Regulation of Superoxide Dismutase (sod) Genes by SarA in Staphylococcus aureus

The scavenging of reactive oxygen species (ROS) within cellsis regulated by several interacting factors, including transcriptionalregulators. Involvement of sarA family genes in the regulationof proteins involved in the scavenging of ROS is largely unknown.In this report, we show that under aerobic conditions, the levelsof sodM and sodA transcription, in particular the sodM transcript,are markedly enhanced in the sarA mutant among the tested sarAfamily mutants. Increased levels of sod expression returnedto near the parental level in a single-copy sarA complementedstrain. Under microaerophilc conditions, transcription of bothsodM and sodA was considerably enhanced in the sarA mutant comparedto the wild-type strain. Various genotypic, phenotypic, andDNA binding studies confirmed the involvement of SarA in theregulation of sod transcripts in different strains of Staphylococcusaureus. The sodA mutant was sensitive to an oxidative stress-inducingagent, methyl viologen, but the sarA sodA double mutant wasmore resistant to the same stressor than the single sodA mutant.These results suggest that overexpression of SodM, which occursin the sarA background, can rescue the methyl viologen-sensitivephenotype observed in the absence of the sodA gene. Analysiswith various oxidative stress-inducing agents indicates thatSarA may play a greater role in modulating oxidative stressresistance in S. aureus. This is the first report that demonstratesthe direct involvement of a regulatory protein (SarA) in controlof sod expression in S. aureus.

INTRODUCTION

Top

INTRODUCTION

Staphylococcus aureus is a major human pathogen that has colonizedmore than one-third of the world's population. It causes a varietyof infections in immunocompromised hosts as well as healthyindividuals. Most S. aureus infections begin as minor colonizationof soft tissue or skin, from which the organism can spread tothe bloodstream and subsequently disseminate into various tissues.Once inside tissues, S. aureus produces a large number of factors,including adhesins, enzymes, toxins, capsular polysaccharides,etc., which facilitate tissue colonization, tissue destruction,and immune evasion. The expression of many of these determinantsis coordinately controlled by regulatory systems, such as agr,saeRS, vraRS, tcaRA, sarA, and nine other sarA paralogs (4,6-10, 17, 24, 27, 30, 32-36, 39, 42, 52).

The sarA locus, an important transcriptional regulatory system,comprises a major open reading frame, sarA, driven by threedistinct promoters, resulting in three overlapping transcriptswith a common 3' end (1). Inactivation of the sarA locus upregulatesfibronectin and fibrinogen binding protein synthesis, hemolysins( ${alpha}$ -, β-, and ${delta}$ -hemolysins), enterotoxins, toxic shock syndrometoxin 1 toxin 1, genes involved in biofilm formation (e.g.,icaRA and bap), and capsule and downregulates proteases, proteinA, and collagen binding protein synthesis (3, 6, 10, 53). TheSarA protein has also been shown to bind to several regulatoryloci (e.g., agr, sarS, rot, sarV, and sarT) to modulate targetgene transcription directly and indirectly (8, 9, 34, 37, 45,46). Transcriptional profiling studies showed that mutationof the sarA gene led to altered expression of roughly 120 genes(76 upregulated and 44 downregulated) (13). Using a combinationof different techniques and genome sequence information, anadditional nine SarA paralogs (SarR to -V, SarX, SarZ, Rot,and MgrA) have been identified. These SarA paralogs have beenshown to regulate expression of target genes, including thoseinvolved in virulence, regulation, biofilm formation, autolysis,antibiotic resistance, and metabolic processes (3, 8, 13, 24,30, 32-37, 39, 42, 46, 52).

Reactive oxygen species (ROS), such as superoxide anion (O₂^–),hydrogen peroxide (H₂O₂), and hydroxyl radicals (OH), are generatedas the by-products of cellular metabolism and have been implicatedin different human diseases. ROS are also important for normalcellular processes such as apoptosis, cell proliferation anddifferentiation, and transcriptional regulation (41, 43). Theyare also involved in bacterial and viral killing by neutrophilsand other phagocyte cells (2, 40). In bacteria, during aerobicgrowth, ROS arise as intracellular natural products from incompletereduction of oxygen during respiration and are highly toxicto nucleic acids, proteins, and cell membrane fatty acids (23,38). In most cases, the harmful ROS are sensed directly by transcriptionalregulators, which upon oxidation increase expression of genesinvolved in the detoxification of ROS (19, 41). In S. aureusfour main iron-dependent global regulators (Fur, PerR, Zur,and MntR) are known to be involved in resistance to oxidativestress (20, 22, 28, 57). Among these, PerR predominantly protectsS. aureus cells against oxidative stress challenge mediatedby H₂O₂. The genes katA (catalase), ahpCF (alkylhydroperoxidereductase), mrgA (Dps-like protein), bcp (bacterioferitin coregulatoryprotein), and trxB (thioredoxin reductase) are members of thePerR regulons that are activated due to perR deletion (12, 20).Another set of enzymes, superoxide dismutases (SODs), are responsiblefor converting superoxides to hydrogen peroxide and oxygen.Thereafter, catalase converts hydrogen peroxide to water andoxygen.

In several bacterial species, SOD has been shown to be importantfor defense against killing by the phagocytes of vertebratehosts. Inactivation of the sod genes in Shigella flexneri andEscherichia coli K-12 results in enhanced sensitivity to killingby serum and neutrophils (15, 38). The sod gene inactivationin Streptococcus pneumoniae, Campylobacter coli, and Haemophilusinfluenzae results in attenuation of virulence, reduced colonizationof the chicken stomach, decreased survival in the spleen orliver of mice, and the inability to colonize the rat nasopharynx(56). In S. aureus, there are conflicting reports regardingthe roles of SODs in virulence and staphylococcal disease. Anearlier study suggested no correlation between SOD activityand lethality in a mouse model of infection (31). An isogenicsodA or sodM mutant of strain 8325-4 showed no effect on virulencein a mouse abscess model (11, 47); however, SOD activity wasfound to be significantly higher in S. aureus strains isolatedfrom patients with staphylococcal disease (25). It has alsobeen reported that isogenic sodA, sodM, and sodA sodM mutantsof SH1000, an rsbU-complemented derivative of 8325-4, had reducedvirulence compared to the wild-type strain in a mouse abscessmodel of infection (26). Therefore, these results suggest thatenzymatic superoxide anion scavenging by SODs is important forthe survival of the pathogen.

S. aureus has two SOD genes, sodA and sodM (11, 55). In Staphylococcus,SodA is the major SOD enzyme, and in fact, coagulase-negativestaphylococci such as Staphylococcus epidermidis lack the sodMgene, suggesting that SodM may have a unique role in S. aureus(56). The regulation of the sod genes in S. aureus is not wellunderstood. Therefore, we explored the possible regulatory roleplayed by staphylococcus-specific sarA family genes in regulationof the sod genes in S. aureus. In this study, we report thatamong the SarA family, SarA is involved in negative regulationof both sodM and sodA genes. SarA specifically binds to thesodA and sodM promoter regions, thus suggesting a direct modeof regulation. Analysis of various single and double sod mutantswith sarA in the presence or absence of different oxidativestress-responsive agents indicates that SarA may play an importantrole in modulating oxidative stress resistance in S. aureus.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Bacterial strains, growth conditions, and oxidative stress treatment. The bacterial strains and plasmids used in this study are describedbelow. Phage ${phi}$ 11 was used as a generalized transducing phagefor S. aureus strains. S. aureus strain RN4220, a restriction-deficientderivative of strain 8325-4, was used as the initial recipientfor the transformation of plasmid constructs. RN6390 (three-prophage-curedderivative from the human isolate NCTC 8325), its isogenic sarAmutant (7), Newman (a human clinical isolate), SH1000 (a single-copyrsbU gene-complemented derivative of 8325-4) (21), and MW2 (ahighly virulent community-associated methicillin-resistant S.aureus human isolate) were used. The S. aureus cells were grownat 37°C with continuous aeration in tryptic soy broth (TSB)or on tryptic soy agar supplemented with antibiotics when necessary(5 µg/ml of erythromycin, 3 µg/ml of tetracycline,and 10 µg/ml of chloramphenicol). The wild-type S. aureusSH1000 and its isogenic mutant strains were tested for susceptibilityto various oxidative stress agents. Overnight cultures grownin TSB were inoculated in 10 ml TSB medium, with or without1 mM methyl viologen (MV), at an initial optical density at600 nm (OD₆₀₀) of $~$ 0.05. The cultures were incubated and grownwith aeration in a shaker at 37°C in 18-mm diameter borosilicateglass tubes. Growth was monitored by measuring the change inturbidity spectrophotometrically (Spectronic 20D+). In someof the cases, the S. aureus strains were allowed to grow toan OD of $~$ 1.1 and the specific compound (1.0 mM MV, 2 mM H₂O₂,0.5 mM tert-butylhydroperoxide, 0.5 mM cumene hydroperoxide,and 2 mM diamide) was subsequently added. The sensitivity ofS. aureus to MV was determined by disc diffusion assays as describedpreviously (26). For microaerophilic growth, the S. aureus cultureswere inoculated in glass tubes as described above and incubatedat 37°C in a candle extinction jar for 16 h. All the chemicalsused in this study were reagent grade and obtained from SigmaChemicals and Fisher Scientific.

Genetic manipulation of E. coli and S. aureus. S. aureus sarA mutants were constructed by transducing the ${phi}$ 11lysate of RN6390 sarA::ermC (7) or 8325-4 sarA::kan (21) intodifferent S. aureus strains, SH1000, Newman, and MW2. The phage ${phi}$ 11 lysate of RN6390 sodA::tet or sodM::ermC (55) was infectedinto SH1000 wild-type and sarA mutant strains to obtain singlesodA and double sarA sodA mutant strains. A ${phi}$ 11 lysate of strainALC812 (7), which contains a single-copy integration of the1.5-kb sarB region of the sarA locus (1) into the lipase (geh)locus, was prepared and transduced into the SH1000 sarA mutantto obtain a single-copy sarA-complemented strain. The 650-bpcontaining P1 sarA region of the sarA locus was cloned intoshuttle vector pSK236 for overexpression of the sarA gene intoa sarA mutant background (29). This recombinant plasmid wasfirst electroporated into S. aureus strain RN4220 and selectedfor chloramphenicol-resistant colonies on tryptic soy agar plates.The plasmid isolated from the RN4220 background was electroporatedinto the wild-type SH1000 and its isogenic sarA mutant. Theauthenticity of complemented strains was validated by Northernhybridization.

Isolation of total cellular RNA and Northern blot hybridization. Total cellular RNA from S. aureus cells was isolated using theTrizol reagent as previously described (32, 33). The concentrationof total RNA was determined by measuring the absorbance at 260nm. Ten micrograms of each RNA sample was resolved on formaldehyde-agarosegels and capillary blotted onto Hybond XL membrane (Amersham).The DNA fragments containing open reading frames of the sodM,sodA, and sarA genes were amplified by PCR or excised from plasmidscontaining the desired gene by employing appropriate restrictionenzymes. For detecting the specific transcripts, the DNA probeswere labeled with [ ${alpha}$ -³²P]dCTP using the random-primed DNA labelingkit (Roche Diagnostics GmbH). Hybridization was performed overnightunder aqueous conditions at 64°C. After hybridization, theblots were washed and autoradiographed. The intensities of differenttranscripts were quantified with the ImageQuant software (Amersham).

DNA binding assays. Gel shift or DNA binding assays were performed to determinethe DNA-protein interaction of purified SarA with the sod promoters.A 253-bp DNA fragment and a 273-bp DNA fragment containing upstreampromoter regions of the sodA and sodM genes, respectively, wereamplified by PCR with suitable primers, and the PCR productswere cloned into the pCR2.1 vector (Invitrogen, CA). The clonedpromoter fragments were excised by restriction digesting theindividual plasmids with restriction endonuclease EcoRI. Thedephosphorylated DNA fragments were end labeled with [ ${gamma}$ -³²P]ATPusing T4 polynucleotide kinase, followed by purification througha Sephadex column (G-25) to remove free nucleotides. The gelshift assays with purified SarA protein were performed as previouslydescribed (32, 33). The cloning and purification of His₆-SarAwere described earlier (9). The SarA protein concentration wasmeasured with a Bio-Rad protein estimation kit, using bovineserum albumin as the standard.

Zymographic analysis. The SodM and SodA activities were determined by negative stainingon native polyacrylamide gels (56). Various S. aureus strainsgrown to different phases of growth in TSB were harvested bycentrifugation and washed once with lysis buffer (25 mM Tris[pH 8.0] and 1 mM EDTA). The cell pellets were resuspended inthe lysis buffer, transferred into 2.0-ml screw cap tubes containingacid-washed 0.1-mm zirconia/silica beads (Biospec), and subjectedto bead beating in a FastPrep FP 120 homogenizer. The cell debriswas separated from the soluble fraction by centrifugation (15min, 14,000 rpm) at 4°C. The concentration of total proteinsfrom clear lysates was determined by using the Bio-Rad proteinestimation kit using bovine serum albumin as the standard (Bio-Rad,Hercules, CA). Equal amounts of protein (20 µg) were resolvedon 12% polyacrylamide gels by electrophoresis in glycine bufferwithout sodium dodecyl sulfate. After the run, the gel was soakedin 1.225 mM of nitroblue tetrazolium solution for 45 min andthen washed with sterile distilled water. Subsequently, thegel was soaked in a solution containing 0.028 mM riboflavinand 28 mM TEMED (tetramethylethylene diamine) for another 45min. The gel was exposed to light on a light box to initiatethe photochemical reaction. The SOD activity was monitored asa clear zone surrounded by a dark blue background.

Western blotting and immunodetection. Whole-cell extracts from S. aureus SH1000 wild-type cells subjectedto 1 mM MV treatment were prepared as described above. Equalamounts of total cellular proteins were separated on a 12% polyacrylamide-sodiumdodecyl sulfate gel and transferred onto nitrocellulose membraneas described earlier (32). The blot was incubated at room temperaturewith a 1:10,000 dilution of anti-SarA polyclonal antibodiesfor 1 h, followed by another hour of incubation with a 1:10,000dilution of anti-mouse-horseradish peroxidase conjugate (Pierce,IL). Immunoreactive bands were detected as described in theECL detection kit (Piece, IL).

RESULTS

Top

RESULTS

Expression of the sodM and sodA genes in various sar family mutant strains of S. aureus. To analyze expression of the sodM gene due to inactivation ofvarious sarA family genes, Northern blot analysis was performedwith an open reading frame gene probe of sodM and total RNAisolated from different regulatory mutants (10 sarA family genes,agr, sigB, and sae) of S. aureus strain RN6390. Results fromthe various regulatory mutants showed that the level of thesodM transcript was considerably enhanced ( $~$ 7-fold) only in thesarA mutant (see Fig. S1 in the supplemental material). Therewas a slight decrease ( $~$ 25%) in sodM expression due to sarR inactivation,but there was no change in the level of sodM transcription ina sarA sarR double mutant compared to sarA single mutant. Expressionof sodA was also enhanced only in the sarA mutant (data notshown). Hence, sarA-mediated sod gene regulation was furtherinvestigated. RN6390 is a derivative strain of the 8325 lineagethat produces less of the alternative sigma factor SigB dueto a natural deletion of 11 bp in the putative regulatory rsbUgene (16). To circumvent possible problems that may arise dueto the rsbU mutation, we have used the rsbU⁺ strain SH1000 (8325-4carrying an intact rsbU gene) for further studies (21). Theresults showed that the level of sodM transcript in differentphases of growth was significantly increased ( $~$ 7- to 8-fold)in the sarA mutant strain compared to the SH1000 parental strain(Fig. 1A). To verify that the sarA mutation had the same effecton sodM expression in different S. aureus clinical isolates,sodM transcription was analyzed in two clinical strains (MW-2and Newman) in which the sarA gene was deleted. As shown inFig. 1A, an enhanced level of the sodM transcript was observedin all sarA-deleted clinical isolates. These results clearlysuggested that sarA repressed sodM expression in S. aureus.In SH1000, sodA expression in wild-type and sarA mutant strainswas similar in early exponential phase (OD₆₀₀ of $~$ 0.7), but inthe exponential (OD₆₀₀ of $~$ 1.1) and postexponential (OD₆₀₀ of $~$ 1.7) phases of growth the sarA mutant showed about 2- and 1.5-fold-elevatedsodA expression, respectively, compared to the wild-type cells(Fig. 1B). S. aureus strains RN6390 and MW2 also showed highersodA expression than the corresponding wild-type strains inpost exponential phase of growth (Fig. 1B). To correlate theeffects of sarA mutation with SOD activity, zymograms with cellextracts of different wild-type and sarA mutant strains wereperformed. The results showed three distinct bands of SOD activity(26, 55), which corresponded to SodA activity (lowermost band),SodM activity (uppermost band) and the intermediate SodA/SodMhybrid activity (Fig. 1C). In S. aureus, SodA activity, themajor SOD activity, did not appear to change significantly inthe sarA mutant in exponential phase of growth, but the middlezone of hybrid SodA/SodM activity was increased substantially(approximately 2.5 to 3.5-fold) in all sarA mutant strains.In postexponential phase samples, the activity of SodA was higherin the SH1000 sarA mutant compared with the wild-type cells(see next section). Interestingly, the zone corresponding toSodM activity was clearly observed only in the sarA mutant andnot in the wild-type strain.

View larger version (57K):
[in this window]
[in a new window]

FIG. 1. Transcription of the sodA and sodM genes in the wild type (Wt) and its isogenic sarA mutant (A^–) in different S. aureus strains. (A and B) Northern analysis of the sodM (A) and sodA (B) transcripts in different strains (as indicated on top of each panel) at various phases of growth. The blots were probed with 600 bp and 650 bp of sodM and sodA DNA fragments containing the entire open reading frames of the sodM and sodA genes, respectively. (C) SOD activity of cell extracts from different S. aureus strains. Extracts from cells grown to the exponential growth phase were separated on a 12% (wt/vol) nondenaturing polyacrylamide gel and stained for SOD activity. Each lane contains approximately 20 µg of protein. Three bands corresponding to SodA, SodM, and a hybrid SodA/SodM activity (54) are indicated. The gel was scanned, and the inverse image generated is shown. (D) Northern analysis of sodM and sodA transcripts in the SH1000 wild-type and sarA mutant strains grown under microaerophilic conditions. The two bands of sodA transcripts are due to transcription from two closely spaced promoters (26). The region of 23S and 16S rRNA of the ethidium bromide-stained gel used for blotting is shown as a loading control in panels A, B, and D. The intensity of different transcripts was quantified with the ImageQuant software (Amersham).

Conditions of high-pressure liquid chromatography aeration areknown to enhance sod expression in S. aureus (54). To furtherelucidate the involvement of O₂ concentration in sod gene expression,the wild type and the sarA mutant of strain SH1000 were grownin microaerophilic conditions. Compared to growth in aerobicconditions, the sodA expression was considerably lower in thewild-type strain grown in microaerophilic conditions, whereasthe sodA expression in the sarA mutant was substantially elevated(3.2-fold) compared to that in the wild type (Fig. 1D). Similarly,the sodM expression was also highly elevated in the sarA mutantcompared to the wild-type cells under the same conditions (Fig.1D). The results from transcription and zymogram analyses suggestthat inactivation of the sarA gene results in substantiallyenhanced SOD expression in S. aureus.

Functional complementation of sod gene expression by introduction of a single copy or multiple copies of the sarA gene into a sarA mutant. To further elucidate the involvement of sarA in sod regulation,complementation analysis with a single copy or multiple copiesof the sarA gene in a sarA mutant was performed. For single-copycomplementation, a functional copy of the sarB DNA fragment(a 1.5-kb fragment consisting of all three sarA promoters) (1)was introduced into the lipase (geh) locus of the SH1000 sarAmutant by ${phi}$ 11 phage transduction. For multicopy complementation,the P1 sarA gene was cloned into shuttle plasmid pSK236 (29)and introduced into the SH1000 sarA mutant by electroporation.The levels of the sodM and sodA transcripts returned to nearthe parental level in a single-copy complemented sarA mutant.Interestingly, the levels of sodM transcript were further reducedin the sarA mutant complemented with the multicopy sarA genecompared to the wild-type SH1000 strain (Fig. 2A and B). Theexpression of sarA in the complemented strains was also monitoredon Northern blots (Fig. 2C). On introduction of a single copyof the sarB fragment of the sar locus on the chromosome of sarAmutant, the content of sar transcripts returned to near theparental level, whereas severalfold higher levels of sarA transcriptwere seen in the same mutant when complemented with a sarA genein a multicopy plasmid (Fig. 2C).

View larger version (45K):
[in this window]
[in a new window]

FIG. 2. Northern analysis of the sodM, sodA, and sarA transcripts and SOD activity in the wild-type (Wt) SH1000, sarA mutant, and complemented strains. The different strains were grown to an OD₆₀₀ of $~$ 1.1 and total RNA isolated. (A to C) The blots were probed with 600-bp sodM (A), 650-bp sodA (B), and 450-bp sarA (C) DNA fragments containing their respective open reading frames. In sarAcpssarB, a 1.5-kb sarB DNA fragment of the sarA locus (1, 7) was integrated into the lipase locus (geh) of the sarA mutant. In sarAcpmsarA, the P1 sarA region was cloned into multicopy shuttle vector pSK236 (29). The sarB region of the sarA locus produces three overlapping transcripts, which are required for optimal expression of SarA (1, 7). The region of 23S and 16S rRNA of the ethidium bromide-stained gel used for blotting is shown as loading control. (D) SOD activities of intracellular cell extracts of different S. aureus strains. The intensity of different bands was quantified with the Gene Tool software from Syngene. Other details are as described for Fig. 1C.

Zymogram analysis of SOD activity showed that the SodM, SodM/SodAhybrid, and SodA zones were reduced to near the parental levelin the single-copy complemented sarA mutant compared with thenoncomplemented sarA mutant (Fig. 2D). Further, these zonesdisappeared completely in the sarA mutant or the wild-type SH1000expressing the P1 sarA gene from the multicopy vector (Fig.2D). The SodA activity was reduced (approximately 20%) in themulticopy-supplemented sarA mutant and the wild-type straincompared to their corresponding parent strains. It is to benoted that the activity of SodA (in postexponential phase) washigher ( $~$ 30%) in the SH1000 sarA mutant than in the wild type,which is consistent with the results from transcriptional analysispresented above. Thus, Northern blots as well as zymographicanalysis suggest that the expression of sod genes is repressedby the sarA gene product in S. aureus.

Binding of SarA protein to the sodM and sodA promoter regions. As the level of sod transcripts increased in a sarA mutant andwas restored to the wild-type level in a single-copy complementedstrain, it was hypothesized that SarA may be directly involvedin binding to the sod promoter regions. The 253-bp and 273-bpDNA fragments containing sodM and sodA upstream promoter regions,respectively, were used for DNA binding assays to test thishypothesis. The sodM and sodA promoter DNA fragments were endlabeled and employed for gel shift assays with various amountsof purified SarA protein (Fig. 3A). The DNA-protein complexescould be clearly observed for both the sodM and the sodA promoterfragments. Unlabeled promoter fragments of the respective genescould compete with the labeled fragments, whereas unlabelednonspecific DNA fragments (e.g., a 185-bp DNA fragment fromthe internal region of the sarX open reading frame) could not(Fig. 3A), indicating that the DNA-protein interactions werespecific. To further demonstrate the importance of DNA bindingability of SarA in regulation of sod genes, a mutant SarA protein(SarAR90A) that is unable to bind to DNA (29) was also analyzed.The SarAR90A protein could not bind to the sodM or sodA promoterfragments in vitro (Fig. 3A, lane 9 in both panels). Interestingly,introduction of the P1 sarAR90A gene construct into a sarA mutantwas unable to repress expression of both sodA and sodM transcriptsin vivo (Fig. 3B). Overall, the data represented here stronglyindicated that the DNA-protein interactions were specific.

View larger version (57K):
[in this window]
[in a new window]

FIG. 3. DNA binding activity of SarA protein. (A) Autoradiograms of 8.0% polyacrylamide gels showing binding of SarA protein to 253-bp and 273-bp promoter fragments of sodA and sodM, respectively. For the sodM promoter, lanes 1 to 3 contain 0 ng, 100 ng, and 200 ng of purified SarA protein, respectively, while lanes 4 to 8 each contain 400 ng of SarA protein. For the sodA promoter, lanes 1 to 3 contain 0 ng, 200 ng, and 400 ng of purified SarA protein, respectively, while lanes 4 to 8 each contain 600 ng of purified SarA protein. Approximately, 0.02 pM of radiolabeled DNA fragments was used in all lanes for both promoters. For competition assays, 25- and 75-fold molar excesses of nonspecific competitor DNA (a 185-bp internal fragment of the sarX gene) were added in lanes 5 and 6, respectively, whereas 25- and 75-fold molar excesses of the respective unlabeled promoter DNA were used in lanes 7 and 8 for each panel. In lanes 9, 1,000 ng of purified mutated SarAR90A protein was used. The arrows indicate the free labeled DNA or the DNA-protein complex. (B) Transcription of sodA and sodM in the SH1000 sarA mutant carrying different P1 sarA gene constructs. The strains were grown microaerophically for 16 h, and total RNA was isolated and hybridized with the sodA or sodM gene probe as indicated. Lane 1, sarA mutant; lane 2, sarA mutant complemented with wild-type P1 sarA construct; lane 3, sarA mutant complemented with P1 sarAR90A construct. The gene constructs used in lanes 2 and 3 were cloned into multicopy shuttle vector pSK236. (C) Nucleotide sequence of the 253-bp DNA fragment upstream of the sodM gene. The –10 and –35 regions of the sodM gene (26) are underlined. The transcriptional start site (+1), ribosome binding site (SD), and translational start (ATG) of the sodM gene are indicated and shown in bold. The consensus SarA binding sites are shown with arrows; solid arrows are for the coding strand of DNA, and broken arrows are for the complementary strand of DNA in the upstream region of translational start site. The consensus regions are deduced from the analysis of the depicted DNA region with the 26-bp SarA consensus binding site (9, 10) using a DNA alignment (LALIGN) program. The deduced six SarA consensus binding regions are located at the different sites on DNA, but four of them are confined within a 47-bp region (shown in bold) on both strands of DNA.

A search with the previously determined 26-bp SarA consensusbinding sequence (5'-ATTTgTATtTAATATTTataTAATTg-3'; lowercaseletters indicate variable residues) (9, 36) revealed the presenceof putative SarA DNA binding sites in both the sodM (Fig. 3C)and sodA (see Fig. S2 in the supplemental material) promoterregions. Six and four regions matching the SarA binding sitewere found within 132 bp of the sodM and sodA promoter regions,respectively (Fig. 3C). Within these regions, the number ofnucleotides that matched those with the consensus SarA bindingsite varied from 16 to 19 out of 26 nucleotides. Most of theSarA binding sites were located immediately upstream of the–35 region of the sodM promoter or within or upstreamof the –35 region of the sodA promoter, suggesting thatSarA protein can bind to the promoter regions of the sodM andsodA genes, presumably to act as a repressor of transcription.

Growth and expression of sod genes on exposure to MV, an internal source of superoxide stress. To determine the effect of internally generated superoxide stresson the expression of sod genes, Northern blot analysis of wild-typeand isogenic sarA mutant strains in the presence of MV was performed.Strains were allowed to grow to exponential phase in the presenceor absence of 1 mM MV, and the expression of sodM or sodA transcriptionwas monitored. The sarA mutant showed sevenfold-increased sodMexpression in the presence of MV compared to the wild type (Fig.4A). The wild type as well as the sarA mutant both showed similarlyhigher levels of sodA transcription in the presence of MV. Inanother experiment, MV was added to the cells at the exponentialphase of growth to determine if any variation in expressionof the sod genes occurred when MV was added at a later stageof growth. Expression of the sod transcripts was monitored overa period of 2 hours. The sodM transcript was markedly elevatedin the sarA mutant, while only a small increase after 2 h ofMV exposure could be seen in the wild-type strain. Both thewild-type and the sarA mutant cells showed a similar increasein levels of sodA transcripts on addition of MV (data not shown).

View larger version (34K):
[in this window]
[in a new window]

FIG. 4. MV-induced expression of the sodM and sodA genes. (A) Northern blots of total RNA extracted from exponential-phase cultures of S. aureus parental strain SH1000 (Wt) and its isogenic sarA mutant grown in the absence (–) or presence (+) of 1 mM MV. The blots were hybridized with 600-bp sodM and 650-bp sodA DNA fragments. (B) Northern blots of the sarA transcripts in the wild-type strain grown to the exponential (OD₆₀₀ of $~$ 1.1) phase of growth and induced with MV for 2 h. The blot was probed with a 450-bp DNA fragment containing the sarA gene. Lower panel, cell extracts from the above-mentioned MV-treated or control SH1000 cells were immunoblotted and probed with anti-SarA polyclonal antibodies at a 1:10,000 dilution. The region of 23S and 16S rRNA of the ethidium bromide stained gel used for blotting is also shown as a loading control in panels A and B. (C) Growth analysis of the wild-type SH1000 and sarA mutant strains in presence or absence of 1.0 mM MV. The different cells depicted are the wild type without MV ( ${circ}$ ), wild type with MV ( ${square}$ ), sarA mutant without MV( ${blacktriangleup}$ ), and sarA mutant with MV (x). The experiments were repeated at least three independent times. Error bars indicate standard deviations.

To determine if MV had a direct effect on expression of sarAin the SH1000 wild-type strain, total RNA was isolated fromcells grown in the presence or absence of MV and hybridizedto the sarA DNA probe. A decrease in expression of sarA transcriptwas observed in the presence of MV. The total cell extractsobtained from the above-mentioned cells were also analyzed forthe production of SarA protein by anti-SarA antibodies. A distinctreduction ( $~$ 45%) in the content of SarA protein was observedin the presence of MV (Fig. 4B), indicating that exposure toMV decreased SarA synthesis in S. aureus.

For phenotypic characterization, growth analysis of wild-typeand sarA mutant strains was performed in the presence or absenceof MV (Fig. 4C). Only a minor decrease in growth ( $~$ 10 to 20%)in the presence of MV was observed in the SH1000 wild-type strain,while the sarA mutant showed more reduced growth (30 to 35%)in the presence of MV. At the end of 6 h, the OD of the SH1000wild-type strain grown in the presence of MV was $~$ 90% of thatof the culture grown without any MV. Meanwhile, at the sametime point, the sarA mutant showed about 70% growth in the presenceof MV compared to the mutant grown without MV. In the presenceof MV (1 mM), the sarA mutant complemented with a single copyof the sarB fragment showed a growth profile very similar tothat of the wild-type strain (data not shown). This indicatedthat the slightly increased sensitivity of the sarA mutant strainto MV was due to absence of a functional sarA gene.

Since SodA is the major SOD in S. aureus (26), the responseof a sodA single mutant and a sodA sarA double mutant (whichoverexpresses SodM) to 1 mM MV was monitored. The sodA genesin SH1000 and an isogenic sarA mutant were disrupted as describedin Materials and Methods. The sodA transcript was absent inthe sodA single and the sodA sarA double mutant strains (Fig.5A). Similarly, the zones of SOD activity corresponding to SodAand the hybrid SodM/SodA were absent in these mutant strainsin a zymogram analysis (Fig. 5B). The sarA sodA double mutantshowed enhanced sodM transcript expression, which was similarto sodM expression in the sarA mutant background (Fig. 5A).The increased expression of sodM transcription correlated withthe enhanced expression of SodM activity on the zymogram gel(Fig. 5B). To analyze the phenotypic similarities and differenceswithin the single and double mutants, growth analysis was performedin the presence or absence of MV. In the presence of MV, thesodA mutant grew to a maximal OD of $~$ 1.2 (Fig. 5C), and subsequentlya reduction in the turbidity of the cells was observed. ThesarA sodA double mutant showed a steady increase in turbiditywith time and, unlike the sodA mutant, could attain an OD comparableto that of the wild-type strain at the postexponential phaseof growth. In disc diffusion assays, the sarA sodA double mutantshowed a smaller zone of inhibition with 0.25 M MV than thesodA mutant (Fig. 5D). A sodA sodM sarA triple mutant was alsoconstructed, and its response to MV was analyzed. The triplemutant was very sensitive to MV and failed to grow in the presenceof MV (data not shown). This suggests that the decreased sensitivityof the sarA sodA double mutant to MV is due to overexpressionof SodM in that background. Overall, the data presented hereindicate that the sarA sodA mutant is more resistant than thesingle sodA mutant to oxidative stress mediated by MV.

View larger version (34K):
[in this window]
[in a new window]

FIG. 5. Genotypic and phenotypic characterization of the sodA single mutant and the sarA sodA double mutant in response to MV. (A) Northern blot analysis of exponential-phase cultures of S. aureus parental strain SH1000 (Wt) and various isogenic single and double mutants (as indicated above each panel) are shown. The blots were hybridized with 600-bp and 650-bp DNA fragments containing open reading frames of the sodM and sodA genes, respectively. The region of 23S and 16S rRNA of the ethidium bromide-stained gel used for blotting is also shown as a loading control. (B) SOD activity of intracellular cell extracts of different S. aureus strains as indicated above the panel. Details of loading and conditions are as described for Fig. 1C. (C) Growth of sodA and sodA sarA mutants in the presence or absence of 1.0 mM MV. The different cells depicted are the sodA mutant without MV ( ${circ}$ ); sodA mutant with MV ( ${square}$ ), sodA sarA mutant without MV ( ${blacktriangleup}$ ), and sodA sarA mutant with MV (x). Error bars indicate standard deviations. (D) Sensitivities of the wild type and various sodA and sarA mutant strains to MV as determined by disc diffusion assays. The zone of clearance was measured as an indicator of sensitivity to MV.

Expression of the sodA and sodM genes in response to different oxidative stress-inducing chemicals and transcription of other oxidative stress-responsive genes in the wild type and the sarA mutant. To investigate the effects of different oxidative stress-inducingagents on the expression of sod genes, Northern blots analysiswas performed with total RNA isolated from the wild type andan isogenic sarA mutant in the presence or absence of variousoxidative agents. The smaller and nonpolar H₂O₂, the moderatelycomplex organic hydroperoxide cumene hydroperoxide (CuOOH),the simple organic hydroperoxide tert-butyl hydroperoxide (tBOOH),and the thiol-oxidizing agent diamide were used. The cells weregrown to the exponential phase of growth (OD₆₀₀ of $~$ 1.1) andthe different compounds added at the concentrations indicated(Fig. 6). The results indicated that exposure for 15 min tohydrogen peroxide (2 mM) and CuOOH (0.5 mM) did not have a substantialeffect on the expression of sodM and sodA transcription. Inthe presence of tBOOH, a relatively small increase in sodA transcriptionwas observed, while no difference was found in sodM transcription(Fig. 6). Interestingly, the presence of 2 mM diamide resultedin a moderate increase in sodM transcription in the wild-typeSH1000, but a substantial enhancement in expression was observedin the isogenic sarA mutant strain. The increase in sodA expressionin response to diamide was similar in both the wild-type andsarA mutant strains (Fig. 6). The enhanced expression of sodMand/or sodA in the presence of diamide or tBOOH was transient,and the expression returned to basal levels after 1 hour ofexposure (data not shown). Overall, the results suggest thatSarA plays a significant role in regulation of the sod genesin S. aureus.

View larger version (81K):
[in this window]
[in a new window]

FIG. 6. Transcriptional analysis of various target genes under different growth conditions in the wild-type (Wt) SH1000 and its isogenic sarA mutant. The different cells were grown to the exponential phase of growth (OD₆₀₀ of $~$ 1.0) and induced with various oxidative stress-inducing agents. After 15 min of exposure, total RNA was isolated and subjected to Northern analysis with sodM and sodA gene probes. The different oxidative stress agents employed are H₂O₂ (2 mM), t-BOOH (0.5 mM), CuOOH (0.5 mM), and diamide (2 mM). The region of 23S and 16S rRNA of the ethidium bromide-stained gel used for blotting is also shown as a loading control.

DISCUSSION

Top

DISCUSSION

All living organisms possess complex regulatory pathways. Thus,insight into how the behavior of a bacterial cell is controlledby its regulatory networks requires a detailed understandingof its individual components. S. aureus exhibits complex patternsof protein expression in response to various diverse environmentalconditions, particularly for proteins that are involved in pathogenesis(42). These protein expression patterns are consistent withthe opinion that a large number of complex regulatory pathwaysare coordinately involved in regulation of determinants. Onesuch regulatory pathway involves the staphylococcus-specificSarA protein family. The SarA protein family is related to thewidely conserved MarR family of transcriptional regulators inboth gram-negative and gram-positive bacteria. MarR family membersregulate genes involved in resistance to antibiotics, solvents,detergents, or oxidative stress agents and, in some pathogens,virulence (14, 49). In prokaryotes, harmful ROS are sensed directlyby transcriptional regulators such as OxyR, PerR, and OhrR,which are oxidized via cysteine or histidine residues and activatethe expression of genes involved in their detoxification (41).OhrR, the organic hydroperoxide transcriptional regulator anda member of the MarR family (41), is absent in S. aureus buthas significant homology with several members of the SarA proteinfamily such as SarZ (63%), MgrA (61%), and SarR (51%). It hasbeen shown that oxidation of a single cysteine residue of MgrAleads to the upregulation of multiple genes involved in antibioticresistance (5).

In this study, the influence of sarA family genes on regulationof the genes involved in scavenging of ROS was examined. Comparedto wild-type cells, enhanced transcription of sodM was observedin a sarA mutant, while decreased transcription was observedin a sarR mutant. SarR is known to repress sarA expression,and the sarR mutant has been shown to produce more SarA proteinthan corresponding wild-type cells (32), so it appears thatthe sarR gene product regulates sodM expression indirectly bymodulating the expression of SarA. Thus, among the SarA paralogs,only SarA appears to directly regulate sodM and also sodA inS. aureus. Both sod genes (sodA and sodM) in S. aureus are transcribedusing ${sigma}$ ^A-type promoters. It has been reported that sigB indirectlyrepresses the expression of sodM and sodA transcripts by anunknown mechanism (26). Our results have identified a new regulator,SarA, for sod gene expression. The results described in thispaper clearly demonstrate that inactivation of sarA enhancesthe expression of sodM in S. aureus irrespective of growth phase.Expression of the sodA gene was also upregulated due to sarAinactivation, but this effect was more prominent under microaerophilicconditions and in exponential to postexponential phases of growth.The results described here show that oxygen availability markedlyaffects the expression of sodA in S. aureus (Fig. 1D). Whenthe oxygen concentration is high (e.g., during aerobic growth),increased levels of superoxide radicals are generated, and henceenhanced SOD expression is required to detoxify them. ElevatedsodA and sodM expression in a sarA mutant, especially undermicroaerophilic growth conditions, suggests that SarA is theprimary repressor involved in keeping the sodA and sodM expressionrepressed when their expression is not required. Previously,uptake of manganese (Mn²⁺), a mineral superoxide scavenger,has been shown to modulate perR regulons and oxidative stressresistance in S. aureus (22). To analyze, whether Mn²⁺ was involvedin SarA-mediated regulation of sod genes, MnSO₄ (25 µM)was added to the cultures and expression of sodA and sodM wasanalyzed. No effect on expression of sodA or sodM in the wild-typeor sarA mutant strain was observed in the presence of Mn²⁺ (datanot shown). This suggests that sarA-mediated repression of sodgenes is independent of Mn²⁺.

In gel shift studies, purified SarA could bind specificallyto the sodM and the sodA promoters, which is supported by thepresence of SarA consensus binding sequences (9) in the promoterregions of the both sod genes and also with in vitro and invivo analyses of the DNA binding of the sarA mutant gene. Aclose analysis of SarA consensus binding sequences reveals thatfour binding sites are organized within a 47-bp region on bothstrands of DNA upstream of the sodM promoter (Fig. 3C), whereasthe organization of four SarA binding sites is more dispersedon the sodA promoter (see Fig. S2 in the supplemental material).A prior study had shown that a 33-bp DNA fragment containingthe 26-bp SarA binding site can bind to SarA protein efficiently(36). In fact, our unpublished observations suggest that theminimum size required for SarA binding is 17 bp, which is consistentwith the structural prediction of the length of the SarA bindingsite and the distance (50 Å) between two wing regionsof SarA (29). It should be noted that the consensus SarA bindingsequences are conflicting, and few alternate binding sites 5'-ATTTTAT-3'(51), ATTATAAAATWT-3' (45) and 5'-AGTTAAG-3' (48) have beenproposed. Interestingly, the proposed 26-bp SarA binding site(9) also shows the presence of the above-mentioned first twobinding sequences.

The increased sensitivity of the sarA mutant to MV was unusual,as higher SOD activity would be expected to correlate with increasedresistance. Previously, E. coli cells transformed with multiplecopies of the gene for FeSOD (i.e., with higher SOD levels)had been shown to be more readily killed by MV. It was suggestedthat enhanced SOD activity increased H₂O₂ formation and resultedin killing of these E. coli cells upon exposure to MV (49).The expression of katA and ahpC (both involved in H₂O₂ detoxification)in the SH1000 wild-type and sarA mutant strains in responseto MV was monitored. Similar levels of katA expression wereobserved in both strains, whereas threefold more ahpC expressionwas seen in the wild-type strain than in the sarA mutant (unpublished).Therefore, like in the high-SOD-producing E. coli cells (49),the MV-associated toxicity of the sarA mutant could be due toSOD overexpression. A previous study has suggested that sodA,but not sodM, mutants showed reduced viability in response toMV (26, 54). Our results also showed the sodA mutant to be moresensitive to MV than the wild type. Interestingly, the sarAsodA double mutant was more resistant to MV than the singlesodA mutant, whereas a triple mutant (sodA sodM sarA) was verysensitive to MV and failed to grow in the presence of MV. Thissuggests that SodM is required for resistance to MV in the sarAsodA double mutant and that its enhanced activity is able torescue the MV-sensitive phenotype associated with the sodA mutantstrain.

Analysis of other oxidative stress-inducing agents showed thatsodM expression remained unchanged in the wild type and thesarA mutant with all tested agents except diamide. Transcriptionof sodA remained unaltered with H₂O₂ and CuOOH, but increasedsubstantially with t-BOOH and diamide in both the strains (Fig.6). The induction of sodM and sodA transcripts in diamide-exposedcells was transient, and transcription of these genes reachedthe basal level by 1 hour after exposure. The transient expressionin response to diamide is not unique to sod genes in S. aureus.In other organisms, such as Aspergillus nidulans, diamide isknown to cause transient induction of certain genes such asthose for glutathione S-transferase, catalase, and superoxidedismutase (SodA) (44). Diamide is also known to induce the expressionof thioredoxin reductase (trxA and trxB) genes in S. aureus(54), and they probably are expressed transiently (unpublished).Thus, it is very likely that sod or trx gene induction helpsS. aureus during the initial adaptation to diamide-induced orthiol-reducing stress.

SodA and SodM both contribute to the survival of S. aureus inan animal model of infection (26). This underscores the importanceof sodM in S. aureus pathogenesis. The expression of regulatorygenes under in vivo conditions during infection can be verydifferent from that observed in vitro (18). Similarly, it ispossible that SarA expression may change during various stagesof infection or conditions (especially under reduced SarA production),and sod gene expression could be enhanced. This would help theS. aureus cells survive oxidative stress occurring during infectionmore efficiently. It appears that direct DNA-protein interactionplays a major role in SarA-dependent regulation of sod genes.Alternatively, it is possible that oxidative stress-inducingcompounds may modify some of the residues of SarA directly orindirectly, which would lead to altered affinity toward variouspromoters and consequently result in differential modulationof target genes in S. aureus. Taken together, the overall resultssuggest that SarA is involved in regulation of SOD transcriptionand may play an important role in regulating oxidative stressresistance in S. aureus. The oxidative stress response is animportant component of the organism's survival machinery, anda deeper understanding of its intricate mechanisms will furtherenhance the understanding of S. aureus physiology and pathogenesis.

ACKNOWLEDGMENTS

We thank Frank DeLeo and Ambrose Cheung for S. aureus strainsMW2, Newman, SH1000, RN6390, and various isogenic mutants. Wethank Keith Weaver for critical reading and comments on themanuscript.

This work was supported by the 2010-initiative start-up fundand SD BRIN (2P20 RR016479) subawarded to A.C.M.

FOOTNOTES

* Corresponding author. Mailing address: Division of Basic Biomedical Sciences, University of South Dakota School of Medicine, 414 E. Clark Street, Vermillion, SD 57069. Phone: (605) 677-6336. Fax: (605) 677-6381. E-mail: amanna{at}usd.edu

Published ahead of print on 13 march 2009.

Supplemental material for this article is available at http://jb.asm.org/.

REFERENCES

Top