Background

In recent years the epidemiology of methicillin-resistant Staphylococcus aureus (MRSA) has changed with the emergence of community associated MRSA (CA-MRSA) strains. Unlike the original healthcare associated MRSA strains, CA-MRSA strains are no longer restricted to the hospital setting and can persist in and be transmitted by healthy individuals in the community. Some of these strains exhibit an enhanced virulence due to the carriage of a number of virulence genes, including the Panton Valentine leukocidin (PVL) lukF-PV and lukS-PV determinants. PVL is a phage-borne, bi-component toxin [1] associated with chronic/recurrent skin and soft tissue infections and with necrotising pneumonia and fasciitis [2].

ST772-MRSA-V, colloquially known as the Bengal Bay Clone [3], is a multiresistant PVL-positive CA-MRSA initially isolated in India in 2004/2005 [4]. Transmission of Bengal Bay MRSA has subsequently occurred in several countries including England [3], Ireland [5], Germany (H.J. Linde, Regensburg, Germany, pers. communication; [6]), Norway (H. Aamot, pers. communication), Italy [7, 8], Abu Dhabi [6, 9], Saudi Arabia [10], Hong Kong [6], Malaysia [11], Australia [12] and New Zealand [13]. Many patients had a travel history or family background suggesting an infection in India, Pakistan or Bangladesh ([3, 5, 12]; H.J. Linde, Regensburg, Germany, pers. communication; author’s unpublished observations), where this strain appears to be increasingly common [14, 15].

In order to identify possible factors promoting its recent emergence and spread we have sequenced the ST772-MRSA-V genome.

Methods

Strains

One isolate of ST772-MRSA-V (07–17048) was selected for next generation genome sequencing. It was isolated from an Indian healthcare worker in Western Australia as part of standard patient care in 2007 and submitted to the Australian Collaborating Centre for Enterococcus and Staphylococcus Species for typing. Thirteen related isolates that were previously submitted for typing purposes to the participating institutions were selected. Their microarray hybridisation profiles where compared to isolate 07–17048 especially with regard to genes associated with resistance or virulence (Table 1, Additional file 1).

Table 1 Overview on typing and microarray hybridisation data for CC1 reference strains (nr. 1 and 2), CC5 reference strains (nr. 3 and 4), the strain described herein (nr. 5), other Bengal Bay isolates (nr. 6 to 14) and other, related strains (nr. 15 to 18)
Full size table

Methods

High-throughput de novo sequencing was undertaken commercially by Geneservice Source BioScience plc (Nottingham, United Kingdom) using the Illumina Genome Analyzer System (Illumina Hiseq 2000 platform, Illumina, Essex, United Kingdom). The average genome coverage was ca. 105. The reads were assembled to contigs using the Velvet de novo genome assembler (vers.1.0.15; Illumina). The project was registered with the NCBI BioProject database under the provisional accession number PRJNA207032 and has been deposited at DDBJ/EMBL/GenBank under the accession number AZBT00000000.

Microarray procedures have been previously described in detail [6].

Analysis

Analysis was performed using automated scripts for full text comparison and BLAST analysis and an in-house database of known, annotated and previously identified S. aureus genomes, genes and gene fragments to the query sequence. This allows determination of identity, clonal parentage and (given the constant order of core genomic genes in S. aureus) position within the genome of each contig (Additional file 2). In parallel, iterated BLAST searches were used for analysis of individual contigs in order to confirm results (http://blast.ncbi.nlm.nih.gov/Blast.cgi; [20]).

Results and discussion

In terms of microarray hybridisation patterns, the sequenced strain represents the typical and most common variant of the Bengal Bay Clone (Table 1 and Additional file 1). The isolate 07–17048 was found to belong to the multilocus sequence type (MLST) ST772 (1-1-1-1-22-1-1), spa type t3387 (RIDOM nomenclature, [21], repeat sequence 26-16-21-17-34-33-34) and dru type dt10ao (5a-4a-0-2d-5b-3a-2 g-3b-4e-3e). According to MLST e-burst analysis, ST772 is considered to belong to CC 1 as it differs from ST1 only in one MLST allele (pta-22). However, because of several differences, as shown below, its inclusion into CC1 needs to be re-assessed.

A total of 340 contigs were obtained after initial assembling. Seventy contigs consisting of 2,741,418 base pairs have been analysed (Additional file 2). The overall G/C content was 33%. 1,946 protein coding sequences have been identified (Additional file 3). 1,234 protein coding sequences were completely identical to previously identified genes from other S. aureus strains. A total number of 416, 239 and 101 protein coding sequences were completely identical to alleles from CC1 genome sequences (from MSSA476, GenBank accession number BX571857 and MW2-USA400, BA000033), CC5 genome sequences (from Mu50, BA000017; ED98, CP001781 and N315, BA000018) and CC8 genome sequences (from COL, CP000046; Newman, AP009351 and NCTC8325, CP000253) respectively. Three-hundred eighty-three genes were identical only to another ST772-MRSA-V sequence (strain 118, whole genome shotgun project AJGE00000000, [22]). Based on identities to previously published gene sequences, ST772-MRSA-V is most closely related to CC1 and CC5. Genes of CC1 and CC5 backgrounds are scattered across the genome, and no evidence for a distinct part of the genome being affected by a genomic replacement (as observed in, for instance, ST239; [23]) can be found. Theoretically, this may be attributed to a high number of recombination events involving CC1 and CC5 strains, or may indicate a common ancestry for both lineages and subsequent accumulation of random mutations in genes that are essentially orthologs in CC1, CC5 and ST772. The question whether the presence of different capsule types and agr groups in otherwise closely related strains (such as ST1 and ST772) can be attributed to recombination or to convergent evolution justifies further study.

Several fundamental differences in the ST772 and CC1 core genomic markers have already been identified by microarray DNA [6] and have been confirmed by sequencing. These differences include agr alleles (group II rather than III), the capsule type (5 rather than 8) and the presence of different allelic variants of hlb, ssl01/set6, bbp, clfB, fnbB, sdrC, sdrD, vwb and hsdS. The egc enterotoxin gene cluster, Q7A4X2 (a hypothetical protein localised close to egc), the metallothiol transferase gene fosB and the enterotoxin homologue ORF CM14 are present in ST772 but absent from other CC1 strains. The genes seh (encoding enterotoxin H), lukD/E (a leukocidin homologue), splA/B/F (serine proteases), ssl11/set2, ssl06/set21 (superantigen-like proteins), and Q2FXC0 (hypothetical protein, located next to serine protease operon) are absent in ST772, but present in other CC1 strains. In lieu of seh, the enterotoxin homologue ORF CM14 was identified in a similar position, i.e., closely downstream of the integration site of the SCCmec element. ORF CM14, absent in the related and possibly parental lineages CC1 and CC5 can be found in a number of different lineages including CC12, ST93, CC121, CC395 and CC705. This may indicate a small scale genomic replacement in a region close/downstream to oriC. Alternatively, ORF CM14 may have been replaced by seh in ancestors of CC1 or entirely deleted in CC5, but retained only in ST772.

Isolate 07–17048 harbours the enterotoxin A and PVL encoding sea and lukF/S-PV genes. Both genes are located on the same contig, and together with several other phage-associated genes, appear to be on a novel prophage. The phage is integrated into a gene of the putative protein A5IT17 which contains an attachment site of the PVL-carrying phage previously identified in other strains (un-truncated in the CC1 strain MSSA476, BX571857.1: SAS1429; truncated in the CC1 strain MW2-USA400, BA000033.2: MW1377/1443). All phage-associated genes identified are shown in Table 2.

Table 2 Genes associated with the PVL prophage in isolate 07-17048
Full size table

Genes sea and sprFG are normally associated with haemolysin beta converting phages rather than with PVL phages. In ST772-MRSA-V, haemolysin beta is interrupted, but there is no complete phage integrated into that gene. Only genes encoding the staphylococcal complement inhibitor scn, the putative membrane protein Q6GFB6 (usually located next to scn on hlb- converting phages, e.g., in genomes of USA300 and USA400, N315, Mu50, NCTC8325, MRSA252, MSSA476) and sprD (coding for small pathogenicity island RNA D) can be found within hlb. A possible assembly error affecting both phage integration sites is highly unlikely. Several other ST772-MRSA-V genome sequences [22, 24] also show this association of lukF/S-PV with sea. Besides, rare naturally occurring variants of ST772-MRSA-V and the related (single locus variant) ST573-MRSA-V (WA MRSA-10, [19]) lack lukS/F-PV and sea while still harbouring scn. Similar constellations can also be observed in ST573/772 MSSA (Table 1). Thus it is more likely that a part of the hlb-converting phage was translocated into a PVL phage that is integrated into a different position of the staphylococcal genome.

Other genes that are associated with mobile genetic elements include enterotoxin genes sec and sel. They are localised in similar position and context as in MW2, where they are also accompanied by phage derived genes and by the gene ear encoding the enterotoxin-linked ampicillin resistance protein. For resistance genes blaZ/I/R, msr(A), mph(C), aacA-aphD, aphA3, sat and aadE it was not possible to unambiguously determine with the given set of contigs whether they were situated on the chromosome or on plasmids. The macrolide/clindamycin resistance gene erm(C), and msr(A), mph(C), aacA-aphD, aphA3 and sat genes have been shown by microarray hybridisation to occur variably in ST772-MRSA-V (Table 1).

The isolate 07–17048 harbours a SCCmec V element. Its terminal sequence towards orfX was also present in another ST772-MRSA-V sequence (strain 118, [22]) and appears to be unique to ST772-MRSA-V. The SCCmec V element consists of a tnpIS431-04, mvaS- SCC (a truncated 3-hydroxy-3-methylglutaryl CoA synthase), a putative protein Q5HJW6, dru (SCC-associated direct repeat units), ugpQ (glycerophosphoryl diester phosphodiesterase), ydeM (a putative dehydratase), a bidirectional rho-independent terminator of mecA followed by mecA in an allelic variant identical to GQ902038 and AM990992 [25], a series of genes coding for putative proteins (Q4LAG7, Q3T2N0, Q4LAG4, Q4LAG3, Q3T2M7), a recombinase homologue “ccrAA” [26], a SCCmec type V recombinase ccrC, further genes encoding putative proteins (Q4LAF9, Q7A206, Q7A207, Q9KX75, A9UFT0), a bidirectional rho-independent terminator of hsdR and three genes (hsdR, hsdS, hsdM2) of a type I restriction-modification system.

Conclusion

In conclusion, ST772-MRSA-V may have emerged from the same root or lineage as the global CC1 and CC5 strains. It has acquired a variety of virulence factors, and, compared to other CA-MRSA strains, it has an unusually high number of genes associated with antibiotic resistance. Whether it evolved in a hospital setting or acquired these genes in community cannot be decided based on a single sequence. Therefore, more epidemiological data and possibly the sequencing of a number of additional isolates are warranted in order to understand the evolution and spread of this conspicuous strain.