Virulence Genes and Genotypic Associations in Nasal Carriage, Community-Associated Methicillin-Susceptible and Methicillin-Resistant USA400 Staphylococcus aureus Isolates

Abstract

It is not well understood why strains of community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA), a major cause of skin and soft tissue infections, became successful so quickly, overtaking the place of methicillin-sensitive S. aureus (MSSA) in many communities. To evaluate the genetic basis of differences in their virulence traits, 293 S. aureus isolates consisting of three cohorts, genotypically defined clinical CA-MRSA (n = 77), clinical MSSA (n = 103), and nasal carriage MSSA (n = 113), collected over a 19-year period in two Midwestern states in the United States, were (i) extensively genotyped and (ii) screened for 40 known virulence genes which included those for enterotoxins, leukocidins, hemolysins, and surface proteins and several newly identified putative toxin genes from the USA400 lineage of CA-MRSA. Genotypically, nasal carriage and clinical MSSA isolates were much more diverse than was the CA-MRSA group, which was found to be of USA400 lineage only. Virulence gene profiles of the three groups showed that CA-MRSA strains harbored significantly higher percentages (≥95%; P value, <0.05) of the sea, sec, sec4, seg2, seh, sek, sel, sel2, ear, ssl1, lpl10, lukSF-PV, lukD, lukE, and clfA genes than did the carriage and the clinical MSSA group (range, 0% to 58%). Genes of the enterotoxin gene cluster, seg, sei, sem, sen, and seo, were present in the clinical and carriage isolates but not in the CA-MRSA group. These results suggest that the presence of additional virulence factors in USA400 CA-MRSA strains compared to the nasal carriage and clinical MSSA strains probably contributed to their enhanced virulence.

Staphylococcus aureus both is a benign commensal and common pathogen in humans and is responsible for a variety of infections, ranging from superficial skin and soft tissue infections to bacteremia, endocarditis, and osteomyelitis (33). Based on its susceptibility to beta-lactams, S. aureus is commonly described as methicillin-susceptible S. aureus (MSSA) or methicillin-resistant S. aureus (MRSA) (15). Infections due to health care-associated MRSA (HA-MRSA) have been a problem since the 1970s; however, starting in the 1990s, new strains of MRSA, referred to as community-associated MRSA (CA-MRSA), appeared in community dwellers and were genetically different from HA-MRSA strains (7). Until recently, individuals who presented with infections due to CA-MRSA typically have had none of the established risk factors associated with HA-MRSA, such as recent hospitalization, surgery, dialysis, long-term care residence, or indwelling medical devices. Lately however, CA-MRSA strains have been reported from both the community and the health care settings (25, 45). Genotyping tools such as pulsed-field gel electrophoresis (PFGE), multilocus sequence typing (MLST), and spa typing have helped in distinguishing the genotypes of CA-MRSA strains from those of other S. aureus strains (2, 14, 26). In PFGE, SmaI-restricted S. aureus genomes are compared to determine their genetic relatedness and also compared against the reference USA genotypes (USA100, USA200, etc., up to USA1200) as described by the Centers for Disease Control and Prevention (CDC) (34, 53). Of these, CA-MRSA isolates mostly belong to USA300 and USA400 clones and in some cases USA1000 and USA1100 clones as well. HA-MRSA isolates generally belong to USA100, USA200, and USA500 (34, 35). One of the two major clones of CA-MRSA, USA400, recognized in the early 1990s and initially referred to as the MW2 clone, was the predominant CA-MRSA clone that initially circulated in the midwestern United States in the 1990s (8, 17, 40, 52). The second and more recent CA-MRSA clone, USA300, was first recognized in 2000 and has since spread throughout the world (54). More than a thousand MLST allelic profiles for S. aureus have been identified so far, of which CA-MRSA strains are primarily represented by sequence type 1 (ST1) (USA400) and ST8 (encompasses USA300 and USA500). Of the several thousand spa types in the Ridom database (http://spaserver.ridom.de/), the most predominant CA-MRSA spa types are t008 and t128.

CA-MRSA strains, besides their distinct PFGE, ST, and spa profiles, almost ubiquitously possess the Panton-Valentine leukocidin (PVL or lukSF-PV) genes, in contrast to only 1% to 5% of S. aureus strains overall having these genes (33, 50, 57). The PVL toxin has been implicated in many skin and soft tissue infections and lethal necrotizing pneumonia, but the exact role of PVL during S. aureus infections remains controversial (10, 28, 31, 58). The genome sequence of the CA-MRSA strain MW2 of the USA400 lineage showed the presence of several additional putative toxin genes (e.g., ear, sec4, sel2, seg2, ssl1 [set16], and lpl10) compared with other S. aureus strains that had been sequenced (1). Some of these toxin genes share homology with classic staphylococcal enterotoxin genes that encode pyrogenic exotoxins (49) typically produced during the post-exponential phase of growth, and the genes encoding these exotoxins are most often carried on plasmids, bacteriophages, or pathogenicity islands. The classic staphylococcal enterotoxin genes (sea, seb, sec, sed, see, seg, seh, sei, sej, sek, sel, sem, sen, and seo) are commonly found in strains of S. aureus (12, 23, 39, 42, 43, 46, 62). Pyrogenic exotoxin genes are common in S. aureus, and as many as 73% of S. aureus isolates carry at least one of the genes encoding a classic pyrogenic exotoxin; however, the distribution among various clonal types differs (3). Strains of USA400 CA-MRSA typically have been shown to possess the sea, sec, seh, and sek enterotoxin genes, whereas HA-MRSA strains usually carry the sed, seg, sei, sej, sem, sen, and seo enterotoxin genes (40). Currently, the distribution of the newly identified putative toxin genes (seg2, sel2, sec4, ssl1, and lpl10) (1) from the MW2 strain has not been reported from among CA-MRSA strains in general or from clinical and nasal carriage MSSA strains. Since CA-MRSA isolates are able to cause disease in humans without predisposing risk factors and have spread rapidly in communities, these strains may possess a greater number of toxin genes than do the other strains of S. aureus.

The aim of this study was to compare the genotypes of clinical CA-MRSA USA400, clinical MSSA, and colonizing nasal carriage MSSA isolates and determine the frequency and distribution of the classic enterotoxin genes as well as the new putative toxin genes in them. Our results showed that MSSA strains were much more diverse in their genotypes than were CA-MRSA USA400 strains. In addition, CA-MRSA USA400 strains possessed a distinct array of toxin genes compared to MSSA strains. These data may provide insight into the success of CA-MRSA USA400 and its ability to cause severe disease in previously healthy people.

MATERIALS AND METHODS

S. aureus strains.

Two hundred ninety-three S. aureus isolates collected over a 19-year period from 1987 to 2005 from outpatient and surveillance cultures from Wisconsin and the Minnesota Department of Health were included in the study. The collection periods for the three groups of the isolates differed but did overlap. These isolates were classified into three groups, (i) nasal carriage MSSA (n = 113), (ii) clinical MSSA (n = 103), and (iii) clinical CA-MRSA (n = 77), based on the collection and phenotypic criteria (Table 1). Sixty-six percent of nasal carriage strains were from Wisconsin and were collected from college-going healthy individuals. Anterior nares of these individuals were screened for S. aureus followed by determination of their resistance to methicillin by MIC testing using oxacillin. Putative MRSA strains were then confirmed using growth on cation-adjusted Mueller-Hinton II agar plates containing 6 μg/ml of oxacillin. None of the subjects were recently hospitalized or had exposure to health care environments. Thirty-four percent of the nasal colonization isolates were from the Minnesota Department of Health and were from a collection of MSSA isolates obtained from outpatients in the upper Midwest. Clinical MSSA strains were strains recovered from patients with confirmed disease due to an MSSA strain. They were mainly recovered from skin and soft tissue infections and in rare instances from invasive cases such as blood. The CA-MRSA strains were primarily (>90%) isolated from a variety of skin and soft tissue infections from outpatient clinics and were genotypically defined. gCA-MRSA strains in this study were strains whose PFGE was related to the MW2 PFGE type and harbored staphylococcal chromosome cassette mec (SCCmec) IV. In addition they were recovered from patients visiting outpatient clinics.

TABLE 1.

Grouping of 293 S. aureus isolates studied and their years of collection

Group	Collection period	No. of isolates	Collection site(s)
Clinical MSSA	1987-2006	103	WI, MN
Nasal carriage MSSA	2000-2005	113	Upper Midwest
CA-MRSA	1990-1999	77	WI

Genotypic methods.

All isolates were typed by PFGE, MLST, and spa typing (2, 14, 26). SCCmec types were determined for MRSA isolates by the method of Oliveira and de Lencastre (41). PFGE clonal groups were defined by the Tenover et al. (53) criteria and ≥80% genetic similarity using the Dice coefficient (1.25% tolerance) and UWPGMA (unweighted-pair group method using average linkages) method. An MLST-based clonal complex (CC) was defined as a group that shared six of the seven MLST alleles through the use of eBURST (http://eburst.mlst.net).

Screening for virulence factor genes.

All isolates were screened for 40 known and putative S. aureus virulence genes located on bacteriophages and pathogenicity and genomic islands on the S. aureus genome. Of the 40 virulence genes, the following have been shown to have clinical relevance: classical staphylococcal enterotoxin genes (sea to see), the toxic shock syndrome toxin gene (tst), enterotoxin gene cluster or egc (seg, sei, sem, sen, and seo), the collagen binding protein gene (cna), fibronectin binding protein genes (fnbA and fnbB), clumping factor genes (clfA and clfB), genes for the serine-aspartate repeat family of proteins (sdrC, sdrD, and sdrE), the intracellular adhesion protein gene (icaA), hemolysin genes (hla, hlb, hld, and hlgB), leukocidin genes (lukSF-PV, lukD, and lukE), and exfoliative toxin genes (eta and etb). Additionally seven putative virulence genes (ear, sec4, sel2, bsa, seg2, lpl10, and ssl1) were screened because of their speculated role in S. aureus virulence (1). The presence or absence of these genes was screened by a combination of singleplex or four multiplex PCRs. The list of primers for the virulence genes, their locations on the genome, and expected product sizes are given in Table S1 in the supplemental material. The PCR primers were commercially synthesized at Integrated DNA Technology (Coralville, IA). All the control strains were obtained from the Network on Antimicrobial Resistance in Staphylococcus aureus (NARSA). The multiplex PCRs included multiplex 1 (sea to see), multiplex 2 (seg to sej), multiplex 3 (sek to seo), and multiplex 4 (seg2, sec4, and ear). Singleplex PCRs were run for the remainder of the genes. Using the genomic DNA as templates, PCR was performed using the HotStarTaq Master Mix kit (Qiagen, Valencia, CA). The sea-to-see multiplex reaction mixtures were set up in a 50-μl PCR mixture and consisted of 25 μl of the HotStarTaq Master Mix; 40 pmol of each of the forward and reverse primers for sea and seb; 10 pmol of each of the forward and reverse primers for sec, sed, and see; 3 μl of 25 mM Mg²⁺; 7 μl of water; and 4 μl of the extracted DNA template. The seg-to-sej multiplex reaction mixture consisted of 25 μl of the HotStarTaq Master Mix; 10 pmol of each of the forward and reverse primers for seg, seh, sej, and sei; 0.5 μl of 25 mM Mg²⁺; 16.5 μl of water; and 4 μl of the extracted DNA template. For sek to seo, the reaction mixture consisted of 25 μl of the HotStarTaq Master Mix; 10 pmol of each of the forward and reverse primers for sek, sel, sem, sen, and seo; 1.0 μl of 25 mM Mg²⁺; 15 μl of water; and 4.0 μl of the template. The first three multiplex reaction mixtures went through an initial activation at 95°C for 15 min, followed by 35 cycles of denaturation at 94°C for 2 min, annealing at 57°C for 2 min, and extension at 72°C for 1 min. A final extension was carried out at 72°C for 7 min followed by a final hold of 4°C. A PCR mix for multiplex PCR 4 included the HotStarTaq Master Mix kit, sterile double-distilled water, 50 pmol of seg2, 40 pmol each of sec4 and ear primers, <1 μg of DNA template to the master mix that contained 2.5 U of HotStar Taq DNA polymerase, and 3.0 mM MgCl₂. The cycling conditions consisted of 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min.

At least two separate PCR amplifications per gene were performed for each S. aureus isolate to ensure reproducibility. The PCR products were electrophoresed on 1.5% agarose gels and stained with ethidium bromide to visualize the DNA bands. A 100-bp DNA ladder (New England Biolabs) was used to determine the size of each PCR product. All PCRs were done in a PE9700 Thermal Cycler (Perkin-Elmer, Wellesley, MA). A positive control for each respective gene and a negative control represented by a strain of E. coli were included for each run. A no-template control was also run for each PCR run to rule out the possibility of contamination of the reagents. A positive PCR was determined by the presence of an appropriately sized DNA band in the gel.

Identity of the amplicons.

At least two randomly chosen amplicons for each of the 40 virulence genes tested were sequenced to ensure that primers amplified the expected genes. Gene identities were confirmed by comparison with the DNA sequence database at GenBank using the BLAST search.

Statistical analyses.

All data analyses were carried out using the Statistical Analysis System (SAS, Cary, NC) and StatXact statistical software (Cytel Software Corp., Cambridge, MA). Fisher's exact test was used to derive the P value. Any P value less than 0.05 indicated statistical significance. Simpson's diversity index was determined as described previously (20).

RESULTS

Range of genotypic diversity in nasal carriage, clinical MSSA, and clinical CA-MRSA isolates.

The nasal isolates (n = 113) were represented by 41 MLST allelic profiles or 15 clonal complexes (6 major and 9 minor). A major CC was defined as a CC that was represented by at least 4% of the isolates in that group. The six major CCs were CC30 (29.2%), CC15 (15%), CC5 (14.1%), CC25 (9.7%), CC59 (9.7%), and CC45 (5.3%). The remaining CCs were CC1, CC8, CC97, CC1021, CC12, CC22, CC50, CC398, and CC509, and their frequency ranged from 1% to 3.5%. These isolates were also represented by 68 unique PFGE types or pulsotypes that could be grouped into 18 PFGE-based clonal groups (CGs) using the criteria of Tenover et al. (53). The distribution of major CGs in order of their percentages was as follows: CG-B = CC30 or USA200 (23.9%), CG-C = CC15 (15.9%), CG-D = CC5 (12.4%), CG-F = CC25 (9.7%), CG-G = CC59 (8.8%), CG-A = CC15 or USA400 (5.3%), CG-E = CC45 or USA600 (5.3%), and CG-L = CC30 (4.4%). The frequencies of the remaining nine minor CGs were below 3%, and they were as follows: CG-H = CC8 and CC30, CG-K = CC9, CG-P = CC9, CG-Q = CC1, CG-R = CC1021, CG-S = CC8, CG-WW = CC398, CG-YY = CC22, CG-Z = CC50, and CC-ZZ = CC509. By our analysis, the non-USA types were represented by 65.5% of the isolates. Fifty-three spa types were identified, of which the seven major ones were t084 (10.6%), t012 (9.7%), t078 (8%), t216 (8%), t021 (4.4%), t548 (4%), and t2917 (3.5%).

Clinical MSSA isolates (n = 103) were represented by 38 MLST allelic profiles, which were grouped into 12 CCs. Seven major CCs were CC30 (39.8%), CC5 (14.6%), CC45 (8.7%), CC121 (7.8%), CC1 (5.8%), CC15 (4.9%), and CC20 (4.9%). The remaining CCs were CC59, CC395, CC8, CC25, and CC97, and their frequency ranged from 1% to 2.9%. This group was also represented by 59 pulsotypes which were grouped into 16 CGs. Three clones matched with the USA types: CG-B = CC30 or USA200 (38.8%), CG-A = CC-1 or USA400 (7.8%), and CG-E = CC45 or USA600 (5.8%). The non-USA types were represented by 47.6% of the isolates. Their breakdowns were as follows: CG-J = CC121 (6.8%), CG-D = CC5 (5.8%), CG-K = CC9 (5.8%), CG-C = CC15 (4.8%), and CG-M = CC20 (4.8%); the frequencies of the remaining 9 CGs (CG-F = CC25, CG-G = CC59, CG-H = CC8 and CC30, CG-L = CC30, CG-P = CC9, CG = Q = CC1, CG-U = CC395, CG-V = CC395 and CC1148, and CG-Y = CC121) were below 3%. Seventy-two spa types were identified in this group, with five major ones being t012 (13.6%), t159 (5.8%), t018 (4.9%), t164 (3.8%), and t209 (3.8%). The percent ranges of the remaining spa types were from 0.9% to 2.9%.

The CA-MRSA isolates were the least diverse and were represented by nine PFGE types. However, they all belonged to CC1 (CG-A or USA400) but were still represented by six spa types: t128 (80.5%), t559 (6.5%), t125 (5.2%), t558 (5.2%), t175 (2.6%), and t555 (2.6%). Using the representative strains from each of the three groups, their comparative genotypic characteristics and diversity are summarized in Fig. 1. Overall, 18 CCs were identified in these three groups of isolates, of which CC20, CC121, and CC395 were not present in the nasal carriage group whereas CC12, CC22, CC50, C398, CC509, and CC1021 were absent in the clinical MSSA group (Table 2 ).

FIG. 1.

PFGE-based dendrogram of S. aureus isolates consisting of at least one unique spa type from each of the three cohorts: clinical CA-MRSA, clinical MSSA, and nasal carriage MSSA. The dendrogram also includes PFGE-based USA genotypes as reference strains. The dendrogram was created using 1.25% tolerance, the Dice similarity coefficient, and the unweighted-pair group method using arithmetic averages. Clonal relatedness was determined by ≥80% genetic similarity. PT, pulsotype.

TABLE 2.

Range of percentages of major CCs in three groups of isolates

CC	% of CC in isolate group:
	Nasal carriage MSSA	Clinical MSSA	CA-MRSA
CC1	3.5	5.8	100
CC5	14.1	14.5	NFa
CC8	2.7	1.9	NF
CC12	0.9	NF	NF
CC15	15.0	4.9	NF
CC20	NF	4.9	NF
CC22	0.9	NF	NF
CC25	9.7	1.9	NF
CC30	29.2	39.8	NF
CC45	5.3	8.7	NF
CC50	0.9	NF	NF
CC59	9.7	2.9	NF
CC97	2.7	1.9	NF
CC121	NF	7.8	NF
CC395	NF	<2.9	NF
CC398	0.9	NF	NF
CC509	0.9	NF	NF
CC1021	2.7	NF	NF
Singletons	0.9	1	NF

^aNF, not found.

Nasal carriage and clinical MSSA strains were genetically more diverse by all three typing methods, unlike CA-MRSA isolates, which showed very little diversity by PFGE and MLST. Interestingly, there was no significant difference in the percentages of unique PFGE types, clonal groups, or MLST allelic profiles between the clinical MSSA strains and nasal carriage isolates. However, clinical MSSA strains were represented by significantly higher percentages of spa types than were the nasal carriage isolates. Overall, S. aureus isolates that were susceptible to methicillin were more diverse than were the MRSA isolates by PFGE. In terms of distinguishing abilities of the three typing methods, spa typing and PFGE were comparable but more discriminatory than MLST for clinical MSSA strains (Table 3). However, once PFGE types were grouped into clonal groups according to the Tenover et al. (53) criteria, spa typing was found to be more discriminatory. Clinical MSSA strains were found to have a higher percentage of spa types than were nasal carriage isolates. In addition to the simple descriptive analysis of diversity of genotypes described above, the diversity indices (DI) of nasal carriage and clinical MSSA isolates were 0.988 and 0.946 by PFGE, 0.963 and 0.973 by spa typing, and 0.931 and 0.897 by MLST, respectively (Table 4). Nasal carriage isolates were genotypically more diverse by MLST and PFGE than were clinical MSSA isolates.

TABLE 3.

Number of pulsotypes, PFGE-based clonal groups, STs, and spa types in study isolates collected during 1987 to 2005

Group (n)	No. (% diversity) of:
	Pulsotypes	Clonal groups	STs	spa types
Nasal carriage MSSA (113)	71 (63)	18 (16)	41 (36)	52 (46)
Clinical MSSA (103)	59 (57)	17 (17)	38 (37)	64 (62)
CA-MRSA (77)	9 (12)	1 (1)	1 (1)	6 (7.8)

TABLE 4.

Simpson's index of diversity of nasal carriage MSSA, clinical MSSA, and CA-MRSA isolates by PFGE, MLST, and spa typing

Genotypic method	Index of diversity for isolate group (n):
	Nasal carriage MSSA (113)	Clinical MSSA (103)
spa typing	0.963	0.973
MLST	0.931	0.897
PFGE	0.988	0.946

Differential distribution of known virulence genes in CA-MRSA USA400 compared to clinical and nasal carriage MSSA isolates.

Representative agarose gels of the four multiplex PCRs are shown in Fig. S1A to D in the supplemental material. Overall, the rate of positivity of the 40 virulence genes showed a wide range, from 0% to 100% across the three groups. Some genes were preferentially present in one group versus the two other groups. Twelve of the 40 genes (ssl1, hla, hlb, hld, hlgB, clfB, cna, fnbA, icaA, sdrD, sdrE, and bsa) were more frequently (≥85%) present across all three groups than were sed, see, sej, etb, and fnbB, which were present in the range of 0% to 24%. In addition, the rate of positivity of genes seg, sei, sek, sel, sen, seo, and lukSF-PV varied widely and ranged from <10% to >80% in the three groups.

Several enterotoxin genes (sea, sec, sec4, seg2, sek, sel, and sel2), leukocidin genes (lukD, lukE, and lukSF-PV), adhesion genes (sdrD and clfA), and putative virulence genes (lpl10 and ear) were significantly associated with CA-MRSA compared to nasal carriage or clinical MSSA strains (P value of <0.05). Enterotoxin genes sed and see, including the enterotoxin gene cluster (egc), seg, sei, sem, sen, and seo; exotoxin genes (tst, eta, and etb); and the fibronectin binding protein gene (fnbB) were present in less than 1% of the CA-MRSA isolates. The toxic shock syndrome toxin gene (tst) and exfoliative toxin genes (eta and etb) were not present in the CA-MRSA group. Enterotoxin genes sed and sej were present at a low frequency, probably because the plasmid pIB485, which harbors these genes, may not be common in these groups of isolates. Genes sea and sek, which reside on ΦSa3mw, were almost always present in CA-MRSA isolates but infrequently present in MSSA isolates. Genes such as staphylococcal superantigen protein-like genes (ssl1), hemolysin genes (hla, hlb, hld, and hlgB), surface protein genes (clfB, cna, fnbB, icaA, and sdrE), and the bacteriocin gene (bsa) were present in nearly all three groups of the isolates. At the same time fnbB, although at a lower rate (14% to 24%), was present exclusively in MSSA strains.

In general, the MSSA clinical and nasal carriage strains have overall virulence profiles similar to those of the CA-MRSA group, except for sel2, tst, hlb, cna, lpl10, and sdrC, which were statistically different (P value, <0.05) when one was used as a reference group (Table 5 ). All of these genes except lpl10 had a significantly higher rate of positivity in nasal carriage isolates than in the clinical MSSA strains. To address the issue of multiple comparisons, we applied the Bonferroni correction to data in Table 5 and found that the results for genes sej, hlb, can, fnbA, and sdrE were no longer statistically significant in each paired comparison among the three groups.

TABLE 5.

Summary of the ranges of virulence gene positivity for each gene and percentages and numbers of gene positivity in each of the three groups along with the P values within each cohort

Gene	% of positivity range	P valuec	% of positivity (no. of positives); P value
			Nasal carriageb	Clinical MSSAa	CA-MRSA
Enterotoxin genes
sea	54-95	—	58 (66); <0.0001	54 (56); <0.0001	95 (73)
seb	3-30	—	26 (29); <0.0001	30 (31); <0.0001	3 (2)
sec	25-95	—	27 (30); <0.0001	25 (26); <0.0001	95 (73)
sec4	25-97	—	29 (33); <0.0001	25 (26); <0.0001	97 (75)
sed	0-13	—	8 (9); 0.0116	13 (13); 0.0007	0 (0)
see	0-3	—	0 (0); 1.000	3 (3); 0.2615	0 (0)
seg	0-90	—	90 (102); <0.0001	89 (92); <0.0001	0 (0)
seg2	27-96	—	27 (31); <0.0001	31 (32); <0.0001	96 (74)
seh	15-100	—	15 (17); <0.0001	17 (17); <0.0001	100 (77)
sei	1-91	—	89 (101); <0.0001	91 (94); <0.0001	1 (1)
sej	1-9	—	9 (10); 0.0300	9 (9); 0.0452	1 (1)
sek	8-96	—	12 (14); <0.0001	8 (8); <0.0001	96 (74)
sel	9-96	—	9 (10); <0.0001	9 (9); <0.0001	96 (74)
sel2	25-97	0.0294	40 (45); <0.0001	25 (26); <0.0001	97 (75)
sem	0-47	—	42 (47); <0.0001	44 (45); <0.0001	0 (0)
sen	0-81	—	75 (85); <0.0001	81 (83); <0.0001	0 (0)
seo	0-82	—	76 (86); <0.0001	82 (84); <0.0001	0 (0)
Exotoxin genes
ssl1	86-100	—	91 (103); 0.0061	86 (89); 0.0003	100 (77)
tst	0-78	0.0021	78 (88); <0.0001	58 (60); <0.0001	0 (0)
eta	0-56	—	54 (61); <0.0001	56 (58); <0.0001	0 (0)
etb	0-22	—	22 (25); <0.0001	18 (19); <0.0001	0 (0)
Leukocidin genes
lukSF-PV	0-100	—	0 (0); <0.0001	3 (3); <0.0001	100 (77)
lukD	44-100	0.0010	66 (75); <0.0001	44 (45); <0.0001	100 (77)
lukE	62-100	0.0279	77 (87); <0.0001	62 (64); <0.0001	100 (77)
Hemolysin genes		—
hla	100	—	100 (113); 1.0	100 (103); 1.0	100 (77)
hlb	91-100	0.0279	98 (111); 0.5154	91 (94); 0.0109	100 (77)
hld	100	—	100 (113); 1.0	100 (103); 1.0	100 (77)
hlgB	100	—	100 (113); 1.0	100 (103); 1.0	100 (77)
Surface protein genes
clfA	70-99	—	76 (86); <0.0001	70 (72); <0.0001	99 (76)
clfB	97-100	—	97 (110); 0.2731	99 (102); 1.0	100 (77)
cna	95-100	0.0234	100 (113); 1.0	95 (98); 0.0721	100 (77)
fnbA	92-100	—	92 (104); 0.0116	95 (98); 0.0721	100 (77)
fnbB	0-24	—	24 (27); <0.0001	14 (14); 0.0003	0 (0)
icaA	99-100	—	100 (113); 0.4053	100 (103); 0.4278	99 (76)
lpl10	55-100	0.0171	55 (62); <0.0001	71 (73); <0.0001	100 (77)
sdrC	17-61	0.0008	82 (93); <0.0001	61 (63); <0.0001	17 (13)
sdrD	85-100	—	88 (100); 0.0010	85 (88); 0.0002	100 (77)
sdrE	90-100	—	96 (109); 0.1481	90 (93); 0.0054	100 (77)
Other genes		—
bsa	100	—	100 (113); 1.0	100 (103); 1.0	100 (77)
ear	18-97	—	18 (20); <0.0001	21 (22); <0.0001	97 (75)

^aP values were determined using CA-MRSA (reference) versus clinical MSSA.

^bP values were determined using CA-MRSA (reference) versus nasal carriage isolates.

^cP values were determined using clinical MSSA (reference) versus nasal carriage isolates. —, data not shown.

We also identified six nasal carriage and eight clinical MSSA isolates whose genotypes were similar to that of the USA400 clone. While the six nasal carriage MSSA strains were isolated in the year 2000 or later, two of the clinical MSSA strains (WSLH-6 and MNDH-77) in this group were recovered in 1987 and 1998, respectively, a period that either predated most of the CA-MRSA outbreaks or overlapped with the early outbreak. The virulence profiles of these 14 isolates were compared against those of the three representative CA-MRSA isolates of the USA400 lineage (Table 6). Notably, both WSLH-6 and MNDH-77 isolates harbored lukSF-PV, lpl10, ssl1, sea, sek, seh, sec4, ear, and sel2, some of the genes more frequently present in USA400 CA-MRSA strains (1, 40). Unlike the WSLH-6 isolate, MNDH-77 also harbored two of the egc genes (seg and sei) besides eta, etb, and tst.

TABLE 6.

Comparison of virulence profiles of nasal carriage MSSA and clinical MSSA USA400 pulsed-field pulsotypes with three representative strains of USA400 CA-MRSA^a

Isolate group and yr	Sample	seh	ssl1	lpl10	sdrC	sdrD	sdrE	ear	sec	sec4	sel	sel2	clfa	hla	bsa	lukSF-PV	lukD	lukE	sea	seg2	sek	hlb	hld	hlgB	fnbB	fnbA	clfB	icaA	cna	seb	sed	see	seg	sei	sej	sem	sen	seo	eta	etb	tst
Nasal MSSA
2003	UWL-57-S	+	+	+	+	+	+	−	−	−	−	−	+	+	+	−	+	+	+	−	+	+	+	+	−	+	+	+	+	+	−	−	−	−	−	−	−	−	−	−	−
2004	UWL-83-S	+	+	+	+	+	+	−	−	−	−	−	+	+	+	−	+	+	+	+	+	+	+	+	−	+	+	+	+	+	−	−	−	−	−	−	−	−	−	−	+
2003	UWL-33-S	−	+	−	+	+	+	−	−	−	−	−	+	+	+	−	+	+	+	−	−	+	+	+	−	+	+	+	+	−	+	−	+	+	+	+	+	+	+	−	+
2005	UWL-92-S	−	+	+	+	+	+	−	−	−	−	+	+	+	+	−	+	+	+	−	−	+	+	+	−	+	+	+	+	−	−	−	−	−	−	−	+	+	−	−	−
2001	MNDH-125-S	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	+	+	+	+	+	+	−	+	+	+	+	+	−	−	+	+	−	+	+	+	+	+	+
2000	MNDH-97-S	−	+	−	+	+	−	+	−	−	−	−	+	+	+	−	+	+	−	−	−	+	+	+	−	−	+	+	+	+	−	−	+	+	−	+	+	+	+	+	+
Clinical MSSA
1998	MNDH-77-S	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	+	+	+	−	−	+	+	−	−	−	−	+	+	+
1987	WSLH-6-S	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	+	+	−	−	−	−	−	−	−	−	−	−	−	−
2004	MC-112-S	+	+	+	+	+	+	−	−	−	+	−	+	+	+	−	−	+	+	+	−	+	+	+	+	+	+	+	+	−	−	−	+	+	+	−	−	−	+	−	+
2002	WSLH-32-S	+	+	+	+	+	+	+	−	−	−	−	+	+	+	−	+	+	+	+	+	+	+	+	−	+	+	+	+	+	−	−	−	+	−	−	−	−	+	−	+
2003	WSLH-35-S	+	+	+	+	+	+	−	−	−	−	−	+	+	+	−	+	+	−	−	−	+	+	+	−	+	+	+	+	+	−	−	−	−	−	−	−	−	+	−	+
2004	MC-127-S	−	+	−	+	+	+	−	−	−	−	−	+	+	+	−	+	+	+	−	−	+	+	+	−	+	+	+	+	−	−	−	+	+	−	+	+	+	+	−	+
1997	MNDH-31-S	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	+	+	+	+	+	+	−	+	+	+	+	+	−	−	+	+	−	+	+	+	−	+	+
2004	MC-118-S	+	+	+	+	+	+	−	−	−	−	−	+	+	+	−	+	+	+	+	−	+	+	+	−	+	+	+	+	−	−	−	+	+	+	+	+	+	+	−	+
CA-MRSA
1992	34-R	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	+	+	−	−	−	−	−	−	−	−	−	−	−	−
1993	88-R	+	+	+	+	+	+	−	−	−	+	−	+	+	+	+	+	+	+	+	−	+	+	+	−	+	+	+	+	−	−	−	−	−	−	−	−	−	−	−	−
1994	169-R	+	+	+	+	+	+	+	+	+	−	+	+	+	+	+	+	+	−	−	+	+	+	+	−	+	+	+	+	−	−	−	−	−	−	−	−	−	−	−	−

^a−, PCR negative; +, PCR positive.

DISCUSSION

This study presents the comparative analysis of genotypes and virulence profiles of 293 S. aureus isolates representing three major phenotypic groups of S. aureus—nasal carriage MSSA, clinical MSSA, and clinical CA-MRSA strains—collected over a 19-year period. Specifically, we determined the breadth of genotypic diversity for each group and assessed whether or not the virulence profile of CA-MRSA strains was significantly different from that of the MSSA groups. Genotyping using PFGE, MLST, and spa typing showed that the clinical and nasal carriage MSSA strains were highly diverse compared to the CA-MRSA group, which was found to be highly clonal. All CA-MRSA isolates belonged to the USA400 lineage only. This result was not entirely surprising as USA400 clones have primarily been reported from the midwestern United States during the collection period of 1990 to 1999 and continued to be reported from the midwestern United States and Alaska (8, 17, 40, 50, 52; S. K. Shukla et al., unpublished data). Unfortunately, we did not have USA300 strains in the study sample due to the time period (1990 to 1999) in which the gCA-MRSA isolates were collected. It should be noted that USA300, which was first identified in 2000, has now become the predominant CA-MRSA strain in most geographic areas in the United States, including the Midwest. Equally important was that we did not identify any clinical MSSA strains belonging to the USA300 clone in our collection.

Our study demonstrated that the virulence profile of the CA-MRSA USA400 strains was significantly different from those of the nasal carriage and clinical MSSA isolates based on the presence of a number of pyrogenic superantigens and surface proteins in addition to the putative toxin genes identified in MW2. Fourteen of the 40 genes were significantly associated with the CA-MRSA group (Table 5). These genes included multiple enterotoxin genes, leukocidin genes, adhesion genes sdrC and clfA, and a few putative virulence genes. Specifically, all genes except the enterotoxin gene see; hemolysin genes hla, hld, and hlgB; and surface protein genes clfB, cna, and icaA (Table 5) had a higher rate of positivity in CA-MRSA USA400 isolates than in either clinical or nasal carriage MSSA isolates. The presence of additional virulence genes, particularly those that seem to provide redundant functions, could help to explain why strains of CA-MRSA USA400 are found to be more virulent than S. aureus strains in general. While expression of virulence genes is likely to be regulated by several factors, including host factors, one can speculate that additional genes within the same functional class (e.g., toxins, adhesion factors, etc.) could provide a synergistic and/or complementary effect.

Several studies have described the evolution and virulence factor genes of HA-MRSA strains (15, 18); however, little has been done to examine these genes in CA-MRSA strains. Of the toxin-related virulence factor genes that have been examined in CA-MRSA strains, much has been done to characterize lukSF-PV, encoding the PVL. lukSF-PV has been shown to be conserved among most CA-MRSA isolates tested, but these same genes have been shown to be present in a low number of other S. aureus isolates (13, 30, 31, 47, 50, 57). Since this gene was present in only 1 to 5% of S. aureus strains in general but almost always in CA-MRSA USA400 strains isolated in the 1990s from the United States, it had become a molecular marker for USA400 isolates isolated during that time period (40, 50, 57). However, some recent studies from the United States and other parts of the world showed that the presence of PVL may not be the most definitive marker for recent CA-MRSA strains, including USA400 (32, 48, 55, 61). The results from our study found that 100% of CA-MRSA USA400 isolates and less than 1% of MSSA isolates possessed lukSF-PV, supporting previous studies (50, 57). The PVL has been shown to cause severe tissue damage and has been implicated in lung damage associated with strains causing severe necrotizing pneumonia (11, 19, 28, 31, 60). However, both a sepsis and an abscess model of infection with isogenic strains demonstrated that PVL-negative strains were as virulent as the PVL-positive parental strains (58). Thus, the role of PVL as a virulence factor is still under investigation. Besides PVL, α-hemolysin, α-type phenol-soluble modulins (PSMs), and the arginine catabolic mobile element (ACME) have been studied for their roles in the pathogenesis of CA-MRSA (9, 11). Lethal pneumonia may be tied to levels of α-hemolysin expressed in USA300 and USA400 strains (5). The PSMs expressed by USA300 and USA400 strains are thought to attack human neutrophils (59). Montgomery et al. (37), using pneumonia and the skin infection model in rabbits, showed that ACME is not associated with enhanced virulence. In contrast, Diep et al. (11) showed that deletion of ACME affects the pathogenicity and fitness of USA300. ACME, which was identified in the USA300 lineage, was not detected in 10 random samples of CA-MRSA strains from this study (data not shown).

Other toxins produced by S. aureus can also cause damage to host tissues. One class includes the staphylococcal pyrogenic exotoxins that can cause direct as well as indirect damage through overstimulation of the immune system (49). These exotoxins function as superantigens that keep the immune system from recognizing and attacking the invading S. aureus. A wide variety of toxins produced by S. aureus are considered to be pyrogenic exotoxins, including staphylococcal enterotoxins and toxic shock syndrome toxin (4).

Fourteen classical enterotoxin genes (sea, seb, sec, sed, see, seg, seh, sei, sej, sek, sel, sem, sen, and seo) have been previously identified in S. aureus and were screened for in this study (12, 23, 42, 43, 46, 62). The sea, seb, sec, seh, and sek genes have been reported to be found in the majority of CA-MRSA USA400 isolates from the United States collected prior to 2003 (17, 40), in contrast to USA400 isolates recovered during 2005-2006 (32). However, our results demonstrated that the CA-MRSA isolates harbored a significantly higher percentage not only of the sea, sec, seh, and sek genes as expected but also of the sec4, seg2, sel, sel2, lukSF-PV, lukD, lukE, clfA, and sdrD genes than did either the nasal carriage or clinical MSSA groups. Interestingly, the clinical and nasal carriage MSSA groups possessed a significantly greater frequency of the seb, sed, seg, sei, sej, sem, sen, seo, tst, eta, etb, fnbB, and sdrC genes than did the CA-MRSA group. The higher frequency of egc in the MSSA cohorts and not in CA-MRSA strains suggests that they are acquired as part of egc as shown previously (24). A study by van Belkum et al. (56) determined that the egc genes were present in 63.7% of carriage isolates and 52.9% of invasive isolates. It is generally believed that the presence of this gene cluster is not associated with severe infections but probably helps in the carriage potential of an S. aureus strain (16, 22, 56). However, a case report of a severe case of necrotizing fasciitis in a leg of a diabetic patient that resulted in the amputation of the leg was caused by an MSSA strain that lacked most of the common virulence genes except the egc genes, suggesting that the products of egc may play a role, in some cases, in severe infections, especially in compromised patients (38). However, the proteins encoded by the egc genes seem to be produced in smaller amounts than are the other well-studied enterotoxins, inducing a lower immunologic response among human hosts (22) that may allow strains possessing the egc element to live a commensal lifestyle in healthy humans. In our study, we did not see any significant difference in the prevalence of egc between the carriage and clinical MSSA groups. The three other enterotoxin genes (seb, sed, and sej) were found in a much smaller percentage of the MSSA isolates but were almost always absent in CA-MRSA isolates. The seb gene encodes an enterotoxin that is known to be a potent superantigen (4, 49) that can overstimulate the host's immune system and be detrimental to a commensal lifestyle for the MSSA isolates. The function of sed and sej is not well established.

The success of the USA400 lineage of CA-MRSA in the midwestern United States and its enhanced virulence could be due to a number of reasons. First of all, this group of isolates harbors more enterotoxin and surface protein genes than do the clinical MSSA strains or the nasal carriage strains in the study. It has been shown that sea and sec are more frequently present in CA-MRSA than in HA-MRSA strains (17, 40), and septic shock appears to be associated with the presence of sea (16). Enterotoxins A and J along with virulence genes fnbA, can, sdrE, sea, sej, eta, hlg, and ica tend to be more common in invasive isolates (44). In a report by Kravitz et al., three of the five cases of purpura fulminans due to either MSSA or MRSA strains produced staphylococcal enterotoxin C (SEC) and PVL (27). SEA and SEC tend to produce higher immunological responses leading to host tissue damage (4, 49) than do other enterotoxins (16). In this context, the presence of the sea and sec genes along with other genes such as seh, sek, sel, sel2, lukSF-PV, lukD, lukE, clfA, and sdrD suggests that several virulence genes may be working in unison to facilitate or prolong an infection. It is also possible that the presence of multiple toxin and adhesion proteins could afford a redundancy of virulence arsenal to the pathogen to mount a more potent infection. Contradictory results of the virulence studies in animal models for PVL and ACME (10, 11, 19, 31, 37, 58) suggest that virulence in S. aureus is not likely to be associated with one toxin or surface protein but a cumulative or synergistic effect of several virulence factors to initiate, maintain, and prolong an infection. It is also very likely that the genomic background of a strain contributes to the overall virulence of an S. aureus strain (29). It was interesting that the nasal carriage and clinical MSSA isolates belonging to the USA400 genotype harbored one or more of the genes egc, eta, and etb, unlike the CA-MRSA isolates (Table 6). Only one of the USA400 MSSA strains harbored fnbB, typically found in USA300 strains. It is possible that a lack of these genes in the USA400 CA-MRSA strains, along with the gain of the SCCmec element, could have aided in the success of the CA-MRSA USA400 clone at the time in the midwestern United States. It is worth mentioning that now USA300 is the predominant CA-MRSA clone in most parts of the United States (54). Interestingly, introduction of USA300 in a correctional facility resulted in a clonal shift in the CA-MRSA clone, i.e., from USA400 to USA300 in that facility (51). This displacement of ST1:USA400 by the ST8:USA300 clone in different parts of the United States may be due to ACME in its genome, which allows enhanced survival on skin (9).

Although the roles that enterotoxins play in skin and soft tissue infections are not known, they have been shown to influence the development of secondary staphylococcal infections due to their proinflammatory effects that trigger increased inflammation (36). Although the other toxin genes have been described in S. aureus, their presence has not been tied to virulence. An exciting new finding that may offer some significant clues to the pathogenicity of the USA400 type of CA-MRSA was the presence of several new putative toxin genes (for example, seg2, sel2, sec4, ear, lpl10, and ssl1) in the CA-MRSA isolates that were present in lower percentages in both cohorts of MSSA strains. Each of the genes was found on pathogenicity islands on the genome of MW2 (1). Pathogenicity islands are distinct genetic elements that encode virulence factors found on bacterial genomes (21). The seg2 gene was found on the ΦSa3 pathogenicity island, while the lpl10 and ssl1 genes were found on vSaα, and the sel2, sec4, and ear genes were found on a separate pathogenicity island labeled vSa3. The functions of these genes have not yet been determined, and yet some sequence homology to the staphylococcal enterotoxins may help to elucidate the functions of the proteins encoded by the new genes (1). Furthermore, a proteomic study has shown antibodies to the Sel2 and Ear proteins in patients infected with CA-MRSA, suggesting a role for both proteins in CA-MRSA pathogenicity (6).

Acquisition of one or more of the new pathogenicity islands may play a role in CA-MRSA infections. Until now, there have been no studies that have looked at the distribution of these pathogenicity islands in CA-MRSA or any other S. aureus isolates. Our results show that the newly discovered toxin genes were found in most CA-MRSA isolates from the USA400 lineage. The MSSA cohorts displayed much smaller detection percentages for the seg2, sel2, sec4, and ear genes, while percentages for ssl1 and lpl10 were relatively higher in these groups but significantly less than those in the CA-MRSA cohort. These data suggest that the vSaα pathogenicity island (ssl1 and lpl10) seems to be widely dispersed throughout S. aureus isolates; therefore, it is unlikely that the presence of this island alone is responsible for the pathogenicity of CA-MRSA USA400 isolates. It seems more likely that the other islands are acting in unison to affect the pathogenicity of CA-MRSA isolates compared to the MSSA populations.

The presence of these genes among the CA-MRSA isolates and their absence among MSSA isolates suggest that the virulence of the CA-MRSA USA400 pulsed-field type may be highly dependent on this group of new putative toxin genes and that the virulence of this CA-MRSA type could be tied to the synergistic activities of the proteins encoded by the variety of these genes. Peacock et al. (44) postulated that the presence of certain virulence genes among S. aureus isolates is dependent upon the rates of gene acquisition and the cost to biological fitness of the organism itself.

In summary, our study showed that isolates belonging to the CA-MRSA USA400 lineage were genetically less diverse than the clinical and carriage MSSA isolates and harbored several additional virulence genes which may have helped them in efficient host acquisition and subsequent infection.