Structural Variations of Staphylococcal Cassette Chromosome mec Type IVa in Staphylococcus aureus Clonal Complex 8 and Unrelated Lineages

Abstract

PCR mapping of staphylococcal cassette chromosome mec type IVa and adjacent mobile elements in 94 methicillin-resistant Staphylococcus aureus (MRSA) strains identified two primary structures (A and B) that could be further classified into two (A1 and A2) and five (B1 to B5) variants, primarily based on structural differences in the orfX-J3 region. While spa type t008 (USA300) invariably contained the A variants, other spa types belonging to clonal complex 8 and unrelated lineages generally contained B variants. These findings have important implications for the typing and identification of MRSA strains containing B variants.

TEXT

One of the most widespread community-associated (CA) methicillin-resistant Staphylococcus aureus (MRSA) clones is USA300 (14, 16). This clone has become the most common cause of CA MRSA infections in North America (15) and has recently spread to western Europe, including Denmark (14). USA300 isolates are characterized by a common pulsed-field gel electrophoresis (PFGE) pattern, generally belong to spa type t008 and multilocus sequence type 8 (ST8, clonal complex 8), and harbor staphylococcal cassette chromosome mec (SCCmec) type IVa. In addition, USA300 isolates typically contain the arginine catabolic mobile element (ACME) and the lukS-PV and lukF-PV genes encoding Panton-Valentine leukocidin (PVL) (5, 14, 16).

Genotypic differences were recently shown among Danish health care-associated MRSA isolates belonging to clonal complex 8 (CC8) and resembling USA300 by PFGE. In particular, t024 isolates were PVL negative and inconsistently carried ACME, whereas all t008 isolates carried both elements (9). A study evaluating the sensitivity and specificity of a commercial PCR-based kit for MRSA detection failed to identify 85% of Danish t024 MRSA isolates due to a structural variation in the orfX J3 region of SCCmec IVa (1). Whole-genome sequencing of two t024 isolates identified a novel J3 region containing two unique clusters of open reading frames (ORFs). One of the isolates contained two additional direct repeats between J3 and orfX. These direct repeats delimited a region containing an SCC element followed by an ACME II variant (SCCmec IVa variant B1, Fig. 1) (2). Based on these findings, we decided to further investigate the J3 region and adjacent structures of SCCmec IVa in a large collection of MRSA isolates.

Fig. 1.

Schematic drawing of SCCmec composite elements detected in the present study. “A” structures have prototype SCCmec IVa (11), and “B” structures have SCCmec IVa with the novel J3 region characterized by two unique clusters of ORFs (reg1 and reg2) (2). A1 (SCCmec IVa), A2 (SCCmec IVa with ACME I downstream), B1 (SCC_M1, ΨSCCarc, SCCmecIVa [2B]; J1, subtype 1-specific ORFs; J3, subtype 3-specific ORFs for M1), and B2 (SCCmec IVa [2B]; J1, subtype 1-specific ORFs; J3, subtype 3-specific ORFs for M1) have been previously described (2, 5, 11) and fully sequenced (GenBank accession numbers NC_003923, NC_007793, HM030720, and HM030721). B3 and B4 are proposed structures of new SCCmec IVa composite elements based on PCR mapping and partial sequencing. Approximate targets of PCRs 1 to 9 are shown by broken lines.

MRSA strains isolated at Hvidovre Hospital are routinely spa typed, SCCmec typed, and screened by PCR for detection of ACME-specific (arcA and opp3AB) and PVL-specific (lukS-PV and lukF-PV) genes (4, 5, 6, 8, 10, 12). For this study, we selected 94 MRSA isolates obtained between 2003 and 2010 from patients with health care-associated or CA MRSA in the greater Copenhagen area. All isolates contained SCCmec IVa and belonged to spa types where arcA had been detected. A representative number of isolates was selected on the basis of the relative isolation frequency of each spa type. When available, arcA-negative isolates were included for each spa type. Ninety-one isolates had spa types belonging to CC8, primarily t024 (n = 50) and t008 (n = 25). Three arcA-positive isolates had spa types related to other MRSA lineages, i.e., CC1, CC5, and CC30. Most isolates were epidemiologically unrelated, except for 36 isolates from persons living or working at nine nursing homes with multiple cases of MRSA. These 36 isolates were included based on the same criteria as other isolates but with the purpose of studying within-outbreak variability. They represented approximately one-third of all MRSA isolates and half of all spa types obtained from these nursing homes between 2003 and 2010.

We investigated all isolates by PCR mapping using previously described primers targeting the SCCmec IVa prototype (3, 7, 13) and novel primers targeting the new orfX-J3 elements identified by our previous work (2) (Table 1; Fig. 1). PCRs were performed with a final volume of 20 μl using the DreamTaq Green PCR Master Mix (Fermentas, Helsingborg, Sweden) and purified DNA from each isolate. The cycling conditions for PCRs 1, 3 to 6, and 9 were initial denaturation at 94°C for 4 min, followed by 25 cycles of 94°C for 30 s, 50°C for 30 s, and 72°C for 60 s and a 5-min final elongation step at 72°C. The cycling conditions for PCRs 2, 7, and 8 were identical, except that annealing temperatures of 55°C (PCRs 2 and 7) and 54°C (PCR 8) were used.

Table 1.

Primers used in the study

PCR no. and primer	Sequence (5′–3′)	Amplicon size (bp)	Reference
1
cR1	AAGAATTGAACCAACGCATGA		7
insert orf	CTTTGATAAGCCATTCATTC	326	This study
2
DR2/IR2_F2	GATTTTCCCAACCGCACC
DR2/IR2_R2	ACCTTTATTCGTTACGCCC	975	This study
3
DR3/IR3_F2	ATATAGGTGAGAGTATGATAAGTGAG
ins-orf-M299	GCTATCAATCTGTCCTCAAA	325	This study
4
Insert_IS431	GTAAATGAACTGCAGCTATTG
IS431F	CTTCTGGTTGGAATGTGTGG	939	This study
5
IS1272F1	GCCACTCATAACATATGGAA		3
IS1272R1	CATCCGAGTGAAACCCAAA	415
6
DCS F2	CATCCTATGATAGCTTGGTC		13
DCS R1	CTAAATCATAGCCATGACCG	342
7
cR1	AAGAATTGAACCAACGCATGA		7
ins-orf-M299	GCTATCAATCTGTCCTCAAA	395	This study
8
cR1	AAGAATTGAACCAACGCATGA		7
DR2/IR2_R2	ACCTTTATTCGTTACGCCC	606	This study
9
DR2/IR2_F2	ACCTTTATTCGTTACGCCC
ins-orf-M299	GCTATCAATCTGTCCTCAAA	764	This study

We identified two main SCCmec IVa structures, A and B (Table 2; Fig. 1). Structure A had the dcs region typical of the SCCmec IVa prototype, and structure B had the novel J3 region comprising two unique clusters of ORFs (2). This variation in J3 has important implications for MRSA diagnostics, since only J3 of structure A is targeted by one of the commercially available PCR kits for MRSA detection (1). Variants of A (A1 and A2) and B (B1 to B5) were defined for the purpose of this study based on the observed structural variations in the regions immediately down- and upstream of SCCmec, respectively (Fig. 1). B5 had the novel J3 region, but the organization of orfX-J3 could not be discerned based on PCR results (Table 2). Elements from J3 were not detected in five isolates, which were therefore considered nontypeable (NT1 and NT2, Table 2). Further analysis of SCCmec composite elements in B5, NT1, and NT2 was attempted by standard- and long-range PCRs using various combinations of primers from Table 1. However, the structures could not be deduced, possibly due to mutations in primer binding sites or because of the presence of novel variants. Further studies sequencing the entire SCCmec composite elements of isolates representing NT1, NT2, and B5 are warranted.

Table 2.

Description of the two SCCmec IVa structures (A and B) detected by PCR mapping^a

SCCmec IVa variant	Strain used as referenceb	spa type(s) (no. of strains)		ccr allotype	ACME gened		PVL	Profile according to PCRs 1 to 9
		CC8	Other CCc		arcA	opp3AB
A1	MW2	t008 (2), t064 (2)		ccrAB2	0	0	1/0e	0 0 0 0 1 1 0 0 0
A2	FPR3757	t008 (23), t068 (1), t121 (1), t211 (1), t3980 (1), t5098 (1)		ccrAB2	1	1	1/0e	0 0 0 0 1 1 0 0 0
B1	M1	t024 (17), t430 (3), t351 (1), t842 (1), t2555 (1)	t012 (1)	ccrAB2 and ccrAB4	1	0	0	1 1 1 1 1 0 0 0 0
B2	M299	t024 (11)		ccrAB2	0	0	0	0 0 0 1 1 0 1 0 0
B3	M15	t024 (16), t430 (1), t846 (1)	t127 (1)	ccrAB2	1	0	0	0 0 1 1 1 0 0 1 0
B4	M1298	t024 (2)		ccrAB2 and ccrAB4	0	0	0	1 0 0 1 1 0 0 0 1
B5	M289		t979 (1)	ccrAB2	1	0	0	0 0 1 1 1 0 0 0 0
NT1f	M620	t024 (3)		ccrAB2	0	0	0	1 0 0 0 1 0 0 0 0
NT2f	M220	t024 (1), t064 (1)		ccrAB2	0	0	0	0 0 0 0 1 0 0 0 0

^aA1 and A2 correspond to the prototype SCCmec IVa cassette and the ACME I-positive variant associated with USA300, respectively. B1 and B2 were described recently (2), and B3, B4, and B5 are novel variants identified in this study.

^bMW2 and FPR3757, isolated in Japan and in the United States, respectively, were kindly provided by T. Ito. The remaining standards were Danish clinical MRSA isolates from Hvidovre Hospital, Hvidovre, Denmark.

^cspa types not associated with CC8: t012 (CC30), t127 (CC1), and t979 (CC5).

^dPossession of both arcA and opp3AB suggests the presence of ACME I, which is typical of the S. aureus USA300 lineage. Possession of only arcA suggests the presence of the recently described ACME II variant ΨSCCarc (2).

^e50% and 83% of the isolates with variants A1 and A2 were PVL positive, respectively.

^fIsolates without any apparent loci in J3 were considered nontypeable (NT1 and NT2).

The SCCmec IVa cassette originally discovered by Ma et al. (11) was identified in four isolates (A1 variant), and the same cassette followed by ACME I (A2 variant), which is characteristic of USA300 (5), was detected in 28 isolates (Fig. 1). The majority (92%) of the t008 isolates were assigned to variant A2 (Table 2; Fig. 1), thus supporting the notion that spa type t008 is strongly associated with USA300 (17). Most of the other isolates possessed B variants. In addition to the two variants recently described by our group (B1 and B2), two closely related new variants (B3 and B4) were described for the first time in this study. B variants were detected in 92% of the t024 isolates but not in t008 isolates (Table 2). Interestingly, B1 and B3 were detected in both CC8 and non-CC8 isolates (Table 2), suggesting either horizontal transfer of these large composite elements or convergent evolution across S. aureus lineages.

B3 differed from B1 only by the lack of ccrAB4. We hypothesized that this difference was due to an excision or insertion of SCC_M1 between DR1 and DR2 (Fig. 1). Sequencing of the 606-bp amplicon generated from this region (PCR 8) confirmed this hypothesis, since DR2 was adjacent to orfX in the B3 variant (Fig. 1). Similarly, B4 differed from B1 only by the lack of ΨSCCarc. Sequencing of the 764-bp amplicon generated from this region (PCR 9) proved that excision or insertion of ΨSCCarc had taken place between DR2 and DR3 (Fig. 1). These results are in favor of our previously described hypothesis that B2 derived from B1 by excision of ΨSCCarc and SCC_M1 (2). The results also indicate that possession of more than the usual two direct repeat regions flanking SCCmec may lead to a greater variety of recombination events and consequently more variants. Further experiments are needed to elucidate which factors influence mobilization of ΨSCCarc and SCC_M1.

Analysis of the 36 isolates originating from nine nursing homes with multiple cases of MRSA showed that in some facilities a single clone occurred over several years, whereas in others different spa types and SCCmec IVa variants developed over the years. In nursing home 1, the B1 variant of SCCmec IVa was detected in three spa types belonging to CC8 (Table 3). This may be suggestive of evolutionary changes in region X of spa or horizontal transfer of composite SCCmec elements between closely related lineages. Detection of both variants B1 and B3 in t024 isolates within three nursing homes (1, 3, and 4, Table 3) may suggest that B3 was generated in one of these places by deletion of SCC_M1, followed by transmission in local hospitals or through staff or residents moving between these nursing homes. A less likely explanation would be that the same excision or insertion event happened in all three nursing homes. Overall, a clear pattern of local SCCmec evolution could not be discerned, since strain isolation dates did not suggest gradual replacement of one variant by another.

Table 3.

Distribution of novel SCCmec IVa variants in MRSA CC8 isolated from nine nursing homes

Nursing home	spa type (no. of strains)	SCCmec IVa variant(s)	Time of isolation
1	t024 (5)	B1, B3	2003–2008
	t430 (3)	B1	2004–2006
	t351 (1)	B1	2004
	t846 (1)	B3	2009
2	t024 (4)	B1	2003–2009
	t842 (1)	B1	2005
3	t024 (5)	B1, B3	2003–2009
4	t024 (5)	B1, B2, B3, NT1	2004–2007
5	t024 (2)	B4	2007–2009
6	t024 (2)	B3	2004
7	t024 (1)	B1	2005
	t2555 (1)	B1	2007
8	t024 (3)	B1	2004–2006
9	t024 (2)	B1, NT1	2006–2008

The present study reveals a wide dispersal of novel SCCmec IVa structures in CC8 and other MRSA lineages in the greater Copenhagen area. The B variants probably evolved from a common ancestral structure through ccr-mediated excision of the mobile elements between J3 and orfX. The variability in the structure of this region has important implications for the typing and identification of MRSA, since strains carrying B variants of SCCmec IVa cannot not be detected by current commercial PCR kits targeting the orfX J3 region. Our PCR mapping approach targeting the unique regions in J3 and upstream of SCCmec can be used for the subtyping of CC8 isolates diverging from the typical USA300 clone by lacking the opp3 gene cluster, being PVL negative, and having spa types other than t008.