Characterization of Staphylococcal Cassette Chromosome mec Elements from Methicillin-Resistant Staphylococcus pseudintermedius Infections in Australian Animals

Characterization of SCCmec elements.Isolates originated from a larger collection of 669 S. pseudintermedius isolates collected during the first Australian survey into antimicrobial resistance in veterinary pathogens (20, 21). All isolates were identified using a BD Bruker MALDI Biotyper and were screened for methicillin resistance using manual oxacillin broth microdilution according to CLSI guidelines (22). The end dilution for MIC testing was 64 mg/liter oxacillin. Oxacillin resistance was confirmed by detection of an oxacillin MIC of ≥0.5 mg/liter using Vitek 2, in accordance with the manufacturer’s instructions, and detection of the mecA gene by whole-genome sequencing as previously described (19). DNA extraction, library preparation, and initial de novo assembly were undertaken as previously described (19). SCCmec typing was undertaken by downloading the sequences of the mec gene complex and ccr elements of the SCCmec elements described by the IWG-SCC (8) from the NCBI online database (https://www.ncbi.nlm.nih.gov/). SCCmec elements previously identified in S. pseudintermedius but not included in the SCCmec working group website were also downloaded (Table 1). Downloaded SCCmec element sequences underwent BLASTn searches against de novo contigs using CLC Genomics Workbench. BLASTn results required more than 85% sequence homology to be assigned a particular ccr gene. mec gene complexes were assigned based on the gene structure of mecA, its regulators, and associated insertion sequences (ISs). If a SCCmec type could not be assigned, contigs were mapped against a scaffold of reference SCCmec types (8) and the reference methicillin-susceptible S. pseudintermedius genome, ED99 (accession no. CP002478.1). The Kruskal-Wallis test was used to determine whether the median oxacillin MICs differed across SCCmec types. The Mann-Whitney U test was used to assess differences between SCCmec types in the median oxacillin MIC. SCCmec types with more than eight isolates were compared as separate entities in analyses; other isolates were grouped together.

Diversity of SCCmec types.Six SCCmec types were identified among 74 of the 77 MRSP from Australia (Table 2). The SCCmec type of the remaining three isolates (two ST497 isolates and one ST71 isolate) could not be determined because of poor sequencing quality (low read coverage and contig breaks) in the region around the mecA gene. Isolates from the same multilocus sequence type (MLST) tended to harbor the same SCCmec type. Four of the SCCmec types have previously been characterized in MRSP or MRSA as follows: SCCmec III (n = 34) and ΨSCCmec₅₇₃₉₅ (n = 7), previously described in MRSP; and SCCmec V_T (n = 10) and SCCmec IVg (n = 5), previously described in MRSP and MRSA (11, 12, 23, 24). The 34 SCCmec III isolates, which were mostly ST71 and ST316, demonstrated 98% to 100% sequence homology to the mec and ccr gene complexes of the III element from MRSP KM1381 (12). ΨSCCmec₅₇₃₉₅ isolates had no ccr genes but had 99 to 100% sequence homology to the region spanning from orfx to IS256 from MRSP 57395 (11). The V_T isolates displayed 96% to 100% homology to the ccrC1 gene from MRSP 06-3228 (23) and 100% homology to its mecA and IS431 genes, but the mecR1 gene was variably truncated from 23 bp to 93 bp. SCCmec IVg isolates had 97% to 100% sequence homology to the entire cassette from MRSA isolated from bovine milk described previously by Kwon et al. (24).

The fifth SCCmec element has 99% sequence homology to the novel element recently reported in MRSP NA45 (15). In our study, this element was identified in nine geographically dispersed isolates from eight different STs (Table 2). These isolates harbored a class C1 mec complex with 99% sequence homology to SCCmec X from MRSA JCSC6945 but did not harbor the same ccr genes as this element. Instead, the isolates harbored a recombinase gene with 99% homology to ccrC6, recently identified in a 43,922-bp SCCmec element in ST84 MRSP (15) and methicillin-resistant S. schleiferi (MRSS) (25). This element is also present in ST398 MRSA RIVM3897 (26), but the RIVM3987 element lacks the final 8,164 bp at the 3′ end of the SCCmec cassette in ST84 MRSP and MRSS. The nine isolates in this study showed 99% homology to the entire SCCmec cassette from MRSP NA45, which contained heavy metal resistance genes arsB, arsC, arsR (arsenic resistance) and copA (copper resistance) but no antimicrobial resistance genes. On the basis of typing recommendations of the IWG-SCC (8), the element in these nine isolates and in MRSP NA45 could be described as SCCmec VII because it harbors a class C1 mec gene complex and ccrC recombinase gene (Fig. S1). However, as the mec gene complex is positioned in reverse orientation in comparison to SCCmec VII, we feel that it is more appropriate to refer to this cassette as “SCCmec_NA45,” a SCCmec VII variant. Most of the isolates harboring the SCCmec VII variant SCCmec_NA45 did not harbor the variable repeat region of the spa gene and therefore could not be assigned a spa type (19).

The final nine MRSP isolates, all ST497, could not be mapped to previously described SCCmec elements. All ST497 isolates were from a geographic cluster in Melbourne, Victoria (19). To characterize the novel element found in the remaining nine isolates, a representative ST497 isolate (KW21) underwent further sequencing by Illumina HiSeq and Nanopore long-read technology using MinIon. MinIon long reads were used to verify the structure of the de novo assembly. De novo assembly of Illumina HiSeq reads using Geneious yielded a 319,216-bp contig that contained the entire putative SCCmec element. This contig was subjected to a blast search against the NCBI database to determine putative components of the element. The element was annotated using PROKKA (27) and manually verified using the BLASTn algorithm in CLC Genomics Workbench.

ΨSCCmec_KW21 element.The novel SCCmec element, designated “ΨSCCmec_KW21,” was integrated at the 3′ end of the orfX gene (rmlH) (Fig. 1). The characteristic direct repeat and insertion sequences (ISs) that typically flank SCCmec elements were absent at the right side (10, 28). The element contained a class A mec gene complex that had 99% BLASTn similarity to the class A mec gene complex SCCmec cassettes from MRSA JKD6008 (29), MRSP KM241 (12) and MRSP AI16 (10) (accession numbers CP002120.1, AM904731.1, and LN864705 respectively). The element did not contain any ccr genes. The 16,711-bp region of ΨSCCmec_KW21 that spanned from orfX to the cadmium resistance operon, cadCAD, had 99% sequence homology to the same region of SCCmec III from ST239 MRSA isolate JKD6008 (Fig. 1). We therefore propose that ΨSCCmec_KW21 is a truncated version of SCCmec III that lost the segment containing ccr genes after the cassette was inserted into the genome. Unlike SCCmec III from ST239 MRSA, ΨSCCmec_KW21 does not harbor ccr genes; instead, cadA and cadC are bordered by a truncated IS30 family transposase. To the right of this transposase is a 10,116-bp region that had 97% BLASTn similarity to a chromosomal region of Macrococcus sp. IME1152 (accession number CP017156.1) that includes the copY gene, encoding a putative copper transport repressor. BLASTn analysis of the de novo contigs of ST497 isolates against the SCCmec element from KW21 revealed that all nine isolates had the same mec gene complex, the genomic region with Macrococcus sequence homology, and no ccr genes. Consequently, they were considered to have the same pseudoelement as KW21. ST239 is an important health care-associated MRSA lineage in humans, so screening of healthy and diseased animals across Australia is now indicated to determine if ST497 and/or other lineages harboring ΨSCCmec_KW21 continue to exist within a geographic cluster or whether this lineage has disseminated across Australia or overseas, as has occurred with ST239 MRSA in humans (30).

• Open in new tab
• Download powerpoint

FIG 1

Alignment of the novel SCCmec element ΨSCCmec_KW21, its flanking regions, and related sequences. The 16,711-bp ΨSCCmec_KW21 element, delineated by orfX and a truncated IS30 transposase, is a truncated version of SCCmec III. The mecA gene complex and cadmium resistance genes (cadCAD) of ΨSCCmec_KW21 have high homology the same region in ST239 MRSA (S. aureus JKD6008). ΨSCCmec_KW21 inserts into the genome of KW21 at the 3′ end of the orfX gene and is bordered by an ∼10,000-bp region with 99% homology to Macrococcus sp. IME1552. This region, which includes a putative copper transport repressor (copY), is then bordered by sequence with 99% homology to the region harboring the spa gene in methicillin-susceptible S. pseudintermedius ED99. Areas of sequence homology, determined by BLASTn, are indicated with gray shading. Green arrows, coding DNA sequences (CDS) shared with MSSP ED99; orange arrows, CDS shared with S. aureus JKD6008; blue arrows, class A mec gene complex; yellow arrows, CDS shared with Macrococcus IME1152; black arrows, unique CDS. The alignment was created in EasyFig V.5 (34).

Variation in oxacillin MICs amongst different SCCmec types.Fig. 2 shows the oxacillin MIC range for the major SCCmec types. The median oxacillin MICs differed significantly depending on SCCmec type (Fig. 2; P < 0.001). The median oxacillin MICs of ΨSCCmec₅₇₃₉₅ and SCCmec IVg isolates (1 mg/liter) and of SCCmec_NA45-VII variant isolates (4 mg/liter) were significantly lower than the median oxacillin MIC of SCCmec III, SCCmec V_T and ΨSCCmec_KW21 isolates (64 mg/liter; P < 0.01). Recently, Kasai and colleagues (18) similarly reported differences in oxacillin MIC on the basis of MRSP SCCmec types. Specifically, they found that isolates with SCCmec III generally had higher oxacillin MICs and were more often associated with suspected hospital-acquired infections than isolates with SCCmec V. Analogous to analyses of MRSA in human medicine, they concluded that oxacillin MIC may give clues as to an isolate’s epidemiological origin, where a high oxacillin MIC may indicate that an MRSP isolate is from a successful “health care-associated” clone whereas isolates with lower MICs may represent “community-associated” clones (18). There were insufficient epidemiological data available in our study to draw similar conclusions, but our results do support the notion that the oxacillin MIC is significantly affected by the SCCmec type and that isolates of the same multilocus sequence type generally harbor the same SCCmec type. Thus, it follows that different MRSP lineages would demonstrate different oxacillin MICs. The mecA gene can be repressed by either mecI or blaI, but MRSA and MRSP isolates with bla regulators typically demonstrate more rapid induction and higher expression of mecA than isolates with mec regulators alone (31–33). To determine whether the oxacillin MIC was influenced by the presence of the blaI and mecI genes rather than simply by the SCCmec type, we screened all isolates for these repressor genes using a BLAST-based command line script (screen_assembly3.py: https://github.com/shimbalama/screen_assembly). Most (45/77) isolates harbored blaI, but only SCCmec III and ΨSCCmec_KW21 harbored both blaI and mecI. The high median oxacillin MIC (64 mg/liter) of III and ΨSCCmec_KW21 isolates could have been due to the fact that blaI attenuates the strong mecA repression expected from the cognate mec regulators (32).

• Open in new tab
• Download powerpoint

FIG 2

Oxacillin MIC for SCCmec types of MRSP isolates from Australia collected in 2013 to 2014. Black dots indicate the median MIC value; error bars indicate the interquartile range.

In summary, we found six SCCmec types among 77 MRSP isolates collected from clinical infections in Australian animals. The oxacillin MIC varied according to SCCmec type. We also described ΨSCCmec_KW21, a novel pseudo-SCCmec element that was found only in a geographic cluster of clinical isolates. This report highlights the utility of nation-wide surveillance studies in unearthing novel and emerging resistance determinants and demonstrates how genomic resistance elements found in significant human pathogens such as S. aureus can also be found in veterinary pathogens such as S. pseudintermedius.

Accession number(s).The genomic sequence of ΨSCCmec_KW21 has been deposited in the National Center for Biotechnology Information (NCBI) database under GenBank accession number MH713898. The contig sequences of the MRSP isolates described in this study were also deposited under BioProject number PRJNA439160 and BioSample accession numbers SAMN08741522 to SAMN08741590.