MRSA isolates from infections in humans and in animals were sent to the National Reference Center for Staphylococci at the Robert Koch Institute, Wernigerode Branch, in Germany, for typing by means of SmaI-macrorestriction pattern as well as spa typing. Selected isolates also underwent MLST. Four human isolates were grown from nasal swabs taken from the staff of a veterinary practice at Veterinary Analytical Center, Geesthacht, Germany. All isolates were primarily grown on sheep blood agar and confirmed by standard procedures as S. aureus. Eleven additional MRSA specimens of lineage ST398 (1 isolate per patient affected) were found among 4,370 MRSA isolates from patients with recognized infections. These isolates were identified by indigestibility of their whole cellular DNA when subjected to SmaI-macrorestriction analysis. Animal isolates were collected from 1 dog and 1 foal at the Veterinary Analytical Center, Geesthacht, Germany; from 1 pig at the diagnostic laboratory of the Institute for Microbiology and Infectious Diseases, School of Veterinary Medicine, Hannover, Germany; and from 2 horses at the Department of Orthopaedics, Veterinary University, Vienna, Austria.

Procedures and primers for DNA extraction and PCR detection of resistance genes were as described previously (6). Macrorestriction patterns were determined by using lysis of cells, deproteinization and digestion of DNA (here by SmaI and ApaI), and pulsed-field gel electrophoresis (7).

The polymorphic X-region of the protein A gene (spa) was amplified and sequenced according to the Ridom StaphType standard protocol (www.ridom.org). The resulting spa-types were assigned by using the Ridom StaphType software package (Ridom GmbH, Würzburg, Germany). The BURP algorithm, implemented in the most recent Ridom StaphType software version, was used for cluster analysis of spa types (7).

Primers used for MLST correspond to the protocol as described previously (8), with the exception of the forward primer for tpi; we used the sequence tpif 5′-GCATTAGCAGATTTAGGCGT. Antimicrobial susceptibility testing was performed by broth microdilution, performed according to DIN 58940, Deutsches Institut für Normung (9). SCCmec elements of types I to IV were characterized by using a PCR approach, including a combination of different PCR reactions (6). To demonstrate SCCmec-elements of type V, we used primers type VF/type VR, as described by Zhang et al. (10), as well primer pair ccrC9f 5′-CACTTAATCCATGTACACAG and ccrC-R (10).

The following set of primers was used for PCR for virulence-associated genes: tst, sea, seb, sec, sed, see, as described by Johnson et al. (11); for lukS-lukF, forward 5′-ATCATTAGGTAAAATGTCTGGACATGATCCA, reverse 5′-GCATCAAGTGTATTGGATAGCAAAAGC; for cna, forward 5′-CGGTTCCCCCATAAAAGTGAAG, reverse 5′-CCCATAGCCTTGTGGATTTG. Annealing temperature was 55° C; cyclic scheme and further conditions were as reported previously (6).

Specimen collection, characterization of the isolates, data processing, and exchange of data were performed within the framework of German public health activities for infection control and prevention of MRSA dissemination. Ethical approval was obtained within this framework as well.

Characteristics of the 20 MRSA isolates investigated are shown in the Table. All isolates share MLST ST398 with the allelic profile 3-35-19-2-20-26-39. Three different spa-types are obviously related (Figure). Types t11 and t34 may have been derived from each other by either deletion or duplication of 2 repeats; t1197 and t11 differ by a single nucleotide polymorphism. BURP analysis of these spa-types groups them as a separate cluster unrelated to other BURP clusters (7). A peculiarity of S. aureus of clonal lineage ST398 is the indigestibility of whole cellular DNA by restriction enzyme SmaI. Therefore, SmaI macrorestriction patterns generate only 1 large fragment because of protection by a novel DNA methylation enzyme (12). We also found poor digestion by the isoschizomeric enzyme XmaI. However, digestion by enzyme ApaI generated similar fragment patterns that differed at most by 3 fragments independent of spa types.

The 2 horse isolates from the Vienna veterinary university contained SCCmec elements of group IVa. For all other isolates investigated, PCR indicated SCCmecV. These findings suggest that MRSA of ST398 from horses are unrelated to the other isolates and probably have evolved independently by acquisition of a different SCCmec element.

In addition to mecA, all investigated isolates contained tetM; isolates from animals and humans from Lower Saxony also contained ermA. The nosocomial human and horse isolates contained ermC; in the horse isolates, aph2″-aac6′–mediating aminoglycoside resistance was demonstrated. PCR was negative for virulence-associated genes and for lukS-lukF (coding for Panton-Valentine leukocidin), tst, sea, seb, sec, and sed, as well as for cna (collagen-binding protein).

Isolates of clonal lineage ST398 seem not to be frequently represented among the S. aureus population. They were not recorded by Grundmann et al. (13) among a population sample of nasal colonizers in the Nottingham area in the United Kingdom and were not found among 108 isolates from carriers in a rural territory in northern Germany (S. Holtfreter et al., unpub. data). Only 2 notations of ST398 are found in the S. aureus MLST database (www.mlst.net), 1 from the Netherlands and 1 from the Cape Verde Islands.

Among 11,250 isolates of various origin (colonization and infections in hospitals as well in the community in humans from all Germany) typed from 1992 through 2003, no isolates refractory to SmaI macrorestrition analysis were seen. Therefore, a rather recent emergence of MRSA ST398 among humans seems likely. However, MRSA of lineage ST398 had been reported from infections in pigs and from nasal colonization in pig farmers in France (14). A more recent report from the Netherlands describes MRSA of ST398 (spa t108, which is in the same BURP cluster as t11 and t34) in pigs and in humans who had contact with pigs (15). A comparison of the allelic profile of ST398 by means of the MLST database does not indicate any relationship to profiles of prevalent clonal complexes of methicillin-susceptible S. aureus (13), of epidemic healthcare-associated MRSA, or of lukS-lukF–containing community-associated MRSA from Europe.

MRSA exhibiting ST398 may colonize and cause infections in humans and in certain animal species such as dogs, horses, and pigs. The isolation of MRSA ST398 showing the same characteristics from a wound infection in a dog and from nasal colonization of the staff of a veterinary practice where this dog had been treated suggests that interspecies transmission may occur. The differences in spa-types between the isolates containing the same PCR results for SCCmec can be explained by a single genetic event. Because isolates taken at the same time from nasal colonization in veterinary staff of the same practice exhibit either spa-type t011 or t034, this difference does not justify discrimination between the two types. Of particular concern was the subsequent detection of MRSA ST398 not only in outpatients but also in inpatients with ventilator-associated pneumonia in the same hospital unit at about the same time (Table).

Future recording of MRSA ST398 from infected and colonized humans (especially when detected by screening at admission to hospitals) will require a thorough analysis with respect to association with animals and routes of transmission. Tracing MRSA carriers among contacts should also include pet animals, horses, and other livestock. Because of the time and labor needed to complete MLST, spa-typing combined with BURP analysis of types is an efficient tool for recognizing this clonal lineage. Furthermore, detection of MRSA by appropriate methods should be implemented into antimicrobial resistance surveillance programs in veterinary medicine.