Comparison of the genomic content of the 10 MRSA isolates used in this study revealed 1,971 invariant genes (see Table S1 in the supplemental material) representing the core genetic content of S. aureus
, which is similar to previous estimates of 1,954 (10
), 2,029 (2
), and 2,198 (4
) common genes. The 21 regions of variability highlighted in this study (Table ) shared some similarities with the 18 regions of difference previously described for S. aureus
). This included regions containing SCCmec
, bacteriophage L54a, cap5
, enterotoxins i and b, exotoxins 2 to 4, epidermin, leukocidin, and the region encoding two accessory regulators, SarT and SarU.
Analysis of the CGH data for each of the MRSA strain types examined determined that there was no specific or set of defined virulence factors whose presence or absence/divergence could definitively differentiate HA-MRSA from CA-MRSA isolates. Similar findings were reported in a recent study comparing the genomic content of an American epidemic community-associated strain type, USA300, with three other S. aureus
). From these studies, it cannot be concluded that such factors do not exist, as genes potentially indicative of epidemicity may not have been represented on the arrays.
A more restricted comparison did reveal the presence or absence/divergence of a few ORFs that might be of significance. For instance, comparison of the two most common HA-CMRSA isolates (CMRSA1 and CMRSA2) with the CA-CMRSA isolates (CMRSA7 and CMRSA10) revealed a gene cluster in the CA-CMRSA isolates encoding epidermin and its modification factors, which yields a lantibiotic against other gram-positive bacteria (18
). The epidermin gene cluster has also been designated bsa
(bacteriocin in S. aureus
) in MW2 (1
) and is similar to those previously reported for S. epidermidis
). This gene cluster is carried on a type II genomic island, νSaβ, and was previously proposed to provide selective advantages for the MW2 CA-MRSA strain in competing with other bacterial species for colonization (1
). However, the epi
gene cluster cannot be used as a specific marker for CA-CMRSA strains, as it was also present in four epidemic HA-CMRSA isolates (CMRSA3, CMRSA5, CMRSA6, and CMRSA9) (Table ).
Previous markers commonly used to detect CA-MRSA strains have included lukF-PV
, which encode the Panton-Valentine leukocidin. However, the presence of Panton-Valentine leukocidin is not an absolute indicator of CA-MRSA, as indicated by its absence in a number of community-associated isolates (16
; data not shown). Further examination of ORFs from our array that might be indicative of CA-CMRSA isolates revealed one ORF, SACOL0046, encoded within the J1 region of the SCCmec
element, which was present only in CMRSA7 and CMRSA10. SACOL0046 encodes a putative protein belonging to a metallo-beta-lactamase family, and its presence in only community-associated epidemic strains is in accordance with data provided in a previous genomic hybridization study comparing epidemic community-associated strains (USA300 and USA400) with epidemic hospital-associated strains (USA100 and USA500) (24
). Therefore, further studies involving the prevalence of SACOL0046 in other CA-MRSA genetic backgrounds should be conducted to determine if this factor is truly specific to established CA-MRSA isolates.
In comparison to the other nine CMRSA isolates used in this study, there were no ORFs represented on the array that could be solely responsible for the high prevalence of CMRSA2 (Fig. ; see Table S1 in the supplemental material). Comparison of CMRSA2 with CMRSA1, which was the prevalent strain type up to 1999, did reveal some differences in genomic content that might be contributing to the higher occurrence of CMRSA2. These included ORFs encoding potential virulence factors involved in capsular biosynthesis, serine proteases, adhesion factors, regulatory functions, leukotoxins, and exotoxins (Table ; see Table S1 in the supplemental material). The comparison of the two community-associated Canadian epidemic strain types (CMRSA7 and CMRSA10) also revealed additional virulence factors in CMRSA10 encoding proteins involved in capsular biosynthesis, exotoxins, regulatory functions, and adhesion (Table ; see Table S1 in the supplemental material), which might be contributing to the prevalence of CMRSA10 in Canadian communities.
In a recent study, 61 invasive S. aureus
isolates associated with community-acquired infections and 100 noninvasive S. aureus
isolates previously collected from healthy blood donors were investigated using CGH (10
). Similar to our study, those authors found many differences in genomic contents between isolates but no set of ORFs that could predict invasiveness. Alternatively, factors such as type of disease, location of acquisition, or epidemicity may be linked with host factors, environmental pressures, or gene expression patterns. Further studies into these additional factors could provide invaluable information in elucidating the potential pathogenicity and/or epidemicity of S. aureus
strains, which could ultimately result in alternative targets for future antibiotic and/or vaccine development.
This is the first study to describe the genomic characteristics of Canadian epidemic CMRSA strain types. CGH comparisons have revealed some interesting genetic differences between hospital- and community-associated isolates, which warrant further examination to better understand the molecular aspects involved in the epidemic and community-associated nature of MRSA.