|Home | About | Journals | Submit | Contact Us | Français|
The BvgAS two-component system positively regulates the expression of the virulence genes of Bordetella pertussis and negatively regulates a second set of genes whose function is unknown. The BvgAS-mediated regulation of the bvg-repressed genes is accomplished through the activation of expression of the negative regulator, BvgR. A second two-component regulatory system, RisAS, is required for expression of the bvg-repressed surface antigens VraA and VraB. We examined the roles of BvgR and RisA in the regulation of four bvg-repressed genes in B. pertussis. Our analyses demonstrated that all four genes are repressed by the product of the bvgR locus and are activated by the product of the risA locus. Deletion analysis of the vrg6 promoter identified the upstream and downstream boundaries of the promoter and, in contrast to previously published results, demonstrated that sequences downstream of the start of transcription are not required for the regulation of expression of vrg6. Gel mobility-shift experiments demonstrated sequence-specific binding of RisA to the vrg6 and vrg18 promoters, and led to the identification of two putative RisA binding sites. Finally, transcriptional analysis and Western blot analysis demonstrated that BvgR regulates neither the expression nor the stability of RisA.
Bordetella pertussis is the causative agent of the respiratory infection of humans known as whooping cough (13). A wide variety of virulence factors have been identified which contribute to the bacterium's ability to colonize the host and cause disease (10-12, 14, 18, 23, 24, 26-28, 40). These factors include adhesins such as filamentous hemagglutinin, pertactin, and fimbriae, as well as several virulence determinants that cause damage to host tissues, including pertussis toxin, adenylate cyclase, and dermonecrotic toxin. The expression of most of these virulence factors is regulated in response to environmental signals by a two-component regulatory system encoded by the bvg locus (1, 29, 33, 34, 37). BvgS is a membrane-spanning protein that presumably acts as a sensor of the external environment, and BvgA is a soluble transcriptional activator. When B. pertussis is grown under normal laboratory conditions in rich media at 37°C, the BvgS protein is autophosphorylated and mediates the phosphorylation of BvgA through a series of phospho-transfer reactions (36, 38, 39). Upon phosphorylation, BvgA binds to the promoters of the virulence genes, inducing transcription of those genes (5-7, 9, 16, 29). Although the signals sensed by BvgS in vivo are unknown, it has been shown that when B. pertussis is grown at low temperatures, or in the presence of MgSO4 or nicotinic acid, the BvgS-mediated phospho-transfer reactions are inhibited, and the expression of the bvg-activated genes is down regulated (a condition referred to as modulation) (36, 38, 39).
In addition to the set of genes that is activated by the bvg locus, a second set of genes have been identified which are repressed by the bvg locus (17). Initial studies identified five bvg-repressed genes: vrg6, vrg18, vrg24, vrg53, and vrg73 (17). A conserved 21-bp sequence, located within the coding region, was identified in each of these genes, and mutations in the conserved region in the vrg6 gene were reported to cause a loss of repression resulting in constitutive expression of the gene (4). Replacement of the vrg6 promoter sequence with that of the nonregulated asd gene was reported to have no effect on the bvg-mediated repression of the gene (2, 3). Southwestern analysis demonstrated the binding of a bvg-activated, 34-kDa protein to the consensus sequence of the vrg6 gene (3). Taken together, these results suggested that the expression of the bvg-repressed genes was repressed by the binding of a bvg-activated repressor to the conserved element found at the 5′ end of each of the bvg-repressed genes.
Transposon mutagenesis studies have identified bvgR, a bvg-activated gene located immediately downstream of bvgS, as the repressor of the bvg-repressed genes (19, 21). In-frame deletions of bvgR result in constitutive expression of the bvg-repressed genes without affecting the regulated expression of the bvg-activated genes (19, 21). A more recent study demonstrated that expression of BvgR is activated by the binding of phosphorylated BvgA to the bvgR promoter (20). Taken together, these studies indicate that BvgA represses the expression of the bvg-repressed genes through the activation of the repressor, BvgR.
In addition to the bvgAS regulatory system that represses expression of the bvg-repressed genes, a second two-component regulatory system that is required for the expression of the bvg-repressed genes was identified. This locus was designated as the risAS locus due to its association with reduced intracellular survival by B. bronchiseptica (15). Two groups independently identified the risAS locus: Jungnitz et al. identified the risAS locus as a region that is required for intracellular survival of B. bronchiseptica, while Stenson et al. identified the same locus as a region that is required for the expression of two bvg-repressed surface antigens, VraA and VraB, in B. pertussis (15, 31, 32).
In this study, we investigated the regulation of the bvg-repressed gene, vrg6, by the bvgASR and risAS regulatory systems. We found that risA is essential for expression of the bvg-repressed genes and that the RisA protein binds to the promoter region of both the vrg6 and vrg18 genes. Although our data clearly demonstrate that BvgR is required for the repression of the bvg-repressed genes, we show that the putative repressor-binding site, conserved in each of the five bvg-repressed genes, is not required for the Bvg- or Ris-mediated regulation of the bvg-repressed genes in B. pertussis.
The bacterial strains and plasmids used in this study are presented in Table Table1.1. Escherichia coli strains were grown on L agar or in L broth supplemented with antibiotics when appropriate. B. pertussis strains were grown on Bordet-Gengou (BG) agar (Difco, Detroit, Mich.) containing 1% proteose peptone (Difco) and 15% defibrinated sheep blood. Unless otherwise noted, the concentrations of antibiotics included in the medium were 10-μg/ml gentamicin sulfate, 10-μg/ml kanamycin sulfate, 100-μg/ml streptomycin sulfate, and 10-μg/ml nalidixic acid.
Prior to mating, B. pertussis strains were grown for 3 days and E. coli strains were grown overnight at 37°C. Matings between E. coli and B. pertussis strains were performed by swabbing bacteria from fresh plate cultures of each strain onto a BG agar plate supplemented with 10 mM MgCl2. After 3 h of incubation at 37°C, bacteria were swabbed onto BG agar plates containing the appropriate antibiotics for the selection of exconjugants, and the plates were incubated at 37°C.
Strain TM1627, which bears an internal, in-frame deletion in risA, was constructed as follows: Oligonucleotide risA-F (5′-GCAGCGGGAAGACGAAGTTTCGA-3′) was used in combination with oligonucleotide risA-R (5′-CCGTATGCGAATAGACCAGGGCCGT-3′) in a PCR using Tohama I chromosomal DNA as template. The PCR product generated by the reaction was cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, Calif.), generating pTM255. Plasmid pTM255 was digested with SacII and was religated, generating pTM266, which has a deletion of a 204-bp fragment from the risA gene. Plasmid pTM266 was digested with EcoRI, and the fragment bearing the truncated risA gene was inserted into the EcoRI site of pSS1577 (35), generating pTM268. E. coli strain S17 bearing pTM268 was mated with B. pertussis strain BP536, and exconjugates in which the plasmid sequences had integrated into the chromosome were isolated by selection with gentamicin. An isolate in which plasmid sequences were lost from the chromosome, but in which the in-frame truncation of risA was retained, was isolated by selection for streptomycin resistance on BG plates and by screening with PCR. This strain was designated TM1627.
Strain TM1793, which bears an internal in-frame deletion in bvgR, was constructed as follows: The SalI restriction fragment that contains the 3′ end of the bvgR gene, and extends 304 bp downstream of bvgR, was excised from pBBR:BgB (19) and inserted into the SalI site in pTM025 (21) to generate pTM119. Plasmid pTM119 was digested with ApaI, followed by treatment with mung bean nuclease to generate blunt ends, digestion with StuI, and finally religation to circularize the plasmid. These manipulations resulted in the loss of an internal 606-bp fragment from bvgR. The plasmid bearing the in-frame internal truncation of bvgR was designated pTM120. E. coli strain S17 bearing pTM120 was mated with B. pertussis strain BP536, and exconjugates in which the plasmid sequences had integrated into the chromosome were isolated by selection with gentamicin. An isolate in which plasmid sequences were lost from the chromosome, but in which the in-frame truncation in bvgR was retained, was isolated by selection for streptomycin resistance on BG plates and by screening with PCR. This strain was designated TM1793.
B. pertussis strains bearing fusions of the vrg6, vrg18, vrg24, and vrg73 promoters to lacZ were constructed as follows. Oligonucleotide pairs vrg6-F1/vrg6-B1, vrg18-F1/vrg18-B1, vrg24-F1/vrg24-B1, and vrg73-F1/vrg73-B1 were used in PCRs using Tohama I chromosomal DNA as a template (Table (Table2).2). The PCR products generated by the aforementioned PCRs were cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, Calif.). The full-length vrg6, vrg18, vrg24, and vrg73 promoter fragments were excised from pCR2.1-TOPO using the restriction enzymes, XbaI and SalI, and the excised fragments were cloned into the previously described reporter plasmid, pSS2809 (8). The pSS2809 derivatives were subsequently transferred by conjugation from E. coli strain S17 into B. pertussis strains BP536, TM1627, and TM1793. Selection for exconjugates was performed by plating onto BG plates containing gentamicin and nalidixic acid. The identities of the resulting strains were confirmed by PCR.
A nested set of 5′ deletions of the vrg6 promoter fused to the lacZ gene were constructed as follows. Oligonucleotide vrg6-B1 (Table (Table2)2) was used in combination with oligonucleotides vrg6-F1, -F2, -F3, -F4, -F5, -F6, -F7, and -F8 in PCRs using Tohama I chromosomal DNA as a template. A nested set of 3′ deletions of the vrg6 promoter fused to the lacZ gene were constructed as follows. Oligonucleotide vrg6-F1 (Table (Table2)2) was used in combination with oligonucleotides vrg6-B2, -B3, -B4, -B5, and -B6 in PCRs using Tohama I chromosomal DNA as a template. The PCR products generated by the aforementioned PCRs were cloned into pCR2.1-TOPO. The full-length promoter fragment and each of the 5′ and 3′ deletions were excised by digesting with XbaI and SalI, and the resulting fragments were cloned into the previously described reporter plasmid, pSS2809 (8). These plasmids were then transferred by conjugation from E. coli strain S17 into B. pertussis strains BP536, TM1627, and TM1793. Selection for exconjugates was performed by plating onto BG plates containing gentamicin and nalidixic acid. The identities of the resulting strains were confirmed by PCR.
β-Galactosidase assays were performed as described by Miller (22) with minor modifications. Bacteria were recovered from the plates with a sterile swab and were resuspended in 3.5 ml of 1 M Tris-HCl, pH 8.0. The A600 was measured. For measurement of β-galactosidase activity, 50 μl of cell suspension was added to 1 ml of Z-buffer (0.1 M sodium phosphate [pH 7.0], 10 mM KCl, 1 mM MgSO4, 50 mM mercaptoethanol). Cells were permeabilized by adding 30 μl of 0.1% sodium dodecyl sulfate and 30 μl of chloroform, followed by vortexing. The remainder of the assay was performed as described by Miller (22). For quantification of β-galactosidase activity, units were defined by the following equation: Units = 1,000 × [A420 − (1.75 × A550)]/(T × V × A600), where T is the incubation time in minutes and V is the volume (in milliliters) of permeabilized cells added to the assay.
Oligonucleotide risAS-F (5′-GCCGGCGCGTGCCAGCAATTCCCGT-3′) was used in combination with oligonucleotide risAS-B (5′-GGCCTCAAGCCCTAAATTCTACGCT-3′) in PCRs using chromosomal DNA from four randomly selected B. pertussis clinical isolates (Bp106, Bp188, Bp509, and Bp10536) as the template. The resulting PCR products containing the risAS locus were sequenced with a BigDye Terminator v1.1 sequencing kit (Applied Biosystems), and the reactions were analyzed on ABI PRISM 3730xl DNA analyzers using Applied Biosystems sequence analysis software. The sequences were edited with Sequencher version 4.1.2, and the data obtained were assembled into contiguous sequences. The sequence of both strands of each of the four amplified DNA fragments was determined and compared to published sequences (15, 25).
An E. coli strain expressing recombinant RisA protein was constructed as follows. Oligonucleotide risA-F1 (5′-CTCGAGATGAACACGCAAAACACCACTCCT-3′) was used in combination with oligonucleotide risA-B1 (5′-CTCGAGACTGCCGCCATCCGGAACGAAAAC-3′) in a PCR using Tohama I chromosomal DNA as a template. The resulting PCR product containing the risA open reading frame was cloned into pCR2.1-TOPO, generating pTM275. The risA open reading frame was excised from pTM275 as an XhoI fragment and was inserted into the XhoI site in the expression vector pET22b (Novagen/EMD Biosciences, Inc., San Diego, Calif.), generating pTM276. Plasmid pTM276 bears a C-terminal fusion of a sequence encoding six histidine residues to the risA open reading frame under the transcriptional regulation of a recombinant T7 promoter engineered to be regulated by the E. coli lac repressor (LacI). Plasmid pTM276 was transformed into E. coli strain BL21(DE3)pLysS (Novagen/EMD Biosciences, Inc., San Diego, Calif.). E. coli strain BL21(DE3)pLysS, bearing pTM276, was grown in Luria broth at 37°C to an optical density of 0.6, and subsequently the expression of the risA gene was induced by the addition of IPTG (isopropyl-β-d-thiogalactopyranoside) to a final concentration of 1 mM. Four hours after induction, cells were harvested by centrifugation for 10 min at 5,000 × g and were lysed with 8 M urea. Denatured protein was purified with the QIAGEN Ni-nitrilotriacetic acid protein purification kit (QIAGEN, Valencia, Calif.) according to the manufacturer's instructions. Refolding of denatured protein was carried out by slow gradient dialysis against dialysis buffer (10 mM HEPES-NaOH, 1 mM EDTA, 0.1 mM dithiothreitol [DTT], pH 7.4). After removal of insoluble protein by centrifugation, the supernatant was concentrated in a Centricon-10 centrifugal concentrator (Millipore, Billerica, Mass.). Glycerol was added to a final concentration of 40%, and the samples were stored at −20°C.
The vrg6 promoter fragment generated by PCR using oligonucleotides F4 and B1 as described above, was excised from pCR2.1-TOPO by digestion with XbaI and SalI. Following gel purification, the fragments were end labeled with 32P by T4-polynucleotide kinase reaction (Lofstrand, Gaithersburg, Md.). Gel-shift reaction mixtures contained 10 ng of probe (2 × 104 dpm/ng), 100 ng of poly(dI:dC), and10 μg of purified protein in binding buffer (10 mM Tris-HCl [pH 7.8], 2 mM MgCl2, 50 mM KCl, 0.2 mM DTT). Reactions were incubated for 20 min at 30°C. Prior to loading samples, the 6% polyacrylamide-Tris-borate-EDTA (TBE) gels (Invitrogen, Carlsbad, Calif.) were pre-electrophoresed for 30 min. Electrophoresis was performed at 15 V/cm, after which gels were exposed to a PhosphorImager screen (Molecular Dynamics, Sunnyvale, Calif.), and the images were visualized with Imagequant software (Molecular Dynamics).
Recombinant RisA was mixed 1:1 (vol/vol) with a 0.65% solution of Alhydrogel (aluminum hydroxide; Superfos a/s, Vedbaek, Denmark) to give a final protein concentration of 500 μg/ml. Each of five female BALB/c mice was injected intraperitoneally with 0.1 ml of the protein-adjuvant solution (50 μg of protein). Booster doses were given 2 and 4 weeks after the initial injection. After the second booster dose, blood was collected from the periorbital artery of each mouse. The serum was collected by centrifugation and was stored at −20°C.
After growth on BG plates in the presence or absence of 50 mM MgSO4, bacteria were resuspended in Laemmli buffer (62.5 mM Tris-HCl [pH 6.8], 2.35% SDS, 100 mM DTT, 10% glycerol, 1 mM EDTA, 0.001% bromphenol blue), and were lysed by boiling. Samples were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (10% polyacrylamide), and proteins were transferred to nitrocellulose membranes using a wet tank immunoblotter (Bio-Rad, Hercules, Calif.). Nonspecific binding sites on the membranes were blocked using 5% (wt/vol) dehydrated milk (Marvel) in phosphate-buffered saline (PBS; blocking solution), and were probed with anti-RisA mouse polyclonal antiserum diluted 1:1,000 in blocking solution. Membranes were washed in PBS, and antigen-antibody complexes were detected with rabbit anti-mouse immunoglobulin G (IgG) antibodies conjugated to horseradish peroxidase diluted 1:500 in blocking solution. Cross-reacting proteins were visualized with the TMB membrane peroxidase substrate system (KPL, Gaithersburg, Md.).
Stenson et al. demonstrated that the expression of two distinct bvg-repressed surface antigens was dependent upon an intact risAS locus (31, 32). Based on this finding, we sought to determine if other bvg-repressed genes in B. pertussis were dependent upon the risAS locus for full expression. We cloned the promoter regions of four bvg-repressed genes (vrg6, vrg18, vrg24, and vrg73) into the promoter assay vector, pSS2809 (8), as described in Materials and Methods. This vector has a multiple cloning site for cloning promoter fragments upstream of a promoterless lac operon. It also has multiple tandem transcription terminators upstream of the cloning sites in order to minimize transcriptional read-through in this region. A 2-kb fragment of B. pertussis genomic DNA in the plasmid enables insertion of the plasmid via homologous recombination into the B. pertussis chromosome at a site distant from the vrg6, vrg18, vrg24, or vrg73 loci. The activity of each promoter fusion was assayed after growth of the bacteria in the absence or presence of MgSO4 in the wild-type, ΔbvgR, and ΔrisA genetic backgrounds (Fig. (Fig.1).1). As expected, all four bvg-repressed genes were expressed in the wild-type strain when bacteria were grown on BG plates in the presence of 50 mM MgSO4 but were not expressed when grown on BG plates in the absence of MgSO4 (Fig. (Fig.1).1). The expression of the four bvg-repressed genes was increased between 6- and 20-fold when the bacteria were grown in the presence of 50 mM MgSO4. When these same promoter fusions were tested in a strain bearing an in-frame deletion of the locus containing bvgR (ΔbvgR), all four loci demonstrated the same high level of expression in the presence of 50 mM MgSO4 (Fig. (Fig.1).1). Although all four loci demonstrated reduced levels of expression upon growth in the absence of 50 mM MgSO4, the level of expression of each of these genes was significantly higher than that observed in the wild-type background in the absence of 50 mM MgSO4. This result indicates that the expression of these four genes is derepressed in the absence of BvgR. When the transcriptional activity of the four bvg-repressed promoters was tested in a strain bearing an in-frame deletion of the risA gene (ΔrisA), the activity of each promoter was reduced to basal levels under all conditions tested (Fig. (Fig.1).1). These results confirm that expression of the four bvg-repressed genes, examined herein, is repressed by the product of the bvgR locus. These data also demonstrate that expression of these four genes is dependent on an intact risA locus.
The B. pertussis genome sequence was determined by Parkhill and colleagues at The Wellcome Trust Sanger Institute (25). According to the Sanger Institute sequence, the B. pertussis risS coding sequence has a frameshift mutation within a tract of three cytosine residues following codon 233. This frameshift was not present in the B. pertussis risS sequence reported by Jungnitz and colleagues (15). In order to determine if risS is a pseudogene in B. pertussis, we amplified and sequenced the DNA region containing the risAS locus from four independent clinical isolates of B. pertussis. The four isolates were selected randomly from the Center for Biologics Evaluation and Research B. pertussis strain collection. Isolates Bp106 and Bp108 were isolated in the 1930s. Isolate Bp509 was isolated in 1982. The isolation date of Bp10536 is unknown. The sequence of the risAS locus in all four clinical isolates was the same as that reported for the B. pertussis risAS locus by Parkhill et al.
We generated deletion mutants in an effort to define the upstream and downstream boundaries of the vrg6 promoter. A nested set of 5′ and 3′ deletions of the vrg6 promoter was constructed by PCR as described in Materials and Methods (Fig. (Fig.2A).2A). In order to determine the activities of these promoter derivatives in vivo, the deletion mutants were inserted into promoter assay vector pSS2809. The promoter deletion constructs were crossed into the wild-type, ΔbvgR, and ΔrisA genetic backgrounds, and the β-galactosidase activity of each reporter strain was determined after growth of the bacteria in the presence or absence of 50 mM MgSO4. In the ΔrisA background, all of the deletion fragments exhibited only a very low level of activity (Fig. (Fig.2B).2B). This observation was in agreement with the previous finding that an intact risA locus is required for expression of the vrg6 promoter (Fig. (Fig.1).1). In the wild-type background, promoter deletions Δ1 to Δ4 showed normal expression when the cells were grown in the presence of 50 mM MgSO4, and approximately sixfold repression when the cells were grown in the absence of MgSO4 (Fig. (Fig.2B;2B; constructs Δ1 to Δ4). These levels are similar to that seen with the full-length promoter. In the ΔbvgR background, promoter deletions Δ1 to Δ4 had the same high level of expression as the full-length vrg6 promoter when the bacteria were grown in the presence of 50 mM MgSO4. In the absence of 50 mM MgSO4, the level of expression observed in the ΔbvgR background was lower than that observed in the presence of MgSO4 but was significantly higher than the expression observed in the wild-type background in the absence of MgSO4. This analysis demonstrated that deletion of all of the sequences upstream of position −271, relative to the transcription start site, does not affect the regulated expression from the vrg6 promoter. Deletion of the sequences up to position −156 resulted in a complete loss of vrg6 promoter activity under all conditions tested, indicating that sequences between −271 and −156 are essential for vrg6 promoter activity (Fig. (Fig.2B;2B; constructs Δ5 to Δ8).
Examination of the 3′ promoter deletions revealed that deletion of all of the sequences downstream of position +24, relative to the transcription start site, does not affect the activation or repression of the vrg6 promoter (Fig. (Fig.2B;2B; constructs ΔC and ΔE). In the wild-type background, promoter deletions ΔC and ΔE showed normal levels of expression when the cells were grown in the presence of 50 mM MgSO4 and approximately sixfold repression when the cells were grown in the absence of MgSO4. In the ΔbvgR background, promoter deletions ΔC and ΔE demonstrated levels of expression similar to that of the full-length vrg6 promoter when cells were grown in the presence of 50 mM MgSO4. In the absence of 50 mM MgSO4, the level of expression observed in the ΔbvgR background was lower than that observed in the presence of MgSO4 but was significantly higher than the expression observed in the wild-type background in the absence of MgSO4. In both the wild-type and ΔbvgR backgrounds, promoter deletion ΔD, which has a 3′ endpoint between deletions ΔC and ΔE, demonstrated a higher level of expression than the full-length promoter and promoter deletions ΔC and ΔE. The expression of promoter deletion ΔD was repressed upon growth in the absence of MgSO4 in the wild-type background, and that degree of repression was significantly reduced in the ΔbvgR background. Deletion of sequences from the 3′ end up to position −63 and beyond resulted in a complete loss of vrg6 promoter activity under all conditions tested (Fig. (Fig.2B;2B; constructs ΔA and ΔB). Taken together, these results indicate that no sequences upstream of position −271 or downstream of position +24 are required for the RisA-mediated activation or the BvgR-mediated repression of the vrg6 gene. The results also indicate that the sequences downstream of +1, previously identified as the BvgR-binding site, are not required for repression of vrg6 expression.
The 5′ and 3′ deletion analyses identified the upstream and downstream boundaries of the vrg6 promoter, defining a 295-bp region (−271 to +24) that was required for the regulated activity of the vrg6 promoter. Since both RisA and BvgR have been shown to affect expression of vrg6, we hypothesized that binding sites for both proteins may be found within this 295-bp region. In order to directly evaluate the interaction between RisA and BvgR with the bvg-repressed promoters, we conducted gel-shift assays. The coding sequences for both BvgR and RisA were cloned into expression vector pET22b, and the expressed proteins were purified as described in the Materials and Methods. Both the BvgR and RisA proteins were expressed in large quantities upon induction and formed inclusion bodies in E. coli. RisA refolded in soluble form upon serial dialysis. However, BvgR remained an insoluble aggregate even after dialysis. Therefore, we conducted the gel-shift assays using only RisA. The addition of purified RisA protein to a reaction mixture containing 32P-labeled vrg6 promoter fragment resulted in a mobility shift of the labeled promoter fragment (Fig. (Fig.3A).3A). Increasing the amount of protein added to a constant amount of 32P-labeled promoter fragment resulted in an increase in the amount of probe that was shifted upward (Fig. (Fig.3A).3A). The binding interaction was specific as demonstrated by competition for binding of RisA to the 32P-labeled fragment by unlabeled vrg promoter fragments (Fig. (Fig.3B).3B). The RisA protein bound to the 32P-labeled vrg6 promoter fragment in the absence of unlabeled DNA competitor. This binding was completely blocked by the addition of unlabeled vrg6 and vrg18 promoter DNA but was not blocked by the addition of unlabeled asd or sodB promoter DNA (Fig. (Fig.3B3B).
We utilized 45-bp, double-stranded oligonucleotides bearing sequences derived from the vrg6 promoter region, to identify RisA binding regions on the vrg6 promoter. We evaluated the ability of the double-stranded oligonucleotides to compete with the 32P-labeled vrg6 probe for binding of RisA. A set of 15 double-stranded oligonucleotides that spanned the vrg6 promoter was constructed (Table (Table3).3). The first eight oligonucleotides (no. 1 to 8) spanned the 295-bp region known to be required for the RisA-mediated activation of the promoter. The next seven oligonucleotides (no. 12, 23, 34, 45, 56, 67, and 78) were constructed to overlap the first eight oligonucleotides in order to account for any binding sites that may be disrupted by linker design. The ability of these linkers to compete with the 32P-labeled vrg6 promoter DNA for RisA-binding was evaluated with gel-shift assays (Fig. (Fig.4).4). Four linkers (no. 2, 5, 12, and 56) were identified that competed with the 32P-labeled vrg6 probe (Fig. (Fig.4).4). Linkers 2 and 12 overlap each other, as do linkers 5 and 56. However, linkers 2 and 12 do not overlap with linkers 5 and 56 (Fig. (Fig.5B).5B). Our results, therefore, define two distinct regions of approximately 25 bp each, which contain a RisA binding site.
An examination of the sequences of the two 25-bp regions that were shown to bind RisA revealed a conserved 7-bp sequence (5′-AAATG/TTA-3′; Fig. Fig.5B).5B). A search of the 500 bp upstream of the start of translation of the vrg6, vrg18, vrg24, and vrg73 genes identified three matches to this sequence in the vrg6 promoter region, two matches to this conserved sequence in the vrg18 promoter region, five matches in the vrg24 promoter region, and a single match in the vrg73 promoter region (Fig. (Fig.5C5C).
The observation that the presumed repressor-binding site in vrg6 does not contribute to repression of gene expression compelled us to consider the alternative mechanisms by which BvgR may repress its target genes. We examined the possible role of BvgR in the expression and stability of RisA. The risA and bvgR promoters were inserted into the promoter assay vector pSS2809. The promoter constructs were crossed into the wild-type, ΔbvgR, and ΔrisA backgrounds, and the β-galactosidase activity of each reporter strain was determined after growth in the presence or absence of 50 mM MgSO4 (Fig. 6A and B). The expression and regulation of the bvgR promoter were the same in the wild-type, ΔbvgR, and ΔrisA backgrounds, indicating that bvgR does not regulate its own expression, nor is its expression regulated by risA (Fig. (Fig.6A).6A). The expression of the risA promoter was the same in the wild-type and ΔrisA backgrounds but higher in the ΔbvgR background, indicating that risA does not regulate its own expression (Fig. (Fig.6B6B).
Polyclonal antiserum to recombinant RisA was raised in mice and was used to probe immunoblots containing cell lysates prepared from wild-type, ΔbvgR, and ΔrisA strains after growth in the presence or absence of 50 mM MgSO4 (Fig. (Fig.6C).6C). As expected, no RisA could be detected in the ΔrisA mutant strain. The amounts of RisA protein present in cell lysates were the same in both the wild-type and ΔbvgR backgrounds regardless of the growth conditions. These results indicate that the level of expression and the stability of RisA are not affected by BvgR- or bvg-mediated regulation.
The mechanism by which the expression of the bvg-repressed genes is regulated is unknown. We demonstrated previously that, in B. pertussis, the expression of the vrg6 and vrg73 genes is repressed by the bvgAS-activated protein BvgR (21). Jungnitz et al. identified a locus in B. bronchiseptica, which they designated as the risAS locus, that was required for survival of B. bronchiseptica in eukaryotic cells and contributed to persistence of the bacterium in a mouse infection model (15). Stenson et al. independently identified the risAS locus in B. pertussis and demonstrated that it was required for the expression of two distinct bvgR-repressed surface antigens (31, 32). In this study, we examined the roles of BvgR and RisA in the regulation of four bvg-repressed genes in B. pertussis (vrg6, vrg18, vrg24, and vrg73). Our results demonstrate that the repression of expression of all four of these genes is dependent upon BvgR and the activation of expression of these genes is dependent upon RisA (Fig. (Fig.1).1). The observation that the expression of all of the bvg-repressed genes examined to date (vrg6, vrg18, vrg24, vrg73, and the bvg-repressed surface antigens described by Stenson et al.) is dependent upon RisA is striking. It suggests that there is a large, if not complete, overlap of the BvgR and RisA regulons. This might be expected if, for example, BvgR represses expression of the bvg-repressed genes by controlling the activity of RisA. Microarray and proteomic approaches should be undertaken to define the BvgR and RisA regulons, in order to evaluate the extent to which they overlap.
Parkhill and colleagues first determined that in B. pertussis the risS gene is disrupted by a frameshift mutation (25). We have confirmed the presence of the same frameshift in the risS gene in four randomly selected clinical isolates. The isolation of these four strains was separated both geographically and by time (separated by nearly 50 years). It is clear from our results that the risA locus is required for expression of the bvgR-repressed genes; yet it is equally clear that the risS gene does not encode a functional RisS protein. Given these facts, we conclude that either RisA is able to activate transcription at its target promoters in the unphosphorylated state or RisA is phosphorylated by a sensor kinase other than RisS.
Once we demonstrated that RisA activates the expression of the bvg-repressed genes, we focused on identifying the cis-acting sequences that are required for RisA activation of the bvg-repressed genes. We noted that Beattie et al. had performed a deletion analysis of the vrg6 promoter (2) and demonstrated that as progressively larger fragments of the promoter were deleted from the upstream side between positions −428 and −221, relative to the transcription start site, almost a fourfold drop in vrg6 promoter activity was observed. Beattie et al. also observed a fivefold drop in promoter activity when the deletions were extended further to position −186, and activity was further diminished as the sequence was deleted even further toward the transcription start site. The activity of all of the constructs generated by Beattie et al. appeared to be repressed upon growth under modulating conditions. These data suggested that an important promoter element lay between positions −221 and −186. However, it also appeared that sequences upstream of position −221 contributed to promoter activity. We noted that Beattie and colleagues performed their analysis with promoter fusions carried on a plasmid present at a concentration of five to seven copies per cell. We hoped a deletion analysis performed with fusions maintained in single copy would yield a clearer picture of the upstream boundary of the vrg6 promoter. We also sought to take advantage of the ability to examine the activity of each fusion in both a ΔbvgR background and a ΔrisA background. Our deletion analysis of the vrg6 promoter demonstrated that all of the sequences required for regulated expression of vrg6 were downstream of position −271 relative to the transcription start site (Fig. (Fig.2).2). All of the deletion fragments with upstream boundaries either at or further upstream than position −271 retained 100% of promoter activity. This activity was dependent upon RisA and was repressed by BvgR. All of the upstream deletions that extended into the promoter to position −159 or beyond lost all RisA-dependent promoter activity, and the basal level of expression observed in those fusions was not repressed by BvgR (Fig. (Fig.2).2). Our results are in general agreement with those of Beattie et al. Our results indicate that the upstream boundary of an important promoter element lies between positions −271 and −159 relative to the transcription start site. However, our results clearly demonstrate that no sequences upstream of position −271 contribute to promoter activity. Finally, we did not see any evidence of BvgR-mediated repression of promoter fusions that had 5′ deletions that extended beyond position −159.
We extended the promoter deletion analysis by generating deletions from the 3′ end (Fig. (Fig.2).2). Our results demonstrate that promoter fragments with deletions from the 3′ end that extended as far as position +24, relative to the transcription start site, retained 100% of promoter activity. The activity of all three of these promoter fusions was dependent upon RisA, and, most significantly, despite the lack of the putative repressor-binding site in two of the constructs, BvgR repressed the activities of all three of these fusions. As expected, deletions from the 3′ end that extended as far as positions −63 and −130, relative to the transcription start site, lost all RisA-dependent promoter activity and BvgR did not repress the basal level of expression observed in those fusions (Fig. (Fig.2).2). Therefore, we have concluded that sequences downstream of the start of transcription are not required for repression of the vrg6 gene and probably are not required for repression of any of the bvg-repressed genes. In retrospect, reexamination of the data that led to the identification of the repressor-binding site within the coding region of vrg6 suggests that what was interpreted previously as a loss of repression of the vrg6 promoter constructs (2, 3) was in fact a loss of vrg6 promoter induction.
Our promoter deletion analysis demonstrated that all of the sequences required for the regulated expression of the vrg6 gene lay between position −271 and +24, relative to the transcription start site. Since all of the promoter fragments that demonstrated activity were repressed by BvgR, we concluded that the sequences required for activation of the promoter by RisA, and repression of the promoter by BvgR, lie within that 295-bp region. We attempted to overexpress and purify both BvgR and RisA for use in an in vitro DNA-binding assay. Although the purification of RisA was relatively straightforward, multiple strategies to purify soluble BvgR were unsuccessful. A gel-shift analysis clearly demonstrated that purified RisA binds to the vrg6 promoter (Fig. (Fig.3).3). Binding of RisA to a 32P-labeled vrg6 promoter fragment was blocked by excess unlabeled DNA fragments containing the vrg6 or vrg18 promoters but not by unlabeled fragments containing the asd or sodB promoters. This result indicated that binding of RisA to the vrg6 promoter was sequence specific and also demonstrated that RisA binds to the vrg18 promoter. By evaluating the ability of a collection of 15 overlapping 45-bp double-stranded oligonucleotides bearing sequences from the vrg6 promoter, to block binding of RisA to the 32P-labeled vrg6 promoter fragment, two distinct RisA-binding regions were identified (Fig. (Fig.4).4). Although none of the linkers completely eliminated the gel mobility shift, four linkers clearly inhibited RisA binding to the vrg6 probe, resulting in a reduced shift in mobility. Interestingly, these four linkers consisted of two pairs of overlapping linkers. This observation suggested that the RisA binding regions are located on the two 22-bp regions defined by the overlap between linkers 2 and 12 and linkers 5 and 56. A comparison of these two sequences revealed a conserved 7-bp sequence that is present in both regions (5′-AAATT/GTA-3′) (Fig. (Fig.5B).5B). Although close approximations of this sequence are found in the vrg18, vrg24, and vrg73 promoter sequences (Fig. (Fig.5C),5C), it should be noted that a nearly perfect match to this sequence (5′-AAATTTG-3′) is found at a third position in the vrg6 promoter, and is incorporated into linker 34, which did not compete with the vrg6 promoter probe for binding of RisA. Although it is tempting to focus on this sequence element, definition of the specific sequences that contribute to the binding of RisA will require a more rigorous analysis of RisA binding to the risA-activated promoters.
To date, attempts at identifying specific cis-acting sequences involved in repression of expression of the vrg6 gene have not been successful. This has led us to consider alternative models for BvgR-mediated repression of vrg6 promoter activity (i.e., models that do not require direct binding of BvgR to the bvg-repressed promoters). One possibility is that BvgR represses the expression of RisA or reduces the stability of RisA once it is synthesized. To address this question, we examined the expression of risA-lacZ fusions in the wild-type, ΔbvgR, and ΔrisA backgrounds (Fig. (Fig.6B).6B). Our results indicated that BvgR does not repress risA transcription. It also indicated that risA transcription is not autoregulated. This conclusion was confirmed by immunoblot analysis which demonstrated the presence of steady-state levels of RisA in wild-type, ΔbvgR, and ΔrisA strains (Fig. (Fig.6C).6C). The observation that the amount of RisA present in cells is constant in the wild-type and ΔbvgR genetic backgrounds, after growth in the presence or absence of MgSO4, demonstrates that BvgR neither regulates the transcription or translation of RisA, nor does it mediate its stability once the protein is expressed. It seems unlikely that BvgR acts independently of RisA at the vrg6 promoter since BvgR does not repress the residual RisA-independent activity of the vrg6 promoter. We speculate that if BvgR acts by binding to the vrg6 promoter, its binding may interfere with the binding of RisA rather than interfere with the binding of the polymerase. Alternatively, BvgR may not directly repress expression of the bvg-repressed genes, but rather, it may exert its effect indirectly through RisA either by binding RisA and preventing its interaction with the promoter or by modifying RisA so that the protein is no longer active. Ongoing work in our laboratory is directed toward distinguishing between these alternative models.
We thank Gopa Raychaudhuri, Michael Schmitt, and Scott Stibitz for their critical reading of the manuscript.