Oxacillin Alters the Toxin Expression Profile of Community-Associated Methicillin-Resistant Staphylococcus aureus

Abstract

The emergence of community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) is a growing cause for concern. These strains are more virulent than health care-associated MRSA (HA-MRSA) due to higher levels of toxin expression. In a previous study, we showed that the high-level expression of PBP2a, the alternative penicillin binding protein encoded by the mecA gene on type II staphylococcal cassette chromosome mec (SCCmec) elements, reduced toxicity by interfering with the Agr quorum sensing system. This was not seen in strains carrying the CA-MRSA-associated type IV SCCmec element. These strains express significantly lower levels of PBP2a than the other MRSA type, which may explain their relatively high toxicity. We hypothesized that as oxacillin is known to increase mecA expression levels, it may be possible to attenuate the toxicity of CA-MRSA by using this antibiotic. Subinhibitory oxacillin concentrations induced PBP2a expression, repressed Agr activity, and, as a consequence, decreased phenol-soluble modulin (PSM) secretion by CA-MRSA strains. However, consistent with other studies, oxacillin also increased the expression levels of alpha-toxin and Panton-Valentine leucocidin (PVL). The net effect of these changes on the ability to lyse diverse cell types was tested, and we found that where the PSMs and alpha-toxin are important, oxacillin reduced overall lytic activity, but where PVL is important, it increased lytic activity, demonstrating the pleiotropic effect of oxacillin on toxin expression by CA-MRSA.

INTRODUCTION

Methicillin-resistant Staphylococcus aureus (MRSA) is a major cause of nosocomial infections worldwide (1, 2). Of increasing concern is the emergence of hypervirulent MRSA strains causing infections in healthy individuals in the wider community, referred to as community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) (2–4). Unlike hospital-associated MRSA (HA-MRSA), CA-MRSA strains are not restricted to health care environments, and in the United States, they appear to be outcompeting HA-MRSA, where they are currently the leading cause of death by any single infectious agent (5, 6).

Resistance to methicillin and oxacillin (the clinically used derivative of methicillin) is conferred by the acquisition of a mobile genetic element, SCCmec (staphylococcal cassette chromosome mec) (7). The elements vary in size (20 to 70 kb) and genetic content, but they all contain the mecA gene, which encodes an alternative penicillin binding protein, PBP2a (7). Regulation of mecA expression is controlled by its own regulators, mecR1 and mecI, also carried on the SCCmec element (8, 9). In the absence of β-lactam antibiotics, mecA transcription is repressed by MecI bound to its promoter region. Detection of β-lactams by the sensory domains in MecR1 removes the repression of mecA transcription by MecI, which leads to mecA transcription, PBP2a translation, and the expression of methicillin resistance (8, 9). The BlaRI/BlaI system also responds to oxacillin to induce mecA expression in MRSA (10, 11). It has been observed by us and others that in the absence of β-lactam antibiotics, HA-MRSA strains express higher levels of PBP2a than do CA-MRSA strains (12, 13).

The expression of virulence factors such as toxins is vital to the pathogenesis of S. aureus infections (1, 2). They are tightly regulated and act to degrade host cells and tissue, subvert the immune response, and enable both intra- and interhost dissemination. The secretion of proteins such as Panton-Valentine leucocidin (PVL) (14, 15), alpha-toxin (16), and phenol-soluble modulins (PSMs) (17, 18) has been associated with the increased virulence of CA-MRSA. In addition to the role of specific effector molecules, their regulation also differs in CA-MRSA relative to HA-MRSA. We previously reported that high levels of PBP2a expression in HA-MRSA strains caused a downregulation in toxicity (13). We showed that PBP2a-induced changes in the cell wall affected the responsiveness of one of the major systems regulating toxin expression, Agr (13). This resulted in low-level expression of toxins and, as a consequence, reduced virulence in a murine model of sepsis. Despite CA-MRSA strains also expressing PBP2a, their toxicity was not affected, which we hypothesized was a result of their relatively lower basal levels of PBP2a production (13).

As the expression of PBP2a can be induced by oxacillin (8–11), and high levels of PBP2a expression render the Agr system unresponsive (13), we hypothesized that oxacillin could be used to reduce the toxicity and, as a consequence, the severity of CA-MRSA infections. There are, however, a number of studies that have shown that subinhibitory concentrations of oxacillin increase rather than decrease the transcription of toxin genes such as alpha-toxin and PVL in S. aureus strains (19–21). As toxin gene transcripts need to be translated and the protein needs to be secreted to have any affect, we sought to test our hypothesis by examining the effect that oxacillin has on both cytotoxicity and the secretion of several toxins in a large diverse collection of MRSA strains. We found that the expression of PSMs can be repressed by inducing higher levels of expression of mecA-encoded PBP2a, following exposure to subinhibitory concentrations of oxacillin. As we reported for HA-MRSA, a high level of expression of PBP2a affects the responsiveness of the Agr system in CA-MRSA, which subsequently affects delta-toxin expression. However, oxacillin also induced an overall increase in exoprotein expression levels by CA-MRSA isolates, including alpha-toxin and PVL, revealing that oxacillin has pleiotropic effects on S. aureus strains, altering their toxin expression profile.

MATERIALS AND METHODS

Strains.

The strains used in this study are listed in Table S1 in the supplemental material.

PBP2a expression.

Total cell protein preparations were made from cultures of bacteria grown overnight in 5 ml of brain heart infusion (BHI) broth in 30-ml tubes at 37°C at 180 rpm in air. The cells were harvested and lysed in H₂O containing 10 μg/ml lysostaphin, 4 units of DNase, 4 units of RNase, and 0.05 mg/ml SDS by incubation at 37°C for 1 h. Insoluble material was removed by centrifugation, and protein concentrations were determined by a Bradford assay (22). Western analyses were performed by using anti-PBP2a antibodies (Abnova) and a protein G-horseradish peroxidase (HRP) conjugate (Sigma) and visualized by using a colorimetric detection system (4CN; Bio-Rad). The Western blots were performed in triplicate, and the bands were scanned and quantified by using ImageJ (http://rsbweb.nih.gov/ij/).

T cell toxicity assays.

T cell toxicity assays were performed as described previously (13, 23). Where relevant, 0.5 μg/ml oxacillin or 50 and 100 ng/ml anhydrous tetracycline were added to BHI broth to induce expression of mecA from its own promoter and from the tetracycline-inducible promoter, respectively.

RNA isolation.

BHI broth (100 ml in 250 ml in an unbaffled, conical flask) was inoculated with 100 μl of cultures, and cultures were grown overnight at 37°C at 180 rpm. RNA was isolated after 10 and 20 h of growth by using the Qiagen RNeasy Midi kit according to the manufacturer's instructions, with the addition of 0.2 μg/μl of lysostaphin to the lysis step. RNA quality and concentration were determined by using the Bio-Rad Experion RNA analysis system.

Reverse transcription and qRT-PCR.

cDNA was generated from mRNA by using the NEB First Strand Synthesis kit in accordance with the manufacturer's instructions, using random hexamers. The following primers were used to amplify the cDNA: RNAIII forward (GAAGGAGTGATTTCAATGGCACAAG), RNAIII reverse (GAAAGTAATTAATTATTCATCTTATTTTTTAGTGAATTTG), gyrB forward (CCAGGTAAATTAGCCGATTGC), gyrB reverse (AAATCGCCTGCGTTCTAGAG), mecA forward (TGCTCAATATAAAATTAAAACAAACTACGGTAAC), mecA reverse (GAATAATGACGCTATGATCCCAA), hla forward (TGGCCTTCAGCATTTAAGGT), hla reverse (CAATCAAACCGCCAATTTTT), lukS forward (TGAGGTGGCCTTTCCAATAC), and lukS reverse (CCTCCTGTTGATGGACCACT).

Standard curves were produced for each primer set on serial dilutions of cDNA to determine primer efficiency. The quantitative reverse transcriptase PCRs (qRT-PCRs) were set up was as follows: 5 μl cDNA, 7.5 μl of SYBR reagent, 0.5 μl forward primer (10 μM), 0.5 μl reverse primer (10 μM), and RNase-free H₂O to a total volume of 15 μl. Threshold cycle (C_T) values were subsequently determined for three biological repeats in triplicate. For each reaction, the ratio of the target gene (X) and gyrB transcript numbers was calculated as follows: 2^{(CT gyrB − CT X)}.

AIP induction assay.

AIP induction assays were performed as described previously (13). Briefly, synthetic AIP-1 was reconstituted in dimethyl sulfoxide (DMSO) to a concentration of 10 mM. Bacterial strains were grown overnight in BHI broth and used to inoculate 5 ml of fresh BHI broth (containing 10 μg/ml of chloramphenicol to retain the expression plasmid) in 30-ml tubes. The bacteria were grown for 2 to 3 h to an optical density at 600 nm (OD₆₀₀) of 0.2 (exponential phase), washed 3 times in phosphate-buffered saline (PBS), and resuspended to an OD₆₀₀ of 0.2 in BHI broth. Either AIP-1 or DMSO (vehicle control) was added to each strain and incubated at 37°C for 3 h. To ensure that the different treatments did not affect cell growth, a measurement of the OD₆₀₀ was performed for each sample before its fluorescence intensity was measured by using a FARcyte Xfluro4 V 4.50 plate reader at between 485 nm and 535 nm. The data are presented as the mean fluorescence intensities of five independent replicates. Error bars represent the 95% confidence intervals.

Alpha-toxin and LukS-PV (PVL) expression.

Cultures of S. aureus grown overnight were diluted 1:1,000 in 5 ml of tryptone soya broth (TSB) or casein hydrolysate yeast (CCY) broth (specifically for PVL expression) and incubated with shaking (180 rpm) for 20 h at 37°C to an OD₆₀₀ of 2.0. Where appropriate, oxacillin was used at a subinhibitory concentration of 0.5 μg/ml. Bacteria were removed by centrifugation at 14,000 × g for 10 min, and supernatant proteins were precipitated by using trichloroacetic acid (TCA) at a final concentration of 20% for 1 h on ice. Samples were centrifuged at 14,000 × g for 15 min at 4°C and then washed three times in ice-cold acetone and solubilized in 100 μl 8 M urea. Proteins (10 μl of each sample) were mixed with 2×-concentrated sample buffer and heated at 95°C for 5 min before being subjected to 12% SDS-PAGE. Separated proteins were electroblotted onto a nitrocellulose membrane using a semidry blotter (25 V for 30 min). Membranes were blocked overnight at 4°C with 5% semiskimmed milk (Marvel) and then incubated with rabbit polyclonal antibodies specific for LukS-PV (1 μg/ml; IBT Bioservices) or alpha-toxin (1:5,000 dilution; Sigma-Aldrich) for 2 h at room temperature. Immunoblots were washed and incubated with horseradish peroxidase-coupled protein G (1:1,000; Invitrogen) for 1 h at room temperature. Proteins were detected by using the Opti-4CN detection kit (Bio-Rad). The Western blots were performed in triplicate, and the bands were scanned and quantified by using ImageJ (http://rsbweb.nih.gov/ij/).

PSM extraction and quantification.

Cultures of S. aureus grown overnight were diluted 1:1,000 into 50 ml TSB in a 250-ml unbaffled, conical flask and grown for 20 h at 37°C to an OD₆₀₀ of 2.0; where appropriate, oxacillin was used at a subinhibitory concentration of 0.5 μg/ml. Bacteria were removed by centrifugation, and 30 ml of cell-free supernatants was mixed with 10 ml 1-butanol. Extraction was performed by shaking mixed solutions for 3 h at 37°C. Samples were then centrifuged for 3 min for complete separation, and the upper organic phase was collected, aliquoted into 1-ml tubes, and concentrated by using a vacuum overnight. Dried samples were dissolved in 200 μl of 8 M urea. Proteins (10 μl of each sample) were mixed with 2×-concentrated sample buffer and heated at 95°C for 5 min before SDS-PAGE (12% acrylamide) was performed. The individual PSMs were quantified by using mass spectroscopy as described previously (24). This assay was performed in triplicate.

RBC and PMN harvesting and lysis assays.

Polymorphonuclear leukocytes (PMNs) were isolated from heparinized venous blood obtained from healthy adult volunteers. Whole-blood samples were layered onto density gradient medium (Lympholyte cell separation medium) and centrifuged as described by the manufacturer (Cederlane). The PMN-containing band was harvested and diluted with an equal volume of culture medium (HyClone Dulbecco's modified Eagle's medium [DMEM]–10% fetal calf serum [FCS] cell medium) at a concentration of 0.5 N to restore normal osmolality. The cells were further diluted with 2 volumes of culture medium and washed by centrifugation for 10 min at 400 × g. The cells were suspended in 5 ml of 0.2% NaCl and incubated for 120 s to lyse erythrocytes by osmotic shock, followed by the addition of 5 ml of 1.6% NaCl for 120 s to normalize the osmolality and centrifugation at 400 × g for 10 min. Purified PMNs were then washed in PBS and enumerated by using a hemocytometer. Purity was assessed by trypan blue and flow cytometric analyses. The final PMN count was adjusted to 1.0 × 10⁶ to 2.0 × 10⁶ cells/ml with culture medium. Thirty microliters of PMNs was incubated with 30 μl of bacterial supernatant for 30 min, and cell viability was assayed by using Guava viability reagent and Guava flow cytometry (Millipore). Red blood corpuscles (RBCs) were isolated from the same blood samples and washed twice by gentle resuspension with a 10× volume of sterile saline (0.9% NaCl) and centrifugation at 1,000 × g for 10 min. RBCs were diluted to 1% (vol/vol), and 200 μl was incubated with 50 μl of bacterial supernatant for 30 min, using free saline as a negative control and 1% Triton X-100 as the positive control. Intact cells and cellular debris were removed by centrifugation at 1,000 × g for 10 min. RBC lysis was assayed by determining the absorbance of the resulting supernatant at 404 nm. These assays were performed in triplicate.

RESULTS

Growth of CA-MRSA with subinhibitory concentrations of oxacillin increases PBP2a expression.

Although specific genetic factors are associated with a strain being either HA- or CA-MRSA, the type IV SCCmec element can be found in both. To avoid confusion, we use the term CA-MRSA only when discussing specific strains for which the infection history is known. Otherwise, we refer to strains as type IV MRSA.

We have previously shown that the high level of PBP2a expression in type II MRSA strains reduces toxin expression (13). In type IV MRSA, the basal level of PBP2a expression is lower, which we believe explains why their toxicity is unaffected (13, 23). There is also an insertion sequence (IS) element in the mecR1-mecI locus and truncation of mecI in type IV SCCmec elements, which may affect its inducibility in response to oxacillin, although BlaRI/BlaI has been shown to have equivalent mecA-inducing activity (7). To determine whether CA-MRSA PBP2a can be induced by oxacillin, we measured the transcription and translation of the mecA gene in the epidemiologically important CA-MRSA USA300 strain LAC and USA400 strain MW2 grown with and without 0.5 μg/ml oxacillin after 10 and 20 h of growth. This concentration of oxacillin was used since it has no effect on growth dynamics over 20 h (see Fig. S1 in the supplemental material). The transcription level of the mecA gene relative to that of gyrB was determined by quantitative reverse transcriptase PCR (qRT-PCR). We observed a 6.96-fold increase in expression levels for LAC at 20 h, but only small and nonsignificant increases in transcription levels were observed for LAC at 10 h or MW2 at 10 or 20 h (Table 1) following growth in oxacillin. Using anti-PBP2a antibodies in Western blots, we also found that the amount of PBP2a in the cell wall increased for both LAC and MW2 (3.01- and 2.3-fold, respectively) (Fig. 1 and Table 2). Despite the presence of the IS element, oxacillin induces increased transcription for LAC and increased translation of PBP2a for both LAC and MW2.

TABLE 1.

Transcriptional changes in response to oxacillin after 10 and 20 h of growth^a

Isolate, locus	Fold change (+ox vs −ox) at 10 h	Significance (P) at 10 h	Fold change (+ox vs −ox) at 20 h	Significance (P) at 20 h
LAC, mecA	↑1.47	0.32	↑6.96	1.35 × 10−8∗
MW2, mecA	↑1.04	0.87	↑1.29	0.56
LAC, RNAIII	↓772.07	0.04∗	↓49.34	2.63 × 10−8∗
MW2, RNAIII	↓3.15	0.006∗	↓3.58	0.02∗
LAC, hla	↓1.81	0.16	↑5.02	0.66
MW2, hla	↓2.90	0.06	↑1.35	0.26
LAC, lukS	↑3.11	0.003∗	↑51.75	0.0002∗
MW2, lukS	↓1.02	0.90	↑2.15	0.16

^aThe arrows indicate the direction of the fold change in the expression level in response to oxacillin (ox). Asterisks indicate statistically significant results.

FIG 1.

Oxacillin-induced expression of PBP2a in USA300 strain LAC and USA400 strain MW2. Western blots of whole-cell lysates show induction of PBP2a expression in CA-MRSA strains LAC and MW2. Lane 1, molecular weight (MW) standards (in thousands); lane 2, LAC grown in antibiotic-free medium; lane 3, LAC grown in medium containing 0.5 μg/ml oxacillin; lane 4, MW2 grown in antibiotic-free medium; lane 5, MW2 grown in medium containing 0.5 μg/ml oxacillin.

TABLE 2.

Protein expression level changes in response to oxacillin after 20 h of growth^a

Isolate, locus	Fold change (+ox vs −ox) at 20 h	Significance (P) at 20 h
LAC, mecA	↑3.01	0.038∗
MW2, mecA	↑2.3	0.025∗
LAC, hla	↑1.8	0.19
MW2, hla	↑1.4	0.29
LAC, lukS	↑3.01	0.008∗
MW2, lukS	↑4.1	0.011∗

^aThe arrows indicate the direction of the fold change in expression in response to oxacillin (ox). Asterisks indicate statistically significant results.

Oxacillin-induced expression of PBP2a represses the toxicity of type IV MRSA strains.

To test the cytolytic activity of S. aureus strains, we used an assay in which immortalized T cells are incubated with bacterial supernatant and T cell survival is measured after 15 min. Using this assay, we previously demonstrated that type IV MRSA strains are highly toxic (13, 23), and to determine whether this is a result of low-level expression of PBP2a, higher levels of expression were induced by growing strains with subinhibitory concentrations of oxacillin. Using a collection of 56 type IV MRSA strains from a diverse range of genetic backgrounds (sequence types [STs] 1, 8, 22, 30, 45, 59, 80, 93, 254, 255, 280, and 296 [see Table S1 in the supplemental material]), the ability of the strains to lyse immortalized T cells was found to be significantly reduced, where T cell survival went from an average of 17.7% in the absence of oxacillin to 47.7% when grown in oxacillin (P < 0.001) (Fig. 2A).

FIG 2.

Oxacillin reduces the toxicity of MRSA strains carrying the type IV SCCmec element. (A) The toxicity of 56 clinical MRSA strains was measured and is represented as their ability to lyse immortalized T cells. The strains were grown in BHI broth with or without 0.5 μg/ml oxacillin for 20 h. The circles represent the toxicity of individual strains, and black bars indicate the means for all 56 strains. Error bars show standard errors of the means. (B) Induced expression of mecA reduces the toxicity of USA300 strain LAC and USA400 strain MW2. The effect of direct overexpression of mecA from a tetracycline-controlled inducible promoter on cytolytic activity was measured in USA300 strain LAC and USA400 strain MW2. Strains were grown in BHI broth with anhydrotetracycline added at 0 ng/ml, 50 ng/ml, and 100 ng/ml. Circles represent individual repeats, and black bars indicate the means. Error bars show standard errors of the means.

Expression of mecA from an inducible promoter represses the toxicity of CA-MRSA.

To ensure that the effect on cytolytic activity upon exposure to oxacillin was a direct result of increased PBP2a expression, two CA-MRSA isolates, LAC (USA300) and MW2 (USA400), were transformed with a plasmid containing the mecA gene under the control of a tetracycline-inducible promoter (13). Anhydrotetracycline, which is a less-growth-inhibitory form of tetracycline, was used at concentrations of 50 and 100 ng/ml, and the increasing repressive effect that this has on the cytolytic activity of these CA-MRSA strains can be seen in Fig. 2B. The uninduced (no-Tet) plasmid did not affect the cytolytic activity of either MW2 or LAC (P > 0.1). For both LAC and MW2, the cytolytic activity was significantly repressed upon induction of mecA expression (P = 0.016 for LAC and P = 0.017 for MW2 for 0 versus 100 ng/ml), demonstrating that increased levels of expression of mecA were directly responsible for the decrease in toxicity for these strains.

Growth in oxacillin decreases RNAIII expression in USA300 and USA400.

We previously reported that the reduced toxicity of type II MRSA strains was as a result of lower levels of expression of RNAIII, the effector molecule of the Agr quorum sensing system in S. aureus (13). To determine whether the oxacillin-induced reduction of toxicity in type IV MRSA strains was a result of reduced levels of RNAIII expression, qRT-PCR analysis was performed after 10 and 20 h of growth. Exposure to 0.5 μg/ml oxacillin reduced expression of RNAIII for both USA300 strain LAC and USA400 strain MW2 (P < 0.001 for both) (Table 1), demonstrating the involvement of the Agr system in the oxacillin-induced repression of toxicity of these CA-MRSA strains.

Oxacillin represses the ability of the Agr system to respond to extracellular AIP.

We previously reported that high levels of PBP2a expression blocked activation of the Agr quorum sensing system in response to synthetic AIP in type II MRSA strains (13). To determine whether this was responsible for the oxacillin-induced reduction of RNAIII expression in CA-MRSA, AIP induction assays were performed. USA300 strain LAC was transformed with a plasmid containing a reporter fusion in which the RNAIII promoter was fused to a gene encoding green fluorescent protein (GFP). This culture was grown (with and without 0.5 μg/ml oxacillin) to mid-exponential phase, when the activity of the Agr system should be low but receptive to inducement with extracellular AIP. Synthetic AIP was added to the culture, and RNAIII::GFP expression was monitored over time. As anticipated, when grown without oxacillin, AIP induced increasing levels of RNAIII::GFP in USA300 strain LAC over the time course of the experiment (Fig. 3). However, when grown in oxacillin, this response was significantly lower (P < 0.05), demonstrating the reduction in responsiveness of the Agr system of USA300 induced by oxacillin.

FIG 3.

Oxacillin interferes with Agr activation in USA300 strain LAC and USA400 strain MW2. Depicted is the response of USA300 strain LAC cells grown with and without oxacillin to the addition of synthetic AIP. An RNAIII::GFP reporter fusion was used to measure the response of USA300 strain LAC to synthetic AIP following growth in BHI broth with or without oxacillin and with either AIP or the DMSO control.

Oxacillin alters the toxin expression profile of CA-MRSA.

Although we report here a clear decrease in lytic activity upon exposure to oxacillin, other groups have reported increases in transcription levels of cytolytic toxins such as alpha-toxin and PVL in response to oxacillin (19–21). The T cells used in these cytotoxicity assays are sensitive to the effects of alpha-, beta-, gamma-, and delta-toxins; PSMα1; PSMα2; and PSMα3 but are resistant to PVL (23). To determine the overall effect of oxacillin on the exoprotein expression profile of CA-MRSA, the supernatants of bacteria grown with and those grown without oxacillin were precipitated by using trichloroacetic acid. Despite a reduction in Agr activity, an overall increase in the expression levels of exoproteins by LAC and MW2 was found to be induced by oxacillin (Fig. 4A). Westerns blots of these extracts demonstrated that the amount of alpha-toxin was increased for both LAC and MW2, although this was not statistically significant (1.8- and 1.4-fold, respectively) (Fig. 4B and Table 2). However, the level of secretion of PVL was significantly increased for both LAC and MW2 (3.01- and 4.1-fold, respectively) (Fig. 4C and Table 2). Using qRT-PCR, we found that oxacillin induced a nonsignificant decrease in translation of the gene encoding the alpha-toxin, hla, at 10 h but induced a nonsignificant increase in transcription at 20 h (Table 1). For PVL, however, there was a significant increase in transcription at both 10 and 20 h of growth (Table 1).

FIG 4.

Oxacillin alters the toxin expression profile of CA-MRSA. Bacteria were grown with and without oxacillin, as indicated. (A) Total exoprotein expression profile (Coomassie stained). (B) Western blot using anti-PVL (LukS) antibodies. (C) Western blot using anti-alpha-toxin antibodies. (D) PSMs extracted from the bacterial supernatants by using butanol (Coomassie stained). (E) Quantitative analysis of the expression of each of the PSMs upon exposure to oxacillin (error bars represent the 95% confidence intervals).

To examine the effect on PSM secretion (including delta-toxin) qualitatively, butanol extractions of these supernatants were performed. These molecules are relatively small (22 to 44 amino acids) and are difficult to separate from each other on an SDS-PAGE gel; however, a clear reduction in band intensity can be seen upon exposure of the bacteria to oxacillin (Fig. 4D). To quantify these changes, we performed mass spectroscopy on these supernatants and demonstrated that oxacillin reduces the secretion of PSMα1, -2, -3, and -4; PSMβ1 and -2; and delta-toxin for both LAC and MW2 (Fig. 4E).

The effect of oxacillin on lytic activity depends on the sensitivity of the cell type used.

As mentioned above, the T cells used in these assays are not sensitive to all the toxins affected by oxacillin. To determine the overall effect of oxacillin on cytolytic activity, we included two additional cell types, human red blood corpuscles (RBCs), which are sensitive to alpha-toxin and the PSMs, and human polymorphonuclear leukocytes (PMNs), which are sensitive to PVL and the PSMs. The effects of oxacillin on the ability of both LAC and MW2 to lyse these cell types were compared across these cell types. The effect on T cells is included for reference (Fig. 5A), and as expected, oxacillin induced an increase in T cell survival upon exposure to bacterial supernatants of both LAC (P < 0.001) and MW2 (P = 0.004). Despite a small increase in the expression level of alpha-toxin, oxacillin decreased the lysis of human RBCs, where the decrease in PSM secretion had a more dominant effect. Contrarily, although human PMNs are sensitive to both PVL and PSMs, the increase in the expression level of PVL had a more dominant effect, with PMN survival decreasing upon exposure to oxacillin.

FIG 5.

Oxacillin differentially alters the sensitivity of cells to lysis. (A) Oxacillin decreased the ability of CA-MRSA to lyse immortalized T cells. (B) Oxacillin decreased the ability of CA-MRSA to lyse fresh human RBCs. (C) Oxacillin increased the ability of CA-MRSA to lyse fresh human PMNs. The data shown are the means, and the error bars indicate the 95% confidence intervals.

DISCUSSION

Our previous work demonstrated that increased PBP2a expression reduced disease severity in vivo by reducing the expression of cytolytic toxins (13, 25). We hypothesized that as oxacillin increases PBP2a expression, it could be used as an antivirulence agent in parallel with other antibiotics to decrease disease severity in infected patients. While our findings show that oxacillin decreases Agr activity and PSM secretion by CA-MRSA, it also increases the secretion of other cytolytic toxins.

Alpha-toxin is the prototype for the class of small β-barrel pore-forming cytotoxins with a specific host cell receptor (26). At high concentrations, it causes pore formation in the host cell membrane and leads to cell lysis, but at lower, sublytic concentrations, it has been shown to alter the cell signaling pathways that govern cell proliferation, inflammatory responses, cytokine secretion, and cell-cell interactions (27). It has been shown to play a direct role in the severity of many types of infections caused by S. aureus, including pneumonia, dermonecrotic skin infection, sepsis, and peritonitis (27). In this study, we show that in concordance with other studies, subinhibitory concentrations of oxacillin increase the expression levels of this toxin, although the effect that we saw was small and was not statistically significant. The classic human cell type used to assay alpha-toxin, RBCs, was more recently shown to also be susceptible to PSMs (28). Using these cells, we found that the overall effect on the ability of CA-MRSA to lyse this cell type was decreased upon exposure to oxacillin, as the decrease in PSM secretion had a greater overall effect than the small increase in alpha-toxin levels.

Like alpha-toxin, the PSM family of lytic peptides has been shown to contribute to many infection types caused by S. aureus, including skin infection, bacteremia, and osteomyelitis (28). The activity of these peptides is not dependent upon a protein receptor but on the lipid composition of the host cell membrane. This allows them to have an effect on a broader range of cell types than the other toxins discussed here. The downregulating effect of oxacillin on PSMα1 and -β1 and delta-toxin secretion by USA300 strain LAC was shown previously (29), and here we show that it also affects the expression of PSMα2, -3, and -4 and PSMβ2 for LAC and MW2.

PVL is a bicomponent leukotoxin whose specific target is the C5a receptor on human PMNs (30–32). This toxin is believed to contribute to necrotic and recurrent skin and soft tissue infections (SSTIs), necrotizing pneumonia, and bone and joint infections (28). PMNs are the primary line of defense against bacterial infections. That both PVL and the PSMs contribute to killing these cells is unsurprising given the importance for bacteria to survive this aspect of host immunity. Oxacillin had opposite effects on the expression of these toxins, and as such, it was important to test which change had the more dominant effect on PMN killing. Our ex vivo assay suggests that the effect on PVL expression was more dominant, with increased PMN killing induced by oxacillin. However, the translation of these findings to infection models that are sensitive to alpha-toxin, PSMs, and PVL is critical to test the gross effect that oxacillin has on toxicity and virulence.

That PSMs are differentially regulated compared to other toxins such as alpha-toxin and PVL has been demonstrated previously (30), and here we provide more evidence in support of this. It is possible that while core regulators like Agr control the expression of all cytolytic toxins in a similar manner, additional regulators might exist as a backup to these, to allow the bacteria adapt to their constantly changing environment. As such, perhaps oxacillin, while downregulating the core Agr regulatory system, also increases the activity of these other regulators. Future work to evaluate the potential of other cell-wall-active antibiotics that can mediate PBP2a-dependent agr repression will be of interest in the context of virulence attenuation, but ultimately, the potential of beta-lactams and related antimicrobials as antivirulence drugs may depend on the site of infection, the S. aureus strain, and the relative contribution of individual toxins to specific infections.