Comparative Genomic Analysis of Staphylococcus aureus FORC_001 and S. aureus MRSA252 Reveals the Characteristics of Antibiotic Resistance and Virulence Factors for Human Infection

Abstract

Staphylococcus aureus is an important foodborne pathogen that causes diverse diseases ranging from minor infections to life-threatening conditions in humans and animals. To further understand its pathogenesis, the genome of the strain S. aureus FORC_001 was isolated from a contaminated food. Its genome consists of 2,886,017 bp double-stranded DNA with a GC content of 32.8%. It is predicted to contain 2,728 open reading frames, 57 tRNAs, and 6 rRNA operons, including 1 additional 5S rRNA gene. Comparative phylogenetic tree analysis of 40 complete S. aureus genome sequences using average nucleotide identity (ANI) revealed that strain FORC_001 belonged to Group I. The closest phylogenetic match was S. aureus MRSA252, according to a whole-genome ANI (99.87%), suggesting that they might share a common ancestor. Comparative genome analysis of FORC_001 and MRSA252 revealed two non-homologous regions: Regions I and II. The presence of various antibiotic resistance genes, including the SCCmec cluster in Region I of MRSA252, suggests that this strain might have acquired the SCCmec cluster to adapt to specific environments containing methicillin. Region II of both genomes contains prophage regions but their DNA sequence identity is very low, suggesting that the prophages might differ. This is the first report of the complete genome sequence of S. aureus isolated from a real foodborne outbreak in South Korea. This report would be helpful to extend our understanding about the genome, general characteristics, and virulence factors of S. aureus for further studies of pathogenesis, rapid detection, and epidemiological investigation in foodborne outbreak.

Introduction

Staphylococcus aureus is a well-known foodborne pathogen that is a leading cause of morbidity and mortality around the world [1,9]. Occasionally, S. aureus acts as an opportunistic pathogen to cause various diseases ranging from wound-related infections (including boils, furuncles, styes, and impetigo) to life-threatening conditions (such as endocarditis, necrotizing pneumonia, osteomyelitis, bacteremia, and sepsis) [13]. S. aureus can be transmitted to humans by skin-to-skin contact. It then expresses a multiple variety of virulence factors that facilitate attachment, colonization, cell-cell interactions, tissue damage, dissemination through the body, and immune evasion [11]. S. aureus is also a common cause of foodborne illnesses such as nausea, diarrhea, and abdominal cramps due to enterotoxin production [3,32]. However, the abuse of various antibiotics, generally in hospitals, has caused the emergence of antibiotic-resistant strains of S. aureus, especially methicillin-resistant strains. In the late 1960s, the methicillin-resistant S. aureus (MRSA) emerged owing to the widespread use of methicillin and semisynthetic β-lactam-containing antibiotics. To date, some Staphylococcus strains have developed resistance to more than 20 different antibiotics [4,17]. The emergence of hypervirulent community-acquired MRSA strains has become a major concern to the global health community, and has highlighted the critical need for novel methods for control and treatment. Therefore, further phenotypic and genotypic understanding of S. aureus is required to prevent foodborne illness outbreaks. To date, 40 complete S. aureus genome sequences have been reported, which has revealed specific gene information regarding the roles of various virulence factors and antibiotic resistance [24]. In addition, comparative genomic analysis demonstrated the presence of various antibiotic resistance-related genes and similar virulence factors among strains. Outbreaks of food-poisoning have been reported in many types of food [12,18,25,30]. However, no genomic study has reported pathogenic S. aureus strains that caused foodborne outbreaks in South Korea.

In the current study, S. aureus FORC_001 was isolated from S. aureus-contaminated parched beans with soybean sauce (Kongjang) in South Korea, and its genome was sequenced and analyzed. Furthermore, its complete genome sequence was compared in silico with those of S. aureus strains reported previously, to further understand the antibiotic resistance and virulence factors involved in the pathogenesis of human infections with S. aureus FORC_001. Moreover, a comparative phylogenetic tree revealed the evolutionary relationship between S. aureus FORC_001 and other S. aureus strains. This is the first report of the complete genome sequence and a comparative analysis of a S. aureus strain that was isolated from a contaminated Korean food. These results will extend our knowledge regarding the virulence and antibiotic resistance of S. aureus in South Korea.

 

Materials and Methods

Sample Collection and Bacterial Strain Isolation

S. aureus FORC_001 was initially isolated from S. aureus-contaminated parched beans with soybean sauce by the Public Health and Environmental Research Institute in Inchon (Isolate No. 1000071), South Korea. It was cultivated routinely at 30℃ with shaking in Brain Heart Infusion (BHI; Difco, Detroit, MI, USA) medium for 12 h. Agar plates were prepared with the supplementation of 1.5% BHI agar (Difco).

Genomic DNA Extraction and Whole-Genome Sequencing

Genomic DNA was isolated and purified using an UltraClean Microbial DNA Isolation Kit (MoBio, Carlsbad, CA, USA) according to the manufacturer’s instructions. The concentration and purity of the extracted DNA were determined using a NanoDrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA). Approximately 5 µg of the extracted genomic DNA was sheared mechanically into 8-12 kb fragments using the Hydroshear system (Digilab, Marlborough, MA, USA). SMRTbell libraries were prepared for each sample by ligating hairpin adaptors to both ends using a PacBio DNA Template Prep Kit 2.0 (3-10 kbp) for SMRT sequencing with C2 chemistry on a PacBio RS II (Pacific Biosciences, Menlo Park, CA, USA). Libraries were purified using 0.45× AMPure XP beads to remove short inserts sized <1.5 kb. The size distribution of the sheared DNA template was characterized using an Agilent 12000 DNA kit (Applied Biosystems, Santa Clara, CA, USA), and the concentration was determined using Invitrogen Qubit (Carlsbad, CA, USA). The sequencing primers were annealed to the templates at a final concentration of 5 nM template DNA, and DNA polymerase enzyme C2 was added according to the manufacturer’s recommendations for small-scale libraries. A DNA/Polymerase Binding Kit P4 (Pacific Biosciences) was used to load the enzyme template-complexes and libraries onto 75,000 zero-mode waveguides (ZMWs) following the instructions in the complex setup and loading calculator provided by the manufacturer. A DNA sequencing reagent 2.0 kit (Pacific Biosciences) was used to sequence SMRT cells using a 120-min sequence capture protocol and a stage start to maximize the subread length with PacBio RS II. The sequencing reads obtained from the RS II sequencing platform were assembled using the SMRT portal system [6]. The whole microbial genome assembly was conducted using the HGAP assembly-3 algorithm with default parameters, except for genome size parameter (3 Mb). After genome assembly using the PacBio RS II sequencing system, one contig was generated with the coverage of about X207. BLAST and MUMmer3 [8] analyses were used to define the orientation and direction of the assembled sequence by comparisons with known reference genomes. The polished sequence was then curated manually using Bioedit software, based on the results of this alignment [16].

Bioinformatics

The Rapid Annotation using Subsystem Technology (RAST) server was used for ORF prediction/annotation and tRNA/rRNA prediction, and the results were stored in a GenBank file format. Subsequently, the SEED viewer [10] was used for the subsystem functional categorization of the predicted ORFs and for visualization of the results [2]. In addition, Clusters of Orthologous Groups(COG) analysis was performed using WebMGA [34]. Genome sequence and annotation data were handled using Artemis16 [26] and selected genomes were compared using ACT13 [5]. A circular genome map was generated using GenVision (DNASTAR, Madison, WI, USA) based on the information provided by the genome annotation, and average nucleotide identity (ANI) values were calculated using JSpecies [15]. A phylogenetic tree was built using housekeeping genes [7].

Comparative Genomics

A total of 40 gapless S. aureus chromosome sequences were obtained from the NCBI database (http://www.ncbi.nlm.nih.gov/genome/genomes/154), and their ANI [15] with S. aureus strain FORC_001 was calculated. To calculate the ANI between two genomes, the query genome sequence was cut randomly into 1,020 nucleotide fragments, and each was BLAST searched against the subject genome; ANI was then defined as the mean identity of all BLASTN matches that showed more than 30% DNA sequence identity over regions at least 70% of their length. The genome of strains with more than 95% ANI values should belong to the same species. Genomic islands (GIs) were identified using Colombo 3.8 [33]. A genome tree was constructed using the R software, and the closest genome sequences were calculated according to ANI values using the unweighted pair group method. Three genome sequences were selected as the closest genomes in a homology group (Group I) based on ANI values, and were compared with S. aureus strain FORC_001. The matched DNA region in the subject contig was extracted and saved as a homolog. The unmatched DNA regions were also determined using the same method, and were confirmed using ACT13 [5]. In addition to the ANI genome tree, an additional phylogenetic tree of the Group I strains was generated using MEGA6 [31] with multilocus sequence typing genes [7]. Nucleotide substitution analysis using Nei’s unweighted method I [27] was performed to determine the evolutionary relationships between specific genes in two S. aureus genomes. The genes were identified as positive selection when the ratio of nonsynonymous nucleotide substitutions (dN) to synonymous nucleotide substitutions (dS) was more than 1 [21]. The genes that encoded virulence factors were identified and categorized into functional groups using annotated genome information.

Nucleotide Sequence Accession Number

The complete genome sequence and annotation data of S. aureus FORC_001 have been deposited with the GenBank Data Library under the accession number CP009554.

 

Results

General Features of S. aureus FORC_001

The genome of S. aureus FORC_001 consists of a single circular DNA chromosome of 2,886,017 bp with a GC content of 32.8% and no plasmid. The genome contains 2,719 predicted open reading frames (ORFs), 57 tRNAs, and 12 rRNAs (Fig. 1 and Table 1). Among the predicted ORFs, 774 genes (25.2%) were predicted to encode hypothetical or uncharacterized genes, and 1,954 (74.8%) were predicted to encode functional genes. Analysis using the SEED subsystem categorization and COG functional categorization are shown in Fig. 2. SEED subsystem categorization of the genome predicted that 1,962 ORFs (72.2% of all predicted ORFs) encoded known functional proteins, whereas 757 ORFs were of unknown function. Among the functionally predicted ORFs, 318 ORFs were responsible for amino acid synthesis, 159 ORFs for carbohydrates synthesis, 210 for protein metabolism, 179 for cofactors, vitamins, prosthetic groups, and pigments, 118 were involved in RNA metabolism, and 104 played roles in DNA metabolism. In addition, 92 ORFs belonged to the category that included virulence-, disease-, and defense-associated genes. Among these, 38 ORFs (41%) were responsible for antibiotic resistance and toxic compounds, 25 ORFs (27%) for adhesion, 14 ORFs (15%) for bacteriocins and ribosomal-synthesized antibacterial peptide production, 9 (10%) played roles in immune invasion and intracellular resistance, and 6 (6%) were associated with toxins and superantigens. In COG functional categorization, 2,308 ORFs (84.8% of totally predicted ORFs) were assigned to COG functional categories. Among these, 875 ORFs (37.9% of the COG-assigned ORFs) belonged to five major COG functional categories: 230 ORFs in category E (amino acid transport and metabolism), 176 ORFs in category G (carbohydrate transport and metabolism), 162 ORFs in category L (replication, recombination, and repair), 155 ORFs in categoryP (inorganic ion transport and metabolism), and 152 ORFs in category K (transcription).

Table 1.Comparison of the choromosomal properties of S. aureus strains in Group I.

Fig. 1.Genome map of S. aureus FORC_001. The outer circle indicates the locations of all annotated ORFs, and the inner circle with the red peaks indicates GC content. Between these circles, sky blue arrows indicate the rRNA operons, and orange arrows indicate the tRNAs. Functional gene clusters are labeled around the outer circle as follows: virulence factors in red, prophages in blue, and other functional gene clusters in black. All annotated ORFs are colored differently according to the COG assignments.

Fig. 2.Functional categorization of all predicted ORFs in the genome of the strain FORC_001 based on (A) COG and (B) SEED databases.

Virulence Factors of S. aureus FORC_001

Virulence factors of strain FORC_001 are listed in Table 2. Common virulence determinants, including coagulase, protein A (SpA), PVL, enterotoxin, and TSST-1, are present in strain FORC_001. These factors are likely responsible for its pathogenicity. In particular, staphylococcal enterotoxins play major roles in tissue invasion, pore formation, cellular damage, toxic shock syndrome, and food poisoning [14,22,23,28,29]. In addition, factors responsible for iron acquisition, such as isd, srtBF, corA, and mgtE, were also present. FORC_001 secretes several proteins, including adhesion proteins (sdrC, sdrD, sdrE), fibronectin-binding B (fnbA), an elastin-binding protein (epbS), a collagen adhesion factor (cna), and an extracellular adhesion protein (eap). These proteins play roles in providing protection against the host innate and adaptive immune systems.

Table 2.aExperimentally confirmed virulence factors.

Comparative Phylogenetic Tree Analysis

Three comparative phylogenetic tree analysis methods were used to compare the complete genome sequences of S. aureus at the strain level: ANI, multilocus sequence typing (MLST), and 16S rRNA gene sequencing. For phylogenetic tree analysis using ANI, 40 complete genome sequences that are available in the NCBI GenBank database, as well as strain FORC_001, were obtained and the ANI values were calculated. A phylogenetic tree was then constructed and visualized based on the metrics of the calculated ANI values (Fig. 3A). The tree identified a subgroup that contained MRSA252, 55_2053, TCH60, and FORC_001, which was designated Group I. Within this group, the genome of MRSA252 was closest to that of FORC_001 (99.87% of ANI value), which suggests that these strains are closely related but evolutionarily distinct.

Fig. 3.Comparative phylogenetic tree analysis. Phylogenetic tree analysis of 40 complete genome sequences of S. aureus using the ANI method (A). Phylogenetic tree analysis of Group I strains using the MLST method (B) and 16S rRNA sequence method (C).

An additional comparative phylogenetic tree was generated using MLST and seven different housekeeping genes: aroE (shikimate dehydrogenase), tpi (triosephosphate isomerase), pta (phosphate acetyltransferase), yqiL (acetyl-CoA acetyltransferase), gmk (guanylate kinase), glpF (glycerol kinase), and arcC (carbamate kinase). The DNA sequences of the seven MLST genes in the Group I strains were compared and aligned to generate a phylogenetic tree (Fig. 3B). Interestingly, the MRSA252 and 55_2053 strains were the closest in Group I, and FORC_001 was closely related to both strains, supporting that the ANI and MLST methods could distinguish them even at the strain level.

However, comparative phylogenetic tree analysis using the 16S rRNA sequences of the Group I strains revealed that strains MRSA252, 55_2053, and TCH60 are the closest in one phylogenetic branch. The FORC_001 strain, however, is closely related to them and the distance in the tree is too close (Fig. 3C), suggesting that the phylogenetic analysis of Group I using the 16S rRNA gene sequences is hardly able to distinguish them on the strain level, probably due to the low resolution of their evolutionary relationships [19].

Nucleotide Substitution Analysis

To further elucidate the evolutionary relationships among the Group I strains, comparative nucleotide substitution analysis was performed using the dN/dS ratio (Table 3). Strain MRSA252 had the highest number of homologous ORFs to strain FORC_001, suggesting that MRSA252 was the evolutionarily closest to strain FORC_001. In addition, strains 55_2053 and TCH60 also had highly close evolutionary relationships with strain FORC_001. Interestingly, the closest order of these strains to strain FORC_001 was very similar to the comparative phylogenetic trees obtained using the ANI and MLST methods, but was quite different from that generated from 16S rRNA sequences, supporting the previous phylogenetic tree analyses results. In the result of comparative nucleotide substitution, only five ORFs (one for MRSA252, two for 55_2053, and two for TCH60) were detected in positive selection (a dN/dS ratio >1), suggesting that the homologous ORFs in the Group I strains were highly conserved (Table 3). Strain MRSA252 had only one positive selection ORF, encoding a hyaluronate lyase precursor (SAR1892; EC 4.2.2.1), which plays a role in polysaccharide cleavage. A DNA sequence alignment of this ORF and FORC1_1755 showed that this ORF was truncated (Table 4). This ORF was also present in the genome of strain 55_2053 (SAAG_01706). Interestingly, DNA sequence alignment of FORC1_1755/SAAG_01706 and SAR1892/SAAG_01706 showed that while the other two ORFs, SAR1892 and SAAG_01706, are highly conserved, FORC_001 is indeed truncated (data not shown). In addition, strains 55_2053 and TCH60 also had two ORFs that encode putative staphylococcal surface-anchored proteins (SAAG_02704 and HMPREF0772_11771, respectively) that play roles in cell wall anchoring as positive controls against FORC1_1331. DNA sequence alignments of these three ORFs revealed that FORC1_1331 was highly conserved with other genes in the GenBank database that encode staphylococcal surface-anchored proteins, whereas the other two ORFs are truncated. This suggests that the staphylococcal surface-anchored proteins are mutated in 55_2053 and TCH60. Furthermore, TCH60 exhibited another positive selection gene, HMPREF0772_11101. This gene encodes the inner membrane protein translocase component YidC, short-form OxaI-like for transmembrane. While FORC1_2033 is highly conserved with other similar genes in the GenBank database, HMPREF0772_11101 is also truncated.

Table 3.Comparative nucleotide substitution analysis using the dN/dS ratio.

Table 4.List of ORFs in positive selection category of nucleotide substitution analysis.

Comparative Genomics with S. aureus MRSA252

Comparative genomic analysis of S. aureus FORC_001 and its closest strain MRSA252 was conducted and the result showed that the number of perfectly matched ORFs is 1,180. In addition, the number of FORC_001 specific genes is 116 and the number of MRSA252 specific genes is 180. There were two unmatched regions, designated Region I (34,461–86,980 nt; Fig. 4A) and Region II (2,152,883–2,167,337 nt for MRSA252; Fig. 4B). Interestingly, the SCCmec gene cluster was detected in Region I of the genome of strain MRSA252 (SAR0039 to SAR0041), but was not present in FORC_001, which suggests that FORC_001 might be highly susceptible to methicillin. To confirm this hypothesis, an antibiotic susceptibility test was conducted using an AST-P601 antibiotic resistance kit (VITEK, Durham, NC, USA). Data revealed that FORC_001 was resistant to clindamycin and erythromycin, but was highly susceptible to 19 other antibiotics, supporting this susceptibility of strain FORC_001 to methicillin (unpublished observations). Furthermore,Region I of MRSA252 had additional antibiotic resistance ORFs, including bleomycin (SAR0032) and kanamycin (SAR0033), suggesting that strain FORC_001 might be susceptible to these antibiotics. However, the ORFs for erythromycin and spectinomycin resistance (FORC1_1603/SAR0050 for erythromycin and FORC1_1604/SAR0051 for spectinomycin) were shared with 100% DNA sequence identities between the genomes of the FORC_001 and MRSA252 strains, although these ORFs were located in a region of FORC_001 (1,768,594-1,770,233 nt) distinct from Region I. An antibiotic susceptibility test showed that FORC_001 was highly resistant to erythromycin (MIC, 8 µg/ml), confirming the genomic data (data not shown). Interestingly, seven mobile elements were scattered around Region I and these ORFs, implying that these antibiotic resistance ORFs might be deleted or acquired under specific environments for host survival.

Fig. 4.Comparative genome analysis of the closely related strains, FORC_001 and MRSA252, which revealed two non-homologous regions. (A) Region I containing various antibiotic resistance genes. (B) Region II: Region II-1 contains prophage regions with very low DNA sequence identity, and Region II-2 contains a pathogenicity island in strain FORC_001. (C) Gene map of Region I, including the SCCmec gene cluster.

Region II contains two different subregions: the prophage Region II-1 (2,123,722-2,138,096 for FORC_001; 2,152,883–2,167,337 for MRSA252), and the pathogenicity island Region II-2 (2,148,461–2,162,549 for FORC_001) (Fig. 4). Sequence alignment of Region II-1 revealed low DNA sequence identity and query coverage (19%) between the genomes of strains FORC_001 and MRSA252, although the size of Region II-1 was similar in both strains, suggesting that these strains might have different prophages. Region II-2, which contains pathogenicity island ORFs, was present in only the FORC_001 genome. This region encodes virulence factors, including toxic shock syndrome toxin 1 (TSST-1), and a number of superantigen-encoding pathogenicity island SaPI proteins (Fig. 4). Interestingly, a region similar to Region II-2 is present in a different area of the MRSA252 genome, but the sequence query coverage was only 46%. However, the region between Region II-1 and Region II-2 (2,138,097–2,148,460 nt) is shared by both strains (Fig. 4), and encodes a bacteriocin gene cluster for leukocidin. This suggests that both strains might produce the natural antibacterial compound bacteriocin.

 

Discussion

Staphylococcus aureus is a common cause of skin infections, respiratory disease, and food-poisoning, and the emergence of antibiotic-resistant strains is a worldwide problem for clinical medicine. Most studies of foodborne pathogens analyze pathogenicity using physiological and molecular techniques. However, microbial genomics and bioinformatics have extended our knowledge and enhanced our understanding of virulence factors at the genomic level. Since the first complete genome sequence of S. aureus was reported in 2001 [20], 39 additional complete sequences have been reported and deposited in GenBank. In the current study, S. aureus FORC_001 was isolated from a contaminated food, kongjang in South Korea, and the complete genome sequence was obtained for the first time. To elucidate the characteristics of this strain, comparative analyses of the complete genomes of closely related S. aureus strains were performed, and phylogenetic trees were constructed.

Various virulence factors in S. aureus FORC_001 were detected in the genome sequence, including the common virulence factors reported previously in the complete genome sequences of pathogenic S. aureus strains (coagulase, protein A (SpA), PVL, enterotoxin, and TSST-1), suggesting that S. aureus FORC_001 could be a real foodborne pathogen. However, further functional analysis of these genes in FORC_001 should be performed to confirm its pathogenicity at the molecular and genomic levels.

Phylogenetic analyses using the 16S rRNA gene have been used widely to represent the evolutionary relationships of bacteria. Although its accuracy is mostly acceptable at the genus or species level, higher strain-level resolution is now required [7]. Recently, novel approaches toward phylogenetic analysis were suggested: ANI and MLST [7,15]. The ANI method was performed using 40 different complete genome sequences of S. aureus, which revealed that S. aureus FORC_001 belonged to Group I. Group I consists of MRSA252, TCH60, 55_2053, and FORC_001 strains. However, this group was not identified in the 16S rRNA phylogenetic tree, suggesting that the ANI and MLST methods can distinguish among strains at a higher resolution than the 16S rRNA method (see the scale bars in Fig. 3). Among the strains in Group I, MRSA252 was the closest strain to FORC_001. Therefore, nucleotide substitution using the dN/dS ratio and comparative whole-genome alignment analysis of strains FORC_001 and MRSA252 were conducted. The nucleotide substitution analysis showed that most ORFs were highly conserved between these two strains, and that only five ORFs were positively selected (dN/dS >1). This suggests that these two strains are very closely related, which supports the presence of Group I identified in the phylogenetic tree. Comparative sequence analysis of the five ORFs identified in the positive selection revealed that they were truncated or partially deleted, but not point-mutated in FORC_001. This suggests that these ORFs were not mutated evolutionarily in FORC_001, but were instead inactivated by gene truncation or deletion. These gene inactivation events probably occurred in the recent period after strain divergence, not in the long period for evolutionary mutation.

Whole-genome alignment between the FORC_001 and MRSA252 strains was conducted to identify genomic differences, particularly regarding antibiotic resistance characteristics. Antibiotic resistance genes, including bleomycin (SAR0032), kanamycin (SAR0033), and the SCCmec gene cluster (SAR0039 to SAR0041), were present in Region I of strain MRSA252, but were not present in FORC_001. This suggests that either MRSA252 might have acquired these antibiotic resistance genes for survival, or FORC_001 might have lost them to adapt to a specific environment. Interestingly, the SCCmec gene cluster was also detected in strains TCH60 and 55_2053, supporting this. Based on this result, although these four strains in Group I have an evolutionarily close relationship, strain FORC_001 might adapt in a different environment unlike the other three strains, indicating its uniqueness in Group I. However, the erythromycin resistance gene (ermA1; FORC1_1603) was present in Region I of both strains. This suggests that this gene might not be involved in the acquisition or loss of specific antibiotic genes for survival or environmental adaptation, but that it is likely still required for both strains. In Region II-1, the prophages were located in a similar position in the genome of both strains. However, they were not homologous, suggesting that they might have different origins. In addition, Region II-2 had a pathogenicity island gene cluster in MRSA252, but not FORC_001. FORC_001 has a similar pathogenicity island gene cluster located in a different region, suggesting indicating that it is still pathogenic. However, the bacteriocin gene cluster encoding leukocidin was present in both strains, that it might be useful for survival in the environmental habitat of S. aureus.

This complete genome sequence of S. aureus strain FORC_001 and the comparative analysis with previously published S. aureus genomes would increase our knowledge about the characteristics of pathogenic S. aureus strains in Korea. In addition, the detection and characterization of virulence factors would provide information useful for epidemiological surveys and also for the development of novel biomarkers for the rapid detection of this pathogen in foods.