Detection of New Methicillin-Resistant Staphylococcus aureus Strains That Carry a Novel Genetic Homologue and Important Virulence Determinants

Abstract

In this study, 18 methicillin-resistant Staphylococcus aureus (MRSA) isolates harboring staphylococcal cassette chromosome mec (SCCmec) type XI, recovered in the Dutch-German Euregio, were characterized by DNA microarrays. In contrast to previous data, we found two MRSA strains of different clonal lineages possessing SCCmec XI that carried important virulence determinants. The worrisome emergence of such toxigenic MRSA strains raises concerns that MRSA strains with enhanced virulence potential and impaired detectability by standard molecular assays may spread in Europe.

TEXT

The emergence of methicillin-resistant Staphylococcus aureus (MRSA) in livestock has caused significant concern during the past decade (4, 11).Very recently, the situation has been complicated by the detection of bovine and human MRSA isolates carrying a new staphylococcal cassette chromosome mec (SCCmec) element (type XI) that leads to phenotypic β-lactam resistance (9, 19). This SCCmec element contains a novel mecA homologue, first isolated from strain LGA251 and designated mecA_LGA251, which is not reliably detectable by conventional molecular approaches and PBP2a agglutination tests (9, 19). MRSA isolates with mecA_LGA251 have been recovered so far in England, Denmark, Germany, and Ireland and mostly belonged to clonal complex 130 (CC130) as determined by multilocus sequence typing (MLST) (5, 9, 12, 19). Other MRSA isolates with SCCmec XI were representatives of MLST CC425, CC705, and CC1943 (9). However, an analysis of the virulence potential of MRSA harboring SCCmec XI was until now only reported for isolates associated with the CC130 clonal lineage (S. aureus protein A gene [spa] sequencing types t373, t843, t1736, and t1773), which demonstrate an absence of pyrogenic toxin superantigen (PTSAg)-encoding genes, including those of the toxic shock syndrome toxin-1 (TSST-1) and staphylococcal enterotoxins, and of exfoliative toxin genes (with the exception of the edinB gene) (5, 19).

In this study, we compare the virulence-associated gene content of 18 MRSA isolates from three distinct clonal lineages containing the novel mecA homologue mecA_LGA251.

A total of 18 nonrepetitive MRSA isolates harboring SCCmec XI with mecA_LGA251 were characterized in this study. Among these, 16 isolates were obtained from a previously described collection already characterized by spa typing, antimicrobial susceptibility testing, and the presence and orientation of mecA_LGA251, mecR1, mecI, blaZ, ccrA, and ccrB, which are related to type XI SCCmec (12). The isolates had spa types t843 (n = 11), t978 (n = 1), t1535 (n = 2), t1773 (n = 1), and t7189 (n = 1). The remaining two isolates were newly recovered from nasal swabs of patients in hospitals participating in the project EurSafety Health-net (www.eursafety.eu) and located in the Dutch-German Euregio Twente-Münsterland. One isolate was recovered on the Dutch side and one isolate on the German side of the border. All isolates were characterized by antimicrobial susceptibility testing by means of the Vitek 2 system (bioMérieux SA, Marcy l'Etoile, France) and the Clinical and Laboratory Standards Institute (CLSI) reference broth microdilution (BMD) method (2). The BMD method was performed with cation-supplemented Mueller-Hinton broth (BBL; BD Diagnostic Systems, Sparks, MD), and the CLSI-recommended reference strain S. aureus ATCC 29213 was used as the control with every set of tests. Susceptibility was determined according to standard CLSI breakpoints (3). The isolates were characterized by using the Xpert MRSA-SA assay on a GeneXpert DX system real-time PCR platform (Cepheid, Sunnyvale, CA). DNA microarray analysis was performed using the StaphyType and PM7Plus kits (Alere Technologies GmbH, Jena, Germany). spa typing was carried out as described by Aires-de-Sousa et al. (1) using StaphType software version 2.2.1 (Ridom GmbH, Münster, Germany) (10). Clustering of the isolates was conducted by the BURP (based upon repeat pattern) algorithm implemented in StaphType, with the calculated cost between members of a group set as less than or equal to 4 (default value). MLST was performed as described by Enright et al. (7). The MLST groups were defined by the eBURST (based upon related sequence types) method, using eBURSTv3 (eburst.mlst.net/v3/mlst_datasets) (8, 20). Isolates with at least six of the seven alleles being identical were classified in the same clonal complex. Total DNA was extracted from all isolates by enzymatic lysis using the buffers and solutions provided with the StaphyType DNA microarray kit (Alere Technologies GmbH, Jena, Germany) and the Qiagen DNeasy blood and tissue kit (Crawley, West Sussex, United Kingdom) as described previously (15). The presence of tested genes encoding species markers, agr types, and virulence factors, as well as the affiliation to clonal complexes, was determined by using the StaphyType DNA microarray kit as described elsewhere (14, 15).

In the testing of 277 phenotypically determined MRSA isolates by microarrays using the StaphyType kit, one isolate of spa type t5930 did not show a signal from the mecA and other probes designed for the detection of the SCCmec elements, with the exception of the probe for ccrA3 recombinase. Furthermore, this isolate was tested by microarrays using the not-yet commercially available PM7Plus kit for confirmation of the presence of SCCmec XI by detecting mecA and blaZ alleles of strains M10 and LGA251. This experiment revealed the presence of SCCmec XI in the tested isolate, because mecA and blaZ alleles of strains M10 and LGA251 were detected.

Using the BURP analysis, we found a single isolate in our collection showing spa type t7603 that had not been previously described as mecA_LGA251 MRSA but which differed from spa type t843 only by one repeat. The microarray approach using the StaphyType and PM7Plus kits revealed that this isolate contained SCCmec XI, because mecA and blaZ alleles of strains M10 and LGA251 were detected.

All isolates were mecA negative in the GeneXpert MRSA real-time PCR assay. Based on Vitek 2 susceptibility determination, all isolates tested were resistant to penicillin and cefoxitin but susceptible to gentamicin, tobramycin, ciprofloxacin, levofloxacin, moxifloxacin, erythromycin, clindamycin, linezolid, teicoplanin, vancomycin, tetracycline, fosfomycin, nitrofurantoin, fusidic acid, mupirocin, rifampin, and trimethoprim-sulfamethoxazole. Using the CLSI BMD method, we found 2 isolates which displayed an oxacillin MIC in the CLSI susceptible breakpoint range (≤2 μg/ml). These 2 isolates had spa types t843 and t7603 with oxacillin MICs of 0.35 and 0.6 μg/ml, respectively. Two other mecA_LGA251-positive isolates showed oxacillin MICs as low as 3 μg/ml. The remaining isolates displayed an oxacillin MIC in the CLSI resistant breakpoint range (≥4 μg/ml) (Table 1). The BMD results showed that all isolates were resistant to cefoxitin and penicillin G but susceptible to all non-β-lactam antibiotics (Table 1).

Table 1.

MICs of β-lactam and non-β-lactam antibiotics for the mecA_LGA251 MRSA isolates, determined by broth microdilution

Antimicrobial class and agent	MIC range (μg/ml)
β-lactams
Penicillins
Penicillin	0.4–8
Oxacillin	0.35–32
Cephalosporins
Cefoxitin	>8–64
Non-β-lactams
Glycopeptides
Vancomycin	0.5–1
Teicoplanin	0.5–1
Lincosamides
Clindamycin	0.125–0.25
Oxazolidinones
Linezolid	0.125–0.5
Rifamycins
Rifampin	0.004–0.008
Fosfomycins
Fosfomycin	1.2–9.6
Folate pathway inhibitors
Trimethoprim-sulfamethoxazole	0.024–0.48
Aminoglycosides
Gentamicin	0.2–0.4
Tobramycin	0.256–0.512
Fluoroquinolones
Ciprofloxacin	0.08–0.16
Levofloxacin	0.2–0.4
Moxifloxacin	0.04–0.08
Macrolides
Erythromycin	0.025–0.4
Monoxycarbolic acids
Mupirocin	0.128–0.256
Tetracyclines
Tetracycline	0.125–1
Nitrofurans
Nitrofurantoin	>2–8
Others
Fusidic acid	0.064–0.128

For all 18 MRSA isolates included in this study, a BURP analysis was performed, which clustered 16 isolates associated with five spa types (t843, t1535, t1773, t7189, and t7603) into the spa clonal complex 843/1535 (spa CC843/CC1535), while 2 remaining isolates (spa types t978 and t5930) were singletons. In previous investigations, isolates of spa types t843 and t1773 were characterized by MLST as clonal complex 130 (CC130) (5, 9, 19). Therefore, the isolates of spa CC843/CC1535 were attributed to MLST CC130. Furthermore, we characterized the isolates of spa types t978 and t5930 by MLST, which exhibited a new sequence type (ST) assigned by the S. aureus MLST curator as ST2361 and an allelic profile which matched that of ST599 in the MLST database (saureus.beta.mlst.net), respectively. To determine the genotypes in the whole S. aureus MLST database that were closely related to the newly identified STs with SCCmec XI, the allelic profiles of ST2361 and ST599 were subjected to an eBURSTv3 analysis. The eBURSTv3 method allocated ST2361 to CC1943, as it differed from ST1943 at only one locus. For ST599, with the exception of ST2179 (single-locus variant of ST599), the eBURSTv3 analysis did not find related STs.

The DNA microarray (Alere Technologies) results for the 18 mecA_LGA251 MRSA isolates are shown in Table 2. Isolates of spa CC843/CC1535 did not carry toxin genes, with the exception of the edinB gene. In contrast, other identified MRSA genotypes containing SCCmec XI were characterized by the presence of higher number of toxin genes: the t5930 isolate possessed three PTSAg genes (tst1, sec, and sel), while the t978 isolate carried eight PTSAg genes (tst1, sec, seg, sei, sel, sen, seo, and seu).

Table 2.

Virulence microarray hybridization profiles of the mecA_LGA251 MRSA isolates assigned to spa CC843/CC1535 complex and spa types t978 and t5930

Gene class	Gene(s)	DNA microarray hybridization profile of isolates assigned toa:
		CC843/CC1535	t978	t5930
Species markers	coa, gapA, nuc, katA, spa	Pos	Pos	Pos
Hemolysins	hla, hlb, hld	Pos	Pos	Pos
Immune evasion	sak, chp, scn	Neg	Neg	Neg
	aur	Pos	Pos	Pos
	splA, splB	Pos	Pos	Pos
	splE	Neg/Posb	Pos	Pos
Superantigens	tst1	Neg	Pos	Pos
	sea, seb, see, seh	Neg	Neg	Neg
	sec-sel	Neg	Pos	Pos
	sed-sej-ser	Neg	Neg	Neg
	seg-sei-sem-sen-seo-seu	Neg	Pos	Neg
	sek-seq	Neg	Neg	Neg
Exfoliative toxins	eta, etb, etd	Neg	Neg	Neg
	edinA	Neg	Neg	Neg
	edinB	Pos	Neg	Neg
Biofilm	ica operon	Pos	Pos	Pos
Capsule		cap8	cap5	cap5
agr type		III	IV	I

^aOnly clinically relevant markers, including exotoxins and resistance genes, are shown. The table does not represent the complete hybridization profile of 334 probes on the array. Pos, positive; Neg, negative.

^bTwo t843 isolates and one t1773 isolate were splE positive.

So far, MRSA isolates with mecA_LGA251 have rarely been detected in cases of human infections. However, the identification of such isolates from joint fluids and infected wound sites (12) indicates that they have the ability to cause infections in humans. Therefore, the emergence of strains with novel mecA_LGA251 harboring numerous toxin genes is worrying, since those isolates are not detectable by either commercially available or in-house nucleic acid detection approaches targeting the classical mecA gene. Moreover, a substantial increase in MRSA isolates with mecA_LGA251 was detected among isolates obtained in Denmark between 2007 and 2010 (9), which raises issues about the necessity of continuous monitoring of these strains to control their resistance to clinically important antimicrobials and the occurrence of new virulence patterns.

In this paper, we do not provide direct evidence that animals may be reservoirs of the more toxigenic MRSA ST599 and ST1946 identified in this study, since the strains were recovered from asymptomatic human carriers. However, as is typical for S. aureus strains from animal origins (13), a lack of the genes sak, chp, and scn that are involved in human immune evasion can show that animals could be the epidemiological source of these strains.

In our study, we found for 2 isolates a lack of association between the presence of mecA_LGA251 and oxacillin resistance. However, isolates with the novel mecA homologue and low-level β-lactam resistance (exhibiting oxacillin MICs of 1.0 μg/ml and 2.0 μg/ml) have been described by Shore and colleagues (19). The discovery of isolates that were both mecA_LGA251 positive and had oxacillin MICs in the susceptible range of the CLSI breakpoint (≤2 μg/ml) increasingly complicates the detection and confirmation of MRSA. Failure to detect such MRSA isolates could have serious implications in the suboptimal implementation of infection control measures or therapeutic strategies.

In this work, we tested 16 MRSA isolates which were investigated previously (12). In both investigations, oxacillin MICs were determined. In general, oxacillin MICs obtained in the current study were lower than those determined in the previous study (12). However, different antibacterial agent susceptibility testing methods were used in the two investigations. BMD was the method for MIC determination in this study, while previously, the Etest (bioMérieux SA, Marcy-l'Étoile, France) was applied. Very recently, Rennie and colleagues (17) compared the MICs obtained by BMD and Etest, and they concluded that the bioMérieux Etest often reported higher MICs than the CLSI reference BMD method.

Among S. aureus isolates carrying the mecA gene, the antibiotic susceptibility patterns most similar to those of MRSA isolates harboring SCCmec XI can be found in community-associated MRSA (CA-MRSA) strains, which are classically characterized by susceptibility to the majority of non-β-lactam antimicrobial agents (6). A number of published studies described CA-MRSA strains of SCCmec IV as susceptible to all non-β-lactam antibiotics (6, 18). Also, some CA-MRSA strains can display low-level oxacillin resistance (16).

The current study showed acquisition of SCCmec XI with mecA_LGA251 by S. aureus isolates with additional genotypes (with spa types t5930 and t7603), which complicates the identification of MRSA. Thus, DNA-based tests, which have been developed to detect MRSA more rapidly, cannot completely replace phenotypic culture-based MRSA screening tests.

Although most of the MRSA isolates with SCCmec XI carried none of the tested virulence factors, the two toxin gene-positive isolates identified in this study showed the genomic plasticity known from “classical” MRSA. This could indicate that we are likely to face further surprises, such as new waves of animal-related MRSA or the emergence of further mecA homologues.