|Home | About | Journals | Submit | Contact Us | Français|
Thioredoxin reductase (encoded by trxB) protects Staphylococcus aureus against oxygen or disulfide stress and is indispensable for growth. Among the different sarA family mutants analyzed, transcription of trxB was markedly elevated in the sarA mutant under conditions of aerobic as well as microaerophilic growth, indicating that SarA acts as a negative regulator of trxB expression. Gel shift analysis showed that purified SarA protein binds directly to the trxB promoter region DNA in vitro. DNA binding of SarA was essential for repression of trxB transcription in vivo in S. aureus. Northern blot analysis and DNA binding studies of the purified wild-type SarA and the mutant SarAC9G with oxidizing agents indicated that oxidation of Cys-9 reduced the binding of SarA to the trxB promoter DNA. Oxidizing agents, in particular diamide, could further enhance transcription of the trxB gene in the sarA mutant, suggesting the presence of a SarA-independent mode of trxB induction. Analysis of two oxidative stress-responsive sarA regulatory target genes, trxB and sodM, with various mutant sarA constructs showed a differential ability of the SarA to regulate expression of the two above-mentioned genes in vivo. The overall data demonstrate the important role played by SarA in modulating expression of genes involved in oxidative stress resistance in S. aureus.
Staphylococcus aureus, a versatile pathogen, has colonized more than a billion people worldwide. An important concern is that the number of methicillin-resistant S. aureus (MRSA) isolates in hospitals and the community has increased in various countries, including the United States (21, 28). S. aureus can cause a wide variety of infections ranging from a local infection (e.g., wound infection, cellulitis, boils, and furuncle) to systemic dissemination (bacteremia) and then to metastatic infections (e.g., pneumonia, endocarditis, septicemia, and osteomyelitis) in healthy as well as immunocompromised hosts. Once inside the host, S. aureus produces a large number of factors that include enzymes, adhesions, toxins, and capsular polysaccharides and which cause tissue colonization, tissue destruction, and immune evasion. The expression of many of these pathogenesis-related factors is controlled by regulatory systems that include transcriptional regulators (i.e., SarA family, SigB) or two-component signal transduction systems (i.e., agr, saeRS, arlRS) (3, 7, 10-16, 27, 31-36, 38, 41, 42, 45, 46).
The SarA family of transcriptional regulators controls the transcription of a wide variety of genes, many of which are important for virulence (13). SarA, the first member of this family to be identified, upregulates the synthesis of fibronectin and fibrinogen binding proteins, hemolysins (alpha-, beta-, and gamma-hemolysins), enterotoxins, toxic shock syndrome toxin 1, and genes involved in biofilm formation (e.g., icaRA, bap), and represses expression of proteases (ssp, aur), protein A (spa), and collagen binding proteins (cna) (7, 11, 13, 15, 46). SarA regulates many genes indirectly by modulating the expression of other regulatory loci (e.g., rot, agr, sarS, sarV, sarT) (13, 15, 33, 35, 36, 46). Deletion of sarA in S. aureus affects large numbers of target genes involved in virulence or metabolic processes, as determined with gene arrays (16). Nine other SarA paralogues (i.e., SarR, SarS, SarT, SarU, SarV, SarX, SarZ, MgrA, and Rot) have been identified and characterized in S. aureus. These paralogues regulate a large number of target genes, including those involved in virulence, biofilm formation, autolysis, antibiotic resistance, and metabolic processes (3, 4, 7, 9, 10, 13, 16, 27, 30-36, 38, 43, 46). The SarA paralogues are homologous with the MarR (multiple antibiotic resistance) family of proteins found in several other bacterial species (13).
Reactive oxygen species (ROS), such as hydrogen peroxide (H2O2), superoxide anion (O2−), hydroxyl radicals (OH·), and organic hydroperoxides (OHPs), are toxic to cells, as they damage cellular constituents such as DNA, lipids, and proteins (51). In spite of their toxicity, ROS are important for several normal cellular processes, such as apoptosis, cell proliferation, and differentiation (39, 44). Host neutrophils and other phagocytes that are involved in immune defense against invading bacteria and viruses produce ROS through special metabolic pathways (6). S. aureus has developed efficient pathways to defend against oxidative stresses when attacked by host phagocytes at the site of bacterial infection. In response to oxidative stress, or any condition that results in an increase of ROS, cells activate expression of oxidative stress-responsive genes (e.g., sod, katA, ahpCF, trxA, and trxB), whose products are involved in detoxification (22, 39). Apart from enzymatic removal of oxidative free radicals, manganese homeostasis is also involved in innate defense mechanism against oxidative stress. The divalent Mn2+ can act as a nonenzymatic superoxide dismutase for the removal of superoxide radicals (O2−) from the S. aureus cells (26). In bacteria, the toxic ROS are sensed by transcriptional regulators (e.g., PerR, OhrR, OxyR), which in turn bring about the desired changes in gene transcription (13, 23). In addition, small thiol-specific proteins such as thioredoxin and glutaredoxin, which maintain the intracellular thiol-disulfide balance and provide reducing power to the key reductive enzymes (e.g., ribonucleotide reductases), are also involved in scavenging ROS in bacteria (52). S. aureus lacks the glutathione pathway; therefore, it requires thioredoxin and alternate thioredoxin pathways to carry out cellular functions (40). Thioredoxin (encoded by trxA) is maintained in the reduced form by thioredoxin reductase (encoded by trxB) (1). In S. aureus, trxB belongs to the PerR stimulon and is negatively regulated by PerR (24). The thioredoxin system is constitutively expressed in S. aureus, but under conditions of disulfide or oxidative stress its expression is enhanced. In fact, the trxB gene has been shown to be essential in S. aureus, thus underscoring the importance of the trxB gene under normal growth conditions (52).
Apart from PerR in S. aureus, the other iron-dependent global regulators that are involved in resistance to oxidative stress are Zur, Fur, and MntR (23, 24, 26, 29, 53). Recently, MgrA and SarZ have been shown to be involved in redox signaling via a thiol-based oxidation mechanism. A conserved cysteine residue located at the N-terminal domain of both proteins has been shown to play a critical role in the redox-sensing mechanism (9, 10). The OhrR, which regulates the expression of the ohr (organic hydroperoxide resistance) genes in Gram-positive bacteria, shows considerable homology with several members of the SarA protein family, such as SarZ (63%), MgrA (61%), and SarR (51%). The OhrR, MgrA, and SarZ proteins possess a unique cysteine (Cys) residue (at their N terminus) which is involved in the redox-sensing mechanism (19). Similar to MgrA and SarZ, SarA has a cysteine residue at N-terminal position 9, but its role in redox signaling is unknown.
Our previous studies showed that SarA can modulate transcription of sodA and sodM (4). To further investigate if SarA plays a broad role in the regulation of oxidative stress-responsive genes, we examined the transcriptional regulation of the thioredoxin (trx) genes, in particular the trxB gene, in sar family mutant strains of S. aureus. The role of the cysteine residue at the position 9 (Cys-9) of SarA was also investigated to elucidate its involvement in redox signaling. Results showed that transcription of trxB was increased in the sarA mutant under aerobic as well as microaerophilic conditions. Northern hybridization results and in vitro gel shift analyses with various mutant forms of the sarA gene showed that the DNA binding ability of SarA was essential for repression of trxB transcription in S. aureus. In this report, we show for the first time that the affinity of SarA to a particular target promoter may determine its ability to bring about repression from that promoter.
The bacterial strains and plasmids used in this study are described in Table Table1.1. Phage 11 was used as a generalized transducing phage for S. aureus strains. S. aureus strain RN4220 was used as the initial recipient for the transformation of plasmid constructs. The S. aureus cells were grown in tryptic soy broth (TSB) or on tryptic soy agar (TSA) supplemented with antibiotics when necessary (5 μg/ml of erythromycin, 3 μg/ml of tetracycline, and 10 μg/ml of chloramphenicol). The cells were grown overnight in TSB and then inoculated in fresh TSB (at an initial optical density at 600 nm [OD600] of ~0.05) and grown with continuous aeration in a shaker at 37°C as described previously (32, 33). Growth was monitored by measuring changes in turbidity at 600 nm in a spectrophotometer (Spectronic 20D+). In some of the cases, the S. aureus strains were allowed to grow to an optical density of approximately 0.7 (early exponential phase of growth), and compounds (1.0 mM methyl viologen, 2 mM H2O2, 0.5 mM tert-butyl hydroperoxide, 0.5 mM cumene hydroperoxide, and 2 mM diamide) were added to induce oxidative stress. For microaerophilic growth, the S. aureus cultures were inoculated in glass tubes, as described above, and incubated at 37°C in a candle extinction jar for 16 h. When required, the microaerophilically grown cells were removed from the candle jar, the earlier-mentioned oxidizing agents were added to the tubes, and the tubes were immediately returned to the candle extinction jar. The cells were harvested 1 h after the addition of oxidizing agents, and the total cellular RNA was extracted from them. All the chemicals used in this study were reagent grade and were obtained from Sigma Chemicals and Fisher Scientific.
S. aureus sarA mutants were constructed by transducing the 11 phage lysate of RN6390 sarA::ermC (12) into different S. aureus strains, SH1000, Newman, and MW2. A 11 lysate of strain ALC812 (12), which contains a single-copy integration of the 1.5-kb sarB region of the sarA locus into the lipase (geh) locus, was prepared and transduced into a SH1000 sarA mutant to generate a single-copy sarA-complemented strain. The 650-bp-containing P1sarA region of the sarA locus was cloned into shuttle vector pSK236 for overexpression of the sarA gene in a sarA mutant background (30). This recombinant plasmid was first electroporated into strain RN4220. The plasmid isolated from RN4220 background was electroporated into the SH1000 sarA mutant. The various sarA mutations we analyzed were cloned into the shuttle plasmid pSK236 and were characterized as described previously (30). These mutations were sarAC9G, sarAY18A, sarAK27A, sarAE29A, sarAK54A, sarAK69A, sarAF80A, sarAR84A, sarAR90A, and sarAQ100A.
Isolation of total cellular RNA and the subsequent analysis by Northern blotting were described previously (32, 33). The DNA fragments containing the open reading frames of the trxB, trxA, perR, sodM, and sarA genes were excised from plasmids containing the respective gene by using appropriate restriction enzymes. The intensities of the signals were quantified by image analysis on a PhosphorImager (Amersham). Each of the experiments was repeated at least three independent times.
The gel shift assays were performed to determine the DNA-protein interaction of the wild-type and the various SarA mutant proteins with the trxB or sodM promoter DNA fragments. DNA fragments of 258 bp and 273 bp containing the promoter regions of the trxB and sodM genes, respectively, were amplified by PCR with suitable primers and cloned into pCR2.1 vector (Invitrogen, CA). The promoter fragments were processed and end labeled with [γ-32P]ATP as described previously (4, 36). The gel shift assays with various SarA proteins were performed as described previously (30, 32, 33). The cloning, purification, and quantification of various mutants with His6-SarA were described previously (30). To determine the effect of oxidizing agents on the trxB promoter binding ability, the wild-type SarA and the SarAC9G proteins were treated with H2O2 (15 mM) and diamide (15 mM) as described previously (9, 10). Briefly, 600 ng of purified protein was incubated in the reaction buffer (25 mM Tris-HCl [pH 7.5], 0.1 mM EDTA, 75 mM NaCl, 5% glycerol, and 20 μg/ml calf thymus DNA) with or without an oxidizing agent at room temperature. After 30 min, 50 mM dithiothreitol (DTT) was added to the desired tubes and incubation continued for another 30 min, after which samples were subjected to gel shift assays with the labeled trxB promoter fragment. The KD values were calculated from densitometry quantification of the free and shifted DNA-protein bands (47). The reported values were calculated based on averages from three independent experiments.
Whole-cell extracts from S. aureus SH1000 wild type or the sarA mutant containing different P1sarA gene constructs were prepared by growing cells under microaerophilic conditions as described previously (4). The concentration of total proteins from clear lysates was determined by using the Bio-Rad (Hercules, CA) protein estimation kit using bovine serum albumin as the standard. Western blotting and detection were performed as described previously (4).
To analyze the effects of inactivation of the individual sarA family genes on expression of the trx genes, Northern analysis was performed with the open reading frames of the trxA and trxB genes, and total cellular RNA was isolated from the 10 sarA family mutants (i.e., sarA, sarR, sarS, sarT, sarU, sarV, sarX sarZ, mgrA, and rot) of S. aureus strain RN6390 in the postexponential phase of growth. Results (Fig. (Fig.1A)1A) showed that the level of the trxB transcript was considerably enhanced (~3-fold) only in the sarA mutant, whereas the level of trxA transcription was not affected in any of the above-mentioned mutants (data not shown). Strain RN6390 is a derivative of the strain 8325-4 lineage that produces less alternative sigma factor, SigB, due to a natural deletion of 11 bp in the rsbU gene, whose product is required for optimal sigB expression (25). To avoid any discrimination in gene regulation that may arise due to a decrease in SigB activity, rsbU+ strain SH1000 (8325-4 carrying an intact rsbU gene ) was used for further studies. Results from transcriptional analysis of SH1000 and its isogenic sarA mutant showed that trxB expression was similar in the early exponential phase (OD600, ~0.7) of growth, but in the exponential (OD600, ~1.1) and the postexponential (OD600, 1.7) phases of growth, the sarA mutant showed severalfold-higher expression of the trxB transcripts than the wild-type SH1000 (Fig. (Fig.1B).1B). To confirm the role of sarA in trxB regulation, complementation analysis with a single copy and multiple copies of the sarA gene in the sarA mutant was performed (Fig. (Fig.1C).1C). Transcription of trxB returned to near-parental levels in the single-copy-complemented sarA mutant, whereas the trxB transcript level was further reduced (compared to wild type) in the sarA mutant complemented with multiple copies of the sarA gene.
Previously we showed that under microaerophilic conditions the sarA mutant displays enhanced transcription of oxidative stress-responsive genes, such as sodA and sodM (superoxide dismutase), compared to the wild-type strain (4). To determine whether the trx genes showed a similar pattern, the expression of trxA and trxB transcripts was monitored in S. aureus under microaerophilic conditions. When compared to the wild-type strain, an approximately 5-fold increase in expression was observed in the SH1000 sarA mutant (Fig. (Fig.1D).1D). The sarA mutants of two clinical isolates (MW2 and Newman) also showed increased trxB expression compared to the respective wild-type strains (Fig. (Fig.1D).1D). Interestingly, no notable difference in expression of the trxA gene was observed in these sarA mutants and their corresponding wild-type cells (data not shown). Transcription of perR (encoding PerR, the negative regulator of trxB ) was also monitored in the SH1000 wild type and the sarA mutant to find out if the SarA regulatory effect occurred indirectly by changes in the expression of perR. Results showed that expression of perR remained unchanged in the sarA mutant as well as in the wild-type strain (Fig. (Fig.1E).1E). These data suggest that sarA does not regulate trxB indirectly by reducing expression of perR; therefore, SarA-mediated trxB regulation appears to be independent of PerR. Overall, the data presented in this section clearly show that transcription of the trxB gene is increased in the sarA mutant in both the laboratory and clinical strains. In S. aureus, several genes, such as superoxide dismutases (sodA and sodM), catalase (katA), thioredoxin (trxA), thioredoxin reductase (trxB), alkyl hydroperoxide reductase subunits C and F (ahpCF), and endopeptidase (clpC) (18), play an important role in combating oxidative stress in vivo. Our previous studies with sodA and sodM (4), and unpublished observations with the katA and ahpC genes, suggest that SarA can modulate transcription of several oxidative stress-responsive genes. Under microaerophilic conditions, the levels of ROS are reduced compared to aerobic growth. Hence, in this environment, it is desirable for the cell to reduce the transcription of genes involved in the detoxification of ROS. The increased transcription of all of these genes in the sarA mutant suggests that SarA may be the principal regulator involved in controlling expression of these oxidative stress-responsive genes.
Because the level of the trxB transcript was increased in the sarA mutant and returned to near the parental level in the sarA-complemented strain (Fig. (Fig.1),1), we hypothesized that SarA may bind to the trxB promoter region to modulate trxB expression. To analyze the direct SarA interaction with the upstream promoter region of the trxB gene, a 258-bp trxB promoter fragment (nucleotide positions −224 to +34 ) was used for gel shift assays with various amounts of purified SarA protein (Fig. (Fig.2A).2A). Retarded DNA-protein complex could be detected with 0.4 μg of SarA. Furthermore, the retarded DNA-protein complexes became more predominant as the concentration of SarA protein increased, with a complete conversion at 0.6 μg of SarA. The unlabeled trxB promoter fragment could compete to a lesser extent with the labeled fragment (Fig. (Fig.2A,2A, lane 7), whereas an unlabeled nonspecific DNA (185-bp DNA fragment from the internal region of the sarX gene ) could not compete effectively with the labeled trxB promoter fragment (Fig. (Fig.2A,2A, lane 8). A prior study employing site-directed mutagenesis identified several residues that were important for binding of SarA to DNA. Among these, the SarA mutants R84A and R90A were shown to completely lack DNA binding ability (30). To verify the importance of DNA binding ability of SarA in trxB repression, we analyzed the above-mentioned two point mutations in vivo and in vitro (Fig. 2B and C). Northern blotting analysis indicated that the levels of the trxB transcript did not decrease in the sarA mutant strains carrying either the P1sarAR84A or P1SarAR90A constructs, despite higher production of the sarA transcript (Fig. (Fig.2B).2B). The sarA mutant complemented with the wild-type P1 sarA or P1sarAK69A could effectively repress trxB transcription. It is worthwhile to remember that SarA is involved in repression of its own transcription (14). The SarA constructs lacking DNA binding activity failed to repress transcription at the sarA promoter, and consequently severalfold increases in sarAR84A and sarAR90A transcript levels were observed compared to wild-type transcript (Fig. (Fig.2B).2B). Gel shift analysis showed the inability of purified SarA R84A and SarA R90A (even at higher protein concentrations) to form protein-DNA complexes with the trxB promoter DNA, suggesting that the DNA binding of SarA to the trxB promoter region is specific.
The consensus DNA binding site for SarA has been determined by DNase I footprinting of several SarA regulatory promoter regions (i.e., agr, sarV, hla, spa, bap, fnb) (13, 15, 33, 35, 46). Therefore, bioinformatics analysis of the 258-bp trxB promoter region was performed to identify the presence of the consensus SarA binding site(s). Analysis showed the presence of three putative SarA binding sites within 258 bp of the trxB promoter region (see Fig. S1A in the supplemental material). Two SarA binding sites were confined within a 25-bp region of DNA (nucleotide positions −133 to −158 ) on both strands of DNA, and the third SarA binding site was located within the mapped “−35” to “−10” region (nucleotide positions −9 to −30) of the trxB promoter (52). Taken together, all of these results strongly suggest that SarA binds to its cognate trxB promoter region to act as a repressor of trxB gene expression. It is useful to recall that the putative PerR binding site was identified upstream of the coding sequences between the ribosome binding site and the “−10” promoter region of the trxB gene (+8 to −8) (see Fig. S1A) (24). It is also worthwhile to remember that the putative PerR binding site (5′-atTAtAATTATTATaAt-3′; lowercase letters show nonconserved nucleotides) was identified based on a homology search using the consensus PerR binding sequences from Bacillus subtilis (8). Except for the mrgA (ferritin-like Dps) gene, where the PerR binding site was found in the “−35” and “−10” promoter region, the PerR binding sites of katA, ahpC, perR, and fur genes were located between the “−10” promoter sequence and the initiation codon of the gene (24). Therefore, we speculate that PerR blocks RNA polymerase and transcription of the perR-regulated genes. The possibility and significance of any cross talk between these two regulators (i.e., SarA and PerR) in regulation of expression of the trxB gene are unknown and will be investigated in the future by analyzing the sarA perR double mutant and by studying possible protein-protein interactions under various oxidative stress conditions.
To determine the effects of different oxidative stress-inducing agents on trxB transcription, Northern analysis was performed with total RNA isolated from the wild type and isogenic sarA mutant exposed to different known oxidative agents: the small and nonpolar H2O2; simple organic hydroperoxide, t-butyl hydroperoxide (t-BOOH); a moderately complex hydroperoxide, cumene hydroperoxide (CuOOH); an internal generator of superoxide stress, methyl viologen (MV); the thiol-oxidizing agent, diamide. During the early phase of aerobic growth (OD, 0.7), a substantial increase in expression of the trxB transcript in response to t-BOOH, MV, and diamide was seen in both the wild-type and the sarA mutant strains (data not shown). The enhanced levels of trxB transcripts returned to basal levels after 1 hour in the wild type as well as the sarA mutant, indicating the effect on transcription was transient, which is consistent with the results obtained in a previous study (52). A different type of transcription pattern for trxB was observed when strains were grown to the exponential or postexponential phase under microaerophilic conditions. At the end of 1 hour of exposure, clear induction of the trxB transcript was seen only in the presence of diamide in the wild type as well as in the sarA mutant and to a lesser extent in the presence of the oxidative stress agent H2O2 (Fig. (Fig.3A).3A). On longer exposure (16 h), only a basal level of trxB expression was observed for all oxidative stress-inducing compounds in both types of strains analyzed (data not shown). Under aerobic conditions, the level of induction of trxB transcription was similar in the sarA mutant and the wild-type strains after a short exposure to diamide, while under microaerophilic conditions, 2- to 3-fold-higher levels of trxB were observed in the sarA mutant compared to the wild-type strain (Fig. (Fig.3A).3A). Taken together, the results suggest that among the oxidative agents, the thiol-oxidizing agent diamide has a significant effect on trxB transcription.
To study further the effect of diamide on SarA function, the N-terminal cysteine residue was mutagenized and analyzed both in vitro and in vivo. The only cysteine residue present at the N-terminal of SarA (Cys-9) is conserved among several SarA family proteins, such as MgrA (Cys-12) and SarZ (Cys-13), as well as in several MarR homologues, like OhrR. The DNA binding affinities of OhrR, SarZ, and MgrA were reduced on exposure to oxidizing agents, in particular, diamide, and H2O2 (9, 10, 19). Therefore, we examined the probable role of the cysteine residue at position 9 of the SarA protein in regulation of trxB expression (Fig. (Fig.3B).3B). The construct containing the sarAC9G mutation (P1sarAC9G) was introduced into the SH1000 sarA mutant, and trxB transcription was monitored under microaerophilic conditions. P1sarAC9G could not repress trxB transcription in the sarA mutant, and the transcript levels were comparable to those found in the sarA mutant (Fig. (Fig.3B),3B), suggesting that Cys-9 of SarA plays an important role in expression of the trxB gene in vivo. In addition to these results, we analyzed the influence of different oxidizing agents on the DNA binding ability of SarA (Fig. (Fig.3C).3C). The wild-type SarA protein was incubated with H2O2 or diamide and subjected to gel shift analysis with the trxB promoter. On exposure to these agents, a partial disruption of the SarA-trxB promoter complex was observed. However, exposure of SarA to DTT (a reducing agent) regenerated the SarA-DNA complexes, similar to that observed with the control (untreated SarA). The purified SarAC9G protein bound to the trxB promoter in a manner similar to the wild-type protein, but unlike the situation with the wild-type SarA the DNA binding ability of this protein was not affected by oxidizing agents (Fig. (Fig.3C).3C). These results suggest that the cysteine residue at position 9 of SarA is important in regulation of the trxB gene. This is consistent with observations that MgrA and SarZ showed reduced binding to target promoters when treated with oxidizing agents (9, 10). Similarly, OhrRC15G protein from B. subtilis also binds to the ohrA promoter but cannot induce expression on exposure to oxidative stress (19). Although both MgrA and SarZ have been shown to possess an oxidation-sensing ability, they have not been implicated in regulation of any oxidative stress-inducing gene (e.g., sodA sodM, katA, trxB, and ahpC). As SarA regulates these genes, it may play a more distinct role in determining the oxidative stress resistance in S. aureus.
Although it is reasonable to predict that overexpression of trxB, or any other ROS scavenging genes, would enhance the ability to resist oxidative stress, this is not always so. Phenotype studies have shown the sarA mutant to be more sensitive to oxidizing agents than wild-type S. aureus (4, 43). Similarly, Escherichia coli strains overexpressing superoxide dismutase were more sensitive than the wild type to oxidative agents such as methyl viologen or ionizing radiation (48-50). Thus, despite high expression of ROS-scavenging genes, the sarA mutant shows decreased resistance to ROS-inducing agents. The reasons behind the oxidative stress sensitivity of the sarA mutant are not clear but may involve additional regulatory systems. The results presented clearly suggest that trxB transcription is mediated by both SarA-dependent and SarA-independent pathways. The increased transcription of trxB in the sarA mutant (over and above the basal levels seen in the sarA mutant) under the influence of the oxidative stress agent diamide is indicative of the SarA-independent mode of trxB regulation. The oxidizing agents diamide and H2O2 may modify SarA at Cys-9, which would result in a reduced affinity to bind to the upstream promoter region of the trxB gene. This would consequently diminish the SarA-mediated repression (Fig. 3B and C). The other known SarA-independent regulator, PerR, would also be modified by the oxidizing agents, leading to a further induction of trxB transcription over and above that seen due to loss of SarA-dependent repression alone. We are not ruling out the possible coordination between SarA and other oxidative stress regulators, in particular PerR, in regulation of the trxB gene. These regulatory interactions merit detailed analyses and will be investigated in the future.
The exact mechanism of target gene regulation by SarA is largely unknown, although the crystal structure of SarA has been determined and various residues have been analyzed for their role in DNA binding function (30). We believe that the presence and the organization of the SarA binding motifs or functional sites may determine the ability of SarA to bind to the respective promoters and regulate their expression. An earlier study showed that mutation of a few residues of SarA (Y18A, E29A, and K54A) decreased the affinity of these mutant proteins to bind target promoter DNA in vitro (30). Recent results from our laboratory have indicated that SarA regulates expression of the sodM gene and binds to the sodM promoter region with high affinity (4). To compare the abilities of various mutant forms of SarA proteins to repress trxB or sodM expression, transcription levels of these genes were assessed in sarA mutants carrying various mutations of the sarA gene in a shuttle vector, pSK236, under microaerophilic conditions (Fig. (Fig.4).4). The in vivo expression of the various mutant forms of the sarA gene was confirmed by Northern and Western analyses (data not shown), which is consistent with findings obtained in an earlier study (30). Transcription of trxB or sodM was severely reduced upon introduction of the P1sarA wild-type construct into the sarA mutant (Fig. (Fig.4A).4A). The P1sarAY18A and P1sarAE29A mutant gene constructs were not as efficient as P1sarA in repressing trxB or sodM transcription. However, both mutant constructs could repress sodM transcription (13-fold) more effectively than trxB transcription (3.4- to 3.8-fold) in the sarA mutant. Surprisingly, in the case of the P1sarAK54A construct, transcription of sodM was reduced by about 10-fold compared to the sarA mutant, while expression of the trxB transcript remained almost unaffected (only 10% reduction). The other constructs analyzed (i.e., P1 sarAK27A, -K69A, and -Q100A) could repress transcription of the trxB or sodM gene in a manner similar to that of the P1sarA wild-type construct in the sarA mutant (Fig. (Fig.4A).4A). These results indicate differential abilities of the mutant SarA proteins to regulate sodM and trxB transcription. In general, the SarA protein appears to repress the sodM promoter more effectively than the trxB promoter. The capacity of the above-mentioned constructs to influence trxB or sodM transcription in vivo in response to oxidative stress was also monitored after exposure to diamide for 1 h (Fig. (Fig.4B).4B). The results suggested that the SarAK54A, SarAY18A, and SarAE29A proteins are able to repress sodM transcription but not trxB transcription like the wild-type SarA. Taken together, all of these results clearly indicate that SarA has the ability to repress transcription of target genes to different extents.
To correlate the in vivo transcription results with the ability of the mutant SarA proteins to bind promoter DNA, DNA binding assays were performed with the promoter regions of the sodM or trxB gene and the purified SarA mutant proteins. As shown in Fig. Fig.5,5, most of the labeled sodM promoter fragment was shifted with 400 ng of the wild-type SarA protein, while only 50% of the trxB promoter DNA was shifted under the same conditions. The binding constants were calculated, and the wild-type SarA protein was observed to bind to the sodM promoter (KD, 5.3 ± 0.5 nM) (mean ± standard error of the mean) with about 3-fold-higher affinity than to the trxB promoter (KD, 16.5 ± 1 nM). The SarAY18A and SarAE29A proteins showed reduced binding to both of these promoter fragments. But the binding affinities of these two mutant proteins for the sodM promoter (KD, 12 ± 0.8 nM for SarA Y18A and 13.5 ± 1 nM for SarAE29A) was higher than the trxB promoter (KD, 37 ± 2.0 nM for SarA Y18A and 45 ± 2.5 nM for SarAE29A). Interestingly, at higher protein concentrations, SarAK54A could bind to the sodM promoter region (KD, 19 ± 1.0 nM), but the same protein failed to bind to the trxB promoter region. Thus, the sarA K54A residue showed a striking difference in regulation of the sodM and trxB genes in vivo, i.e., it failed to repress trxB transcription, whereas it was very effective in repressing sodM transcription in the sarA mutant. Analysis of the consensus SarA binding sequences present near the promoter region of the sodM gene indicates that four SarA binding sites are organized within a 47-bp region on both strands of DNA upstream of the sodM promoter (see Fig. S1B in the supplemental material), whereas two SarA binding sites are confined within a 25-bp region on both strands of DNA 90 bp upstream of the trxB promoter region (see Fig. S1A in the supplemental material). Thus, differential organization or the presence of multiple SarA binding sites could lead to a 3-fold increase in the DNA binding affinity of SarA to the sodM promoter compared to that to the trxB promoter. Even a 2-fold difference in affinity of a regulator protein for two different promoters can lead to considerable differences in the levels of transcription at those promoters. For example, the multidrug recognition repressor (TtgV) from Pseudomonas putida binds to the ttgD gene promoter with 2-fold-higher affinity than the ttgG gene promoter. In vivo, in P. putida, the ttgD is silent while ttgG is expressed at a high basal level (17). Therefore, these observations for the first time suggest that the ability of SarA to regulate a particular target promoter depends on the binding affinity of SarA to that promoter. The binding affinity is probably dependent on the presence or organization of the SarA binding motifs and also the lysine (K) residue at position 54 of SarA, which is probably involved in multi-SarA dimer formation. It has been shown that expression of the regulatory genes under in vivo conditions can be very dissimilar from that observed in vitro (20). Although SarA is uniformly expressed in S. aureus during growth under laboratory culture conditions (5), it is possible that SarA expression may change during different stages of clinical infection. Conceivably, this could lead to differential regulation of target genes in S. aureus.
In conclusion, this study has demonstrated the role of SarA in transcriptional repression of the trxB gene. To further define the precise role of SarA in global regulation of the oxidative stress-responsive genes, we are in the process of analyzing more oxidative stress-responsive genes and analyzing the probable interactions with other oxidative stress regulators, such as PerR. Oxidative stress tolerance is an important component of the S. aureus survival mechanism, and a detailed understanding of this process will augment the knowledge of S. aureus physiology and pathogenesis.
We thank Michael S. Chaussee for critical reading and comments on the manuscript.
This work was supported by the 2010 Initiative Start-Up Fund, Sanford School of Medicine Research Fund, and grant R21 AI077671 from the National Institutes of Health.
Published ahead of print on 23 October 2009.
†Supplemental material for this article may be found at http://jb.asm.org/.