig 2A and 2B show the result of spectral analysis of Rvi. gelatinosus mutant strains; ΔppsR strain (blue line) and C436A mutant (red line) and wild-type (black line) under two different growth conditions. This analysis was performed three times, and similar results were obtained. Rvi. gelatinosus has two types of light harvesting complexes; they are light harvesting (LH) 1 and LH2 which absorb 875 nm and 800-850 nm wavelengths of light, respectively. As shown in Fig 2B, the ΔppsR strain synthesized more LH1 and LH2 complexes than wild type under anaerobic (photosynthetic) conditions. In contrast, C436A mutant showed similar photosynthetic apparatus levels as compared with wild-type in anaerobic growth conditions. When grown aerobically, the ΔppsR strain exhibits significant increased synthesis of LH1 and LH2 as compared to wild-type and C436A mutant (Fig 2A). Interestingly the C436A mutant exhibited levels of the photosynthetic apparatus that are similar to that of wild type with just slightly higher synthesis of LH1 (Fig 2A, inset).
We also used quantitative real-time PCR to assess mRNA levels of select individual photosynthesis genes. Under anaerobic conditions, the ΔppsR strain exhibited 1.7-to 2-fold elevated pucB and crtI expression as compared with wild-type cells (Fig 3B). In contrast, C436A mutant exhibited similar crtI expression level when compared to wild-type cells although pucB expression was elevated to 1.7-fold ( Fig 3B). Under aerobic conditions, the ΔppsR strain exhibited significant up-regulation of pucB (~190-fold) and crtI (~25-fold) transcription (Fig 3A). C436A mutant also exhibited small (~1.5-fold) but significant elevated pucB and crtI transcription compared with wild type under aerobic condition (Fig 3A).

DNA-binding activity of purified Rvi. gelatinosus PpsR
For in vitro studies, we purified recombinant Rvi. gelatinosus PpsR using E. coli overexpression system followed by affinity chromatography and size exclusion chromatography. The molecular weight of the purified PpsR was estimated to be 223.5 kDa, based on elution profiles of the size exclusion chromatography (Fig 4A). Given the molecular weight of PpsR, calculated from its amino-acid sequence, is 52.4 kDa, the data indicated that PpsR exists as tetramer in solution. Only monomeric 53 kDa band was observed on the SDS-PAGE of the purified PpsR under oxidizing (no treated) and DTT-reducing conditions (Fig 4B), showing that PpsR does not form any intermolecular disulfide bonds. DNase I footprint analysis was undertaken with purified PpsR to assay the extent of protection to the pucB and crtI promoter regions (Fig 5). The protection pattern showed good protection to both promoter regions centered around the PpsR binding recognition sequences as described above. In the pucB promoter region, PpsR binding region overlaps to the -35 and -10 σ-subunit recognition sequences (the promoter recognition sequences in the crtI promoter region has not been determined) [21]. These binding patterns are very similar to that reported for the CrtJ/PpsR binding pattern to the bchC promoter region in Rba. capsulatus [22]. Interestingly, neither protection region exhibited large differences between oxidizing (no treated) and reducing (presence of DTT) conditions. The one notable difference is a slight increase in DNase I sensitivity of a nucleotide in the pucB promoter region as high-lightened by an asterisks in Fig 5A and 5B.
We also assayed the binding affinity by performing gel mobility shift analysis with the purified PpsR and DNA fragments for pucB and crtI promoters. As shown in Fig 1B and 1C, PpsR clearly binds to both promoters in a concentration-dependent manner. We then performed gel mobility shift analysis with small incremental increases (0.007-0.1 μg) of PpsR to generate binding isotherms for determining EC 50 values (effective concentration of PpsR for 50% response) of PpsR under oxidizing (no treated) and reducing (in the presence of 5 mM DTT) conditions (Fig 6A-6D). We used no treated PpsR as oxidized PpsR, because purified PpsR is almost oxidized in the absence DTT in Rhodobacter species [7]. In fact, purified PpsR of Rvi. gelatinosus showed similar binding to pucB promoter with or without H 2 O 2 treatment (S1 Fig). The binding affinity of PpsR to the pucB promoter under reducing conditions was estimated to be EC 50 = 63.26 nM which was a modest 1.27-fold larger than that under oxidizing conditions (49.82 nM) (Fig 6A, Table 1). Furthermore, the binding curve under oxidizing conditions showed more sigmoidal than under reducing conditions (Fig 6A). On the other hand, binding affinity of PpsR to the crtI promoter region showed no significant difference between reducing (EC 50 = 86.46 nM) and oxidizing (EC 50 = 80.46 nM) conditions ( Fig 6B, Table 1). It should be noted that the binding affinity of PpsR to the crtI promoter region was lower than that observed for the pucB promoter region (Table 1).
We next constructed two variants of the crtI promoter (MT1 and MT2) that exhibited a sequence structures that are more typical of previously characterized PpsR-binding sites (TGT-N 12 -ACA). For probe MT1 the Rvi. gelatinosus two PpsR recognition sequences in the crtI promoter was changed from [(TGa-N 12 -ACA)-N 8 -(cGT-N 12 -ACA)] to [(TGT-N 12 -ACA)-N 8 -(TGT-N 12 -ACA)]. This probe thus contains a PpsR binding site with the caveat that the   (Fig 2). (D) The binding isotherm of PpsR to the MT2 mutant crtI promoter (Fig 2). (E) The binding isotherm of C436A mutant PpsR to the pucB promoter region.
spacing between the two TGa-N 12 -ACA recognition sequences is 8 instead of the 9 nucleotides. For probe MT2, the two divergent PpsR recognition sequences remained as present in the wild sequence with the caveat that the spacing between the two recognition sequences was changed from 8 to 9 nucleotides as in the pucB promoter (Fig 1A). The results of gel mobility shift analysis indicates that the binding affinity of PpsR to the MT1 probe [(TGT-N 12 -ACA)-N 8 -(TGT-N 12 -ACA)] was higher than that to native promoter probe. Furthermore as is the case of the wild type probe, there was no significant difference in binding observed between reducing (EC 50 = 31.89 nM) and oxidizing conditions (EC 50 = 30.81 nM) (Fig 6C, Table 1) with the MT1 probe. On the other hand, the binding affinity of PpsR to the MT2 probe that contains the 9-bp-spacing between the degenerate recognition sequences was significantly higher under oxidizing conditions (EC 50 = 972.02 nM) than under reducing conditions (EC 50 = 1478.50 nM) (Fig 6D, Table 1). Furthermore, the binding curve under oxidizing conditions changed into more sigmoidal than under reducing conditions ( Fig 6D).
We also measured the binding affinity of PpsR C436A mutant in which the conserved putative redox responding Cys in the HTH domain was replaced with Ala ( Fig 6E-6G). The binding affinity of PpsR C436A to the pucB promoter region was estimated to be EC 50 = 94.13 nM and 93.25 nM under oxidizing and reducing conditions, respectively ( Fig 6E, Table 1). To the crtI promoter region, the binding affinity was estimated to be EC 50 = 146.36 nM and 146.59 nM under oxidizing and reducing conditions, respectively ( Fig 6F, Table 1). Thus, EC 50 values for C436A mutant PpsR to the wild type pucB and crtI promoter regions are no significant difference between oxidizing and reducing conditions. Furthermore, there was also no significant difference in binding observed between oxidizing (EC 50 = 4515.14 nM) and reducing conditions (EC 50 = 4646.31 nM) (Fig 6G, Table 1) with the MT2 probe.

Discussion
Whole cell analysis of photopigment accumulation shows that PpsR has an important role in controlling photosystem synthesis in Rvi. gelatinosus not unlike that observed in Rhodobacter (F) The binding isotherm of C436A mutant PpsR to crtI promoter region. (G) The binding isotherm of C436A mutant PpsR to the MT2 mutant crtI promoter. The analysis was performed at least three times, and similar results were obtained, although the binding isotherm of each replicate was shown. species. This conclusion is based on the observation that pigment accumulation was significantly larger in the aerobically grown ΔppsR strain than in wild type Rvi. gelatinosus (Fig 2A). The transcript levels of pucB and crtI in the aerobically grown ΔppsR strain were also higher than that observed in wild type cells (Fig 3A). These results suggest that PpsR in Rvi. gelatinosus acts as a repressor of photosynthesis gene expression under both aerobic and anaerobic conditions. This is quite different from that observed with PpsR homologs from Rhodobacter species where they primarily function as aerobic repressors [4,5]. PpsR function in Rvi. gelatinosus as an semiaerobic repressor and activator has also been reported [23]. What is less clear is the mechanism of controlling the DNA binding activity of PpsR in this α-proteobacterium.
In Rhodobacter species, mutation of a conserved Cys located in the HTH domain of PpsR/CrtJ results in a disruption of redox sensing [9]. However, an equivalent Cys436Ala mutant of Rvi. gelatinosus PpsR has almost no discernable phenotype in vivo (Figs 2 and 3). This suggests that Cys436Ala may have a minimal role in the control of PpsR activity in vivo. Assuming that PpsR has only minor redox control in this species, then PpsR may control photosystem gene expression under both aerobic and anaerobic conditions with inputs other than redox regulating its activity. In Rba. sphaeroides it has been shown that PpsR is controlled by the redox state of a conserved Cys in the HTH domain as well as by the antirepressors AppA and PpsA [24,10]. The antirepressor activity of the AppA is controlled by light excitation of a bound flavin as well as the presence of heme [7]. In Rba. capsulatus, it was recently shown that a PpaA homolog called AerR controls the DNA binding activity of the PpsR homolog CrtJ in response to light-dependent binding of cobalamin (vitamin B 12 ) [25]. The characterization of a potential antirepressor of Rvi. gelatinosus PpsR is beyond the scope of this study, but our in vivo results indicate regulation of PpsR activity other than the redox state of Cys436 is likely very important in this species. Our in vivo results are also supported by several of our in vitro studies with isolated PpsR. Specifically, we observed excellent DNase I footprint protection of PpsR to the puc and crtI promoter regions under both oxidizing and reducing conditions (Fig 5). The only observed difference was a minor enhancement of a single peak in the puc promoter under reducing conditions. Gel mobility shift results also shown that the binding affinity of Rvi. gelatinosus PpsR to the pucB promoter was only slightly higher (1.3-fold) under oxidizing conditions than under reducing conditions (Table 1). This minor effect is in stark contrast to Rhodobacter PpsR which exhibits a 19-fold difference in redox control in vitro [8]. Cys436 does appear to have a role in this minor redox control as the binding affinity of the Cys436Ala mutant PpsR to the pucB promoter region showed no significant difference between oxidizing and reducing conditions (Table 1).
Interestingly the binding affinity of wild type PpsR to the crtI promoter region showed no significant difference between oxidizing and reducing conditions (Table 1). When the space between the two crtI PpsR-binding sites was changed to 9-bp, the binding of PpsR to this site gained redox control as evidence by higher binding under oxidizing conditions than under reducing conditions (Table 1). Furthermore, the binding affinity of PpsR C436A to this changed promoter showed no significant difference between under oxidizing and reducing conditions ( Table 1). Collectively these results indicate that PpsR can modestly affects DNA binding based on the redox state of the conserved Cys436. However, this modest level of redox control is apparently only operative when the space between the two PpsR-binding sites 9-bp than 8-bp. Interestingly, the binding curve of PpsR to the crtI promoter is more sigmoidal indicating more cooperative when the spacing is 9-bp than when it is 8-bp. This indicates that cooperative binding of PpsR to its target sequence is dependent on appropriate spacing between the PpsR binding sites and that the observed redox control may be affecting such cooperative interactions. Finally, in another Rubrivivax species (Rvi. benzoatilyticus), spaces between the two PpsR-binding sites in the pucB and crtI promoters are the same as those in Rvi. gelatinosus, specifically 9 vs 8-bp (Fig 1A). Therefore, different spacing in the pucB and crtI promoters may be characteristics to Rubrivivax species. Why would the binding affinity of Rvi. gelatinosus PpsR differ between pucB and crtI promoters? If photosystems are synthesized in the presence of oxygen, then light excitation of bacteriochlorophylls would generate ROS. In order to avoid production of ROS, puc genes encoding LH complex apo-proteins, seem to be repressed under aerobic conditions. In contrast, carotenoids themselves work as a ROS scavengers where they can decay reactive singlet oxygen back to its ground state by absorbing energy from singlet oxygen [26]. Therefore, it is possible that carotenoid synthesis is by design not repressed under aerobic conditions. The lower affinity of Rvi. gelatinosus PpsR to crtI promoter region than pucB promoter region (Table 1) might indicate such necessity of carotenoid production under aerobic conditions. The reason for repression of carotenoid synthesis in Rhodobacter species under aerobic conditions might be related to their more enhanced capability to repress synthesis of their photosynthetic apparatus in response to oxygen.
β-proteobacteria are thought to have obtained a photosynthesis gene cluster by horizontal gene transfer from α-proteobacteria [14,27]. Thus, Rvi. gelatinosus might have adapted their regulatory genes to coordinate photosynthesis and respiration that is unique to this species. Cleary, PpsR orthologs in purple bacteria have species specificity in their redox sensing mechanisms that differ from that described for the Rhodobacter paradigm.