Virulence and Resistance Determinants of German Staphylococcus aureus ST398 Isolates from Nonhuman Sources

Methicillin-resistant Staphylococcus aureus (MRSA) of sequence type 398 (ST398) has gained particular attention during recent years because of its association with pigs and its ability to colonize pig farmers and other people in close contact with pigs (7, 12, 47). The MRSA isolates of ST398 usually lack important virulence determinants that are typical in other community and hospital MRSA isolates. The majority of the ST398 isolates analyzed so far carry only hemolysin-encoding genes (13, 21, 31, 32), although a small number of cases in which the isolates carried the bicomponent leukotoxin Panton-Valentine (lukPV genes) (43, 49) or staphylococcal enterotoxins (SEs, se genes) (21, 26) have also been reported. Genes for other toxins, like exfoliatins (ET, et genes), leukotoxins, and toxic shock syndrome toxin (TSST-1, tst gene) have not been found yet in ST398 isolates (13, 21, 31, 32, 44). The regulation of the expression of most extracellular virulence factors in S. aureus is under the control of a two-component signaling system called the accessory gene regulator (agr), which is polymorphic and divided into four distinct genetic groups (I to IV). A correlation exists between some agr groups and certain pathotypes and clonal complexes (CCs) (48), and CC398 seem to be associated with agr group I (agrI) (31, 32).

Typically, ST398 strains display a multiresistant phenotype. The majority of the isolates have been reported as MRSA, but in most of the epidemiological studies, the results were influenced by the use of selective isolation methods (8, 17). Methicillin-susceptible S. aureus (MSSA) strains have also been described (3, 17, 28, 43), and in the latest S. aureus surveillance studies of humans in Europe, all ST398 isolates detected were MSSA (16). Resistance to methicillin and other β-lactam antibiotics is mediated by mecA carrier elements called staphylococcal cassette chromosome mec (SCCmec). Together with methicillin resistance in ST398 isolates from swine, resistance to tetracycline, erythromycin, clindamycin, quinupristin-dalfopristin, ciprofloxacin, sulfonamides, trimethoprim, sulfamethoxazole-trimethoprim, and aminoglycosides has been reported (5, 9, 21, 38, 39).

The aim of the present work was to study the virulence and antimicrobial resistance genetic repertoire of a collection of 100 S. aureus strains previously ascribed to ST398 which were recovered from different sources (mainly livestock and food of animal origin) in Germany.

The emerging MRSA ST398 clonal group is widespread in the domestic pig population and can also be identified in other livestock species, such as veal calves, dairy cattle, and poultry (1, 12). Farmers and other livestock professionals handling carrier animals are at high risk of being colonized (7). Although infrequently, clinical infections with MRSA ST398 have been described, being in most cases related to these high-risk groups (12, 27, 45, 47). Human infections with MSSA ST398 have also been recorded, and they seem to be more frequent than those caused by MRSA ST398 (16). While the prevalence of the pathogen in farm animals and food has been studied extensively during the last few years, there is not much information on the molecular epidemiology of virulence and resistance determinants in this clonal group. The present work has investigated the virulence and resistance gene contents in a collection of 100 ST398 strains of nonhuman origin, representative of the German NRL-Staph collection (years 2004 to 2008), which was well characterized by several molecular methods previously (3).

The tested isolates of spa types assigned to ST398 lack several clinically important S. aureus-associated virulence factors regardless of their source. This is in line with the results of other studies (13, 21, 31, 32) and with the fact that clinical disease in swine and humans colonized by ST398 MRSA is rarely observed (16, 45, 47). However, cases of clinical and subclinical mastitis associated with this type of MRSA in dairy herds have been reported (13, 40), as have pathological lesions of pigs (28). From the 37 genes tested in this study that are related to virulence, the presence of pathogenicity islands, and regulators, we could only detect hemolysin-encoding genes (hla, hlb, hld, and hlg) and the agr group. None of the German isolates carried enterotoxins, but seb was present in two of the Dutch ST398 isolates. Low levels of occurrence of either seb sek and seq, or sed and seg were also reported for MRSA ST398 isolates from Germany (21) and France (26). It has been speculated that the environment contributes to the low presence of virulence determinants in this clone. However, some studies focused on the pig population have shown how in other S. aureus isolates (i.e., MSSA isolates from undetermined CCs), the presence of enterotoxins can be high (33). The Panton-Valentine leukocidin (PVL) was absent in our isolates. Genes encoding this toxin, which is frequently present in community-acquired S. aureus isolates, have also been detected in ST398 isolates from hospitalized patients (49) and, at a low frequency, in pig isolates (43). The majority of PVL-positive ST398 isolates described so far has been obtained from humans without exposure to animal husbandry (46, 49).

The low number of virulence factors in ST398 isolates is in clear contrast to the high number of resistance determinants. In this respect, type I and II restriction-modification systems detected in the genome of ST398 (36) may have interfered so far with the acquisition of virulence factors, which are mainly encoded by pathogenicity islands or phages (34). However, this does not explain the high number of resistance genes carried by the clone. In S. aureus, resistance genes are mainly located in plasmids and, depending on size and GC content, their chances to escape restriction might be high. Moreover, selection pressure on ST398 isolates induced by the use of antimicrobials in livestock production could have potentiated positive selection for resistance determinants, while infrequent contact with pathogenic bacteria, more common in hospitals, could explain why the clone still lacks many virulence determinants.

In different countries, different antimicrobial usage habits can influence the antimicrobial resistance patterns (regarding phenotypes and mobile genetic determinants) observed in the bacterial populations. The nonhuman German ST398 isolates analyzed in this study were highly heterogeneous with regard to antimicrobial resistance properties. This may be due to the livestock being raised with different types of antimicrobials, but data on the antimicrobial usage on the farms from which the isolates came are not available. A high percentage (59%) of multiresistance was present in the series. Almost 37% of the isolates, 43% of them recovered from healthy carrier pigs, showed resistance to five or more classes of antimicrobials, whereas only 28% did in a study of 55 German ST398 isolates from diseased swine (21). All 100 isolates were resistant to tetracycline, a property of the ST398 clone also found by other authors (9, 41). This resistance is mostly related to the presence of the tet(M) gene located in transposons Tn5801 and Tn916. Only a few ST398 isolates lacking tet(M) have been described so far (21, 22). Nearly all isolates (91%) in this study also carried additional tetracycline resistance genes, like tet(K) and/or tet(L), that are normally harbored on plasmids. The latter gene, which has rarely been described in non-ST398 MRSA isolates (14, 42), was relatively common in our series (40%), as well as in other isolates from the ST398 clone (24%) (13, 21). The prevalence of trimethoprim resistance was also high (65%), and the dfrK and dfrG genes, which have rarely been detected among other MRSA isolates, seem to be frequent in German ST398 isolates (13, 21). Colocalization of the tet(L) and dfrK genes on the pKKS2187 plasmid was described in a German MRSA ST398 isolate from a diseased pig (22). In 36% of our isolates, both genes appeared together, with 4% and 6% being positive for only tet(L) or only dfrK, respectively. Chloramphenicol resistance was expressed in nine isolates (9%), but only two of them were positive for fexA, a gene that has rarely been detected in staphylococci from animal sources (25). Among these fexA-positive isolates, one isolate also carried the multidrug resistance determinant cfr, which was previously found, associated with fexA, in individual ST398 and ST9 isolates (24). This indicates that, currently, both genes are infrequent in the German ST398 population.

In the present study, the percentage of resistance to gentamicin (14%) was similar to that reported for diseased swine by Kadlec et al. (21) but, in comparison to that report, the percentage of resistance to kanamycin was significantly higher (29% versus 7%; χ² test, P = 0.003) in our study. High percentages of gentamicin and/or kanamycin resistance (22% and 41%, respectively) were reported in ST398 isolates from cases of bovine mastitis (13). The genes implicated in the aminoglycoside resistance of the isolates from our series were aac(6′)-Ie-aph(2″)-Ia and/or ant(4′)-Ia, very frequently present in S. aureus isolates (14) and ST398 isolates from other series (13, 21). The aph(3′)-IIIa gene, infrequent in S. aureus isolates, was found in only 2% of our series. In 16 isolates (16%) that had repeatedly been shown to be susceptible to kanamycin in disc diffusion assays (corresponding to MICs of ≤16 μg/ml), ant(4′)-Ia could be detected. With their MICs having been tested by broth microdilution and considering the EUCAST cutoff values (susceptible range, ≤8 μg/ml), the same isolates showed either susceptibility or resistance (various MICs between 32 and 8 μg/ml). This phenomenon has also been noticed in other studies (11, 19, 30) and could be due to a requirement for the induction of transcription or the presence of an inactive ant(4′)-Ia. It has been demonstrated that a rearrangement between the blaZ gene and a Tn4001-IS257 hybrid structure has developed an aac(6′)-Ie-aph(2″)-Ia gene inducible by β-lactams (20). In our series, all except one of the aac(6′)-Ie-aph(2″)-Ia-positive strains were also positive for blaZ, but in the present work, no further investigations into the presence of this hybrid structure were conducted. Regarding the MLS_B resistance, several erm genes were found (prevalences of 30 to 40%), appearing either alone or in all possible combinations. This is interesting because, in S. aureus, the ermA gene is the most frequent determinant that confers constitutive resistance to MLS_B, while ermC is more common in strains with an inducible phenotype (14, 15, 18). The ermB gene has been identified in multiple bacterial genera (35). It has been found more frequently in staphylococci from animal sources than from human sources (14, 15, 18, 37). None of the nine (9%) quinupristin-dalfopristin-resistant isolates carried the genes coding for resistance to type A and B streptogramins tested in the study. The novel gene vgaC found in the ST398 clone (23) was also absent. This suggests that the clone possesses other mechanisms responsible for this resistance.

When looking at the clustering analyses (Fig. 2), there was a correlation between the resistance profiles of the isolates, their Cfr9I PFGE profiles, and their SCCmec types previously described by Argudín et al. (3). Isolates with SCCmec type V usually carried tet(K) and clustered together in subcluster B1a. These isolates also clustered together at a similarity of 0.63 when analyzed by Cfr9I PFGE (3). Conversely, subcluster B1b grouped mainly isolates with SCCmec IVa or V* without tet(K), and their PFGE profiles were also grouped within a different cluster (similarity of 0.64) (3). Some of the SCCmec V carriers also have dfrK, like most of the SCCmec IVa or V* carriers. Despite the differences, all of the ST398 Cfr9I PFGE profiles clustered together and were separate from the outgroup strains of different CCs.

So far, the virulence gene content of the ST398 clone appears to be low. However, continuous surveillance is needed in order to clarify whether the pathogenicity potential of the clone will evolve in the coming years. Moreover, MRSA ST398 is a reservoir for multiple determinants (13, 21, 26, 29) conferring resistance to several antimicrobials of clinical relevance. For the same phenotypic resistance, one or more genes, frequently found on plasmids [blaZ, ant(4′)-Ia, tet(K), tet(L), ermC, dfrD, and dfrK] or transposons [aac(6′)-Ie-aph(2″)-Ia, aph(3′)-IIIa tet(M), ermA, ermB, and dfrS1], are present in these isolates. Many of them may be colocalized on the same genetic element (22, 34). The high prevalence of this type of S. aureus in livestock and food, together with the carriage of multiple resistance determinants with a high risk of spread, underlines the importance of controlling this emergent bacterium in the animal population and its spread to exposed humans.

We thank J. Kowall, U. Kämpe, W. Barownick, and J. Beutlich (Federal Institute for Risk Assessment, BfR, Berlin) for their support. We also are grateful to the different German laboratories, as well as to D. Mevius and K. Veldman (Central Veterinary Institute of Wageningen, Lelystad, Netherlands) for providing the isolates or samples that were included in this study, to W. Witte and G. Werner (Robert Koch Institute, Wernigerode, Germany) and F. Aarestrup and H. Hasman (Food-DTU, Lyngby, Denmark) for the control strains, and to the Department of Biological Safety of the BfR for their hospitality toward M. A. Argudín during a short-term visit.

The project was supported by the BFR (grants BfR-45-004, BfR-46-001, and BfR-41-001), the German Ministry for Food, Agriculture, and Consumer Protection (grant 2808HS032), and the Fondo de Investigaciones Sanitarias (grant FIS-PI080656) of the Spanish Ministry of Science and Innovation. M. A. Argudín, Ph.D. student, was the recipient of grant FPU AP-2004-3641 from the Ministry of Science and Innovation, Spain, cofunded by the European Social Fund. She performed a short stay at the BfR supported by the same grant.