In the current study we have identified two distinct ST93-IV clades circulating concurrently in Australia. The identification of the two clades by SNP analysis of the core region was supported by the PCA based on the absence and presence of genes matrix. The clade 1 isolates were primarily isolated in the northern regions of Australia spread over three states/territories (Western Australia, Northern Territory and Queensland) whilst the clade 2 isolates were distributed across the country. Based on the genomic data of the van Hal et al. historic ST93-IV isolates that were located at the root of the phylogenetic tree, we believe the two clades recently diverged from a common ancestor.
Clade 1 isolates differed from the clade 2 isolates by having acquired up to seven additional accessory genes. The known biological significance of these accessory genes varies. The entA and entE genes, which encode the superantigen enterotoxins A and E respectively, play an important role in serious staphylococcal infections by triggering an overexpression of inflammatory mediators. The ohrR gene, which has previously been identified in Pseudomonas aeruginosa (11) and Bacillus subtilis (12), is known to increase an organism’s resistance to organic hydroperoxides. The ability to resist peroxide provides the organism a growth advantage and increases its survival in host cells (13). The soj gene, a parA homologue involved in chromosome segregation during DNA replication, is not normally found in S. aureus (14). Typically, chromosome segregation in S. aureus is performed by the parB homologue spo0J, which was identified in all ST93-IV genomes. In Bacillus subtilis, soj and spo0J are present and work together to prevent premature midcell Z ring assembly (15). By having acquired soj, clade 1 isolates may have an advantage over non-clade 1 isolates as they have a more efficient DNA replication system. The role of the three remaining accessory genes, acul (a putative protein), and ypuA and hutl (both hypothetical proteins) is not known. The addition of the seven accessory genes, which are likely to have originated on mobile genetic elements, may explain the high rates of ST93-IV skin infections amongst indigenous children living in the northern regions of Australia (16). Further studies are required to determine if clade 1 has become the predominant ST93-IV strain in the region’s indigenous communities.
Based on the variability of the ST93-IV accessory genes over time and location we attempted to identify clade 1 or 2 specific subclades. Although minor accessory gene variations occurred in a small number of isolates, (for example, four isolates contained qacA [antiseptic resistance protein], qacR [HTH-type transcriptional regulator] and tnsB [transposon] which were all located on the same contig), no significant differences in the absence or presence of accessory genes related to specific subclades were observed.
GWAS for Bacteraemia vs Non-bacteraemia MRSA
In 2017 a GWAS performed by Lilje and colleagues was not able to identify genetic differences between S. aureus bacteraemia and non-bacteraemia genomes (9). Their results however may have been influenced by studying a variety of S. aureus lineages and clonal complexes. To identify if specific genetic factors are harboured by S. aureus bacteraemia genomes our study was limited to a single S. aureus lineage. GWAS identified two genes associated with the ST93-IV bacteraemic isolates. The hsdM gene has recently been shown to be a hotspot for chromosome rearrangements in staphylococcus causing phenotype switching associated with persistent infections (17). The clfA gene, which mediates staphylococcal binding to fibrin coated surfaces, has previously been linked to bacteraemia (18).
Chromosome rearrangements of genes may lead to altered gene expression (19). The 23 bacteraemic genomes that did not harbor hsdM and clfA all carried rearrangements of the pls, sdrF, setC and sftA genes. The pls and sdrF genes encode surface proteins. pls, which mediates bacterial aggregation and binding to glycolipids and human epithelial cells (20, 21) , has been shown in mice models to be an important factor in causing sepsis (22). sdrF, which is a microbial surface components recognising adhesive matrix molecule (MSCRAMM), allows staphylococcus to attach to and colonise host cells (23). The sdr gene in the ST93 genome JKD6159 has previously been reported to be the most diverse amongst different S. aureus clones, suggesting acquisition by horizontal transfer has occurred. In the Huping et al. study the sdr in the ST93 genome was classified as sdrC (24). However, an updated annotation database has identified the gene as sdrF which had previously only been reported in S. epidermidis. sdrF adheres to human keratinocytes and epithelial cells facilitating S. epidermidis colonisation of the skin (25). The sftA gene encodes for a DNA translocase. DNA translocases play an important role in allowing an organism to convert to the “L-form” cell wall deficient state which is resistant to cell wall-targeting antibiotics (26). DNA translocases have also been shown to be necessary in the formation of biofilms (27). Although the role of sftA in S. aureus is believed to be performed by SpolllE, in Bacillus subtilissftA and SpoIIIE have been shown to have different functions during cell division. sftA aids in moving DNA from the closing septum while SpoIIIE translocates septum-entrapped DNA(28). We therefore hypothesise the rearrangement of sftA in some ST93-IV causing bacteraemia may provide the organism with a more efficient cell division pathway. setC, which encodes a sugar efflux transporter, is not a known bacteraemia factor. However, the gene is co-located with sftA.