Methionine Sulfoxide Reductases Protect against Oxidative Stress in Staphylococcus aureus Encountering Exogenous Oxidants and Human Neutrophils

Abstract

To establish infection successfully, Staphylococcus aureus must evade clearance by polymorphonuclear neutrophils (PMN). We studied the expression and regulation of the methionine sulfoxide reductases (Msr) that are involved in the repair of oxidized staphylococcal proteins and investigated their influence on the fate of S. aureus exposed to oxidants or PMN. We evaluated a mutant deficient in msrA1 and msrB for susceptibility to hydrogen peroxide, hypochlorous acid and PMN. The expression of msrA1 in wild-type bacteria ingested by human PMN was assessed by real-time PCR. The regulation of msr was studied by screening a library of two-component regulatory system (TCS) mutants for altered msr responses. Relative to the wild-type bacteria, bacteria deficient in Msr were more susceptible to oxidants and PMN. Upregulation of staphylococcal msrA1 occurred within the phagosomes of normal PMN and PMN deficient in NADPH oxidase activity. Furthermore, PMN granule-rich extract stimulated the upregulation of msrA1. Modulation of msrA1 within PMN was shown to be partly dependent on the VraSR TCS. Msr contributes to staphylococcal responses to oxidative attack and PMN. Our study highlights a novel interaction between the oxidative protein repair pathway and the VraSR TCS that is involved in cell wall homeostasis.

Introduction

Staphylococcus aureus is an important human pathogen that is responsible for a broad spectrum of infections, ranging from mild cellulitis to life-threatening endovascular disease. The rising prevalence of antibiotic resistance among staphylococci exacerbates the clinical challenges imposed by staphylococcal disease.

Optimal host defense relies on the capacity of polymorphonuclear neutrophils (PMN) to phagocytose and kill S. aureus with antimicrobial agents delivered to or generated in phagosomes. Especially potent among phagosomal oxidants is hypochlorous acid (HOCl), which is generated by the myeloperoxidase-hydrogen peroxide (MPO-H₂O₂)-chloride system [1]. PMN-derived oxidants readily damage bacterial proteins, lipids and DNA. Methionine is highly susceptible to oxidation, and oxidation in proteins can disrupt their normal function [2]. Methionine can be oxidized to either of two enantiomers, methionine-S-sulfoxide and methionine-R-sulfoxide, and each can be selectively repaired by a specific class of methionine sulfoxide reductase (Msr); MsrA and MsrB repair the S and R stereoisomers, respectively. The S. aureus genome harbors three msrA genes (msrA1, msrA2 and msrA3) and a single msrB [3]. The adjacent msrA1 and msrB genes are cotranscribed and encode enzymes that provide the majority of the Msr activity in S. aureus [4, 5].

To persist and subsequently establish infection, invading bacteria need to evade, resist or tolerate insults inflicted by PMN. Given the role of Msr in the defense against physiological oxidative stress [6], we speculated that S. aureus might employ these enzymes to repair and thus limit damage caused by PMN-derived oxidants, thereby providing a mechanism to survive within phagosomes.

In this study, we investigated the contribution of Msr to the fate of S. aureus exposed to oxidants and to human PMN. We also explored the regulation of msr in S. aureus ingested by PMN and the role of staphylococcal two-component regulatory systems (TCS) in regulating the msrA1 response.

Methods

Bacterial Strains and Culture Conditions

The S. aureus strains used in this study were constructed in the community-associated S. aureus USA300 LAC strain that is cured of the native plasmid pUSA03 which confers erythromycin resistance [7] unless otherwise indicated. S. aureus was cultured in tryptic soy broth (TSB; BD Biosciences) at 37°C with shaking at 200 rpm. For strains harboring empty vector or complementation plasmids, the media was supplemented with 10 μg/ml of chloramphenicol (Sigma-Aldrich; table 1). To obtain stationary phase organisms, bacteria were inoculated from glycerol stocks into TSB and grown for 16-18 h. For mid-log phase bacteria, the stationary bacteria were subcultured for approximately 2.5 h from a starting optical density at 550 nm (OD₅₅₀) of 0.05. In experiments involving PMN, 0.01% of human serum albumin (Talecris Biotherapeutics) was added to the subculture media. For experiments measuring msrA1 expression in the presence of both oxidants and vancomycin, an overnight starter culture of S. aureus grown in TSB was washed once with HEPES-buffered Hank's balanced salt solution (HBSS) before being used to innoculate minimal media (7.6 mM NH₄SO₄, 33 mM KH₂PO₄, 60 mM K₂HPO₄, 11 mM NaCl, 3 mM KCl, 0.2% glucose, 1 mM MgSO₄, 0.5 µg/ml nicotin, 0.5 µg/ml thiamine, 0.5 µg/ml panthothenate, 3 ng/ml biotin, and 25 µg/ml of each of the following amino acids: Gly, Val, Leu, Thr, Phe, Tyr, Cys, Met, Pro, Arg, and His). The bacteria were then cultured between 4 and 6 h before being treated.

Table 1.

S. aureus strains and plasmids used in this study

Strain or plasmid ID	Background strain	Plasmid or deletion	Details	Reference
Wild-type	USA300-0114 MRSA ΔpUSA03		erms	[7]
Wild-type pCM28	USA300-0114 MRSA ΔpUSA03	pCM28	erms Cmr	this study
Wild-type pEPSA5	USA300-0114 MRSA ΔpUSA03	pEPSA5	erms Cmr	this study
ΔmsrA1B	USA300-0114 MRSA ΔpUSA03		erms	this study
ΔmsrA1B pCM28	USA300-0114 MRSA ΔpUSA03	pCM28	erms	this study
ΔmsrA1B pCM28-msrA1B	USA300-0114 MRSA ΔpUSA03	pCM28-msrA1B	erms Cmr	this study
ΔvraSR pEPSA5	USA300-0114 MRSA ΔpUSA03	pEPSA5	erms Cmr	this study
ΔvraSR pEPSA5-vraSR	USA300-0114 MRSA ΔpUSA03	pEPSA5-vraSR	erms Cmr	this study
Δ nre	USA300-0114 MRSA ΔpUSA03	SAUSA300_2337-38Δ	erms	this study
Δ 0217-8	USA300-0114 MRSA ΔpUSA03	SAUSA300 0217-18Δ	erms	this study
Δ 1219-20	USA300-0114 MRSA ΔpUSA03	SAUSA300_1219-20Δ	erms	this study
Δ kdp	USA300-0114 MRSA ΔpUSA03	SAUSA300_2035-36Δ	erms	this study
Δ 2558-9	USA300-0114 MRSA ΔpUSA03	SAUSA300_2558-59Δ	erms	this study
Δ 1798-9	USA300-0114 MRSA ΔpUSA03	SAUSA300_1798-99Δ	erms	this study
Δ lyt	USA300-0114 MRSA ΔpUSA03	SAUSA300_0254-55Δ	erms	this study
Δ pho	USA300-0114 MRSA ΔpUSA03	SAUSA300_1638-39Δ	erms	this study
Δ hss	USA300-0114 MRSA ΔpUSA03	SAUSA300_2308-09Δ	erms	this study
Δ vra	USA300-0114 MRSA ΔpUSA03	SAUSA300_1865-66Δ	erms	this study
Δ arl	USA300-0114 MRSA ΔpUSA03	SAUSA300_1307-08Δ	erms	this study
Δ 1441-2	USA300-0114 MRSA ΔpUSA03	SAUSA300_1441-42Δ	erms	this study
Δ agr	USA300-0114 MRSA ΔpUSA03	Δagr:: tetM	tetr	[8]
sae:: spec	USA300-0114 MRSA ΔpUSA03	sae:: spec	specr	[9]
graS:: Tn-mariner	USA300-0114 MRSA ΔpUSA03	graS:: Tn-mariner	ermr
Pcm28		empty	ampr Cmr	[11]
pEPSA5		empty	ampr Cmr	[12]
pCM28-msrAlB		pCM28-msrAlB	ampr Cmr	this study
pEPSA5-vraSR		pEPSA5-vraSR	ampr Cmr	this study
pJB38			ampr Cmr	[10]
pJMB176		pJB38ΔmsrA1B	ampr Cmr	this study
pJB204		pJB38ΔvraSR	ampr Cmr	this study

For antibiotic sensitivity to ampicillin, erythromycin, spectinomycin and chloramphenicol:^s = sensitive;^r = resistant.

Construction of ΔmsrA1B, ΔvraSR and TCS Mutants

S. aureus strains with mutations in the graRS [7], agr [8] and saeRS [9]. TCS were available as part of laboratory collections (table 1). Other TCS chromosomal mutations, ΔmsrA1B and ΔvraSR strains were created by homologous recombination using pJB38 [10]. To create the complementing plasmid and strains, the msrA1 and msrB or vraSR genes were amplified by PCR and cloned into the multicopy plasmids pCM28 [11] or pEPSA5 [12]. All clones were passaged through RN4220 [13] and subsequently transduced into the appropriate strains using bacteriophage 80α [14]. For oligonucleotide (Integrated DNA Technologies) sequences, refer to table 2. See online supplementary information for further details (for all online suppl. material, see www.karger.com/doi/10.1159/000355915).

Table 2.

Oligonucleotides used in this study

Oligonucleotide	Sequence 5′ to 3′	Application
gapdh primer1	TACACAAGACGCACCTCACAGA	real-time PCR
gapdh primer 2	ACCTGTTGAGTTAGGGATGATGTTT	real-time PCR
msrA1 primer t	TCCAGGCATCAAGTCAGTCG	real-time PCR
msrA1 primer 2	TTGTCAACATGACCGCCACTA	real-time PCR
vraS primer t	CGGCAAGTATGATGCTATCTGCTA	real-time PCR
vraS primer 2	TAACTGCGAATCTTGAACCATTTT C	real-time PCR
msrA1_5EcoRI	GGGGAATTCGGGCA-ATTGAATCAATTAATTAAAGGTGGCAG	cloning
msrA1_up3 NheI	ACGCGTGGTACCGCTAGCGCTAGCATGAATTACCTCCTCTATCTATCTAATTAT	cloning
msrB_dwn5mluI	GCTAGCGGTACCACGCGTACGCGTGTGGAAAGTATGTTTAAAAAATTATTCGGAAAAGG	cloning
msrB_dwn3BamHI	CCCGGATCCCACATCAATCACTTTTGTTTCACCTTTAATCACAGC	cloning
msrA_veri5	GAGAAAACATTAAATTATTTGTAG	verification
msrB_veri3	CATTGTCACATCAATCAC	verification
msrA1B_comp5EcoRI	GGGGAATTCGGCAATTGAATCAATTAATTAAAGGTGGCAG	cloning
msrA1B_comp3BamHI	GGGGGATCCCTACTTATCAAAATGTGATATTAAATCGCC	cloning
1865up5EcoRI	GGGGAATTCCGGCGAAACATGACACACAAATATATATCAACGC	cloning
1865up3fuse	ACGCGTGGTACCGCTAGCGCAGCTAGCATGCTATATACTAAGATGAGCATTG	cloning
1865dwn5fuse	GCTAGCGGTACCACGCGTCGCCAATCACAATATAACATCAAATAGACACC	cloning
1865dwn3SalI 1865veri5	CCCGTCGACGCCGTAATGGAATCGTATTAAAACTACTAAGTGGAGAGCAACAGTAAGAAATAATA TCAAGCACTACC	cloning
1865veri3	GCACCCGCTGAAACATCTACTCTAATTAAG	verification
1865 comp5BamHI	CCCGGATCCTTAGGAGGATGATTATTTATGAACCACTACATTAGAACAATTGGTTCAATGCTCATC	cloning
1865 comp3SalI	CCCGTCGACGGTGTCTATTTGATGTTATATTGTGATTGGC	cloning

Neutrophil Isolation

Informed consent was obtained from each individual following a protocol approved by the institutional review board for human subjects at the University of Iowa. Peripheral blood was drawn from normal healthy volunteers or individuals with X-linked chronic granulomatous disease (CGD), and PMN were purified as previously described [15].

Feeding PMN to S. aureus and PMN Killing Assay

Mid-log phase bacteria were pelleted for 5 min and resuspended in 20 mM HEPES-buffered HBSS containing Ca²⁺ and Mg²⁺. The OD₅₅₀ of the suspension was measured and this was converted to colony-forming units (CFU)/milliliter: an OD₅₅₀ reading of 1.0 was equivalent to 3 × 10⁸ CFU/ml. Bacteria were used immediately or held on ice. Bacteria were opsonized by incubation at 37°C for 20 min in the presence of 10% pooled human serum. Opsonized bacteria were fed to PMN or PMN pretreated with 10 μM diphenylene iodonium (DPI) at the desired multiplicity of infection (MOI) and tumbled end over end at 37°C to allow phagocytosis to proceed. The PMN were then pelleted at 380 g for 5 min, and the extracellular bacteria were aspirated away. The cell pellets were resuspended to the original volume with 20 mM HEPES-buffered HBSS. These samples were processed to quantitate CFU to assess PMN-mediated killing or prepared for gene expression experiments. For PMN killing experiments, a 50-μl sample of cell suspension was removed at the 0 min (time 0). The remaining samples were tumbled at 37°C, and 50-μl aliquots were removed at specified time points. For viability testing, the 50-μl samples were diluted and incubated with 2.5 ml of pH 11 water for 5 min to lyse the PMN. These samples were vortexed and serially diluted into saline. 10 μl of each diluted sample was spotted onto tryptic soy agar plates in at least triplicate. After 16 h of growth at 37°C, the CFU were enumerated. From these, the CFU in the total PMN suspension for each sample were calculated. To assess the viability at each time point, the total calculated CFU for each strain were expressed as a percentage of the starting CFU for that strain. The starting CFU for each strain were taken as the CFU recovered from bacteria fed to DPI-treated PMN at time 0, in which there was minimal bacterial killing (data not shown). For gene expression experiments, the samples were tumbled at 37°C and at the desired time point, and the PMN that had ingested bacteria were pelleted, lysed and processed for real-time PCR analysis.

Real-Time PCR

RNA was extracted from bacteria or PMN-ingested bacteria as described previously [11]. cDNA was synthesized with AMV Reverse Transcriptase (Roche) and used as a template for real-time PCR. Reactions were conducted in PerfeCTa SYBR Green fast mix ROX (Quanta Biosciences) with each primer at 200 nM. Expression of gapdh was used to normalize all the gene expression data as described previously [11]. The gapdh primers amplify S. aureus gapdh but do not amplify human gapdh (data not shown). After denaturation at 95°C for 30 s, the samples were cycled at 95°C for 15 s, then at 60°C for 30 s, for 40 cycles using the ABI 7000 system. For primer sequences, refer to table 2.

Bacterial Exposure to Oxidants

In experiments used to determine the susceptibility of bacteria to oxidants, S. aureus strains in the mid-log phase were resuspended in PBS to an OD₅₅₀ of 0.250. Bacteria (0.5 ml volume) were either left untreated or treated with reagent grade H₂O₂ or HOCl (Sigma-Aldrich) at predetermined concentrations for 1 h at 37°C. Residual H₂O₂ was quenched by diluting samples 1:100 in PBS supplemented with catalase (1,300 U/ml; Sigma). Where H₂O₂-MPO was employed, 75 nmol H₂O₂ was added to each 0.75-ml sample with 0, 1.5, 3.8, 7.5 or 15 pmol of recombinant MPO. In parallel, samples without S. aureus were assayed for HOCl generation (n = 3), as described in Dypbukt et al. [16]. The average amount of HOCl generated from 3 experiments was used to plot the data. After the 1-hour incubation, samples were diluted and spot-plated onto tryptic soy agar (BD Biosciences). CFU for each strain were enumerated after overnight growth at 37°C and then calculated as a percentage of non-treated bacteria for the same strain at time 0. For gene expression studies, S. aureus was resuspended to an OD₅₅₀ of 0.25 in minimal media before being challenged with H₂O₂ or HOCl at the specified concentrations in the presence or absence of 7 µg/ml of vancomycin for 1 h at 37°C.

Bacterial Exposure to PMN Granule-Rich Extract

PMN were resuspended in cold relaxation buffer (RB; 10 mM PIPES, pH 7.3, 100 mM KCl, 3 mM NaCl₂, 3.5 mM MgCl₂, 1.25 mM EGTA, 1 mM ATP) and sonicated on ice. Unbroken cells and nuclei were pelleted (200 g at 4°C for 5 min) and the supernatant was centrifuged at 10,000 g at 4°C for 20 min. The granule-enriched pellet was washed with ice-cold RB, spun at 10,000 g, resuspended in RB and lysed by freeze thawing 3 times in a methanol-dry ice bath. Granule proteins (10⁷ cell equivalents) and 1.5 × 10⁸ bacteria were incubated together in HBSS with calcium and magnesium for 30 min at 37°C.

Treatment with Cell Wall-Active Antibiotics

In experiments involving TCS mutants, S. aureus were diluted to an OD₅₅₀ of 0.3 in TSB, in the presence or absence of 7 μg/ml of vancomycin or 25 μg/ml of D-cycloserine (Sigma-Aldrich), and shaken at 200 rpm for 1 h at 37°C.

Msr Activity Assay

Bacteria were washed with assay buffer (50 mM TRIS-HCl, pH 7.5, 145 mM NaCl), pelleted and resuspended into 200 μl of assay buffer. Bacteria were lysed with 25 µg of lysostaphin (Sigma-Aldrich) at 37°C for 10 min. Lysates were sonicated on ice and then incubated with 2 mM dabsyl-methionine sulfoxide (a mixture of the S and R enantiomers; Anaspec) and 20 mM DTT in assay buffer. After 1 h at 37°C, reactions were quenched with acetonitrile. Samples were analyzed on an LC-Dabs HPLC column (Supelco) to separate dabsyl-methionine sulfoxide from dabsyl-methionine. For further details, refer to the online supplementary material.

Statistical Analysis

Statistical analyses were performed using Graphpad Prism 5 software. Data were analyzed using paired t tests or one-way ANOVA with post-tests as indicated in the figure legends. p values <0.05 were considered to be statistically significant.

Results

Validation of the msrA1 msrB Deletion Mutant

To study the role of Msr in S. aureus, we created a mutant in the wild-type background in which both the msrA1 and msrB genes were deleted (ΔmsrA1B). The growth of the ΔmsrA1B double-mutant strain was indistinguishable from that of the parent strain (online suppl. fig. 1).

To assess the impact of msrA1B deletion on Msr activity in S. aureus, we studied bacteria in the absence or presence of vancomycin, since Msr expression increases upon exposure to cell wall antibiotics such as vancomycin [17]. In the absence of vancomycin, Msr activity in the wild-type lysates was low, and in ΔmsrA1B, it was below the level of detection (fig. 1). Msr activity was restored when both msrA1 and msrB were reintroduced into the ΔmsrA1B strain via a multicopy plasmid (pCM28). Upon treatment with vancomycin, Msr activity increased in the wild-type and the complemented deletion mutant (ΔmsrA1B with pCM28-msrA1B), but not in ΔmsrA1B. Thus, the data confirm that under these conditions, MsrA1 and MsrB support the majority of Msr activity in S. aureus, which was eliminated in our msrA1B deletion mutant.

Fig. 1.

Msr activity in wild-type and msr mutants. Conversion of dabsyl-methionine sulfoxide to dabsyl-methionine by lysates from wild-type, ΔmsrA1B and complemented ΔmsrA1B mutants were measured. Bacteria were treated with buffer (non-treated) or with vancomycin for 1 h. Data represent the mean ± SEM (n = 3 experiments).

Susceptibility of Wild-Type and msr Mutants to Oxidants and PMN

Reasoning that the capacity to repair oxidant-mediated damage by Msr would provide a survival advantage, we compared the abilities of the wild-type, ΔmsrA1B and complemented mutant strains to grow after exposure to H₂O₂ for 1 h in PBS. The number of culturable bacteria present after treatment with H₂O₂ was expressed as a percentage of the initial number of bacteria present prior to treatment.

At the lowest concentration of H₂O₂ tested (25 mM), there was no significant difference in the growth of the wild-type or ΔmsrA1B mutant after exposure to the oxidant. At higher concentrations of H₂O₂ (50, 75 and 100 mM), the viability of the wild-type mutant was greater than that of ΔmsrA1B mutant (p < 0.05; fig. 2a). Viability of the complemented mutant treated with 25-75 mM of H₂O₂ was likewise better than that of the ΔmsrA1B strain, suggesting that msrA1B contributed to better survival of bacteria in the presence of H₂O₂ at these concentrations. At 100 mM H₂O₂, ΔmsrA1B and its complement were equally susceptible to killing by H₂O₂, suggesting that expression of msrA1B driven from the complementation plasmid where gene dosage is not precisely controlled may be limited.

Fig. 2.

Susceptibility of wild-type and msr mutants to oxidants and PMN. Wild-type, ΔmsrA1B and complemented ΔmsrA1B strains were grown to the mid-log phase, resuspended in PBS and treated with H₂O₂ (a), HOCl (b) or HOCl generated by the MPO-H₂O₂ system (c) at the stated concentrations for 1 h. After 1 h at 37°C, the bacteria were plated and CFU enumerated after overnight growth. Data represent CFU/ml as a percentage of non-treated control for each strain at time 0. Data are the mean ± SEM (n = 6 experiments for a and b, and n = 3 experiments for c). At each concentration of the oxidant, a paired t test was used to compare the wild-type and ΔmsrA1B strains (* p < 0.05, ** p < 0.01). d Each bacterial strain was opsonized in pooled human serum and then fed to normal or DPI-treated PMN (MOI of 5:1). Ingested bacteria were recovered at the indicated time points and CFU were enumerated. At each time point, the viability of a strain is expressed as bacteria recovered from normal PMN as a percent of the same strain recovered from DPI-treated PMN at time 0. A paired t test was used to compare the wild-type and ΔmsrA1B strains at each time point (* p < 0.05). Data represent the mean ± SEM (n = 6 experiments).

Although H₂O₂ contributes to oxidative stress imposed on bacteria within PMN phagosomes, most of the oxygen consumed by activated PMN is converted into HOCl [18, 19, 20], a more potent antimicrobial agent [21]. To determine whether Msr-deficient S. aureus was more vulnerable to damage by the physiologically relevant oxidant HOCl, we subjected S. aureus to sublethal amounts of HOCl or a HOCl-generating system (MPO-H₂O₂-chloride). The latter system generates toxic oxidants continuously and better mirrors conditions that exist within PMN phagosomes.

The number of bacteria recovered after treatment with HOCl was expressed as a percentage of the initial number of bacteria prior to treatment (fig. 2b). Wild-type bacteria were more resistant to killing by 100 µM HOCl than was ΔmsrA1B (p < 0.01; fig. 2b). In the presence of the HOCl-generating system, the wild-type and ΔmsrA1B strains also exhibited a significant difference in survival at the high end of concentrations of HOCl produced (p < 0.05; fig. 2c).

Given that Msr can influence the susceptibility of wild-type S. aureus to the antimicrobial effects of H₂O₂ and HOCl, oxidants that are generated by activated PMN, we reasoned that Msr-deficient bacteria would be more vulnerable to PMN killing. We fed opsonized wild-type, ΔmsrA1B and complemented mutants to PMN for 10 min to allow maximum ingestion. Subsequently, extracellular bacteria were removed and the fate of ingested S. aureus assessed immediately (time 0) and after 30 and 60 min (fig. 2d). Because PMN begin to kill bacteria soon after the ingestion of bacteria and during the 10 min that are allowed for phagocytosis to occur (data not shown), we measured in parallel the viability of bacteria fed to DPI-treated PMN. The pharmacological inhibition of the NADPH oxidase by DPI [22] ensured that minimal killing of bacteria was achieved by time 0, thus providing us with the total number of ingested bacteria at the start of the time course. To quantitate PMN killing of ingested S. aureus over time, we enumerated viable bacteria recovered from normal PMN at each time point relative to the number of bacteria that was recovered at time 0 from the same strain fed to DPI-treated PMN. There was a reduction in the viability of all strains over time, but viable bacteria persisted even 1 h after ingestion. The ΔmsrA1B mutant was more susceptible to PMN killing than was the wild-type mutant at 30 and 60 min (p < 0.05).

Together, these data indicate that msrA1B contributed to the recovery of bacteria, not only after exposure to oxidants in isolation but also in the more complicated context of the PMN phagosome, where oxidants and granule proteins synergize to damage ingested targets.

msrA1 Induction in S. aureus Exposed to PMN

To examine the expression of msr in PMN-ingested bacteria, we recovered S. aureus from PMN that had been fed bacteria. After 10 min of phagocytosis by PMN, extracellular bacteria were aspirated away and the time course initiated. At time 0, the level of msrA1 in the ingested bacteria was higher in organisms grown to the stationary phase relative to those grown to the mid-log phase (fig. 3a, b), with levels in both increasing further at 15 min and remaining high at 30 and 60 min. Incubation in the absence of PMN did not promote the expression of msrA1.

Fig. 3.

msrA1 expression in wild-type mutant ingested by normal and oxidase-deficient PMN. Serum-opsonized wild-type S. aureus in the mid-log (a) and stationary phase (b) were incubated alone or fed to PMN at an MOI of 1:1. At 0, 15, 30 and 60 min after phagocytosis, both ingested bacteria had recovered from PMN and bacteria incubated alone were prepared for real-time PCR. The expression of msrA1 was normalized to the expression of gapdh, and fold change in gene expression was relative to that in opsonized S. aureus prior to exposure to PMN. Data represent the mean ± SEM (n = 3 experiments). At each time point, a paired t test was used to analyze the difference between the ingested and non-ingested bacteria (* p < 0.05, ** p < 0.01). c Opsonized mid-log phase wild-type S. aureus was incubated alone or fed at an MOI of 1:1 to suspended normal PMN, DPI-treated PMN or PMN derived from 1 of 4 unrelated individuals with X-linked CGD. Data represent the mean ± SEM (n = 4 experiments). Data after 60 min were analyzed by 1-way ANOVA followed by a Tukey post-test (* p < 0.05).

Given that Msr participates in cellular oxidant defense, we sought to determine whether oxidants generated within PMN phagosomes could drive msrA1 expression in ingested S. aureus. To that end, we compared msrA1 expression in S. aureus fed to normal PMN, PMN pretreated with DPI and PMN isolated from individuals with CGD, thereby providing PMN with pharmacologic or genetic absence of a functional phagocyte NADPH oxidase as a source of reactive oxygen species [23]. We detected high levels of msrA1 expression in S. aureus ingested by DPI-treated and CGD PMN, despite the absence of PMN-generated oxidants in those PMN. The magnitude of the induction was highest in bacteria ingested by DPI-treated PMN, followed by bacteria ingested by CGD PMN at 60 min (fig. 3c).

The Role of Oxidants in the Expression of msrA1

Given that the induction of msrA1 was more marked in PMN that were unable to produce oxidants, we studied whether oxidants may have a role in dampening the msrA1 response. We treated bacteria grown in minimal media with vancomycin to induce msrA1 expression in the presence or absence of H₂O₂ or HOCl. Since HOCl is such a potent antimicrobial agent [1], much lower concentrations of HOCl (10 and 30 µM) were used relative to H₂O₂ (100 mM). At the tested concentrations, neither oxidant induced msrA1 expression when used alone, but vancomycin alone induced msrA1 more than 15-fold (fig. 4). In the presence of 10 or 30 µM of HOCl, the level of msrA1 induced by vancomycin was similar to that observed with vancomycin treatment alone. However, in the presence of 100 mM H₂O₂, the effect of vancomycin on msrA1 induction was markedly reduced (fig. 4). These data suggest that although oxidants did not directly induce msrA1, exposure to H₂O₂ but not to HOCl may limit the msrA1 response to stimuli arising from disruption of the bacterial cell wall. The observed decrease in vancomycin-induced msrA1 response in the presence of H₂O₂ did not reflect a global inhibition of transcription by oxidant treatment, as the expression of the H₂O₂-responsive gene alkyl hydroperoxide reductase was increased as expected [24] (online suppl. fig. 2).

Fig. 4.

msrA1 expression in S. aureus treated with oxidants and vancomycin (vanc). Wild-type S. aureus cultured in minimal media was treated with H₂O₂ or HOCl in the presence or absence of vancomycin (7 µg/ml) or with vancomycin alone for 1 h at 37°C. Bacteria were then assessed for msrA1 expression by real-time PCR and the data were normalized to gapdh expression. The data represent the fold change in msrA1 expression relative to the non-treated control (mean ± SEM, n = 6). A paired t test was used to compare sample conditions (* p < 0.05, ** p < 0.01).

Modulation of msrA1 by Staphylococcal TCS

We sought to determine which pathways in S. aureus participate in the detection and relay of signals that modulate the induction of msrA1. To sense and respond to the environment, prokaryotes utilize TCS [25], which typically consist of a cell surface histidine kinase as sensor and a cytoplasmic regulator to coordinate a transcriptional response. Only a few of the 16 TCS encoded in the S. aureus genome have been thoroughly characterized.

In a screen of a library of 15 nonessential TCS mutants for impaired msrA1 induction in response to vancomycin, deletion of vraSR resulted in loss of msrA1 induction under our experimental conditions (fig. 5a). To confirm the observed relationship between VraSR and Msr, we quantitated Msr activity in the vraSR deletion mutant (ΔvraSR) and its isogenic wild-type mutant after treatment with a cell wall-active antibiotic structurally unrelated to vancomycin. We chose D-cycloserine for its ability to induce msrA1 and, more importantly, because its antistaphylococcal activity, in contrast to that of vancomycin, is unchanged in the absence of a functional VraSR system [26, 27]. At a concentration of D-cycloserine that modulated Msr activity with the least impact on viability (data not shown), D-cycloserine-treated wild-type mutants showed approximately 2-fold more Msr activity than did the non-treated controls. In contrast, treatment of ΔvraSR resulted in less Msr activity relative to the treated wild-type strain (p < 0.05; fig. 5b). These data suggested that VraSR contributed to the regulation of msrA1 in response to cell wall-active antibiotics.

Fig. 5.

Regulation of msrA1 by TCS. a Wild-type and mutant strains of S. aureus that were each deficient in one TCS were grown to the mid-log phase and treated with vancomycin for 1 h (table 1). TCS mutants are labeled in the graph with the gene locus or TCS name where known. The expression of msrA1 measured by real-time PCR was normalized to the expression of gapdh. Fold change in msrA1 is expressed relative to the untreated wild-type strain. The data represent the mean of two independent experiments. b Wild-type and ΔvraSR grown to the stationary phase were diluted to an OD₅₅₀ of 0.3 and then treated with buffer or with D-cycloserine. Msr activity in bacterial lysates was measured. Data shown represent the mean ± SEM (n = 4). A paired t test was performed between the cycloserine-treated wild-type and mutant samples (* p < 0.05). c Serum-opsonized wild-type, ΔvraSR and complemented ΔvraSR grown to the mid-log phase were fed to PMN (MOI of 1:1), and msrA1 expression in the ingested bacteria was measured and normalized to gapdh expression. Fold change is expressed relative to opsonized wild-type bacteria at time 0. Data represent the mean ± SEM (n = 9) and are analyzed by 1-way ANOVA with a Tukey post-test (* p < 0.05, *** p < 0.001).

To determine whether VraSR participated in the regulation of msrA1 in PMN-ingested bacteria, we measured msrA1 expression in wild-type, ΔvraSR and ΔvraSR mutants complemented with a multicopy plasmid (ΔvraSR with pEPSA5-vraSR) after phagocytosis by PMN. Phagocytosis of S. aureus by PMN was the same for all the strains (data not shown). Both wild-type and complemented mutants significantly upregulated msrA1 in response to ingestion by approximately 11- and 15-fold, respectively. In contrast, ΔvraSR exhibited a relatively muted response, with only an approximately 7-fold induction in msrA1 (p < 0.05; fig. 5c). Similar to the trend observed in msrA1, vraS expression in S. aureus ingested by DPI-treated PMN and CGD PMN was higher than that in normal PMN (online suppl. fig. 3). Taken together, these data provide support for a link between VraSR, a TCS that can be activated by perturbations to the cell wall, and Msr, one of the oxidant repair systems. However, they also highlight the complexity of msrA1 regulation in a biological setting composed of multiple interacting antimicrobial toxins, as occurs in PMN phagosomes, and implicate factors other than VraSR.

PMN Granules Induce the Expression of msrA1

Optimal PMN antimicrobial action relies not only on oxidants from the NADPH oxidase but also on granule proteins released into phagosomes during degranulation [28]. Granule proteases (e.g., elastase, cathepsin G) [29, 30, 31] and cationic antimicrobial proteins (e.g., defensins and cathelicidins) [32] can compromise the stability of bacterial membranes that in turn may disrupt the cell wall [33, 34, 35]. Because our data demonstrate that upregulation of msrA1 in ingested S. aureus could be mediated by factors other than PMN-derived oxidants (fig. 3c) and could be induced by cell wall stress, we hypothesized that granule proteins could trigger a response that was relayed through VraSR. To test this hypothesis, we measured msrA1 in bacteria exposed to PMN granule-rich extracts in the absence of exogenous reactive oxygen species. The level of msrA1 expression was lowest at the start of the experiment prior to any treatment in all the strains. After a 30-min incubation period, msrA1 expression was slightly higher in bacteria that were not treated or in the RB control. This increase in msrA1 expression may reflect changes in the growth status of the bacteria, since msrA1 expression varies during different growth phases [3]. In bacteria treated with granule extract, msrA1 expression was significantly higher in the wild-type when compared to the RB-treated bacteria at 30 min (p < 0.05; fig. 6). In contrast, the level of msrA1 in granule extract-treated samples in ΔvraSR was not significantly elevated compared to its RB control. These data suggest that granule proteins alone promoted increased msrA1 expression and that VraSR contributed to the modulation of msrA1 in response to granule proteins.

Fig. 6.

msrA1 expression in S. aureus treated with PMN granule proteins. Wild-type, ΔvraSR and complemented ΔvraSR strains were treated with PMN granule extracts (Gex), PBS (negative control) or RB (vehicle control). The levels of msrA1 in the samples at 0 and 30 min were assessed by real-time PCR. The expression levels were normalized to the expression of gapdh. Data represent the mean ± SEM (n = 3 experiments). To compare the RB sample with the Gex sample at 30 min, a paired t test was used for each strain (* p < 0.05).

Discussion

Optimal antimicrobial action in PMN relies heavily on phagocyte-generated oxidants. In such an inhospitable environment, ingested S. aureus rapidly initiates a program of transcriptional responses that promote its survival and dissemination [11, 36], making it likely that the net balance between sustained damage and repair dictates the fate of ingested bacteria. Defensive strategies employed by bacteria in response to oxidants involve both genes encoding proteins that target oxidants, such as catalase, superoxide dismutase, thioredoxin, and alkyl hydroperoxide reductase [36], as well as those that repair oxidant-mediated damage and restore normal function. Msr are key enzymes in the defense against damage mediated by reactive oxygen species that are produced as a byproduct of aerobic respiration, and their evolutionary conservation across nearly all species underscores their importance [37, 38]. HOCl oxidizes protein methionines in Escherichia coli in a dose-dependent fashion, and methionine oxidation closely correlates with loss of bacterial viability [39]. Conversely, the capacity of Msr to repair damaged methionines parallels the increased resistance of E. coli to killing by HOCl [39] and likewise contributes to the defenses of Mycobacterium tuberculosis[40] and Helicobacter pylori against HOCl [41]. Thus, we reasoned that oxidation of key proteins in S. aureus may also result in reduced viability when the Msr-mediated repair system is inadequate to meet demands.

We demonstrate that the absence of Msr activity increased the susceptibility of S. aureus to the oxidants H₂O₂ and HOCl. We found that the ability of the Msr system to temper the effects of oxidant damage was less marked when S. aureus was treated with low doses of H₂O₂ or HOCl, which may indicate that the role of Msr in oxidant defense becomes more important when S. aureus is subjected to higher levels of oxidant stress and greater levels of protein damage (fig. 2b). In addition, we observed that even in the absence of any oxidant treatment, the percentage of recovered ΔmsrA1B was sometimes slightly lower than that of the wild-type and complemented strain (fig. 2a-c). Although this difference did not reach statistical significance, it suggested that ΔmsrA1B may be slightly more sensitive to the aerobic environment during the 1-hour incubation in PBS than the other strains. This behavior contrasts with our observation that all the strains showed similar growth in TSB. Since TSB is a rich source of substrates that could serve as oxidant sinks, it is likely that the enriched media both promotes microbial growth directly and neutralizes some of the environmental oxidants, thus limiting their deleterious effects on the bacteria.

The absence of Msr activity increased the susceptibility of S. aureus not only to reagent H₂O₂ and HOCl but also to the antimicrobial effects of PMN. In contrast to the immediate exposure to high concentrations of oxidants imposed by adding H₂O₂ or HOCl to suspensions of S. aureus, oxidant generation within phagosomes occurs gradually, following the kinetics of NADPH oxidase assembly and activation. As a result, oxidants within PMN may be generated sufficiently slowly to allow reparative responses such as the induction of msr to be engaged. Given the role of Msr in oxidant repair, we speculated that PMN-derived oxidants may drive the upregulation of msrA1 that we observed in S. aureus ingested by PMN. Our data show that bacteria ingested by normal PMN upregulated msrA1 rapidly after phagocytosis. S. aureus ingested by PMN that have impaired NADPH oxidase activity and were not exposed to phagosomal oxidants also upregulated msrA1 after phagocytosis and to higher levels. These data suggest that the msrA1 response to ingested S. aureus may be muted by the presence of oxidants in the phagosome, either directly by oxidant attack on bacteria or indirectly by oxidative modifications of the host-derived phagosomal contents. In the absence of PMN, S. aureus exposed simultaneously to H₂O₂ and a cell wall-active antibiotic (vancomycin) had a blunted msrA1 transcriptional response relative to that seen in vancomycin alone. Unexpectedly, HOCl, a much more powerful oxidant than H₂O₂, did not influence vancomycin-induced msrA1. The mechanistic basis for the differential responses to the two oxidants is unknown and merits further study. These findings illustrate how oxidants and granule proteins present in phagosomes can act synergistically on target bacteria.

The S. aureus genome encodes three other msr genes besides msrA1, including msrB, msrA2 and msrA3. Since msrB is cotranscribed and regulated by the same promoter, the pattern of expression follows that of msrA1 and it is responsive to the same stimuli. The other two msr genes are located at different genomic loci and under the regulation of promoters that are responsive to differing stresses than that of msrA1 and msrB. Not only do msrA2 and msrA3 show a different pattern of expression to msrA1 and msrB during varied phases of growth, they also do not respond to the same stress signals [3]. For these reasons and also because MsrA1 and MsrB account for almost all the Msr activity in S. aureus (fig. 1), the expression of msrA2 and msrA3 was not evaluated in this study.

The regulation of msrA1 in S. aureus is not yet fully elucidated, although msr is one of a set of genes known as the cell wall stress stimulon, which is regulated by stresses inflicted by cell wall-active antibiotics such as β-lactams, vancomycin and D-cycloserine [42]. Genes in the cell wall stress stimulon overlap with those regulated by VraSR, a TCS that influences susceptibility of S. aureus to vancomycin by modulating cell wall biosynthesis [26]. The VraSR system senses disruptions of cell wall biosynthesis in target bacteria, which may explain why both cell wall-active antibiotics and antimicrobial peptides induce the VraSR system [43]. In our study, PMN granule proteins alone stimulated msr expression in a VraSR-dependent fashion. The mechanism by which VraSR is activated and how it subsequently modulates msrA1 remain to be elucidated, including whether VraR directly or indirectly regulates msr. We noted that the introduction of the complementation plasmid into ΔvraSR increased msrA1 expression to higher levels than that observed in ΔvraSR both in the presence and absence of granule extracts but did not restore it to wild-type levels. The levels of msrA1 were also similar in the RB and granule extract-treated samples for this strain. Complementation of a mutation can be challenging and, given that the restoration of vraSR expression was driven from a multicopy plasmid and from an inducible promoter rather than from its native promoter, the gene dosage and regulation of vraSR may not accurately emulate the wild-type strain under these conditions. Nonetheless, the levels of vraSR in ΔvraSR with pEPSA5-vraSR were higher than in ΔvraSR and allowed for greater msrA1 expression.

The transcriptional changes that occur in phagocytosed bacteria reflect a coordinated response to the multitude of stresses experienced within the host environment. Given the complexity of this environment, multiple signaling pathways underpin the net global response, and their cross-talk can contribute to the modulation of stress response genes such as msr. Indeed, our data show that there are factors other than VraSR that can modulate the expression of msrA1 in S. aureus within the PMN environment. The identification and characterization of signaling systems in addition to VraSR in the regulation of msrA1 modulation remain an area of interest.

Our study highlights Msr as one of the mechanisms by which S. aureus may resist the consequences of the multifaceted antimicrobial attack that occurs in PMN phagosomes and reemphasizes that the relationships between stress and repair pathways are not necessarily straightforward. We speculate that Msr and functionally related proteins in the antioxidant pathway serve important roles in cell wall biosynthesis and homeostasis. Further understanding of the relationship between oxidant stress, both exogenous and endogenous, and cell wall damage in the context of elucidating how S. aureus survives within PMN may provide insights into the pathogenesis of staphylococcal disease and into the development of novel therapeutic approaches.

Disclosure Statement

None of the authors have conflicts of interests.

Supplementary Material

Supplementary data