INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The purple photosynthetic bacterium Rhodovulum sulfidophilum is a marine bacterium that can grow by either photosynthesis or respiration on a wide range of organic compounds (20). Like those of other purple bacteria, the photosynthetic apparatus of R. sulfidophilum is composed of three membrane-spanning pigment protein complexes known as the reaction center (RC) and the light-harvesting 1 and 2 complexes (LH1 and LH2, respectively). The light energy captured by LH1 and LH2 is transferred to the RC, where the primary photochemical reaction takes place. The RC complexes of most purple bacteria are known to consist of at least L, M, and H subunits. The light-harvesting complexes are composed of two membrane-spanning polypeptides,  and  subunits, which bind bacteriochlorophylls (BChls) and carotenoids (13, 15, 48). It is known that these proteins are encoded by three operons: the puf operon, which encodes the LH1  and  polypeptides and the RC-L and -M polypeptides; the puc operon, which encodes the LH2  and  polypeptides; and the puh operon, which encodes RC-H polypeptide (9, 26, 44, 46, 47). Some purple bacteria have a pufC gene in the puf operon that encodes an RC-bound cytochrome subunit with four c-type hemes. Recently, we sequenced the whole puf operon of R. sulfidophilum and showed that this bacterium has a pufC gene in the puf operon encoding a unique cytochrome subunit that contains only three possible heme-binding sites (30).

In some purple bacteria such as Rhodobacter species, the synthesis of the photosynthetic apparatus is regulated by oxygen concentration and light intensity. Many studies using Rhodobacter capsulatus and Rhodobacter sphaeroides have demonstrated that photosystem synthesis is controlled at a number of different levels in the cells (2-4, 11, 15, 24, 34, 45). Much of the environmental regulation appears to be exerted at the level of photosynthesis gene transcription (2, 3, 34, 45). The importance of the posttranscriptional mRNA degradative processes and posttranslational effects in determining final levels of membrane-bound photosynthetic apparatus has also been reported (15, 24). In contrast to Rhodobacter species, R. sulfidophilum synthesizes the photosynthetic apparatus even under highly aerated conditions in the dark (14). A study of the puc operon of R. sulfidophilum showed that puc mRNA synthesis is weakly repressed by oxygen but markedly suppressed by high-intensity light (19). This is distinctly different from the results shown for the Rhodobacter species, which show a high degree of repression by oxygen and a weak repression by light (2, 3).

The RegA-RegB two-component regulatory system was identified as a transcriptional factor that anaerobically activates the expression of the puf, puc, and puh operons in R. capsulatus (31, 37). Genes homologous to regA and regB were also found in R. sphaeroides and named prrA and prrB, respectively (17, 36). Recently, we have found and characterized the RegA-RegB regulatory system in R. sulfidophilum and showed that it controls the photosynthetic gene expression in this bacterium (29), as it does in R. capsulatus (37). We also demonstrated that the species-dependent difference in the repression of photosynthetic gene expression under aerobic conditions is not the result of altered redox sensing by the sensor kinase protein, RegB (29). These observations suggested the presence of another redox-responding protein that affects the RegB activity.

To date, high-resolution mRNA mapping for the identification of transcription initiation sites in combination with detailed promoter deletion studies for the puf operon has been undertaken only with Rhodobacter species, although the location of the 5' ends of puf mRNA has been reported for several purple bacteria (1, 5, 6, 12, 22, 27, 32, 33, 41). Genetic analyses of the puf operon promoter from R. capsulatus have indicated that transcription initiation sites and a cis-acting regulatory site involved in oxygen and light control of promoter activity are located far upstream from the 5' end of the stable puf mRNA (1, 5). Our previous study showed that R. sulfidophilum has a pufQ gene as the first gene of the operon, as in R. capsulatus, and that the DNA sequences upstream of the operons of R. sulfidophilum and R. capsulatus showed high similarity (30). Thus, comparative studies of the R. sulfidophilum puf operon promoter would be useful for understanding the molecular basis of the regulation of puf operon expression with respect to the oxygen-independent synthesis of the photosynthetic apparatus in this bacterium.

In this study, we identified and characterized the R. sulfidophilum puf operon promoters. The results suggest that the R. sulfidophilum puf operon has three possible transcription initiation sites, one of which is oxygen and regA independent. The promoter region was compared with that of the R. capsulatus puf promoter.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and cultures. The bacterial strains and plasmids used in this study are listed in Table 1. R. capsulatus was anaerobically grown at 30°C in 30-ml screw-cap bottles filled with RCV medium (39). Wild-type R. sulfidophilum W4 and the regA-disrupted strain RESA1 were grown under the same conditions as R. capsulatus in RCV medium supplemented with 0.35 M sodium chloride. Aerobic growth of R. capsulatus and R. sulfidophilum was achieved by shaking a 10-ml culture in a 100-ml conical flask at 200 rpm. Illumination was provided by 60-W tungsten lamps. Escherichia coli strains DH5 and S17-1 were grown at 37°C in Luria-Bertani medium. Antibiotics were added at the final concentrations given: to E. coli cultures, ampicillin (100 µg/ml) or tetracycline (20 µg/ml), and to R. sulfidophilum cultures, kanamycin (50 µg/ml) or streptomycin (20 µg/ml), where necessary.

Conjugation techniques. The pCF1010-derived plasmids were mobilized into R. sulfidophilum cells by conjugation with the mobilizing strain E. coli S17-1 (38).

Plasmid construction. The plasmid pCF1010 (28) was used to make transcriptional fusions of the R. sulfidophilum puf operon promoter region to a promoterless lacZ gene including a ribosome-binding site. First, a 2,580-bp EcoRI-BamHI fragment containing pufQ, pufB, pufA, and part of pufL as well as 1,073 bp upstream of pufQ was cut out from the plasmid pUFS101 (30) and cloned into EcoRI/BamHI-cut pUC119, resulting in pUFS001. This plasmid has a unique XbaI site just downstream of the BamHI site (in the polycloning site). DNA fragments digested by HincII-XbaI, NruI-XbaI, SmaI-XbaI, and BalI-XbaI from pUFS001 (see Fig. 3) were cloned into StuI/XbaI-cut pCF1010 to construct the puf-lacZ fusions named pPS001, pPS002, pPS003, and pPS005, respectively. A PstI-XbaI fragment from pUFS001 was cloned into PstI/XbaI-cut pCF1010 to construct pPS004. A HincII-BalI fragment from pUFS001 was cloned into StuI-cut pCF1010 and named pPS100. For the construction of pPS200 and pPS300, two DNA fragments were amplified by PCR using pUFS001 as a template DNA. The fragments were amplified by the combination of forward primer M13-F (5'-CGACGTTGTAAAACGACGGCCAGT-3') and reverse primer Z1200RXbaI (5'-TTTCTAGACCGAGCGGCAGGATATGAG-3') and forward primer M13-F and reverse primer Z940RXbaI (5'-TTTCTAGAGCAAAGGCGCAGGGCAGCC-3'). Both Z940RXbaI and Z1200RXbaI primers have additional polynucleotides at the 5' ends to give the XbaI sites in the fragments (underlined). The DNA fragments amplified by M13-F and Z1200RXbaI primers and M13-F and Z940RXbaI primers were digested with HincII and XbaI and then ligated into the StuI/XbaI-cut pCF1010, resulting in pPS200 and pPS300, respectively. The sequence accuracy of the amplified fragments was confirmed by sequencing using a 310A Genetic Analyzer (Applied Biosystems).

-Galactosidase assays. Cell extracts were prepared from 10-ml cultures showing an optical density of 0.5 to 0.6 at 660 nm, and the -galactosidase activity of the extracts was determined as described by Young et al. (43). Protein content determination was performed with a Bradford assay kit (Bio-Rad Laboratories). Final results were obtained as the amount of o-nitrophenyl--galactoside (ONPG) hydrolyzed per minute per milligram of total protein. Results are the averages based on at least three independent assays.

Northern hybridization analysis. Total RNA of R. sulfidophilum and R. capsulatus cells grown till the logarithmic growth phase, showing an optical density of 0.5 to 0.6 at 660 nm, was extracted using an RNeasy kit (Qiagen). The total amounts of RNA obtained from the same amounts of the cells estimated from the optical density varied by less than 15% among the various conditions. Electrophoresis of the total RNA was performed in 1.0% agarose gels containing formaldehyde (40 mM MOPS [morpholinepropanesulfonic acid], 10 mM sodium acetate, 1 mM EDTA, and 2.2 M formaldehyde, pH 7.0). After electrophoresis, the RNA was transferred to positively charged nylon membranes (Boehringer Mannheim) or a Hybond-N+ positively charged nylon membrane (Amersham). Four probes were used (probes A, B, C, and D, shown in Fig. 1a). Probe A is a PCR product amplified with forward primer ZENDF (5'-AGAGGGAGCTCGCATGT-3') and reverse primer L350R (5'-CCGGGTTTGTAGTGGAA-3') from cell lysates of R. sulfidophilum or R. capsulatus as described previously (30). Probe B is a 1.2-kb DNA fragment excised from pUFS101 by ApaI endonuclease. Probe C is a PCR product amplified with forward primer Z0F (5'-CGGAGTTCATGGTCTATT-3') and reverse primer B140R (5'-CCACGGGCGCCACTGCCA-3') using pUFS001 as a template DNA. Probe D is a PCR product amplified with forward primer Z590F (5'-CCATGTCGCCGAGATGCG-3') and reverse primer Z0R (5'-AATAGACCATGAACTCCG-3') using pUFS001 as a template DNA. These DNA fragments were labeled with digoxigenin-dUTP as instructed by the manufacturer (Boehringer Mannheim). Probe A of R. sulfidophilum and R. capsulatus was also labeled with [-³²P]dCTP using a Random Primer DNA labeling kit (Takara). RNA Molecular Weight Marker I (Boehringer Mannheim) was used as a molecular weight standard in some cases. Hybridization was carried out with DIG Easy Hyb Granules (Boehringer Mannheim) at 50°C for over 12 h. After hybridization, the membrane was washed twice with 2× SSC (20× SSC is 3 M sodium chloride and 0.3 M sodium citrate, pH 7.0) and 0.1% sodium dodecyl sulfate (SDS) for 5 min at room temperature and twice with 0.1× SSC and 0.1% SDS for 15 min at 65°C. For quantification of the radiolabeled bands, a BAS 2000 photoimaging system (Fuji film) was used.

Half-life measurement of puf mRNA by Northern blot analysis. Cells of R. sulfidophilum were grown till the logarithmic growth phase, showing an optical density of 0.5 to 0.6 at 660 nm. Then, rifampin was added to the cultures (500 µg/ml), and cells were collected at various time points. Total RNA was isolated with an RNeasy kit. Total RNA (3 µg per lane) was electrophoresed on 1.0% formaldehyde agarose gels and transferred to a Hybond-N+ positively charged nylon membrane. Probe C was used as a puf-specific DNA probe (see "Northern hybridization analysis" and Fig. 1a). The probe was labeled with [-³²P]dCTP using a Random Primer DNA labeling kit. DNA-RNA hybridization was carried out with DIG Easy Hyb at 50°C as instructed by the manufacturer. After hybridization, the membrane was washed twice with 2× SSC and 0.1% SDS for 5 min at room temperature and twice with 0.1× SSC and 0.1% SDS for 15 min at 65°C. For half-life measurement, the radiolabeled bands were quantified using a BAS 2000 photoimaging system.

Primer extension experiment. Total RNA of R. sulfidophilum grown till the logarithmic growth phase, showing an optical density of 0.5 to 0.6 at 660 nm, was extracted with an RNeasy kit from aerobically and photosynthetically grown cells. Primers used in this experiment were Z1100R (5'-GTCGAGCAGTG CCTGGGCGTCCTC-3'), Z0R (5'-AATAGACCATGAACTCCG-3'), and AENDR (5'-TCATGGGCGTGATGATCC-3') (nucleotide numbers of the 5' ends of the primers in the sequence with GenBank accession no. AB020784 are 975, 1218, and 1353, respectively). These primers were labeled at the 5' end with [-³²P]ATP using a DNA MEGALABEL labeling kit (Takara). To determine the 5' end of the puf operon mRNA, a mixture containing 20 µg of RNA, 50 mM Tris hydrochloride (pH 8.3), 50 mM KCl, 10 mM MgCl₂, and 0.5 pmol of primer in a final volume of 50 µl was incubated at 80°C for 2 min and then at 60°C for 45 min for DNA-RNA hybridization. Then, four deoxynucleoside triphosphates (final concentration, 0.5 mM each) and 20 U of reverse transcriptase (Rous-associated virus-2) were added to the mixture, and the primer extension reaction proceeded at 42°C for 1 h. The DNA synthesized was extracted with phenol-chloroform (1:1), precipitated with ethanol, and suspended in Tris-EDTA buffer. DNA was electrophoresed on a 6% polyacrylamide gel containing 8 M urea with a sequencing ladder, using the same DNA primers. Sequencing gels were dried and exposed to film with an intensifying screen at 80°C.

Measurement of BChl content. Membranes were obtained by sonicating cells grown until mid-logarithmic phase, showing an optical density of 0.5 to 0.6 at 660 nm, followed by differential ultracentrifugation as described previously (30). BChl contents in the membranes were determined using acetone-methanol (7:2) extract as described previously (10). Membrane protein content was determined with a Lowry assay kit (Bio-Rad Laboratories). Then, the relative values of BChl content per membrane protein in each sample were calculated.

Materials. Restriction endonucleases and reverse transcriptase were purchased from Takara. Synthetic oligonucleotide primers were purchased from Life Technologies, Inc. [-³²P]ATP and [-³²P]dCTP were purchased from Amersham.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Northern blot analysis. Figure 1a shows the structure of the R. sulfidophilum puf operon, which consists of six puf genes, pufQ, -B, -A, -L, -M, and -C (30). The puf operon transcripts were analyzed by Northern blot hybridization with the four puf-specific probes (probes A to D). As shown in Fig. 1b, probe A, corresponding to pufQBA and part of pufL (Fig. 1a), was hybridized with a 0.5-kb mRNA and also weakly with an approximately 3.5-kb mRNA. Probe B, corresponding to pufC (Fig. 1a), was hybridized with the 3.5-kb mRNA but not with the 0.5-kb mRNA. These two bands were not detected with probe D, which consisted mainly of the 3' end of bchZ and a small 5' segment of the pufQ gene, but they were detected with probe C, which consisted of pufB and part of the pufQ gene (Fig. 1b). The transcription of the puf operon appeared to be terminated after pufC because the signal nucleotide sequence for the typical Rho-independent transcriptional termination was found downstream of the puf operon (30). These observations indicated that the large transcript (3.5 kb) and the small transcript (0.5 kb) in Fig. 1b encode pufBALMC and pufBA, respectively. The small transcript (0.5 kb) was more abundant than the large transcript (3.5 kb). This difference may be a factor that is thought to adjust the ratio of LH1 peptides to RC proteins, as previously suggested for R. capsulatus (7, 25, 47).

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Physical map and expression of the puf operon of R. sulfidophilum. (a) Gene arrangement of the R. sulfidophilum puf operon. The genes are indicated by open boxes: Q, pufQ; B, pufB; A, pufA; L, pufL; M, pufM; C, pufC. The positions of putative hairpin structures are indicated by open circles. DNA probes used in Northern hybridization experiments are indicated by thick lines. (b) Northern hybridization analysis of mRNA transcripts with four different specific probes for the R. sulfidophilum puf operon. Total RNA was extracted from R. sulfidophilum grown under anaerobic light conditions. The digoxigenin-dUTP-labeled probe A, probe B, probe C, and probe D were used. The lengths of the standard RNAs are indicated on the left. (c) Effects of oxygen and light on puf operon transcription in R. sulfidophilum and R. capsulatus. Three micrograms of total RNA isolated from cells grown under four different conditions (LL-ANA, anaerobic low-intensity light [3 W/m2]; HL-ANA, anaerobic high-intensity light [100 W/m2]; DR-AER, aerobic dark; and HL-AER, aerobic high-intensity light [100 W/m2]) was loaded on each lane. Northern hybridization was performed with the 32P-labeled probe A of R. sulfidophilum or R. capsulatus. Ethidium bromide staining of 16S rRNA (lower panel, rRNA) was used to confirm equivalent loading of total RNA.

Primer extension analysis. To determine the possible transcription initiation site of the R. sulfidophilum puf operon, we performed RNA 5'-end mapping by primer extension analysis. The ³²P-labeled primers (Z1100R, Z0R, and AENDR: see Materials and Methods) were hybridized to the total RNA extracted from cells grown under anaerobic light and aerobic dark conditions (Fig. 2,  and +, respectively). Using a Z0R primer, two 5' ends of puf mRNA were observed. One had a less intense signal corresponding to 11 bp downstream from the pufQ start codon (pufP3), and the other showed a more intense signal corresponding to 132 bp upstream from the pufQ start codon (pufP2) (Fig. 2B). Results obtained with the AENDR primer also showed the presence of the two 5' ends of the puf mRNA at the corresponding positions (data not shown). A third 5' end of the puf mRNA was detected by the Z1100R primer (pufP1) (Fig. 2A). This site was resolved to a position much farther upstream (332 bp) from the pufQ start codon. These 5' ends of puf mRNA are indicated in Fig. 3 as P1, P2, and P3. All three transcripts were detected in both anaerobically and aerobically grown cells, but the relative band intensity between the conditions was different depending on the initiation sites.

View larger version (60K): [in this window] [in a new window]   FIG. 2.   Primer extension mapping of the 5' end of puf operon transcripts. Total cellular RNA isolated from R. sulfidophilum grown under aerobic dark conditions (+) and anaerobic light conditions () was hybridized with 5'-end-labeled Z1100R primer (A) or Z0R primer (B) (see Materials and Methods). Lanes G, A, T, and C show sequence ladders using the same synthetic primers. The numbers of 5' and 3' ends shown in the nucleotide sequences are the nucleotide numbering in the GenBank sequence with accession no. AB020784.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Deletion analysis of R. sulfidophilum puf operon. 5' and 3' deletions were constructed by using DNA-restricted and PCR-generated fragments (see Materials and Methods). These deletions were fused to the lacZ gene to construct various puf-lacZ transcriptional fusions. -Galactosidase activities associated with the transcriptional fusions on plasmids present in R. sulfidophilum cells were determined under different growth conditions (LL-ANA, anaerobic low-intensity light [3 W/m2]; HL-ANA, anaerobic high-intensity light [100 W/m2]; DK-AER, aerobic dark; HL-AER, aerobic high-intensity light [100 W/m2]). Values in parentheses are the standard deviations of at least three independent assays. No activity was detected for the sequence without insertion of the lacZ fusion (vector only). The positions of three 5' ends of puf mRNA determined in experiments in Fig. 2 are indicated by arrows (P1, pufP1; P2, pufP2; P3, pufP3).

Deletion analysis of the puf operon promoter. In order to analyze the R. sulfidophilum puf operon promoter, a nested set of deletions was constructed in the region of DNA upstream from the pufL gene. The deleted DNA fragments were cloned into the promoter testing vector, pCF1010 (28), to construct puf-lacZ transcriptional fusions (Fig. 3). These constructs were introduced into R. sulfidophilum cells and assayed for -galactosidase activity levels under anaerobic low-intensity-light (3 W/m²), anaerobic high-intensity-light (100 W/m²), aerobic dark, and aerobic high-intensity-light (100 W/m²) conditions (LL-ANA, HL-ANA, DR-AER, and HL-AER, respectively, in Fig. 3). The largest fragment, which extended from 584 bp upstream of pufP1 to pufL (pPS001), showed the highest lacZ expression under anaerobic low-intensity-light conditions. -Galactosidase activity of the plasmid was reduced by aeration and high-light illumination, suggesting that the transcriptional activity assayed with this lacZ fusion was repressed by oxygen and high-intensity light. High-light repression of lacZ activity was more apparent under aerobic conditions than under anaerobic conditions for this construct. Deletion of 303 and 540 bp from the 5' end of the largest fragment (pPS002 and pPS003, respectively) resulted in different levels of lacZ activity. A comparison of the activities with the largest fragment (pPS001) indicates that the values were reduced under anaerobic low-light, anaerobic high-light, and aerobic dark conditions (~30 to 50%) but increased under aerobic high-light conditions (~280 to 480%).

Decay rate of puf mRNA in R. sulfidophilum. In order to examine the posttranscriptional control of puf operon expression in R. sulfidophilum, the half-lives of both small (0.5-kb) and large (3.5-kb) puf transcripts were determined under anaerobic and aerobic growth conditions by Northern hybridization with the ³²P-labeled probe C (Fig. 1a) corresponding to pufB and part of the pufQ gene. Results are shown in Fig. 4. The half-lives of 3.5-kb puf mRNA were about 10 and 13 min in aerobic and anaerobic conditions, respectively, whereas the half-lives of 0.5-kb mRNA isolated from aerobically and anaerobically grown cultures did not show a significant difference, with values of about 40 min. We also performed corresponding measurements with R. capsulatus (data not shown). The results obtained were similar to those of the previous study, which showed that the half-lives of the larger puf transcript (2.7 kb) were about 3 and 8 min in aerobically and anaerobically grown cultures, respectively, and that those of the smaller RNA (0.5 kb) were about 30 min under both aerobic and anaerobic conditions (23). The larger transcript was degraded about twice as fast under aerobic conditions in R. capsulatus but only about 1.3 times faster under aerobic than under anaerobic conditions in R. sulfidophilum. The stability of puf mRNA was only weakly influenced by oxygen tension in R. sulfidophilum.

View larger version (49K): [in this window] [in a new window]   FIG. 4.   puf mRNA decay kinetics in R. sulfidophilum. Three micrograms of total RNA isolated from rifampin-treated R. sulfidophilum cells grown under aerobic dark and anaerobic light conditions was loaded on each lane. Hybridization was performed with the 32P-labeled probe C (Fig. 1a). The optical density of the bands was plotted against the time of RNA isolation (squares, aerobic conditions; circles, anaerobic conditions). The half-lives calculated from these blots for the 3.5-kb mRNA were about 10 min under aerobic conditions (dashed line) and about 13 min under anaerobic conditions (solid line). For the 0.5-kb mRNA, the half-lives were about 40 min under both aerobic conditions (dashed line) and anaerobic conditions (solid line).

Influence of regA mutation on the control of transcription of R. sulfidophilum puf operon. We previously constructed a regA deletion mutant (RESA1) from R. sulfidophilum and showed that the RegA-RegB regulatory cascade has an important role for the expression of photosynthesis genes in this bacterium, as in the Rhodobacter species (29). To test the effects of the regulatory mutant on the promoter activity of puf operon in R. sulfidophilum, we assayed the -galactosidase activity of various puf-lacZ fusions in RESA1. As shown in Fig. 5, the effects of -galactosidase activity caused by regA mutation fell into three types. In one type, the transcriptional pufL-lacZ fusions containing all three promoters (pPS001, pPS002, and pPS003 [Fig. 3]) exhibited nearly baseline levels of lacZ activity in RESA1 (Fig. 5A for pPS001; data for pPS002 and pPS003 are not shown). In contrast, a construct containing only the pufP3 promoter (pPS004 [Fig. 3]) retained the -galactosidase activity in the regA mutant as much as in the wild-type cells (Fig. 5B). The third type is a pufQ-lacZ transcriptional fusion from which the 3' end was deleted from a DNA fragment of pPS001 (pPS100 [Fig. 3]) and which showed a low level of -galactosidase activity even in the mutant while still containing the same 5' end of pPS001 (16 and 28% of those in wild-type cells under anaerobic and aerobic conditions, respectively [Fig. 5C]).

View larger version (15K): [in this window] [in a new window]   FIG. 5.   Measurement of puf operon transcriptional activity in wild-type (W.T.) and regA-disrupted R. sulfidophilum strains. Bars represent -galactosidase activities associated with the puf-lacZ transcriptional fusions, pPS001 (A), pPS004 (B), and pPS100 (C) (Fig. 3), in cells grown under anaerobic high-intensity light (100 W/m2) (O2) or aerobic dark conditions (+O2). Data for the wild type are taken from the results shown in Fig. 3. Units of -galactosidase activity represent micromoles of ONPG hydrolyzed per minute per milligram of total protein.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Three possible puf operon promoters in R. sulfidophilum. In this study, the R. sulfidophilum puf operon and its promoters were analyzed to clarify the mechanism of the oxygen-independent expression of photosynthetic genes. Primer extension analysis showed that there are three 5' ends of puf mRNA (pufP1, pufP2, and pufP3 in Fig. 2). By employing deletion analyses, two of the three 5' ends of puf mRNA were mapped as the possible transcription initiation sites at positions 332 bp upstream and 11 bp downstream of the pufQ start codon (pufP1 and pufP3, respectively [Fig. 2 and 3]), because the deletion of pufP1 or pufP3 from the inserted fragment of lacZ fusions resulted in significant reduction of the lacZ activity (Fig. 3). The other 5' end found in the primer extension analysis (pufP2; 132 bp upstream from pufQ) was not directly supported as a transcription initiation site by the deletion analysis because of the lack of appropriate restriction sites. In R. capsulatus, two puf transcription initiation sites have been reported and mapped to positions 316 and 129 bp upstream of the pufQ start codon. The nucleotide sequence alignment of puf operon promoter regions of R. capsulatus and R. sulfidophilum revealed that the two puf transcription initiation sites of R. capsulatus are located near the corresponding sites of pufP1 and pufP2 of R. sulfidophilum (separated by 9 and 4 bp, respectively) (data not shown) (1, 5). The regions are highly conserved between these two species, suggesting that the pufP2 of R. sulfidophilum is also a transcription initiation site, as in R. capsulatus.

Oxygen regulation on the puf operon expression in R. sulfidophilum. Northern blot hybridization experiments showed that the mRNA levels of the R. sulfidophilum puf operon were mostly maintained even under aerobic dark conditions (Fig. 1c), which was consistent with the relative BChl content in the cells (Table 2). This observation is in contrast to that in the Rhodobacter species (Fig. 1c) (7, 26, 46, 47), indicating that the aerobic synthesis of the photosynthetic apparatus in R. sulfidophilum seems to be mostly due to the puf operon expression, which is weakly sensitive to oxygen and regulated at the transcriptional level.

Light regulation of puf operon expression in R. sulfidophilum. Light regulation of puf operon expression is apparent under aerobic conditions in R. sulfidophilum (Fig. 1c and Table 2). pufQ- and pufL-lacZ activities were highly repressed by high-intensity light under aerobic conditions (~6 to 20% of those under aerobic dark conditions). Similarly, high-light illumination also resulted in lower levels of chromosomally derived puf mRNA under such conditions (~24 to 45% of those under aerobic dark conditions). Since the major fraction of puf mRNA seems to be derived from the readthrough transcription under aerobic conditions as discussed above, the readthrough activity may also be repressed to some extent by high-intensity light under aerobic conditions.

RegA-RegB regulatory system for puf operon expression in R. sulfidophilum. As mentioned previously, two promoters in the R. capsulatus puf operon have been reported. The R. capsulatus upstream promoter corresponding to pufP1 of R. sulfidophilum is highly expressed and regulated (1, 5). It was recently reported that the response regulator RegA binds to this promoter region (8, 16). It is clear from puf-lacZ fusion analysis that RegA is involved in activating puf operon transcription in R. sulfidophilum through association with the upstream region of the PstI site located between pufP1 and pufP3 (Fig. 5A and B). The promoter activity of pufP1 may be regulated by the RegA-RegB regulatory system in R. sulfidophilum, as in R. capsulatus (8, 16).