|Home | About | Journals | Submit | Contact Us | Français|
Methicillin-resistant Staphylococcus aureus (MRSA) isolates carry the methicillin resistance gene (mecA) on a horizontally transferred genetic element called the staphylococcal chromosome cassette mec (SCCmec). Community-acquired MRSA (CAMRSA) isolates usually carry SCCmec type IV. We previously reported that 76% of 17 CAMRSA isolates (multilocus sequence type 59) obtained from pediatric patients with skin and soft tissue infections (SSTI) from Taipei did not carry SCCmec types I to IV. We used DNA sequence analysis to determine that the element harbored by these nontypeable isolates is a novel subtype of SCCmec V called SCCmec VT. It contains a ccrC recombinase gene variant (ccrC2) and mec complex C2. One SSTI isolate contained molecular features of SCCmec IV but also contained ccrC2 (a feature of SCCmec VT), suggesting that it may harbor a composite SCCmec element. The genes lukS-PV and lukF-PV encoding the Panton-Valentine leukocidin (PVL) were present in all CAMRSA SSTI isolates whether they contained SCCmec type IV or VT. SCCmec VT was also present in 5 of 34 (14.7%) CAMRSA colonization isolates collected from healthy children from Taipei who lacked MRSA risk factors. Four (80%) of the these isolates contained lukS-PV and lukF-PV, as did 1 of 27 (3.7%) SCCmec IV-containing colonization isolates. A total of 63% (10 of 16) of the SSTI isolates and 61.7% (21 of 34) of the colonization isolates tested were resistant to at least four classes of non-β-lactam antimicrobials. SCCmec VT is a novel SCCmec variant that is found in multiply resistant CAMRSA strains with sequence type 59 in Taipei in association with the PVL leukotoxin genes.
Methicillin-resistant Staphylococcus aureus (MRSA) was first recognized as a healthcare-associated pathogen in the 1960s (22). Since the early reports of community-associated MRSA (CAMRSA) (13, 27), infections caused by MRSA in patients lacking traditional MRSA risk factors with onset outside health care settings have been increasing globally (36). Clinical syndromes caused by these CAMRSA isolates have ranged from skin and soft tissue infections (SSTIs) to necrotizing pneumonia (1, 4, 9, 12, 25, 33), severe sepsis (1, 4, 25), and necrotizing fasciitis (24). Asymptomatic colonization with MRSA among healthy subjects, considered to be rare until recently, has also been documented in various populations, especially from the same geographic location from where CAMRSA infections have been reported (10, 14, 37). In CAMRSA infections in children without predisposing risk, the clinical syndromes resemble those of CA methicillin-susceptible S. aureus (MSSA) infections (13), and the responsible isolates lack multiple resistance to antimicrobials other than β-lactams (13, 41).
Two important genotypic characteristics have been associated with CAMRSA. The staphylococcal chromosome cassette mec (SCCmec) (15, 16, 21), the genetic element that carries the methicillin resistance gene, mecA, integrates into the orfX gene in the S. aureus genome in a site specific manner. SCCmec type IV (21) has been associated with CAMRSA in a variety of genetic backgrounds (23, 41). Outbreaks in the United States have been associated with isolates with sequence types (ST) 8 and 1 (23), determined by multilocus sequence typing (MLST) (11). Also associated with CAMRSA are the lukS-PV and lukF-PV genes that encode the two subunits that comprise the Panton-Valentine leukocidin (PVL), a synergohymenotropic cytotoxin associated with furunculosis, severe necrotizing hemorrhagic pneumonia, necrotizing fasciitis, and other lesions involving the skin or mucosa in both CAMRSA and CAMSSA strains (12, 18, 24). A high carriage rate of the lukS-PV and LukF-PV genes (i.e., the PVL locus) among CAMRSA has been documented in isolates associated with SSTIs, severe sepsis, necrotizing fasciitis, and necrotizing pneumonia (12, 24, 25, 41). Notably, the PVL locus is infrequently found among healthcare-associated MSSA (18) or MRSA isolates (41). Few studies have documented the prevalence of the PVL locus among CAMRSA colonizing asymptomatic individuals in nonoutbreak settings.
We previously analyzed 17 CAMRSA isolates obtained from patients with SSTIs from the Tri-Services General Hospital (TSGH), a tertiary care, military medical school-affiliated institution in Taipei, for antimicrobial susceptibility patterns, genotyping by MLST and SmaI genomic fingerprinting, the presence of the PVL locus (42), and the SCCmec type. The resistance phenotypes differed from CAMRSA described from other locations in that the isolates uniformly had the constitutive macrolide, lincosamide, and streptogramin B resistance (MLSBc) phenotype. Furthermore, the CAMRSA isolates were uniformly from the ST 59 genetic background and carried the PVL locus. The SmaI genome fingerprints of these CAMRSA isolates were similar to each other but differed from those of healthcare-associated isolates from the same institution. Moreover, only 3 of the 17 SSTI isolates studied carried SCCmec IV. The remaining isolates were nontypeable in that they lacked SCCmec types I to IV.
We examined here the SCCmec type in 13 of the 14 nontypeable SSTI isolates. This was accomplished by sequencing the ccr and mec complexes contained in one of the CAMRSA SSTI isolates and by screening the remaining isolates by PCR with type- and subtype-specific primers. The “nontypeable” isolates harbor a variant of the newly described SCCmec V element (16) that we have called SCCmec VT. We also examined antibiotic resistance profiles, SCCmec types and PVL locus prevalence among 48 CAMRSA isolates asymptomatically colonizing healthy children from Taipei.
All isolates were confirmed as S. aureus by a positive agglutination reaction using the Staphaurex Plus system (Remel) and by Gram staining. Isolates were frozen in skim milk (Difco) at −70°C. Two groups of CAMRSA isolates from Taipei were studied. The first group, TSGH 1 to 17, were consecutive CAMRSA isolates from children hospitalized at TSGH for SSTIs during the 5-year period from September 1997 to August 2002 (42). A case was considered community acquired if the isolate was obtained from a patient within 72 h of admission to TSGH. None of the SSTI patients had any of the selected risk factors for MRSA infection as described previously (42) as follows: (i) hospitalization within 6 months of the date of MRSA isolation, (ii) history of any surgical procedure, (iii) antimicrobial therapy within 6 months of the date of MRSA isolation, and (iv) household contact with an individual with an identified risk factor or a worker in a healthcare environment. One isolate (TSGH 6) was mecA negative and was excluded.
The CAMRSA colonization isolates (C1 to C48) were obtained in a 1-year period (January to December 2003) by culturing the nares of 640 healthy children. Subjects enrolled were 12 years of age or younger with no acute medical problem who either presented for a well-child healthcare visit or attended one of three kindergartens in Taipei near the TSGH. Isolates were stratified by whether or not they were from subjects with one or more of the risk factors listed above.
Strains ATCC 29213 (methicillin susceptible) and ATCC 43300 (methicillin resistant) were used as controls for oxacillin susceptibility testing. Control strains used for SCCmec typing—NCTC10442 (SCCmec I), N315 (SCCmec II), 85/2082 (SCCmec III), and WIS (SCCmec V)—were kindly provided by Keiichi Hiramatsu and Teruyo Ito (Juntendo University, Juntendo, Japan). Strain MW2 (SCCmec IV) was obtained from the Network for Antimicrobial Resistance in S. aureus (www.narsa.net).
Statistical comparisons were performed by using the chi-square test with a web-based chi-square calculator (http://www.georgetown.edu/faculty/ballc/webtools/web_chi.html). In the case of a comparison group with ≤5, the Fisher exact test (two-tailed) was used. Comparisons were considered significant if the P value was ≤0.05.
Strains were tested for susceptibility using the Vitek 2 system (bioMérieux Vitek, Inc., Hazelwood, MO) with a gram-positive card according to the manufacturer's recommendations in the Clinical Microbiology Laboratories at The University of Chicago Hospitals. When oxacillin susceptibility testing was performed by broth MIC analysis, procedures recommended by the Clinical and Laboratory Standards Institute (CLSI [formerly the National Committee for Clinical Laboratory Standards]) (30) were used. Trimethoprim-sulfamethoxazole (SXT) testing was performed by disk diffusion according to CLSI guidelines (30). The non-β-lactam antibiotics in the Vitek 2 panel were erythromycin (ERY), clindamycin (CLI), fluoroquinolones (ciprofloxacin, norfloxacin, ofloxacin, and levofloxacin), gentamicin (GEN), tetracycline (TET), chloramphenicol (CHL), and rifampin (RIF).
PCR was performed to detect mecA using the primer pair mecAF and mecAR as described previously (13) (Table (Table1)1) . SCCmec elements were distinguished by the molecular architecture of the ccr and mecA complexes as described previously (15, 25, 26). PCR typing of SCCmec types I to IV was performed under conditions previously described (21, 26). SCCmec type II (ccrAB complex type 2 and mec complex class A), SCCmec type III (ccrAB complex type 3 and mec complex class A), and SCCmec type IV (ccrAB complex type 2 and mec complex class B) were assigned according to previously described criteria (21). PCR primers used to detect mecI (primers mI3 and mI4), the mecR1 membrane-spanning region (MS) (primers mcR3 and mcR4), and the mecR1 penicillin-binding region (PB) (primers mcR1 and mcR5) were originally reported by Suzuki et al. (38) (Table (Table1).1). Screening for ccrAB complex types 1, 2, and 3 (ccrAB 1, 2, and 3) was accomplished with a new multiplex PCR assay which uses a mixture of four primers consisting of a common forward primer (β2) and reverse primers, α2, α3, and α4 specific for ccrAB complexes 1, 2, and 3 (15). Thermocycler conditions used were as follows: 94°C for 1 min, followed by 30 cycles of 94°C for 30 s, 63°C for 1 min, and 72°C for 1 min, followed in turn by a single extension at 72°C for 5 min. The presence of the ccrAB gene complex type 4 (ccrAB4) was assessed in a separate reaction that used the primer pair ccrA4F and ccrB4R (Table (Table1).1). Screening for the ccrC complex (ccr5) was performed by using a forward primer (γF) in combination with either the reverse primer γR described by Ito et al. (16) or CDS15-R designed in the present study (Table (Table1).1). Prototype strains used for SCCmec typing were NCTC10442 (SCCmec I), N315 (SCCmec II), 85/2082 (SCCmec III), MW2 (SCCmec IV), and WIS (SCCmec V). The control strain used for detection of ccrAB4 was S. epidermidis strain ATCC 12228 that contains ccrAB4 in the non-mec containing SCC composite island we recently described (26).
MLST was performed by PCR amplification and sequencing of seven housekeeping genes using the primer pairs designed by Enright et al. (11). Denville Taq-Pro Complete (Denville Scientific) or the Taq DNA polymerase (Promega) was used for the PCRs. PCR products were evaluated on an agarose gel and purified by using Millipore 96-well Montage plates according to the manufacturer's instructions. The purified templates were sequenced at the University of Chicago Core Sequencing Facility. Each sequence was submitted to the MLST database website (http://www.mlst.net) for assignment of the allelic profile and sequence type.
Agarose plugs were prepared containing intact bacterial cells and digested with SmaI as reported (7). The restriction fragments were resolved on a Chef DR-III pulsed-field gel electrophoresis (PFGE) apparatus (Bio-Rad) as described previously (23). On each gel, a SmaI digest of S. aureus strain 8325 was the molecular size standard. The USA pulsed-field type (PFT) to which the isolates belonged was determined by submitting images of pulsed-field gels in tagged image file format (Tiff) to Linda McDougal at the Centers for Disease Control and Prevention (CDC), who used Bionumerics software (Applied Maths, Austin, TX) to compare the images with those in the national database as described previously (23).
Pulsed-field gels were blotted onto GeneScreen Plus Hybridization transfer membranes overnight in a 10× SSC buffer (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) (2). The gel was washed in 0.25 M HCl for 15 min, in 1.5 M NaCl-0.5 M NaOH for 30 min, and in 3 M NaCl-0.5 M Tris for 30 min prior to blotting. Before hybridization, the blot was cross-linked using UV light with a UV Cross-linker FB-UVXL-1000 (Fisher Biotech/Fisher Scientific) and the “optimal cross-link” setting. The Promega Prime-a-Gene kit was used to radiolabel the appropriate probes with [α-32P]dATP (Amersham). Hybridization was performed at 70°C overnight with constant rotation in a hybridization buffer consisting of 7% sodium dodecyl sulfate, 1% bovine serum albumin, and 1 mM EDTA (pH 8) in 0.25 M Na2HPO4. Membranes were probed with a mecA gene probe produced by PCR with the primers mecAF and mecAR (Table (Table1).1). After overnight exposure on a Fuji Imaging Plate and scanning of the hybridization image with a PhosphorImager (Molecular Dynamics), membranes were stripped of the mecA probe by boiling in distilled water and rehybridized with an orfX gene probe produced by PCR, with the primer pair orfXprobeF-orfXprobeR. The aroE probe for detecting chromosomal DNA was produced by PCR with the same primers used to produce the aroE template for MLST analysis.
Sequencing was initiated by producing a λEMBL3 library from strain TSGH 17 and screening recombinant plaques for the presence of mecA by hybridization. This was accomplished by isolating genomic DNA from strain TSGH 17 with the Genomic Tip kit (QIAGEN, Inc.) and performing a Sau3AI partial digestion (2), followed by purification of ~20-kb products after agarose gel electrophoresis using the Qiaex II gel extraction kit (QIAGEN). These products were inserted into the λEMBL3 cloning vector (LambaGEM-11 BamHI Arms; Promega) by using an in vitro packaging kit (Promega Packagene Extract) as recommended by the manufacturer with the bacterial host strain LE392. Plaques containing the mecA gene were identified by hybridization with two [α-32P]dATP-labeled mecA gene probes produced by PCR with the primer pairs mecAF-mecAR and mecA aminoF-mecA aminoR (Table (Table1)1) in a final concentration of 2 μM in a standard PCR cocktail. DNA was isolated from mecA-hybridizing purified plaques with the use of the Lambda Midi kit (QIAGEN) and phenol-chloroform-isoamyl alcohol extraction, followed by ethanol precipitation to concentrate the sample using standard procedures (2). Inserts from lambda DNA were subcloned into the ClaI site of a dephosphorylated pBluescript (Stratagene) cloning vector by using T4 DNA ligase (NEB, Beverley, MA). Ligation products were transformed into Electromax STBL4 cells (Stratagene) by using a Bio-Rad Gene Pulser II and a 0.l-cm cuvette with settings of 1.2 kV, 25 μF, and 200 Ω. Transformed cells were plated onto LB agar supplemented with ampicillin (100 μg/ml) and overlaid with 100 μl of 2% X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) dissolved in dimethyl formamide. Plasmid DNA from white colonies was isolated with the QIAprep Spin Miniprep kit (QIAGEN, Inc.) and sequenced with the M13/pUC universal forward and reverse primers (5′-GTAAAACGACGGCCAGT-3′ and 5′-CACACAGGAAACAGCTATGACCAT-3′, respectively).
To obtain the sequence of the ccr complex, primers were designed from the sequences obtained from λEMBL3 subclones mentioned above in conjunction with ccrC-specific primers. In addition, primers were designed to extend the sequence obtained from the cloned fragments using a primer-walking strategy. Raw sequence data consisting of high-pressure liquid chromatography chromatograms were evaluated, edited, and assembled into contigs using software packaged within the VectorNTI suite (version 8; Informax, Inc., Bethesda, Md.). For completion of the sequence of the mecA complex, a primer-walking strategy with an initial set of mecA- and IS431-specific primers (V3, mA2, and mA3  was used.
Isolates were screened for the lukF-PV and lukS-PV genes encoding the PVL toxin by PCR amplification with the primer pair PVL-1 and PVL-2 (final concentration, 10 μM) that produces a 3.5-kb product (25) encompassing both lukF-PV and lukS-PV open reading frames (ORFs) and flanking DNA. The PVL locus was occasionally screened by using primer pair luk-PV-1 and luk-PV-2 (Table (Table1)1) (final concentration, 10 μM) that amplifies a 433-bp product (18) that includes a portion of both the lukS-PV and lukF-PV ORFs. The thermocycler conditions used with the PVL-1-PVL-2 primer pair consisted of 94°C for 1 min, followed by 35 cycles of 94°C for 30 s, 63°C for 30 s, and 72°C for 1 min, with a final extension performed at 72°C for 7 min. The thermocycling conditions used with the luk-PV-1-luk-PV-2 primer pair were 95°C for 1 min and then 30 cycles of 95°C for 1 min and 68°C for 1 min, with a final extension performed at 68°C for 1 min.
The DNA sequences of the mec complex and the ccrC2 gene from TSGH 17 have been deposited in GenBank under accession numbers AY894415 and AY894416, respectively.
Of the 640 healthy children who were screened, 157 (24.5%) had a culture yielding S. aureus. Forty-eight (31%) of these were MRSA. Thus, the overall MRSA colonization rate was 7.5%. However, 14 of the isolates were from subjects that had at least one designated risk factor for MRSA; these were excluded from further analysis unless otherwise indicated. Thus, the MRSA colonization rate in subjects lacking the designated risk factors for MRSA was 5.3%.
To assess resistance to non-β-lactam antimicrobials, 16 SSTI CAMRSA isolates (TSGH 1 to 5 and TSGH 7 to 17 [excluding TSGH 6, which was methicillin susceptible]) and the colonization isolates were subjected to susceptibility testing by using the Vitek 2 system (Table (Table22).
Unlike the usual pattern for CAMRSA, resistance to the non-β-lactam antimicrobials was common among both groups of CAMRSA from Taiwan (Table (Table2).2). Consistent with our previous results obtained by disk diffusion (42), 100% of the SSTI isolates were resistant to ERY and CLI (Table (Table2).2). Of the 34 colonization isolates from patients lacking risk factors, 94.1 and 91.2% were resistant to ERY and CLI, respectively (Table (Table22).
Multiple resistance to non-β-lactam antimicrobials was also highly prevalent among the SSTI and colonization CAMRSA isolates (Fig. (Fig.1).1). Figure Figure11 shows the percentage of isolates with resistance to ≤2, 3, or ≥4 non-β-lactam antimicrobials. One (6.3%) SSTI isolate was resistant to five non-β-lactam antimicrobials. A total of 63% (10 of 16) of the SSTI isolates and 62% (21 of 34) of the colonization isolates were resistant to ≥4 non-β-lactam antimicrobials (counting the four quinolones as a single antibiotic) (Fig. (Fig.1).1). Using a less stringent definition, 94% (15 of 16) of the SSTI isolates and 85.3% (29 of 34) of the colonization isolates were resistant to ≥3 non-β-lactam antibiotics. Of the 34 colonization isolates, 1 (2.9%) was susceptible to all of the non-β-lactam agents tested. Both the SSTI and colonization isolates were uniformly susceptible to the fluoroquinolones and SXT.
Table Table33 shows the results of SCCmec typing for the colonization and the SSTI isolates. Of the colonization isolates that were from subjects that lacked an MRSA risk factor, a majority (27 of 34 [79.4%]) harbored SCCmec IV. SCCmec IV was also carried by all 14 colonization isolates that were from the subjects with at least one MRSA risk factor. Thus, 34% of the 41 SCCmec IV-containing colonization isolates were from subjects who had an MRSA risk factor.
One colonization isolate (2.9%) harbored SCCmec II, one (2.9%) harbored SCCmec III, and five (14.7%) harbored an element that was nontypeable when assessed for SCCmec types I to IV.
Among the 34 colonization isolates, the predominant genotype (91.2%) was ST 59 (Table (Table4).4). Only three of the isolates had unique backgrounds: ST 89, ST 508, and ST 239. In addition, all 14 isolates that were associated with an MRSA risk factor were ST 59.
The CDC has recently reported a system for normalizing pulsed-field patterns determined from different laboratories and assigning a PFT (designated with the prefix USA) based on similarity calculations and clustering with the use of Bionumerics software (23). To determine whether the PFT of the ST 59 isolates from Taipei clustered with ST 59 isolates in the U.S. national database, representative SmaI pulsed-field patterns from a sample of the colonization (every third isolate up to C28) and SSTI isolates (42) (Fig. (Fig.2B)2B) were submitted to the CDC. The SSTI isolates and the colonization isolates clustered with PFT USA 1000 (73% similarity). ST 59 isolates from Taiwan were more closely related to each other (76% similarity). One pattern from the ST 59 lineage (isolate TSGH 5) did not match any known PFT in the national database and had >6 band differences compared with that of the other ST 59 isolates. The PFT of the ST 89 isolate did not correspond to any PFT in the CDC database.
To gain insight into the size and architecture of the nontypeable SCCmec element, Southern blotting of SmaI-digested DNA from CAMRSA isolates from SSTIs and an ST 59 MSSA isolate was performed with mecA (Fig. (Fig.2C)2C) and orfX (Fig. (Fig.2D)2D) hybridization probes. The hybridization pattern was similar among all strains containing a nontypeable element, suggesting they all contained a similar element.
Two findings distinguished the SCCmec IV-containing isolates from those containing a nontypeable element. First, there was a striking difference between the sizes of the mecA- and orfX-hybridizing bands in the isolates carrying a nontypeable SCCmec (Fig. 2C and D, TSGH 1, 2, 3, 4, 7, 8, and 9). This is best illustrated in Fig. Fig.2D,2D, where the mecA-specific band is smaller than the orfX-hybridizing band. In contrast, the orfX and mecA gene probes cohybridized in the SCCmec IV-containing strains (Fig. (Fig.2D,2D, lane 5).
The separation of the mecA and orfX hybridizing bands in the ST 59 strains containing a nontypeable element suggested that (i) the nontypeable SCCmec element was not inserted into the attB insertion site in the orfX gene, as are all other known SCCmec elements; (ii) the nontypeable element was present on a plasmid; or (iii) the nontypeable element contained a SmaI recognition sequence between mecA and orfX.
Southern blotting of undigested genomic DNA from one isolate (TSGH 3) demonstrated that the mecA gene cohybridized with the chromosomal gene probe aroE (data not shown). A plasmid could not be detected on agarose gels. These data suggested that the mecA and orfX Southern blotting results can be explained by the presence of an internal SmaI restriction site in the nontypeable element integrated into the genome and argued against the presence of an SCCmec element on a plasmid.
To determine the SCCmec type of the nontypeable isolates, the DNA sequences of the ccr and mecA complexes of one of the SSTI isolates, TSGH 17, were determined (see Materials and Methods). By performing a BLAST search with sequences deposited in GenBank, one 5,753-bp contig was found to contain highly significant sequence similarity with the ccrC complexes from three other SCC elements (expectation scores ranged from 0 to 3.8 with sequence identities of 80 to 100%): (i) SCCmec III (strains 85/3907 and 85/2082), (ii) SCCmec V (strain WIS), and (iii) SCCcap1 (strain M). The last element does not contain mecA but does contain a ccrC homologue and encodes the type 1 capsular polysaccharide biosynthesis gene cluster (20). Individual nucleotide sequence alignments between the ccrC ORF from strain TSGH 17 and the corresponding 1,677-bp region from each of the SCC elements mentioned above revealed nucleotide identities of 97.7% (SCCmec III, 85/3907), 94.8% (SCCmec III, 85/2082), 90.3% (SCCmec V, WIS), and 88.6% (SCCcap1, strain M). Based on these data we designated these separate ccrC alleles as ccrC1 (strain WIS), ccrC2 (strain TSGH 17), ccrC3 (strains 85/2082), and ccrC4 (strain M).
The sizes of the ccrC ORFs differed among the five elements due to nucleotide polymorphisms that created stop codons in different locations (Fig. (Fig.3A).3A). A deletion at nucleotide 1612 of the ccrC ORF in strain TSGH 17 abolished a stop codon and extended the ccrC ORF by 57 nucleotides compared to that in strain WIS (Fig. (Fig.3A),3A), making the ccrC ORF from TSGH 17 the largest of the ccrC alleles described to date (1,677 bp). Interestingly, a truncated ccrC (ccrC′) and a unique overlapping ORF at the 3′ end are present in SCCmec III in strain 85/3907 but not in strain 85/2082 (Fig. (Fig.3A)3A) due to a premature stop codon in the former.
A multiple sequence alignment was performed between the 1,677-bp ccrC ORF from strain TSGH 17 and the corresponding regions encompassing the ccrC homologues from all four SCC elements mentioned above (data not shown). The sequence alignment tree (Fig. (Fig.3B)3B) illustrates that the ccrC ORF from SCCmec V from strain WIS is most closely related to that of SCCcapl and more distantly related to SCCmec VT.
The dendrogram in Fig. Fig.3D3D is based on a multiple sequence alignment between four ccrC alleles and ccrA1, ccrA2, ccrA3, ccrB1, ccrB2, and ccrB3. This tree clearly illustrates how the four ccrC homologues form distinct branches in their cluster and how the ccrA, ccrB, and ccrC homologues form separate clusters.
The mec complex of strain TSGH 17 consists of mecAΔmecRI flanked by IS431 elements positioned in opposite orientation that point toward the outside of the element (Fig. (Fig.4).4). This architecture conforms to that of mec complex class C2 (17), similar to that in SCCmec V of strain WIS (16, 17). The overall similarity between the mec complex of TSGH and that of strain WIS was 99.4%. However, a single nucleotide polymorphism in the IS431 transposase gene upstream of mecA (C144 in strain WIS is G144 in strain TSGH 17) converts a Tyr codon to a premature translational stop codon in strain TSGH 17. An additional distinguishing feature of the mec complex of strain TSGH 17 is the presence of an extra direct repeat unit (dru) in the intergenic hypervariable region upstream of the glycerophosphoryl diester phosphodiesterase gene (Fig. (Fig.4).4). This dru is responsible for the length polymorphism of SCCmec elements (34). There are four such dru's in strain TSGH 17 and three in strain WIS.
Thus, the SCCmec element in TSGH 17 carries the ccrC2 recombinase gene and the mec complex C2, with signature molecular features in both ccrC2 and the mec complex that clearly distinguish this element from the SCCmec type V element in strain WIS. The presence of mec complex C2 in combination with a ccrC2 homologue indicates that TSGH 17 carries a variant of SCCmec V, which we call SCCmec VT.
To determine whether the remaining SCCmec nontypeable isolates contained SCCmec VT, primers described by Ito et al. (16) to characterize SCCmec V (WIS) and those derived from the ccrC2 and mec C2 complexes of TSGH 17 were used. To detect the presence of ccrC, a forward primer, γF, was used in combination with either the reverse primer γR described by Ito et al. (16) or a reverse primer (CDS15-R) designed for the present study (Table (Table1).1). When using either of these primer pairs, products of the expected sizes (0.52 and 2.2 kb, respectively) were produced from strain WIS, strain TSGH 17, and all remaining nontypeable SSTI strains (Table (Table3).3). None of the prototype strains containing SCCmec types I to IV gave a product using these primers, validating the use of γF/CDS15-R as SCCmec V screening primers. In addition, no product was detected from the ST 59 strains that contained SCCmec II, III, or IV. To determine whether the ccrC ORF from the nontypeable isolates was similar to that of TSGH 17, the ccrC nucleotide sequence was determined for all nontypeable isolates starting at position 1249 of the ccrC ORF (relative to that of TSGH 17) and ending with the TAG stop codon (Fig. (Fig.3C).3C). All but one of the sequenced strains had an identical ccrC sequence to that of TSGH 17 within this region, including the nucleotide polymorphisms that resulted in an extended ccrC ORF (Fig. (Fig.3C).3C). The one different strain had only one nucleotide polymorphism (data not shown). These data strongly suggest that all 18 nontypeable strains (from both SSTI and colonization isolates) contain an element that is highly similar to SCCmec VT.
To screen for the unique architecture of mec complex C2 (17), which is flanked by two complementary inverted repeats of IS431, we developed a PCR assay using a single IS431-specific primer (IS-5) (Table (Table1)1) that reads toward the inside of mec complex C2. A 5.5-kb product was produced from TSGH 17, all other nontypeable strains and, as expected, from WIS, the SCCmec V prototype strain (Table (Table3).3). This primer did not amplify a product from the prototype strains containing SCCmec types I to IV. Another distinguishing feature of the mec complex from strains TSGH 17 and strain WIS was the absence of the mecR1 membrane-spanning domain (MS) and the mecR1 penicillin-binding domain (PB) (Table (Table3).3). Similarly, a product for neither the mecR1 (MS) nor the mecR1 (PB) was detected from any of the nontypeable isolates in the present study (Table (Table3).3). This is in contrast to mec complex class B that contains the mecRI (MS), although it lacks the mecR1 (PB) (15, 17) (Table (Table33).
While amplifying templates for sequencing, we found a primer (ccrC-FR) (Table (Table1)1) that produced a 4-kb product when used as the only primer in a PCR with TSGH 17 genomic DNA. No product was formed from the prototype strain WIS containing SCCmec V (Table (Table3).3). ccrC-FR also produced a 4.0-kb product from all of the nontypeable strains but not from any of the SCCmec type I- to IV-containing isolates.
Thus, from the ccrC sequence data, the PCR screening assays, and the Southern blotting data, we conclude that the nontypeable elements in these Taiwan isolates are uniformly SCCmec VT. Therefore, 5 (14.7%) of the 34 colonization isolates contained SCCmec VT and 26 (76.4%) contained SCCmec IV (Table (Table4).4). In contrast, all 13 nontypeable SSTI isolates studied had SCCmec VT (81.3%) and 2 SSTI isolates (12.5%) had SCCmec IV.
Table Table44 also shows the STs of the colonization isolates stratified by the number of isolates containing each SCCmec type. Of the 31 ST 59 isolates, 26 (83.9%) carried SCCmec IV. All four SCCmec VT isolates were ST 59. Each of the isolates with a unique ST carried a different SCCmec type: ST 89 (SCCmec II), ST 239 (SCCmec III), and ST 508 (SCCmec IV). Thus, one SCCmec IV-containing strain had an ST other than ST 59.
TSGH 10 contained mec complex type B (Table (Table3).3). In the ccr assays, both ccrC and ccrAB2-specific products were detected (Fig. (Fig.5).5). The entire sequence of the ccrC ORF from TSGH 10 was determined and is identical to that of TSGH 17. These data suggest that TSGH 10 contains SCCmec IV (ccrAB2 complex and mec class B), but the detection of the ccrC gene from SCCmec VT suggests the presence of a composite island.
Only 5 (14.7%) of the 34 CAMRSA colonization isolates from healthy children harbored the PVL locus (Table (Table5).5). The PVL locus was more frequent among the colonization isolates that harbored SCCmec VT (Table (Table5)5) (P = 0.0005 [Fisher exact test]). This is unlike the SSTI isolates, which all harbored the PVL locus (42) (Table (Table5)5) irrespective of whether they harbored SCCmec type IV or VT.
We previously described the molecular epidemiology and resistance patterns of CAMRSA isolates that caused SSTIs in patients from the TSGH in Taipei (42). In that study we found that all of the SSTI CAMRSA isolates studied were of the ST 59 genetic background, and a majority harbored a nontypeable SCCmec element with a few harboring SCCmec IV. All of them carried the PVL locus associated with SSTIs, severe sepsis, necrotizing pneumonia, and necrotizing fasciitis. We have now designated the nontypeable element as SCCmec VT, a variant of the SCCmec V element described recently (16). We have also identified SCCmec VT-containing isolates among a group of CAMRSA colonization isolates from the same geographic location. This is the first report that documents a high prevalence of an SCCmec element other than SCCmec IV among a group of CAMRSA isolates (41).
The colonization isolates that carried SCCmec VT were more likely to carry the PVL locus than those carrying SCCmec IV. This is also a departure from the situation elsewhere in which the PVL locus has been found exclusively in SCCmec IV-containing isolates (41) or in MSSA isolates. Among the SSTI isolates from Taipei, the PVL locus was uniformly present among both SCCmec types IV and VT-containing isolates. It has been suggested that the PVL leukotoxin may be the determinant that is favoring the spread of MRSA isolates in the community (5) since the PVL locus has been less often associated with nosocomial MRSA or MSSA infections. In the present study, a majority of CAMRSA isolates asymptomatically colonizing healthy individuals lacked the PVL locus, suggesting that the PVL toxin is not required for the successful spread of CAMRSA, at least of the ST 59 genetic background. These data also suggest the need to design studies to determine whether CAMRSA isolated from patients with an SSTI, necrotizing pneumonia or necrotizing fasciitis are more likely to contain the PVL locus than CAMRSA that colonize asymptomatic individuals.
This is also the first report documenting such a high prevalence of a single genotype carrying an SCCmec V element. SCCmec V has also not been previously identified in the ST 59 (PFT USA 1000) genetic background or in an isolate from a continent other than Australia (6, 16). The predominant clone of CAMRSA currently circulating in the United States is from the ST 8 genetic background (PFT USA 300) (5). ST 1 (PFT USA 400) has also been reported among CAMRSA (41), especially among patients presenting with severe sepsis syndrome with necrotizing pneumonia (1, 25). We have also reported MSSA ST1 isolates in association with severe sepsis and necrotizing pneumonia (1, 25). MRSA isolates with ST 59 have been reported infrequently, mainly from San Francisco (3, 8, 32), and have usually contained SCCmec IV (3, 8). Sporadic ST 59 isolates with SCCmec II (3), SCCmec III (32) or a nontypeable SCCmec element (NT1) (3) have also been reported. Considering the high prevalence of ST 59 in CAMRSA isolates from Taiwan, it is tempting to speculate that the ST 59 isolate carrying the nontypeable element circulating in San Francisco might have originated from Taipei.
The genetic backgrounds that SCCmec V has been found in previously were all from Australia and are ST 45 (strain WIS, unpublished data), ST 8, and ST 152 (6). Thus, the ST 59 background is the fourth into which an SCCmec V-like element has been introduced (but the first in which SCCmec VT has been found).
SCCmec V is characterized by the presence of a ccrC recombinase complex and mec complex type C2 (16). The SCCmec VT variant we identified contains signature features in both its mec and ccr complexes. The nucleotide polymorphisms we found in the ccr region of SCCmec VT extended the ccrC ORF compared with that found in strain WIS and led us to distinguish it from the other ccrC homologues by naming it ccrC2. We have also designated the other ccrC alleles as ccrC1 (SCCmec V), ccrC3 (SCCmec III), and ccrC4 (SCCcap1). Future studies will reveal whether the longer CcrC2 recombinase has a different activity than the other ccrC alleles. Interestingly, the sequence of the ccrC2 gene and flanking sequence is more closely related to that found in SCCmec III than to that in the prototype strain SCCmec V. Also, the ccrC1 gene in SCCmec V is most similar to that in SCCcap1. These observations suggest that SCCmec VT and SCCmec V were formed by independent recombination events with ccrC genes from separate sources or that the ccrC sequences underwent divergent evolution. In addition, in one isolate, we have provided evidence for the existence of a novel composite SCCmec element that contains all of the features of SCCmec IV but also contains a ccrC2 homologue. Whether the origins of SCCmec V will be found in that isolate is the subject of ongoing investigation.
SCCmec type IV was the fourth allotype of the integrated genomic island (SCCmec) found to carry the mecA gene (7, 21). The type IV element has associated with CAMRSA in multiple genetic lineages, even in narrow geographic locations (7), a finding that suggests ease of horizontal transfer, likely to be facilitated by its small size compared to the types I to III SCCmec allotypes (21). Since SCCmec IV is usually found among CAMRSA isolates, it is interesting that approximately one-third of the SCCmec IV-containing colonization isolates were associated with an MRSA risk factor.
The SCCmec V element described from strain WIS was similar to SCCmec IV in that it did not harbor any antibiotic resistance genes other than mecA and in its relatively small in size. In contrast, strain TSGH 17, the prototype SCCmec type VT strain, was one of the CAMRSA isolates resistant to four non-β-lactam agents (ERY, CLI, TET, and CHL). Completion of the sequence of SCCmec VT from this isolate will reveal whether it harbors any of these antibiotic resistance determinants.
The CAMRSA colonization isolates were obtained by screening 640 healthy children from Taipei at locations near the TSGH. The CAMRSA colonization rate among subjects lacking traditional risk factors was 5.3%, a high rate compared with 0.24% calculated in a meta-analysis (36), 0.6% among children in Chicago (14) in 1999, 3.5% among healthy people in southern Taiwan (19), and 1.4% in Hong Kong (31). It remains to be determined whether this high MRSA colonization rate is widespread in Taipei and other locations in Taiwan or can be found among adults from the same geographic location. We have avoided making direct comparisons between the SSTI and colonization isolates since the SSTI isolates were collected in a nonoverlapping time frame. Thus, any differences may simply reflect a temporal change in the epidemiology of CAMRSA isolates circulating in this area of Taipei. Nevertheless, the observation that five SCCmec VT-containing MRSA isolates were among the colonization isolates indicates that the SCCmec VT-containing clone was still circulating in 2003. The differences found in SCCmec VT and PVL locus prevalence between the SSTI and colonization isolates warrant a comparative study in which the SSTI and colonization samples are gathered during the same time period.
The rate of multiple resistance to at least four non-β-lactam antimicrobials among the CAMRSA in this study population was unusually high for CAMRSA, which have tended to be “non-multiply resistant” to non-β-lactam agents in the United States, Europe, Australia, and elsewhere (6, 13, 28, 29, 41). This trend was true for both SSTI and colonization isolates in association with either SCCmec IV or SCCmec VT. These data are consistent with a recent report documenting high prevalence of multiple resistance in CAMRSA from Taiwan (19) and may reflect high antimicrobial usage in the community.
Despite the high rates of multiple resistance among our CAMRSA isolates, they were still less often multiply resistant than the healthcare-associated isolates from the same institution (42) and from colonization isolates reported elsewhere in Taiwan (19). Our isolates were uniformly susceptible to fluoroquinolones and SXT, and only one isolate was resistant to RIF. Although we also found uniform vancomycin susceptibility, Vitek 2 testing is unreliable for detecting intermediate or vancomycin resistance in S. aureus (39, 40).
Thus, the presence of the multiply resistant ST 59/PFT USA 1000 CAMRSA clone containing the PVL genes circulating in the community has the capability of sharply limiting therapeutic options should it become widespread.
This study was supported by grants from the CDC (R01 CCR523379), the National Institutes of Health (R01 AI40481-01A1), the TSGH (TSGH-C92-77), and The Grant Healthcare Foundation (Lake Forest, IL).
We are grateful to Linda McDougal from the CDC (Atlanta, GA) for comparing the pulsed-field types of the isolates described herein with those in the national database. We are grateful to Daniel Kim for providing assistance with and maintaining the MRSA database.