Identification of a Novel Staphylococcus aureus Two-Component Leukotoxin Using Cell Surface Proteomics

Staphylococcus aureus is a prominent human pathogen and leading cause of bacterial infection in hospitals and the community. Community-associated methicillin-resistant S. aureus (CA-MRSA) strains such as USA300 are highly virulent and, unlike hospital strains, often cause disease in otherwise healthy individuals. The enhanced virulence of CA-MRSA is based in part on increased ability to produce high levels of secreted molecules that facilitate evasion of the innate immune response. Although progress has been made, the factors that contribute to CA-MRSA virulence are incompletely defined. We analyzed the cell surface proteome (surfome) of USA300 strain LAC to better understand extracellular factors that contribute to the enhanced virulence phenotype. A total of 113 identified proteins were associated with the surface of USA300 during the late-exponential phase of growth in vitro. Protein A was the most abundant surface molecule of USA300, as indicated by combined Mascot score following analysis of peptides by tandem mass spectrometry. Unexpectedly, we identified a previously uncharacterized two-component leukotoxin–herein named LukS-H and LukF-G (LukGH)-as two of the most abundant surface-associated proteins of USA300. Rabbit antibody specific for LukG indicated it was also freely secreted by USA300 into culture media. We used wild-type and isogenic lukGH deletion strains of USA300 in combination with human PMN pore formation and lysis assays to identify this molecule as a leukotoxin. Moreover, LukGH synergized with PVL to enhance lysis of human PMNs in vitro, and contributed to lysis of PMNs after phagocytosis. We conclude LukGH is a novel two-component leukotoxin with cytolytic activity toward neutrophils, and thus potentially contributes to S. aureus virulence.


Introduction
Staphylococcus aureus is a leading cause of human bacterial infections worldwide. The organism can cause a wide range of diseases, including superficial skin and soft tissue infections, as well as invasive diseases such as pneumonia, bacteremia, endocarditis, and joint infections (reviewed in [1]). The high prevalence of infections is confounded by the ability of the pathogen to readily acquire genetic elements that confer resistance to antibiotics. Although methicillin-resistant S. aureus (MRSA) remains a significant problem for healthcare facilities in most industrialized countries, an MRSA strain known as USA300 is the most abundant cause of bacterial infections outside of healthcare facilities in the United States [2][3][4]. The ability of USA300 to cause infections in otherwise healthy individuals suggests the strain has enhanced capacity to circumvent killing by the innate immune system-a notion confirmed by studies with human neutrophils [5]. The ability of CA-MRSA strains such as USA300 to produce relatively high levels of secreted molecules such as phenol-soluble modulins (PSMs) provides an explanation in part for the enhanced ability of these pathogens to avoid destruction by neutrophils [6,7].
Despite this recent progress, our understanding of CA-MRSA virulence mechanisms is incomplete, largely because S. aureus produces many molecules that can potentially contribute to immune evasion and virulence [8]. To gain a comprehensive view of the molecules that potentially promote USA300 virulence, we evaluated the surface proteome of this pathogen. Unexpectedly, we identified an uncharacterized two-component leukotoxin-herein named LukS-H and LukF-G-bound to the surface of USA300, and used USA300 wild-type and isogenic lukGH mutant strains to verify its ability to function as a leukotoxin.

Results and Discussion
Isolation of peptides from USA300 surface proteins We optimized the approach described by Rodriguez-Ortega et al. to determine the cell surface proteome (surfome) of USA300/ LAC [9]. Because S. aureus cell wall-associated proteins are expressed maximally during exponential growth [10], we used cultures grown to late-exponential phase of growth. USA300 surface proteins were removed by digestion with trypsin in the presence of sucrose and DTT, which make surface proteins more accessible to trypsin [9]. One concern with this approach is that proteolytic digestion of the bacterial surface proteins has the potential to cause cell lysis, either through destabilization of the membrane as a result of loss of membrane proteins or through activation of bacterial autolytic and/or proteolytic proteins. To determine if trypsin decreased viability of LAC, we plated the bacteria before and after digestion with trypsin (F. S1A). There was a slight increase in CFU/ml after exposure to trypsin; this phenomenon was likely due to decreased clumping of bacteria following removal of the proteinaceous surface clumping factors (Fig. S1A). Although exposure of bacteria to trypsin for an extended period of time (21 h) resulted in lysis of bacteria (data not shown), there was no decrease in bacterial viability following the relatively short trypsin incubation period used here to remove surface proteins (Fig. S1A). Separation of a representative 1-h tryptic digest by Tricine SDS-PAGE demonstrated that the surface proteins were digested into small peptides (Fig. S1B).

Identification of USA300 surface proteins
Peptides recovered from the surface of USA300 were subjected to fractionation and tandem mass spectrometry (LC-MS/MS) as outlined in Fig. 1. Using this approach, we identified 113 proteins associated with the surface of USA300 during late-exponential phase of growth (Fig. 2, and Tables S1 and S2). Of these proteins, 17 are predicted to be associated with the cell wall using a combination of PSORT [11] and previous studies (Table S1). In addition to the 17 cell wall-associated proteins, we recovered 19 extracellular and 14 membrane proteins from the surface protein preparations of USA300. Three of the 19 putative extracellular proteins (LukS-H, IsaA, and LukF-G) were also found in high abundance on the cell surface during exponential or stationary phase of growth ( [12]; this study). As shown in Fig. 2A, 54% of the peptides (''queries'' in Table S2) identified by our MS/MS analysis could be assigned to only 50 proteins (17 cell wall-associated, 19 extracellular, and 14 membrane); the number of peptides assigned to a protein is a direct measure of the abundance of that protein in the sample. In contrast, Mascot assigned the remaining 46% of the identified peptides to 63 proteins that normally reside in the cell cytoplasm (Table S1 and Fig. 2A). As such, the confidence with which these proteins were identified was significantly lower than that for cell wall-associated and/or extracellular proteins (Fig. 2B).

Cell wall-associated proteins
The classical Gram-positive peptide, LPxTG, which covalently anchors proteins in the cell wall peptidoglycan, is present in 12 of the 17 cell wall-associated proteins identified by our analysis (Table S1). The LysM domain, which non-covalently links proteins to the cell surface, is present in 2 of these 17 proteins, as well as one membrane protein (EbpS). The most abundant protein on the surface of USA300 was immunoglobulin G binding protein A (protein A or Spa), based upon the number of unique peptides identified (46), the number of total peptides matched (273), and the combined Mascot score (6910) (Tables S1 and S2). This finding is in agreement with previously published studies of S. aureus surface proteins [13][14][15][16]. Interestingly, two previously unidentified/uncharacterized proteins encoded by open reading frames annotated in the FPR3757 genome as SAUSA300_1975 and SAUSA300_1974 were the 2 nd and 7 th most abundant proteins on the surface of USA300 (Table S1). SAUSA300_1974 and SAUSA300_1975 are predicted to encode LukF and LukS subunits of a two-component leukotoxin, which is present and highly conserved among all sequenced S. aureus strains ( Fig. S2A and S2B). Here we designate SAUSA300_1974 as lukF-G (or lukG) and SAUSA300_1975 as lukS-H (or lukH), and the two-gene operon as lukGH. Neither LukF-G nor LukS-H (hereafter called LukG and LukH, respectively) contains LPxTG or LysM cell wall anchoring domains, though both were highly abundant on the surface of S. aureus during late-exponential growth in vitro (Table S1). We verified by immunoblot analysis of subcellular fractions of USA300 that LukG is associated with the detergent-soluble membrane fraction (Fig. 3A).
The majority of the cell wall-associated proteins identified by our analysis have been described previously (reviewed in [17]). Interestingly, iron-regulated surface determinants A (IsdA) and B (IsdB) were the 6 th and 21 st most abundant proteins on the cell surface, as predicted by Mascot score (Table S1). Expression of the Isd proteins has been shown to occur only in iron-limited conditions [18], unlike the rich TSB medium in which USA300 was cultured for this study.

Extracellular/secreted proteins
The methodology we employed to remove proteins from the surface of USA300 was not specific to conventionally defined surface proteins (i.e., cell wall-associated and membrane proteins). Thus, we identified proteins in the process of being secreted into the culture medium as well. Five of the 10 most abundant proteins are putative or proven secreted proteins, and 3 others are associated with both the cell wall and the extracellular milieu. LukG and LukH are among the most abundant extracellular proteins identified by our analysis (in this category ranked 1 st and 3 rd in abundance by Mascot score) (Table S1). Gamma-hemolysin subunit A (HlgA) was the only other two-component leukotoxin subunit identified as associated with the cell surface (Table S1). Expression of delta-hemolysin and other secreted proteins, many of which were identified here, is largely dependent on upregulation of the global gene regulator agr, which occurs in vitro at the transition from exponential to stationary phase of growth [10]. Thus, our finding that delta-hemolysin is associated with the surface of S. aureus suggests bacteria were beginning to make the transition from exponential growth (high production of cell wallassociated proteins) to stationary phase of growth (production of secreted proteins).

Membrane proteins
We identified 14 membrane proteins on the surface of S. aureus in our surfome analysis. Membrane topology analysis using HMMTOP showed that many of the peptides assigned to membrane proteins were located on the exterior of the cell membrane (Fig. 3B). Six of the membrane proteins are putative components of as yet uncharacterized ABC transport systems, and one is predicted to be an amino acid efflux protein. In addition, we found MecA (SAUSA300_0032) on the surface of USA300 strain LAC (Fig. 3B). MecA is the penicillin-binding protein that confers resistance to b-lactam antibiotics in MRSA strains. We also identified CapA (SAUSA300_2598), which is encoded by the first gene in the staphylococcal capsule operon (Table S1). S. aureus capsule operon expression is tightly regulated by agr and is usually not present during the exponential phase of growth [19]. These findings indicate that the proteomic approach utilized here was sufficient to identify surface-exposed peptides from membrane proteins.

Cytoplasmic proteins
Based on the number of proteins identified by surfome analysis, the majority are putative or proven cytoplasmic proteins (Table  S1). However, as mentioned above, these proteins were significantly less abundant on the surface than were those predicted to be cell wall-associated and/or extracellular ( Fig. 2B and Tables S1 and S2). Ribosomal proteins comprise 25 of the 63 (40%) cytoplasmic proteins identified by our analysis, presumably as a result of normal cell turnover. Although there was no decrease in CFU/ml following digestion of surface proteins with trypsin, it is likely that some cell lysis occurred during digestion or subsequent processing, perhaps accounting in part for presence of cytoplasmic proteins in our cell surface proteome analysis.

LukGH is a novel S. aureus leukotoxin
Using the combined Mascot score as a measure of relative abundance, the 2 nd and 7 th most abundant USA300 proteins identified by surface proteomics were LukH and LukG, respectively (Table S1). As noted above, lukG and lukH are present in all sequenced S. aureus strains and there is limited allelic variation among the strains (97-100% nucleotide identity) (Fig. S2A). USA300 strains contain up to 4 operons encoding known or predicted two-component leukotoxins [20]. LukGH shares significant homology with other staphylococcal leukotoxins, including PVL and HlgABC (see figure 1 in ref. [21]) (Fig. S2C). LukG is 37% identical to HlgB and 36% identical to LukF-PV, and LukH is 26-28% identical to LukS-PV, HlgA and HlgC, based on a CLUSTALW2 analysis of inferred amino acids from strain FPR3757 (Fig. S2C). However, LukGH has not been previously characterized and its function remains to be determined. Based on homology and predicted function as a secreted toxin, LukGH should be secreted into the extracellular milieu during post-exponential growth of S. aureus in vitro [10]. However, Luong et al. describe expression of lukG (N315-1812) and lukH (N315-1813) during late-exponential phase of growth [22]. Further, we demonstrated previously that lukH was highly expressed during PMN phagocytosis (labeled as aerolysin/ leukocidin family protein or lukM in those reports), and following exposure to neutrophil azurophilic granule proteins [5,23,24]. As a first step toward determining the function of LukGH, we constructed isogenic lukGH (DlukGH) and lukGH/lukSF-PV (DlukGH/Dpvl) deletion mutants in USA300 (LAC) using the counterselectable marker system developed by Bae and Schneewind ( Fig. 4A) [25]. We generated the USA300DlukGH/Dpvl strain in the genetic background of a previously described USA300Dpvl strain [23]. PCR analysis of genomic DNA isolated from the wildtype and mutant strains confirmed lukGH were deleted from DlukGH and DlukGH/Dpvl, and all strains exhibited virtually identical growth in TSB media ( Fig. 4B and C). Since LukGH has identity to other two-component leukotoxins that are normally secreted into culture media, we assessed exoprotein profiles of each strain cultured to early-stationary phase of growth in CCY medium, which promotes high production of PVL (Fig. 4D). Protein bands corresponding to LukF-PV and LukS-PV (yellow arrows) were clearly present in the media in which LAC wild-type and DlukGH strains were cultivated, and absent in the media from the Dpvl and DlukGH/Dpvl cultures (Fig. 4D). Similarly, bands corresponding to LukG and LukH, as indicated by black arrows, were present in the LAC wild-type and Dpvl supernatants, and absent from supernatants derived from DlukGH and DlukGH/Dpvl cultures (Fig. 4D). Rabbit IgG specific for LukG identified the protein in culture supernatants, indicating the leukotoxin is at least in part secreted (Fig. 4E). DlukGH and DlukGH/Dpvl were complemented with lukGH encoded on a plasmid (DlukGH::plukGH and DlukGH/Dpvl::plukGH, respectively), thereby restoring production of the leukotoxin (as assessed by immunoblot for LukG) (Fig. 4E). All strains produced equivalent amounts of alpha-hemolysin or protein A, and had comparable levels of betahemolysis on blood agar (data not shown), suggesting that the global regulator agr remained intact during strain passage and mutagenesis.

LukGH has cytolytic activity toward human PMNs
Because we detected LukGH in the culture media, we compared the ability of CCY culture supernatants from LAC wild-type, Dpvl, DlukGH, and DlukGH/Dpvl strains to cause formation of plasma membrane pores in human PMNs (Fig. 5A). Culture supernatants from Dpvl, DlukGH or DlukGH/Dpvl strains had significantly decreased capacity to cause formation of membrane pores compared with those from the wild-type strain (Fig. 5A). Using conditions in which pore-forming capacity is retained in either of the single operon deletion strains (30 min incubation using more concentrated supernatants), culture supernatants from the DlukGH/Dpvl strain typically caused little or no formation of membrane pores (Fig. 5A). Complementation of the DlukGH/Dpvl strain with a plasmid containing lukGH restored a significant level of pore-forming capacity, thereby demonstrating that LukGH contributes to neutrophil plasma membrane poreformation (Fig. 5A, far right panel).
Although we observed significantly reduced pore formation in PMNs exposed to the highest concentration of culture supernatants from Dpvl or DlukGH strains (1:250 dilution at 5 or 15 min) (Fig. 5A), there was only a limited corresponding decrease in PMN lysis (as measured by LDH release) (Fig. 5B). For example, supernatants from cultures of Dpvl or DlukGH strains retained most  of their cytolytic capacity after 3 h of incubation with human PMNs-although there was a trend for decreased LDH release (Fig. 5B). This finding is consistent with our studies indicating that PMN pore formation does not necessarily correlate with cytolysis [26]. By contrast, culture supernatants from the DlukGH/Dpvl strain had zero cytolytic capacity, indicating synergy or cooperativity between PVL and LukGH (Fig. 5B). We note that culture supernatants from the Dpvl strain had significantly reduced cytolytic capacity by 6 h (LDH release was 58.067.9% for wildtype supernatants and 33.865.4% for those from the Dpvl strain, P,0.05), but the differential supernatant lysis between the two mutant strains is likely due to the selective overproduction of PVL in CCY media (Fig. 4D) [26,27]. The finding that there is synergy between PVL and LukGH merits further investigation.
Inasmuch as we identified LukGH by surface proteome analysis, we tested the hypothesis that surface-associated LukGH contrib-utes to rapid lysis of human PMNs after phagocytosis of USA300 [5]. Lysis of PMNs 6 h after phagocytosis of the Dpvl strain was comparable to that caused by the wild-type USA300 strain (LDH release was 63.2612.0% and 59.3617.0% for the wild-type and Dpvl strains, n = 10) (Fig. 5C). These findings are consistent with our previous studies demonstrating PVL does not contribute to lysis of PMNs after phagocytosis [5,23]. By comparison, DlukGH or DlukGH/Dpvl strains had significantly reduced capacity to cause lysis after uptake by PMNs (e.g., LDH release was 35.5615.9% following phagocytosis of DlukGH, P,0.01 versus LAC wild-type) (Fig. 5C). The complemented mutant strains, DlukGH::plukGH and DlukGH/Dpvl::plukGH, caused PMN lysis at a level similar to that of the wild-type strain (Fig. 5C). Although it was not possible to differentiate the relative contribution of surface-associated versus freely secreted LukGH in this process, there is a relative high abundance of surface-associated LukGH, whereas the other two-  (Table S1). Taken together, these data demonstrate that LukGH has potent cytolytic activity toward human neutrophils.

Concluding remarks
Rodriguez-Ortega et al. used a surface proteomics approach to identify proteins present on the surface of group A Streptococcus, some of which are potential vaccine candidates [9]. More recently, Solis et al. evaluated the surface proteome of S. aureus strain COL using methodology that reduces contamination with proteins from the cytoplasm [28]. Indeed, we recovered a higher percentage of peptides corresponding to cytoplasmic proteins than did Solis et al., but Mascot scores for the corresponding cytoplasmic proteins were significantly lower than those of cell-wall associated, extracellular, or membrane proteins (Fig. 2B). Many of the proteins identified here by our surface proteome analysis, such as Atl, ClfB, MecA, IsaA, IsdH, SasF, and Spa, were also reported to be surface-associated by Solis et al. [28]. However, to our knowledge, LukG and LukH have not been identified previously as exoproteins.
Herein we used a surface proteomics approach to identify this previously uncharacterized leukotoxin (LukGH) as an abundant exoprotein of USA300. LukGH was both surface-associated and present in USA300 culture media. Using isogenic USA300 mutants, we discovered that this novel leukotoxin has potent cytolytic activity toward neutrophils and can work synergistically with PVL to cause PMN lysis. Given the high abundance of LukGH in our surface protein preparations, and the critical role that neutrophils play in host defense against S. aureus infection, it is tempting to speculate that the toxin contributes to the enhanced virulence phenotype of USA300. Studies in animal infection models to demonstrate a role for LukGH in S. aureus virulence are ongoing.

Ethics statement
Human neutrophils were isolated from heparinized venous blood of healthy donors in accordance with a protocol approved by the NIAID Institutional Review Board for human subjects. Donors were informed of the procedure risks and provided written consent prior to enrollment.

Bacterial strains and culture conditions
We used USA300 strain LAC (Los Angeles County clone) for surfome analysis. LAC has been described and characterized previously [5] and is representative of the USA300 epidemic clone [47]. LAC was grown in trypticase soy broth (TSB) or CCY media (3% yeast extract, 2% Bacto-casamino acids, 0.21 M sodium pyruvate, 44 mM dibasic sodium phosphate, 3 mM monobasic potassium phosphate, pH 6.7) in a flask-to-media volume ratio of 5:1 at 225 rpm at 37uC. LAC was cultured overnight in TSB, diluted 1/200 into fresh media, and subcultured in either TSB or CCY using the same conditions until the desired growth phase was obtained (late-exponential phase in TSB for proteomics, earlystationary phase in CCY for culture supernatant assays with PMNs, and mid-exponential phase in TSB for PMN lysis assays with intact bacteria). Growth curve analysis in TSB was performed in triplicate for each strain, and bacteria were plated for enumeration on trypticase soy agar plates.
Sample preparation for proteomics analysis USA300/LAC was cultured in TSB to late-exponential phase of growth (OD 600 = 0.8521.0) in 25-or 100-ml volumes. Bacteria were harvested by centrifugation at 4,000 g for 15 min at 4uC, and washed 3 times with a volume of chilled 50 mM Tris, pH 7.5, that was equivalent to the original culture volume. The pellet was resuspended in 1 ml (25 ml culture) or 3 ml (100 ml culture) of chilled 50 mM Tris, pH 7.5, and aliquoted (1 ml/tube) into 1.5ml Protein Lo-Bind tubes (Eppendorf North America, Westbury, NY). Bacteria were pelleted at 16,100 g for 5 min at 4uC and resuspended in 0.3 ml digestion buffer (0.6 M sucrose buffered with 50 mM Tris, pH 7.5,) containing 2 mM dithiothreitol (DTT), and 10-12 mg mass spectrometry-grade trypsin (Promega, Madison, WI). Sucrose was included in the digestion buffer to facilitate swelling of the bacterial cells, which enhances trypsin digestion of surface proteins [9]. Bacteria were digested with trypsin for 1 h at 37uC. A 256 solution of SIGMAFast protease inhibitor cocktail (Sigma-Aldrich Company, St. Louis, MO) was prepared in digestion buffer; the cocktail was added to a final concentration of 16 following trypsin digestion. An aliquot of the digested sample before and after incubation at 37uC was diluted serially in 50 mM Tris, pH 7.5, and plated on trypticase soy agar plates to assess bacterial viability. Following trypsin digestion, bacteria were pelleted by centrifugation at 8,000 g for 10 min at 4uC. The supernatant containing tryptic peptides from USA300 surface proteins was clarified/sterilized by using a Microcon YM-100 centrifugal filter unit (Millipore, Billerica, MA) at 4uC for 25 min. The filtrate was frozen at 280uC until further processing.

Fractionation of tryptic peptides
Several approaches were used to prepare enriched peptide fractions. Trifluoroacetic acid (TFA) was added to the tryptic peptides from three 25-ml cultures to a final concentration of 0.2% and extracted using a reverse phase polymer peptide trap cartridge (Michrom Bioresources, Auburn, CA) to remove interfering salts and sucrose. The dried sample was dissolved in 50 ml of 50 mM formic acid/75% isopropanol and applied to polyhydroxyethyl A 10-ml solid phase extraction tip (Glygen Corporation, Columbia, MD). The analytes were eluted with 10 ml of 15 mM ammonium acetate, pH 3.5/3% acetonitrile, 0.1% TFA and fractionated by reverse phase chromatography using a Zorbax XDB 5 mm C18, 0.5 mm 6150 mm column on an 1100 series capillary HPLC system (Agilent Technologies, Santa Clara, CA). Linear gradients were run from 0.1% TFA/3% acetonitrile to 60% acetonitrile. Column effluent was monitored at 220 and 280 nm. The 12-ml column fractions were dried and dissolved in 4.5 ml of 0.1% acetic acid/50% methanol for analysis by mass spectrometry. In other experiments, the tryptic peptides from five 25-ml cultures were adjusted to pH 2.8 by the addition of 10% formic acid and applied to a Polysulfoethyl A 10-ml solid phase extraction tip (Glygen Corporation). Following a 100-ml wash with 0.2% formic acid/5% acetonitrile, the resin was eluted with 20 ml of 0.01, 0.05, 0.1, 0.5 and 1 M NaCl in the same solvent. The batches were dried, dissolved in 10 ml of 0.1% TFA/3% acetonitrile and each submitted to reverse phase chromatography as described. Ion exchange chromatography was also performed with a Zorbax SCX, 5 mm 2.1 mm 650 mm column on a 1100 series analytical HPLC system (Agilent Technologies) using the peptides from eleven 25-ml cultures. The column was developed with a gradient of NaCl from 0.01 to 1 M in 0.2% formic acid/5% acetonitrile. The 0.2-ml fractions were combined into seven separate pools which were qualitatively based on the UV peak patterns at 220 nm. After drying under vacuum 5 of the 7 pools were dissolved in 10 ml of 0.1% TFA/3% acetonitrile and submitted to reverse phase chromatography. The two higher salt pools were dissolved in 200 ml 0.1% TFA and desalted by reverse phase cartridge as described prior to capillary HPLC.

Protein identification by mass spectrometry
Protein identification, for 1D and/or 2D LC resolved peptides, was performed on reduced and alkylated, trypsin-digested S. aureus prepared as described above. HPLC recovered peptide pools were injected by direct infusion with a Nanomate (Advion BioSciences, Ithaca, NY), an automated chip-based nano-electrospray interface source coupled to a quadrupole-time of flight mass spectrometer, QStarXL MS/MS System (Applied Biosystems/Sciex, Framingham, MA). Computer-controlled data-dependent automated switching to MS/MS provided peptide sequence information. AnalystQS software (Applied Biosystems/Sciex) was used for data acquisition. Data processing and databank searching were performed with Mascot software (Matrix Science, Beachwood, OH) and a database containing forward and decoy sequences (RSUE database) that was constructed using the USA300_FPR3757 sequence.
To generate the complemented gene deletion strains, DNA encoding the ribosome binding site and lukGH ORF were PCRamplified from the USA300 genome and ligated into BamHI and NarI sites of the S. aureus expression vector pTX-15 [53]. The resulting plasmid (plukGH) was transformed into DlukGH and DlukGH/Dpvl to produce DlukGH::plukGH and DlukGH/ Dpvl::plukGH. Primers used to construct the lukGH complementation vector are as follows: pTX15_lukGH_F: TTATTCACATGTCGAGGATCCTCAACAAATATCA pTX15_lukGH_R: ATTCTATGTAGGGCGCCACTTTTATTACTTATTTC LukGH expression was confirmed by SDS-PAGE and immunoblot analysis as described below.

SDS-PAGE and immunoblotting
LAC wild-type and isogenic Dpvl, DlukGH, and DlukGH/Dpvl deletion strains were cultured to early-stationary phase of growth in CCY media (OD 600 = 2.0). Bacteria from 10 ml of culture were pelleted by centrifugation. Culture supernatant was aspirated, filter-sterilized, and stored at 280uC until use. Samples were boiled in Laemmli sample buffer [54] for 5 min and separated by 12.5% SDS-PAGE using Criterion precast polyacrylamide gels (Bio-Rad, Hercules, CA). Proteins were transferred to PVDF membranes using the iBlot transfer system (Invitrogen, Carlsbad, CA). Membranes were blocked using 10% goat serum in Trisbuffered saline containing 0.05% Tween-20 and antibody incubations were performed in blocking buffer diluted 1:4 in TBS. Washes were performed in triplicate for 15 min each between antibody incubations using wash buffer (250 mM NaCl, 10 mM Hepes, and 0.2% Tween-20). LukG was detected using affinity-purified rabbit IgG directed against a peptide region of the protein (LWAKDNFTPKDKMP) (GenScript Corporation, Piscataway, NJ) and a donkey anti-rabbit IgG-HRP secondary antibody (Jackson ImmunoResearch, West Grove, PA). PVL was detected as described previously [26]. Proteins were visualized using a Supersignal West Pico Horseradish Peroxidase Detection Kit (Pierce Biotechnology, Rockford, IL) and X-ray film (Phenix Research Products, Chandler, NC).

Subcellular localization of LukG
Bacteria from early-stationary phase (8 h) cultures were harvested by centrifugation (8,0006 g for 10 min), washed with deionized water, and disrupted by high pressure (18,000 psi) using a French Press. The cell lysate was treated with Protease Inhibitor Cocktail and Nuclease Mix (Amersham Bioscience Corp, Piscataway, NJ) and the particulate fraction, which includes cell envelope components, was isolated by centrifugation (25,0006 g for 30 min at 4uC). The clarified lysates, which contain cytoplasmic proteins, were stored at 280uC until used. The resulting pellet, containing cell envelope components, was resuspended in rehydration buffer (7 M urea, 2 M thiolurea, 2% triton X-100, 2 mM tributylphosphine, and 1% bromophenol blue) and incubated at ambient temperature for 2 h. These samples were clarified by centrifugation (25,0006 g, 30 min) and proteins were analyzed by SDS-PAGE and immunoblotting (labeled as ''membrane fraction'' in Fig. 3A). Detergent-insoluble material (pellet from above) was incubated for 3 h in TE buffer (50 mM Tris-HCl and 10 mM EDTA, pH 8.0) with lysostaphin (40 mg/ml). These samples were clarified by centrifugation (25,0006 g, 30 min) and soluble material was analyzed by SDS-PAGE (labeled as ''lysostaphin extract'' in Fig. 3A).
Culture supernatant proteins were precipitated with 9 vol of trichloroacetic acid:acetone (1:8 v/v for 2 h at 220uC) and then analyzed by SDS-PAGE and immunoblotting as described above (labeled as ''extracellular proteins'' in Fig. 3A).

Human PMN assays
Human neutrophils were isolated from heparinized venous blood of healthy donors using a published method [55]. PMN plasma membrane permeability (pore formation) was assessed by ethidium bromide (EtBr) uptake as described previously [23,26]. Neutrophil lysis either after incubation with CCY culture supernatant or following phagocytosis of serum opsonized USA300 was measured by the release of lactate dehydrogenase (LDH) using a Cytotoxicity Detection Kit (Roche Applied Sciences, Pleasanton, California) as described previously [23,26].

Statistical analyses
All statistical analyses were performed using GraphPad Prism. Comparison of Mascot score distribution across subcellular components was performed using the Kruskal-Wallis test with Dunn's post-test. One-way or repeated-measures ANOVA followed by Tukey's post-test was used to assess significance in PMN pore formation and lysis assays.

Table S2
The surface proteome (surfome) of USA300 strain LAC contains cell wall-associated, extracellular, membrane, and cytoplasmic proteins. A total of 113 proteins were identified in the LAC surfome. The peptide abundance ranking in Column A was determined by ranking the identified proteins by 1) Combined Mascot Score (Column G), then 2) Queries matched (Column E), then 3) Unique peptides matched (Column F). The FPR3757 number in Column C refers to the gene number in the SAUSA300_FPR3757 genome. The Percent Coverage (Column H) was determined by dividing the number of peptides identified by the total number of tryptic peptides in a given protein. SignalP 3.0 and PrediSi were used to determine if a Gram-positive signal sequence was present in each identified protein (Columns I and J). The GRAVY index (Column M) is a measure of the hydrophobicity of each identified protein. The PSORT localization (Column N) was assigned based upon the PSORT 2.0 localization for FPR3757. The experimental/knowledge localization (Column O) was assigned based upon 1) published data that demonstrate localization of the protein, or 2) PSORT 2.0 data if published data were not available. Whether a protein contained an anchoring domain was determined using AUGUR. The cell wall anchoring domain (Column P) was then identified by visual inspection of the primary protein sequence. The signal peptide cleavage site (Column Q) was identified using SignalP-NN. Found at: doi:10.1371/journal.pone.0011634.s004 (0.06 MB XLS)