Multilocus sequence typing (MLST) identified TW20 as sequence type 239 (ST239), an international health care-associated (HA) MRSA lineage prevalent in Asia (19, 38), South America (2, 37), and Eastern Europe (5, 33), which includes EMRSA-1, -4, -7, and -11 and the Brazilian, Portuguese, Hungarian, and Viennese clones (24). To investigate the genetic basis for increased resistance and transmissibility, the TW20 genome was completely sequenced, assembled, and finished and annotated as described previously (16, 25). The final finished genome (10) assembly contained 64,087 capillary reads, giving an average coverage of 13.3. At 3,075,806 bp, the TW20 genome is the largest S. aureus genome sequenced thus far. It consists of a single chromosome of 3,043,210 bp in size (Fig. ​(Fig.1)1) and 2 plasmids (pTW20_1 and pTW20_2), of 29,585 bp and 3,011 bp.

TW20 belongs to clonal complex 8 (CC8), which contains strains NCTC8325 (ST8) (14), Newman (ST8) (3), USA300 (ST8) (11), and COL (ST250) (13). Comparative genomic analysis with these strains by reciprocal Fasta analysis (36) revealed that between 83.7 and 82.7% of protein coding sequences (CDSs) in the TW20 chromosome have reciprocal matches with CC8 members. The highest numbers of matches in any sequenced S. aureus strain, however, was with MRSA 252 (85.9% of CDSs). MRSA 252 (ST36) belongs to CC30 and is a representative of EMRSA-16 that has been a dominant MRSA clone in United Kingdom hospitals for more than 10 years (16). In comparison to CC8, most of the additional matches to MRSA 252 are to CDSs in horizontally acquired mobile genetic elements (MGEs) rather than to orthologous CDSs. A significant component of the S. aureus genome is derived from MGEs that contribute to the accessory genome (21). In the TW20 genome, 16.2% of the CDSs (12.6% of the total genomic DNA) are found in MGEs (Fig. ​(Fig.1).1). Both TW20 and MRSA 252 are representatives of successful hospital-associated MRSA lineages and have large accessory genomes that contain many of the CDSs associated with drug resistance.

Methicillin resistance is conferred by a mecA gene on a type III staphylococcal cassette chromosome (SCC) mec element (SCCmecIII). TW20 has a composite SCC region of two SCC elements, SCCmercury and SCCmecIII, identical in structure to the type III SCCmec region found by Ito et al. (18) in an isolate from New Zealand in 1985. The SCCmecIII region is present in a part of the chromosome hypothesized to have been transferred from CC30 into a CC8 background as part of a large block of DNA (26). The approximate boundaries of the recombination were identified from pairwise comparisons of the TW20 chromosome with MRSA 252 (CC30) and USA300 TCH1516 (CC8). A marked shift in DNA percent identity of approximately 1 percentage point was observed across the approximate recombination breakpoints (data not shown), demonstrating that 635 kb (~20.6% of the TW20 chromosome; SATW20_26800 to SATW20_03960) may have been transferred from a CC30 donor. This transfer event also contributes to the high level of reciprocal Fasta matches between TW20 and MRSA252 (ST36).

The origins of SCCmecIII in the TW20 genome are unclear, since SCCmecIII has not been found in the CC30 lineage. Each of the SCC elements contains further MGEs: SCCmercury contains Tn554, encoding a streptomycin 3′-adenylyltransferase and an erythromycin resistance protein, ErmA1, and SCCmec contains an integrated plasmid, pT181, and ΨTn554, containing cadmium resistance CDSs. In addition to Tn554 and ΨTn554 in the SCCmec region, the TW20 chromosome contains an additional Tn554 and a Tn552 transposon, encoding the β-lactamase BlaZ, within an integrative conjugative element (ICE) (31).

Further resistance determinants are found on plasmid pTW20_1. Importantly, it carries a gene encoding an antiseptic resistance protein, QacA, that confers resistance to antiseptics such as cationic biocides, quaternary ammonium salts, and diamidines via an export-mediated mechanism (29). In addition, part of the plasmid is highly similar (98 to 100% DNA identity) to the mer operon of the SCCmercury region found on the chromosome (Fig. ​(Fig.2).2). pTW20_1 also contains a homologue of the gene encoding the cadmium-transporting ATPase CadA, found in ΨTn554 of SCCmec. This region in pTW20_1 is bordered by IS431 elements, as it is in the chromosomal copy of SCCmercury. Notably, upstream of the SCCmercury mer operon, there is a CDS that encodes a putative NADH-binding protein. A fragment homologous to the 3′ end of this CDS is also present on pTW20_1 upstream of the mer operon and is truncated by an IS431 element. The absence of the 5′ region of this CDS on pTW20_1 suggests that this region, including the mer operon, may have arisen on the plasmid by recombination between chromosomal and plasmid IS431 elements. It is therefore possible that IS431-mediated recombination plays a role in the evolution of the SCC region.

Two other drug resistance genes, encoding a tetracycline resistance protein, TetM, and a trimethoprim-resistant dihydrofolate reductase, DfrG, are found in a 31.3-kb region (Tn5801-like), similar to transposons/ICE found in the genomes of S. aureus strains Mu50 (20) and Mu3 (23) and Streptococcus agalactiae strain COH1 (35). In comparison to the Tn5801 elements in Mu50 and Mu3, the TW20 element contains an additional four CDSs, including dfrG, in the central region of the element.

There are three prophage within the TW20 genome, two of which are similar to those previously found in sequenced S. aureus genomes: Sa1(TW20) is 43.3 kb in size, is integrated within the 5′ region of a lipase gene, and does not carry CDSs with homology to known virulence factors; Sa3(TW20) is 44.7 kb in size, is integrated in the phospholipase C gene, and carries the staphylococcal complement inhibitor SCIN (28), staphylokinase (30), and enterotoxin A (7) genes associated with virulence. Genes for two other enterotoxins, enterotoxins K and Q, are carried on a Staphylococcus aureus pathogenicity island (SaPI), SaPI1.

At 127.2 kb, the third prophage, SPβ-like(TW20), is markedly larger than the other two and does not display similarity with other S. aureus prophage. SPβ-like(TW20) exhibits extended similarity with the SPß-like region in the Staphylococcus epidermidis RP62a genome (13) (Fig. ​(Fig.3).3). Comparison of the two sequences reveals a region of sequence divergence and rearrangement in the center of the prophage. In SPβ-like(TW20), this region contains CDSs associated with aminoglycoside resistance and streptothricin resistance (Fig. ​(Fig.3).3). In addition, SPβ-like(TW20) contains a CDS that may have a role in promoting persistence of TW20 in the hospital setting. S. aureus possesses many surface-anchored proteins with the LPxTG motif, which bind host molecules (27). SATW20_21850 encodes an LPxTG motif surface-anchored protein that does not have orthologs in any of the genomes of the other sequenced S. aureus strains currently available. A highly similar CDS (95.1% amino acid identity), sesI (8), is present in the S. epidermidis SPβ-like region (Fig. ​(Fig.3).3). A recent study by Söderquist et al. found that sesI was absent from normal S. epidermidis flora of healthy individuals without any health care association but was found in approximately 50% of clinical isolates causing invasive infections, leading them to suggest that this gene was a potential marker of invasive capacity (32). The presence of an LPxTG motif surface-anchored protein on an MGE in TW20 suggests that this strain has augmented its array of this family of functionally important proteins through a recent acquisition event and therefore this LPxTG motif surface-anchored protein may not be widely distributed in related strains. Genome sequencing of a global collection of ST239 strains revealed only 7% (3/42) of isolates were positive for orthologs of this CDS (14a). Work is under way to survey the wider distribution of this gene in the S. aureus population and investigate the function of the encoded protein.

Evidence of adaptation to survive in a health care environment is also found in the core genome. Several housekeeping genes have alleles associated with antibiotic resistance. The TW20 DNA gyrase subunit A (GyrA) contains a leucine residue at position 84. The more-widespread residue in S. aureus GyrA proteins is serine, suggesting this is the plesiomorphic amino acid at this position. In vitro studies have demonstrated that substitution of Ser84Leu generates resistance to quinolones in S. aureus (34). TW20 exhibits low-level resistance to mupirocin. The TW20 isoleucyl-tRNA synthetase contains a phenylalanine residue at position 588. The substitution of Val588Phe has been shown to confer chromosomal low-level mupirocin resistance in S. aureus without significantly affecting fitness (17).

In conclusion, genomic analysis of TW20 provides evidence of its adaptation to survive in a health care setting through acquisition of drug and antiseptic resistance genes carried on MGEs, large chromosomal insertions, and point mutations in housekeeping genes. The large size of the TW20 genome reflects the ability of the ST239 lineage to undergo prolonged and continuing evolution to adapt to the hospital environment. Further studies are under way to elucidate the components of the genome that promote transmission and interaction with the host.