| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 43, 44407-44416, October 22, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
**
From the
Laboratoire des Symbioses Tropicales et Méditerranéennes (UMR113-IRD-CIRAD-AGRO-M-INRA-UM2), TA 10/J, Campus de Baillarguet, 34398 Montpellier Cedex 5, France, the ¶CEA/Cadarache DEVM-Laboratoire de Bioénergétique Cellulaire, UMR 6191-CNRS-CEA, Univ-Méditerranée CEA 1000, 13108 Saint Paul lez Durance, Cedex, France, and the ||Laboratoire de Biochimie et Biophysique des Systèmes Intégrés (UMR 5092 CNRS-CEA-UJF), CEA-Grenoble, 38054 Grenoble cedex 9, France
Received for publication, July 16, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The genes necessary for the formation of the photosynthetic apparatus are clustered in a 45-kb region designated as the photosynthesis gene cluster (PGC).1 The PGC includes several operons involved in 1) the photopigment synthesis (bacteriochlorophyll (bch), and carotenoid (crt)); 2) the light-harvesting polypeptides (pucBAC and pufBA); 3) the reaction center subunits (puhA, pufLM); and 4) various regulators (ppsR, tspO, and aerR) (4, 5). The fine control of photosystem synthesis by oxygen relies on the combined action of several transcription factors that activate or repress the expression of photosynthesis genes in response to redox conditions (2, 6). One well characterized regulator is the aerobic repressor PpsR in Rhodobacter sphaeroides or its homolog CrtJ in Rhodobacter capsulatus (7, 8). Both proteins share the same mode of action: under oxidizing conditions, PpsR/CrtJ blocks transcription by binding as a tetramer to a palindromic (TGTN12ACA) motif, which is found in tandem in the bch, crt, and puc promoters (9–11). In addition, biochemical studies have recently shown that CrtJ/PpsR have redox sensing capabilities thanks to the formation of an intramolecular disulfide bond promoted by oxygen, which stimulates their binding to target promoters (12, 13).
Light intensity also controls the expression of photosynthesis genes with repressive effects less drastic than oxygen. The molecular mechanism of this regulation, described in R. sphaeroides, implicates the antagonistic actions of the repressor PpsR and the flavoprotein AppA (14). Masuda and Bauer (13) have recently shown that AppA is a blue light photoreceptor that modulates the DNA binding activity of PpsR in response to oxygen and light signals, using two distinct mechanisms: it may break the disulfide bond in oxidized PpsR, or form an inactive AppA-PpsR2 complex.
A new type of light regulation has been described for the two closely related species Bradyrhizobium strain ORS278 and Rhodopseudomonas palustris (15). In these two bacteria, depending on the oxygen tension, far-red light is necessary to trigger the expression of photosynthesis genes. The mechanism of this light regulation has been partially elucidated and involves a bacteriophytochrome. This chromoprotein exerts its regulatory activity by switching between a red light absorbing form (Pr) and a far-red light absorbing form (Pfr) (16). It is noteworthy that the bacteriophytochrome genes (bphP) responsible for this regulation are found in the PGC region contiguous to a ppsR gene in both bacteria. Mutation in bphP of Bradyrhizobium strain ORS278 results in the absence of photosystem formation irrespective of the light conditions, whereas mutation in ppsR leads to the opposite phenotype. Based on phenotypes of these bphP and ppsR mutants and the action spectrum of the photosystem synthesis, it was proposed that the Pr form of the bacteriophytochrome antagonizes the repressive effect of PpsR. More recently, similar results have been obtained for Rps. palustris strain CEA001 (17).
In this study, we reveal the presence of a second ppsR gene linked to the PGC of Bradyrhizobium strain ORS278. Examination of the genome sequence of Rps. palustris (available at http://spider.jgi-psf.org/JGI-microbial/html/) also reveals the presence of a second ppsR gene located in the PGC region. The unexpected presence of two ppsR genes in the same bacterium raises several questions. Why two different PpsR? Do both of them exert the same repressive effect as usually observed in purple bacteria? On the same photosynthesis genes? Are both PpsR involved in redox and light regulation circuit? Here, we specify the roles and the mechanisms of action of both PpsR from Bradyrhizobium, by combining genetic and biochemical approaches.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Determination of Photosynthetic Activity—The amount of the photosynthetic apparatus in intact cells was estimated as previously described (18).
Construction of ppsR1, ppsR2, and ppsR2/BrbphP Mutant Strains— Construction of the ppsR2 mutant has been previously described (15). For the construction of a ppsR1 mutant, a region of about 2.3 kb containing the ppsR1 gene was amplified by PCR using the primers 5'-TCAGGATCCTGACCGAACGGCGATAGCTTTAG-3' and 5'-ATCG-GATCCTGCGCCGGACGGCCCCACACAATGAG-3' and subsequently cloned in the pGEM-T vector (Promega, Madison, WI). The 4.7-kb BamHI lacZ-Kmr cassette of pKOK5 (19) was then inserted in the unique BglII site of ppsR1. The resulting 7-kb BamHI insert containing the mutated ppsR1 gene was cloned into the pJQ200mp18 suicide vector linearized by BamHI digestion (20). For the construction of the double mutant ppsR2/BrbphP, a region of about 5 kb containing the contiguous ppsR2 and BrbphP genes was amplified by PCR using the primers 5'-TGCGGATCCGCACCCCGTCCTGTGCCAGCGTATC-3' and 5'-TAGGGATCCATAACCACCGCCGCCTGTGATGATAAAAC-3' and subsequently cloned into the pGEM-T vector. A 3.2-kb region containing a part of the ppsR2 and BrbphP genes was deleted by XhoI digestion and replaced by the 4.7-kb SalI lacZ-Kmr cassette of pKOK5. The resulting 6.6-kb BamHI insert containing the mutated ppsR2 and Brb-phP genes was cloned into the pJQ200mp18 suicide vector linearized by BamHI digestion. The pJQ200 derivatives obtained, which encoded a counter selective sacB marker, were transformed into E. coli S17-1 for mobilization into ORS278 as previously described (18). Double recombinants were selected on sucrose and the insertion was confirmed by PCR.
-Galactosidase Assay—For analysis of ppsR1 and ppsR2 expression, -galactosidase activity was assayed as previously described (15).
Cloning, Protein Expression, and Purification—The ppsR1 and ppsR2 genes from Bradyrhizobium strain ORS278 were amplified by PCR using Pfu DNA polymerase and the following pairs of primers: PpsR1.pBAD.f, 5'-GGATCCGTGAGGGCGTTCAGAGCTCCGAAAGA-G-3' and PpsR1.pBAD.r, 5'-AAGCTTCTATTCCAACTGACTGTCTTCT-TCGCTTG-3'; PpsR2.pBAD.f, 5'-ATGGATCCATGGCCGAGTTTCAC-GGTCCACGA-3' and PpsR2.pBAD.r, 5'-ATGGATCCCTAGCTCCCCT-TTTCGGTTTCCT-3'. The PCR products were digested (BamHI/HindIII for ppsR1; BamHI for ppsR2) and ligated into the expression vector pBAD/HisB (Invitrogen). The recombinant PpsR proteins were overexpressed in E. coli LMG194 (Invitrogen) by induction with L-arabinose as previously described (17). Purification of His6-appended PpsR proteins was performed using nickel chelate affinity chromatography as described (17). The purified protein were quantified using the Bradford assay (Bio-Rad) and then aliquoted and stored at –80 °C in the presence of 20% (v/v) glycerol.
Reduction and oxidation of both PpsR were tested using various reagents: 10 mM dithiothreitol (DTT), 0.1 and 1 mM H2O2, 0.5 and 5 mM K3Fe(CN)6 (potassium ferricyanide), or a mixture of 0.5 and 5 mM K3Fe(CN)6 with 10 µM TMPD (N,N,N',N'-tetramethylphenylenediamine dihydrochloride) diluted in water, and air by exposing the sample to a constant flow of air gas during 3 h. After addition of the oxidizing or reducing reagents, the proteins were incubated overnight on ice and subsequently analyzed by nonreducing SDS-PAGE and gel filtration chromatography, or used for DNase I footprint experiments.
Site-directed Mutagenesis—Single mutations were introduced in PpsR1 and PpsR2 amino acid sequences using the QuikChangeTM site-directed mutagenesis kit (Stratagene) according to the manufacturer's recommendations. To construct the PpsR1-C429S mutant, the plasmid pBAD::ppsR1 was used as a template with the following primers: PpsR1-C429S sense (5'-GTGATCGAGCGGATGTCGATCGAGACCGC-GCTC-3') and PpsR1-C429S antisense (5'-GAGCGCGGTCTCGATCG-ACATCCGCTCGATCAC-3'). The Y407C mutation on the PpsR2 sequence was generated using the plasmid pBAD::ppsR2 as a template and the primers PpsR2-Y407C sense (5'-GTCGAGCAGCATTGTGTCC-GCGCCGCG-3') and PpsR2-Y407C antisense (5'-CGCGGCGCGG-ACACAATGCTGCTCGAC-3').
DNase I Footprint Analysis—Four probes corresponding to the promoter regions of bchC, crtE, aerR, and cycA were used for studying the DNA binding of PpsR1 and PpsR2. Probes were prepared by PCR using 32P5'-end labeled oligonucleotide primers, as previously described (17).
The bchC promoter region was obtained using the primers: 5'-CTC-GAGGATCCGCGGAATTCAAGCTTCTACGACGACCCGCTGCCTA-C-3' and 5'-CTCGAGGATCCGCGGAATTCAAGCTTCTACGACGACC-CGCTGCCTAC-3'; the crtE promoter region was obtained using the primers: 5'-GCCGATTGCGATCTGGCGCATCTTG-3' and 5'-AGGGA-TTTTTCGATCCGGGCCATCAC-3'; the cycA promoter region was obtained using the primers: 5'-GAGAGTCTTGAGAATGTCGTTAG-3' and 5'-GTCGCCGCCGTCCAGCGCGAATG-3'; the aerR promoter region was obtained using the primers: 5'-CGGGCACGCACTCGGGGA-TTTG-3' and 5'-TGTCGCGCCGGCGCTTTTCCTC-3'. DNase I footprint experiments were performed as previously described (17). PpsR1, PpsR1-C429S, or PpsR2 DNA binding isotherms to the bchC promoter were generated by DNase I titration assays using different protein dilutions treated by 10 mM DTT or a mixture of 0.5 mM K3Fe(CN)6 with 10 µM TMPD. The gel was then analyzed for PpsR binding isotherms using a PhosphorImager (Storm, Amersham Biosciences) to quantify the level of PpsR protection of a single band in the upstream and downstream palindromes. Values were corrected for loading by normalization with a band from an unprotected region.
Gel Filtration Chromatography—The oligomerization states of PpsR1 and PpsR2 were estimated by gel filtration chromatography. The Superdex 200 26/60 column (Amersham Biosciences) was equilibrated with 20 mM HEPES (pH 8.5), 50 mM NaCl in the presence of either 1 mM DTT or 1 mM K3Fe(CN)6. Before injection of the samples, the purified proteins were previously incubated under different redox conditions as described above. Each experiment was carried out in duplicate. The column was previously size calibrated using commercial gel filtration standards (Amersham Biosciences).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
Role of PpsR1 and PpsR2 in Bradyrhizobium Strain ORS278—To specify the role of PpsR1 and PpsR2 identified in Bradyrhizobium strain ORS278, we constructed isogenic mutants (278
ppsR1, 278ppsR2) and compared their photosynthetic phenotypes to the wild type strain after growth under different light and oxygen conditions.
The synthesis of the photosynthetic apparatus is highly enhanced by far-red light when the WT strain is grown under semiaerobic conditions (Fig. 2A). When the oxygen tension is increased (aerobic conditions), the stimulatory effect of far red light is not observed anymore (Fig. 2B) and the photosystem synthesis remains very low irrespective of the light conditions. This suggests that oxygen exerts a repressive effect that dominates the activation by far-red light.
|
ppsR2 mutant constitutively synthesized the photosystem under semiaerobiosis irrespective of the light conditions (Fig. 2A). We therefore concluded that PpsR2 acts as a repressor and that far-red light antagonizes its repressive effect under semiaerobic conditions via the action of the bacteriophytochrome BrBphP (15). Under aerobic conditions, the phenotype of the 278ppsR2 mutant is puzzling. Indeed, we observed that the synthesis of the photosystem is high in darkness and low under far-red light illumination (Fig. 2B). To specify the nature of the pigment involved in this light inhibition, we cultivated the 278ppsR2 mutant in aerobiosis under a series of light-emitting diodes of different wavelengths (between 590 and 870 nm). Bacteriochlorophyll fluorescence was used as a specific marker of the presence of the photosynthetic apparatus (Fig. 2C). We observed a clear negative effect of light in the 700–770 nm region, with a maximum close to 750 nm (Fig. 2D). The action spectrum of this repressive effect of light under aerobic conditions is very similar to the absorption of the Pfr form of a bacteriophytochrome. To test whether this effect of light under aerobic conditions was controlled by BrBphP, we constructed a PpsR2 and BrBphP double mutant. The same repressive effect of far red light was observed in this mutant (data not shown) demonstrating that BrBphP is not involved in this light inhibition. Therefore, a second bacteriophytochrome is likely involved in this light regulatory circuit. Such an hypothesis makes sense because sequence analysis of the closely related Rps. palustris genome revealed the presence of 6 genes encoding putative bacteriophytochromes, four of them being located close to photosynthesis genes (21).
The phenotype of the 278ppsR1 mutant was also unexpected as we observed a very low photosystem synthesis whatever the light and oxygen conditions (Fig. 2, A and B). This clearly shows that contrary to PpsR2 and PpsR/CrtJ of the Rhodobacter species, PpsR1 is an activator of transcription that plays a key role because its presence appears essential for the synthesis of the photosynthetic apparatus in Bradyrhizobium.
Oligomerization States of PpsR1 and PpsR2 in Response to Redox Conditions—The His-tagged versions of both PpsR were successfully overexpressed in E. coli LMG194 using the pBAD-HisB vector. They were subsequently isolated to a high level of purity (>95% as judged by Coomassie Blue staining) (Fig. 3, A and B) by affinity binding of the His6 tag to a Ni2+ column.
|
To ascertain that the 100-kDa band results from the formation of an intermolecular disulfide bond, we constructed a mutated PpsR1 protein (PpsR1-C429S) in which the unique Cys residue is replaced by a Ser residue. In this case, SDS-PAGE did not reveal any dimeric structure whatever the conditions (Fig. 3C).
CrtJ was previously shown to be folded as a tetramer, irrespective of the redox conditions (12). To test whether the formation of an intermolecular disulfide bond in PpsR1 induces a more considerable change in its quaternary structure, we performed size exclusion chromatography of PpsR1 after preincubation under different redox conditions. Under reducing conditions, we observed a single peak elution volume corresponding to a molecular mass of 220 kDa (Fig. 3, D and E). Because the calculated molecular mass of His-tagged PpsR1 monomer is 54 kDa, we can conclude that PpsR1 is folded in solution as a stable tetramer as previously reported for CrtJ or PpsR in Rhodobacter species (12, 13). In contrast, the elution volume is shifted under oxidizing conditions to a lower value that corresponds to a calculated molecular size of 
400 kDa. As a result, the oligomeric state of PpsR1 varies from a tetrameric to an octameric form under oxidizing conditions (Fig. 3, D and F). This redox-dependent modification of the PpsR1 quaternary structure can be caused by the formation of two Cys429-Cys429 intermolecular disulfide bonds connecting two tetramers. In agreement with this interpretation, gel filtration analysis carried out on the PpsR1-C429S mutant revealed a tetrameric structure, irrespective of the oxidizing treatment (Fig. 3, E and F). The gel filtration procedure was also applied to PpsR2, confirming the results obtained by SDS-PAGE experiments: PpsR2 displayed the same elution volumes under reducing or oxidizing conditions (Fig. 3, E and F). It is thus folded as a tetramer that is not affected by the oxidizing treatment.
We further investigated the correlation between the presence of an intermolecular bridge and the redox sensing capacity of both PpsR, by constructing a PpsR2 mutant (PpsR2-Y407C) in which Tyr407 was replaced by a Cys residue. Indeed, in the sequence alignment of the proteins, the residue 407 of PpsR2 corresponds to Cys429 in PpsR1, the residue responsible for the oligomerization change. As shown by SDS-PAGE analysis (Fig. 3C), this single point mutation renders PpsR2 redox sensitive, appearing as a band of 100 or 50 kDa depending on the redox conditions. Nevertheless, two fundamental differences are observed with PpsR1: (i) in its fully oxidized form, PpsR2-Y407C migrates on SDS-PAGE as a single band corresponding to a dimer (100 kDa), whereas PpsR1 was equally partitioned between monomer and dimer; and (ii) PpsR2-Y407C is fully oxidized after the elution step of the purification process, thus much easier to oxidize than PpsR1. These differences clearly indicate that in addition to the crucial role of the Cys residue in redox sensing, other patterns in the primary amino acid sequences define structural specificities of each PpsR.
DNA Binding Characteristics of PpsR1 and PpsR2—PpsR/CrtJ of Rhodobacter species is known to recognize the TGTN12ACA palindromic motif found in tandem in the promoter region of some photosynthesis genes (7, 9, 11). In addition, CrtJ displays a structural flexibility in DNA recognition, because it is also able to bind to the TGTN12ACG sequence located in the promoter region of crtI (10). A close examination of the partial PGC sequence available for Bradyrhizobium ORS278 revealed the presence of four putative PpsR binding sites (Fig. 4E). Two of them are found in the promoter regions of photosynthesis genes sensu stricto (bchC and crtE-crtD intergenic region). They correspond to a pair of perfectly conserved palindromic TGTN12ACA. The two other putative binding sites are located in the promoter regions of cycA (encoding the soluble cytochrome c2) and aerR (coding for a regulator of photosynthesis genes in Rhodobacter species (22, 23)) and contain only one palindromic motif TGTN12ACA associated to an incompletely conserved pattern.
|
The DNA binding activity of the PpsR proteins was further investigated by quantifying the footprints obtained with the bchC promoter, varying either the amount of added protein at a fixed DNA concentration or the reverse. Panels A and B in Fig. 5 show typical results obtained for PpsR2. The binding isotherm (A) is clearly steeper than would be expected if a single ligand was involved (dashed line). A fit of the data with the function 1/(1 + (K/x)n) gives a Hill coefficient n 
2, suggesting the cooperative binding of (at least) two ligands (solid curve). The same conclusion applies to PpsR1 (see panel D) and its C429S mutant (data not shown). For all cases (except oxidized PpsR1, see below), the protein concentration (expressed in terms of the monomeric subunit) for 50% of the maximum protection (EC50) was in the range of 60–75 nM.
|
We next addressed the question whether the DNA binding of both PpsR is redox dependent as observed for PpsR and CrtJ in Rhodobacter species. Fig. 5C shows that the oxidizing treatment did not affect significantly the binding properties of PpsR2, in agreement with the biochemical results presented previously. In contrast, the redox conditions did modulate the binding of PpsR1 (Fig. 5D). The EC50 of the oxidized protein was increased more than 3-fold (226 nM). This effect was not observed with the PpsR1-C429S (data not shown), supporting the view that it is caused by the formation of an octameric form because of the establishment of a disulfide bond.
Interestingly, the binding isotherm for oxidized PpsR1 (Fig. 5, open circles in panel D) is also characteristic of a cooperative two-ligand process, although in this case the ligand species should be an octamer. If one compares the EC50 expressed for the effective oligomerization state of the ligand (l6 nM for the tetrameric, reduced PpsR1 and 28 nM for the oxidized octamer), it appears that the overall decrease of the affinity is contributed both by the aggregation of two tetramers to form a single ligand (i.e. the protein-DNA complex now consists of 16 rather than 8 monomeric subunits) and by a 1.75-fold lower affinity of this bulkier ligand.
Expression Levels of PpsR1 and PpsR2 in Regard to Oxygen and Light Conditions—Interestingly, the ppsR1 gene of the Bradyrhizobium strain ORS278 is located immediately downstream from the regulator aerR, whose promoter is recognized by PpsR2. Such an organization is also found in R. capsulatus for which it has been shown that the promoter of aerR contributes significantly to the expression of the downstream ppsR gene (22). It is therefore tempting to think that PpsR2 could control the level of expression of ppsR1. To test this hypothesis, we used 278ppsR1 and 278ppsR2 mutants carrying lacZ fusions (see "Experimental Procedures") cultivated under various light and oxygen conditions. As shown in Fig. 6A, the expression level of ppsR2 remains constant irrespective of the light and oxygen tension indicating that PpsR2 is constitutively synthesized. In contrast, the expression level of ppsR1 is enhanced under far-red light and semiaerobic conditions, i.e. the optimal conditions for photosystem synthesis. A close examination of the light effect using a series of light-emitting diodes of different wavelengths (from 590 to 870 nm) confirmed that the synthesis of PpsR1 is controlled by a bacteriophytochrome, the action spectrum corresponding to the Pr form of BrBphP (Fig. 6B). Taken together with the footprint analysis, these results strongly suggest that the expression level of ppsR1 is controlled by the PpsR2/BrBphP repressor/antirepressor system.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
In this study, we have demonstrated that only PpsR1 is redox sensitive. As it is observed for PpsR/CrtJ from the Rhodobacter species, the Cys residues play a key role in the redox response of PpsR1 via the formation of a disulfide bond. Nevertheless, contrary to PpsR/CrtJ, disulfide bond formation in PpsR1 is intermolecular and leads to a global change of the quaternary structure, switching from a tetramer to an octamer. The mechanism of PpsR1 oxidation also appears different from that of PpsR/CrtJ proteins because oxidation of PpsR/CrtJ occurs mainly in response to oxygen, whereas oxidation of PpsR1 remains very limited in such a condition. A mediator (TMPD) appeared necessary in our different assays to oxidize efficiently the protein in the presence of ferricyanide. Bauer's group has also studied the oxidation state of PpsR/CrtJ in vivo (12, 13) and showed that they formed a disulfide bond specifically under aerobic conditions. It has been also observed that the R. sphaeroides cytoplasmic redox poise remains constant, around –220 mV, under both aerobic and anaerobic conditions. Because this is lower than the midpoint potential for oxidation of CrtJ (–180 mV); it was proposed that molecular oxygen itself stimulates disulfide bond formation in CrtJ (12). The absence of an oxygen response on the redox state of PpsR1 in vitro leads us to suggest that PpsR1 senses in vivo another signal redox than the oxygen tension. The redox state of the quinone pool that has been shown to play a key role in the regulation of photosystem synthesis in R. sphaeroides (25) could be this signal. As observed in PpsR/CrtJ (12, 13), disulfide bond formation affects the DNA binding affinity of PpsR1. However, contrary to PpsR/CrtJ, the reduced form presents the higher affinity. PpsR1 exhibits a 3.5-fold modulation of its affinity for DNA, depending on the redox conditions, which is in the same range as the difference determined for CrtJ (5-fold) or PpsR (2.2-fold).
Contrary to PpsR1, PpsR2 contains no Cys residue and is thus not redox sensitive. Interestingly, a single point mutation (PpsR2-Y407C) induces a redox sensitivity of PpsR2. This result raises the possibility of a spontaneous mutation of PpsR2 in the ORS278 strain, rendering it redox insensitive. In the same way, a double spontaneous mutation in the Rps. palustris CGA009 has been recently revealed: one frameshift mutation in the bacteriophytochrome rpa1537 gene, and one single point mutation in the ppsR2 gene had rendered the two corresponding proteins inactive (17). To assess the possibility that the absence of a Cys residue in PpsR2 might result from the accidental selection of a spontaneous mutant, we sequenced the corresponding gene from two other strains (Bradyrhizobium Btai1 and Bradyrhizobium ORS285). Both strains have been isolated from different regions and different host plants. A slight polymorphism is observed between the various amino acid sequences of PpsR2, the identity percentage ranging from 95.4 to 96.5%. However, in all three photosynthetic strains, the PpsR2 contains no Cys residue. In addition, sequencing the ppsR1 genes of the Bradyrhizobium Btai1 and Bradyrhizobium ORS285 strains confirmed the presence of only one conserved Cys residue in the corresponding protein. These data clearly indicate that the presence of two PpsR that differ in their Cys content and, in all likelihood, in their redox sensing behavior, is a typical feature of photosynthetic Bradyrhizobium strains.
The opposite phenotypes observed for the 278ppsR1 and 278ppsR2 mutants suggest that both PpsR play dual roles in the regulation of photosystem formation in Bradyrhizobium, with a classical repressor role for PpsR2 and an unexpected activator role for PpsR1. Two other pieces of evidence are in agreement with the activator role of PpsR1: 1) its expression level is correlated with the amount of photosynthetic apparatus and 2) its DNA binding affinity is higher under reducing conditions, as expected for promoting photosynthesis under low oxygen tension conditions. Such an activator role has been very recently described in Rubrivivax gelatinosus (26). In this bacterium, a single PpsR displays both a negative and a positive activity: it represses the expression of the crtI gene on the one hand and it stimulates the expression of pucBA genes on the other hand. It is believed that the difference of spacing observed between the two putative PpsR binding boxes of the crtI and pucBA promoters could be responsible for this modulation of PpsR behavior. In the case of Bradyrhizobium, the situation is different because the activator PpsR1 and the repressor PpsR2 protect exactly the same DNA region of the bch and crt promoters. Interestingly, primer extension analysis of the crtE gene (data not shown) indicates that the two PpsR binding palindromes overlap the –35 and –10 promoter regions (see Fig. 4B). PpsR2 is then likely to exert its repressive effect by preventing the RNA polymerase from binding to the DNA, or by blocking another step of the transcription initiation, like, for instance, the formation of the open complex. The PpsR1 activator role, when bound to the same sites, could also be explained by different mechanisms. PpsR1 could be a class II activator, activating transcription by recruiting the RNA polymerase via its interaction with the 
subunit. Alternatively, it might accelerate promoter clearance or alter the promoter conformation leading to a better recognition by the RNA polymerase, as observed for the transcriptional regulators of the MerR (27). Further experiments are required to understand the molecular mechanisms used by PpsR1 and PpsR2 to control transcription, in particular for assessing the probable role of their PAS domains, which is likely to be a determinant in their repressor/activator behavior toward RNA polymerase.
The two PpsR interact with the same photosynthesis promoters while they play antagonistic roles. This implies that the regulation of photosystem formation in Bradyrhizobium results from a competition between the two PpsR for the binding to these target genes. The EC50 observed under reducing conditions are similar for the two PpsR proteins (65 ± 5 nM for PpsR1 and 76 ± 5 nM for PpsR2). However, redox and light conditions modulate their efficiency, via the change of the oligomerization state of PpsR1 and via the action of BrBphP, which antirepresses PpsR2, respectively. The latter mechanism is still unknown, but one may suppose that it is similar to that of AppA (13), involving the formation of an inactive, light-dependent, BrBphP-PpsR2 complex. An essential parameter to consider in this competition process is the relative amounts of PpsR1 and PpsR2. Using ppsR::lacZ fusions, we demonstrated that ppsR2 is constitutively expressed, whereas the ppsR1 expression level is strongly dependent on oxygen and light. The light control of ppsR1 expression surely involves the antagonistic action of PpsR2/BrBphP. In contrast, the aerobic circuit remains unknown and may involve an aerobic repressor not yet identified.
Besides the fact that the two PpsR display similar binding properties toward the photosynthesis genes, we showed that PpsR2, but not PpsR1, recognized the aerR and cycA promoters. This indicates that the DNA recognition process is more flexible for PpsR2 than for PpsR1 leading to a broader panel of target genes. In the cycA and aerR promoters, the upstream palindrome is not perfectly conserved and we observed that it was not protected by PpsR2 from DNase I digestion. Its role in the process of recognition of PpsR2 is therefore debatable. If this region is not able to interact tightly with the protein, one may hypothesize that the binding of PpsR2 on these promoters is weaker and will exert a weaker repressive effect on the downstream gene. This could be beneficial in the case of the cycA gene, which encodes the cytochrome c2. This soluble electron carrier is implicated in both photosynthesis and respiratory processes. A not too drastic control of the cycA gene by PpsR2 would permit the synthesis of cytochrome c2 to a basal level sufficient for the functioning of the respiratory chain and also boost its production when photosynthesis takes over. On the other hand, the absence of control of the cycA gene by the activator PpsR1 makes sense, because in aerobic conditions, where PpsR1 is probably inactive, the synthesis of cytochrome c2 is still indispensable for the respiratory process.
The regulator PpsR has been identified in all purple bacteria studied so far, suggesting that it is a common regulator for this bacteria family. However, as illustrated in this article, its mode of action has evolved to respond to specific needs in various species. Indeed, the mechanism of regulation of photosynthesis, implicating the combined action of two PpsR in Bradyrhizobium appears more sophisticated than in Rhodobacter species. Presumably, this sophistication, which integrates both light and redox signals permits better control of photosystem formation. This fine control seems particularly adapted to Bradyrhizobium, which contrary to the Rhodobacter species is a strict aerobic bacterium and therefore must carry out photosynthesis under conditions where harmful reactive oxygen species are generated.
|
FOOTNOTES |
|---|
Supported by a French Minister of National Education, Higher Education and Research doctoral grant.
** Both authors contributed equally and should be considered co-senior authors.
To whom correspondence should be addressed: IRD, LSTM TA 10/J, Campus de Baillarguet, 34398 Montpellier Cedex 5, France. Tel.: 33-467593783; Fax: 33-467593802; E-mail: Giraud{at}mpl.ird.fr.
1 The abbreviations used are: PGC, photosynthesis gene cluster; DTT, dithiothreitol; TMPD, N,N,N',N'-tetramethylphenylenediamine dihydrochloride; BrBphP, Bradyrhizobium bacteriophytochrome photoreceptor; Rps., Rhodopseudomonas.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Bruscella, J. M. Eraso, J. H. Roh, and S. Kaplan The Use of Chromatin Immunoprecipitation to Define PpsR Binding Activity in Rhodobacter sphaeroides 2.4.1 J. Bacteriol., October 15, 2008; 190(20): 6817 - 6828. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Jaubert, L. Vuillet, L. Hannibal, J.-M. Adriano, J. Fardoux, P. Bouyer, K. Bonaldi, D. Fleischman, E. Giraud, and A. Vermeglio Control of Peripheral Light-Harvesting Complex Synthesis by a Bacteriophytochrome in the Aerobic Photosynthetic Bacterium Bradyrhizobium Strain BTAi1 J. Bacteriol., September 1, 2008; 190(17): 5824 - 5831. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. D. Swingley, S. Sadekar, S. D. Mastrian, H. J. Matthies, J. Hao, H. Ramos, C. R. Acharya, A. L. Conrad, H. L. Taylor, L. C. Dejesa, et al. The Complete Genome Sequence of Roseobacter denitrificans Reveals a Mixotrophic Rather than Photosynthetic Metabolism J. Bacteriol., February 1, 2007; 189(3): 683 - 690. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Conlan, C. Lawrence, and L. A. McCue Rhodopseudomonas palustris Regulons Detected by Cross-Species Analysis of Alphaproteobacterial Genomes Appl. Envir. Microbiol., November 1, 2005; 71(11): 7442 - 7452. [Abstract] [Full Text] [PDF]
|
|
AP2, where P stands for the protein and A for DNA). The dashed line is a best fit when assuming that the protection involves the binding of a single ligand. B, the concentration of PpsR2 was kept fixed (100 nM) and a variable concentration of DNA was added. The ordinate is the concentration of protected DNA (corrected for the amplitude factor of 0.85 mentioned above). The experimental scatter increases at high [DNA] where the protected DNA becomes a small fraction of the total. The solid line was computed assuming the binding of two tetramers, using the dissociation constant estimated from the fit in A. The dashed line is the predicted behavior for single ligand binding. C, fraction of protected DNA as a function of the concentration of reduced (solid circles) or oxidized (open circles) PpsR2. The horizontal scale is logarithmic. D, same as C, for PpsR1. The lines in C and D are best fits with the function a/(1 +(K/X)2), yielding K = 76 and 107 nM for reduced and oxidized PpsR2, 65 and 226 nM for reduced and oxidized PpsR1, respectively. All protein concentrations are expressed with respect to the amount of the monomeric subunit.