The genome of S. aureus consists of many accessory elements scattered throughout the genome accounting for approx. 25% of the whole genome sequences. Within the genomes a wide range of putative mobile DNA elements have been identified, including pathogenicity and genomic islands, up to 30 insertion sequences, five transposons, and several bacteriophages. Methicillin-resistant isolates typically carry the methicillin resistance encoding chromosomal cassette mec (SCCmec). Most of these elements have been acquired by horizontal gene transfer and some of them are mobile. Accessory genes often carry virulence genes which have substantial implications for disease types. For example, various toxin genes are associated with certain strains or lineages including toxic shock syndrome toxin tst, Panton-Valentine leukocidin, serine-protease splB, and staphylococcal superantigens sea, seg, and sei. Interestingly, many toxin genes are located on genomic and pathogenicity islands.
Pathogenicity islands are large chromosomal structures encoding virulence traits which have been acquired by horizontal gene transfer first described in the gramnegative bacterium E. coli [58, 59]. A typical pathogenicity island is characterized by a GC content differing from that of the core genome, the presence of one or more virulence-associated genes, the presence of mobility genes like transposases or recombinases, direct repeats at the flanking borders, and putative mobility . In gram-negative bacteria, pathogenicity islands often integrate into tRNA genes and are lost with varying efficiency. Pathogenicity islands can be regarded as a subgroup of so-called genomic islands (GEIs) which have the same characteristics as pathogenicity islands but do not necessarily carry putative virulence genes. Probably, SaPIs (S. aureus pathogenicity islands) have derived from bacteriophages as they carry some phage-related genes and show interaction with phages. Importantly, most SaPIs carry superantigenic toxins therefore the presence or absence of a particular SaPI has great implications for the pathogenetic potential of a strain. Different schemes of classification of genomic islands in S. aureus have been suggested [16, 17, 29]. SaPIs or alternatively designated islands of S. aureus can be classified on the basis of integrase homology and insertion site into specific groups. The first pathogenicity island of S. aureus was discovered by Lindsay et al. and was the first pathogenicity island identified in gram-positive pathogens . This 15.2-kb pathogenicity island, named SaPI1, encodes the toxic shock syndrome toxin-1 (TSST-1) which is a potent superantigen associated with most cases of menstrual toxic shock syndrome. The island is flanked by 17-bp-long direct repeats and in addition to tst it carries other genes whose products are presumably involved in virulence. Like other pathogenicity islands, SaPI1 carries a locus which is homologous to members of the integrase family of recombinases of bacteriophages. Interestingly, this type of staphylococcal island has the capacity to integrate into target sites which are identical to the directly repeated sequences at the ends of SaPIs in different tst-negative strains. The SaPI1 can be excised and circulated by the helper bacteriophages U13 and 80a; following excision, the islands are transduced to other strains with high frequency. The integration of SaPI into the recipient genome by a Campbell mechanism near the tyrB locus is recA-independent, requiring the SaPI1-encoded integrase . Obviously, SaPI1 uses excision, replication, and encapsidation functions of the helper phage. It has been shown that SaPI1 interferes with phage growth after its excision and is encapsidated into phage heads. Upon transduction SaPI1 integrates into the attc site, for which the SaPI1-coded integrase is necessary. SaPI1 was first identified in the clinical isolate RN4282; however, a similar type of island is also present in strain COL but not in the other five sequenced S. aureus genomes . The island in strain COL was originally designated SaPI3, but it is virtually identical to SaPI1
as both islands share many island-specific genes, are integrated into the same target site, and carry identical 17 nt flanking directed repeats [17, 61]. An important difference, however, is that SaPIl of strain COL carries the staphylococcal entero-toxin B gene (seb) instead of the tst gene on the island. The fact that tst and seb are located on different islands of a similar type may account for the fact that TSST-1 and SEB are presumably never coproduced .
SaPI2 also carries the tst gene but is located at another site of the S. aureus genome within the tryptophan locus. This island is found in many menstrual toxic shock syndrome (TSS) strains [25, 63]. In addition to tst, usually two other superantigens are located on SaPI2, staphylococcal enterotoxin L (sel) and enterotoxin C3 (sec3). This island is alternatively designated vSa4 , SaPIn1 (strain N315), or SaPIm1 (Mu50) . As has been shown for SaPI1, SaPI2 can be induced to be excised and transduced by different phages .
A third type of staphylococcal island has been identified in the genomes of strain MW2, where it is designated vSa3, and Mu50, where it is designated SaGIm . Although this island does not carry tst, it is closely related to SaPI1 and SaPI2 as many ORFs of unknown function are common to all of these islands. However, their integrases show significant differences in homology and vSa3 is integrated at a completely different locus in the S. aureus genome . Characteristic genes on vSa3 include a siderophore transporter JhuD homologue in strain Mu50 and ear (putative b-lactamase type protein), enterotoxin L2 (sel2), and enterotoxin C4 (sec4) in strain MW2. vSa3 spontaneously excises from the chromosome and forms extracellular circles. Transduction by specific helper phages has not been discovered so far.
A fourth class of island was identified in the genome of the epidemic MRSA252 strain. This island was designated SaPI4 and contains the same gene order and homologues of pathogenicity island proteins of SaPI1 (including "SaPI3" of strain COL, later classified as SaPI1), and SaPIbov. SaPI4 contains no protein with homology to characterized virulence genes. The island integrates downstream of the ribosomal protein gene rpsR  and comprises 25.1 kbp.
In addition to these four islands, two genomic islands which also carry patho-genicity factors and two islands in bovine isolates have been described. The two genomic islands vSaa and vSab are present in all sequenced isolates, and it is very likely that they are also present in many other clinical strains, suggesting that these elements are very stable [15-17]. vSaa was originally designated SaPIn2 in strain N315 and SaPIm2 in strain Mu50, and vSab was termed SaPIn3 in strain N315 and SaPIm3 in strain Mu50. vSaa and vSab possess some specific features in comparison to other staphylococcal islands, especially to the tst island family. Mobility has not yet been described for these islands, which is consistent with the fact that an inactive transposase is located on vSaa and vSab. Moreover, both islands encode a restriction modification system hsdS and hsdM (host specificity determinants) which might be important for stabilization of the islands in the S. aureus genome . Most impressively, on vSaa an exotoxin gene cluster (set) and a lipoprotein gene cluster (lip) is carried, and on vSab an enterotoxin (seg, sen, sei, sem, seo) and serine protease (splA-F) gene cluster [15, 16, 29, 64]. vSaa and vSab exist in different allelic combinations in the S. aureus genomes. For example, the superantigen gene cluster carried by vSab of N315, Mu50, and MRSA252 is missing in MW2 and MSSA476, which instead have a bacteriocin gene cluster bsa . The presence of this bacteriocin might be of advantage for strains in the community as these strains compete with many other species on the skin and mucosal surfaces.
Three families of vSaa and vSab genomic islands have been classified on the basis of allelic variations of the hsdS gene which determines target-specific methy-lation and occurs in three allelic forms. The HsdS protein shows less than 66% amino acid identities. Amino acid variations are mainly found in nucleotide sequence recognition regions, suggesting an evolutionary advantage of different methylation patterns. The vSaa and vSab genomic islands may serve as an important reservoir of genome flexibility and variations in S. aureus forming new pathotypes. Extensive gene duplications along with recombination and gene loss have resulted in different sets of superantigens and exotoxins and may serve as a hot spot for evolutionary events in the future, leading to new S. aureus lineages.
Recently, a new type of genomic island has been identified in strain COL named vSay. This island is found in all S. aureus genomes, but is also present in the S. epi-dermidis RP62A and ATCC12228 genomes and designated vSey. The S. aureus vSac island contains a cluster of two genes of the phenol-soluble modulin (PSM) family and a small cluster of exotoxin genes similar to those in vSaa .
Two pathogenicity islands have been found in bovine isolates. The SaPIbov is similar in size (15.9kbp) and gene order to SaPI1. It also encodes TSST-1, as SaPI1 and SaPI2; however, in contrast to these islands from human isolates, SaPI-bov is flanked by a 74-nt direct repeat, one copy of which has also been shown to be present in a TSS-negative bovine strain . Parts of the bovine island have also been found in human isolates, but the entire sequence has not been identified so far in human strains. SaPIbov2, the second bovine island, was recently identified and characterized in detail . The 28-kb island is integrated at the 3' end of the GMP synthetase and is flanked by 18-bp direct repeats. Comparison of the sequences of the island revealed extensive similarities to other SaPIs. Importantly, the toxin genes present in other SaPIs were exchanged for a biofilm-asso-ciated adhesin called Bap carried on a transposon-like structure. Bap is regarded as an important virulence determinant of bovine S. aureus strains causing mastitis. For the SaPIbov2 transfer of the island was demonstrated without helper phage using a construct of the island-specific Sip integrase flanked by the left and right attachment sites. This module was excised, circularized, and integrated RecA-independently in the attachment site of S. aureus but also on an E. coli plasmid. SaPIbov2 is present in bovine isolates but not in human strains, indicating the presence of an ancestor S. aureus strain in animals. Interestingly, DNA flanking the bap gene in other animal pathogens including S. xylosus, S. chromogenes, and S. simulans is very similar, which suggests that the bap gene was horizontally transferred between different animal pathogens, probably via a composite trans-poson-like structure such as was found on SaPIbov2 .
The exact role of a particular pathogenicity island on the outcome of an infection remains a task for future research. For example, it is not presently known whether the extensive gene duplications found on vSaa and vSab genomic islands impacts on the pathogenicity of S. aureus isolates. Moreover, most of the ORFs on SaPIl have homology to proteins of unknown function. Since 22 of the 24 ORFs of SaPIl of strain COL are expressed in vitro, they might be important for maintenance of the island or regulation of enterotoxin expression . Staphylococcal pathogenicity islands could have evolved by specialized transduction events and recombination of pathogenicity island modules along with deletion and addition of genetic material. Probably, the origin of the pathogenicity islands is a common ancestral genetic element derived from a phage. This hypothesis is supported by the fact that genes with high homology to phage genes are present in all staphylo-coccal pathogenicity islands.
The classification of staphylococcal islands is currently a matter of debate. The term "genomic island" may best denote the characteristic of an island structure which is defined as a putative foreign genetic element that was acquired by horizontal gene transfer. In addition, most authors also include staphylococcal phages in the group of genomic islands, reflecting the fact that many of the staphylococ-cal genomic islands are very similar in structure and gene content to bacterio-phages and some of them can be mobilized by phages. The genomic islands identified in the sequenced S. aureus genomes are shown in Table 9.3.
Type of island 
Characteristic genes on the island
Was this article helpful?