INTRODUCTION
Staphylococcus pseudintermedius is part of the normal microbiota of dogs and cats, but can cause pyoderma and other opportunistic infections [Reference Sasaki1]. However, it rarely causes zoonotic infections in humans [Reference Chuang2]. Methicillin-resistant S. pseudintermedius (MRSP) strains have recently been reported [Reference Bannoehr3], and are increasingly being isolated from dogs [Reference Kawakami4]. MRSP isolates are not only resistant to β-lactam antibiotics, but show little or no susceptibility to various other antimicrobials, including aminoglycosides, macrolides, tetracycline, and fluoroquinolones [Reference Chanchaithong5–Reference Perreten7], limiting the treatment options for MRSP infections. Molecular analysis of MRSP isolates using multilocus sequence typing (MLST) based on four housekeeping genes in addition to the 16S rRNA gene (MLST-4) [Reference Bannoehr3], spa typing [Reference Moodley8], and staphylococcal cassette chromosome mec (SCCmec) typing [Reference Perreten7, Reference Kondo9] revealed that sequence type (ST)71-spa t02-SCCmec II–III and ST68-spa t06-SCCmec V are the major MRSP clones in Europe and North America, respectively [Reference Perreten7]. SCCmec II–III and SCCmec V MRSP isolates have also been obtained from dogs and veterinarians in Japan [Reference Bardiau10, Reference Ishihara11]. MLST-4 analysis of MRSP isolates from dogs and cats with dermatitis demonstrated that the ST71 lineage of SCCmec II–III MRSP is widespread in Japan [Reference Bardiau10]. MLST-4, which was developed for discrimination of S. intermedius groups, was applied in these previous studies on MRSP [Reference Perreten7, Reference Bardiau10, Reference Ruscher12]. A new S. pseudintermedius-specific MLST method based on seven housekeeping genes (MLST-7) has been established to increase discrimination between isolates [Reference Solyman13].
To compare the characteristics of MRSP isolates from Japan with those from Europe, North America, and other countries, the current study examined canine MRSP isolates from Japan using MLST-7, SCCmec typing, pulsed-field gel electrophoresis (PFGE), spa typing, antimicrobial susceptibility testing, and detection of antimicrobial resistance genes.
METHODS
Bacterial isolation and identification
Methicillin-resistant staphylococci were isolated from buccal mucosal samples from 292 dogs (225 dog patients brought to veterinary clinics for veterinary care or health maintenance. Twenty-two dogs and nine dogs out of 225 dog patients had a major complaint of dermatosis and external otitis, respectively; these dog patients included those admitted for vaccination and prevention of filariasis; 35 blood donor dogs; 13 dogs owned by veterinary staff; and 19 healthy dogs brought to veterinary clinics for purposes other than veterinary care) using CHROMagar MRSA (Kanto Kagaku Co., Japan). The samples were collected as part of a previous study [Reference Ishihara14] from 69 private veterinary clinics in the Ishikari region around Sapporo, Hokkaido Prefecture, Japan, during April and June 2008. As part of the previous study, mecA-positive isolates were confirmed by polymerase chain reaction (PCR) [Reference Ishihara14]. All isolates containing mecA, other than methicillin-resistant Staphylococcus aureus (MRSA) isolates, which were examined previously [Reference Ishihara14], were tested using the ID32 STAPH system (Sysmex bioMérieux Co., Japan) according to the manufacturer's instructions. DNA was extracted from cultures using InstaGene Matrix (Bio-Rad, USA). Isolates classified as S. intermedius by ID32 STAPH were also analysed by PCR–restriction fragment length polymorphism (RFLP) of their pta genes, which can discriminate S. pseudintermedius from S. intermedius and S. aureus, as described previously [Reference Bannoehr15]. Confirmed S. pseudintermedius isolates were then examined using the following tests.
Analysis of antibiotic resistance
Minimum inhibitory concentration (MIC) analysis was performed as described in the Clinical and Laboratory Standards Institute guidelines [16] using the broth micro-dilution method on Eiken Frozen Plates (Eiken Chemistry Co., Japan). The following antimicrobials were tested: oxacillin, cefazolin, cefotiam, imipenem, streptomycin, kanamycin, gentamicin, arbekacin, erythromycin, tetracycline, minocycline, chloramphenicol, ciprofloxacin, vancomycin, teicoplanin, quinupristin-dalfopristin, and linezolid.
The following antimicrobial resistance genes were screened by PCR using Go Taq Green Master Mix (Promega, Japan), as described previously for the detection of MRSP isolates [Reference Perreten7]: mecA [Reference Ishihara11], blaZ [Reference Milheiriço17], aac(6′)-Ie-aph(2′)-Ia [Reference Vakulenko18], aph(3′)-III [Reference Vakulenko18], ant(6′)-Ia [Reference Perreten19], sat4 [Reference Perreten19], ermB [Reference Lim20], dfrG [Reference Argudín21], lnuA [Reference Gholamiandehkordi22, Reference Lina23], tetK [Reference Tenover24], tetM [Reference Ng25], and cat pC221 [Reference Schnellmann26]. The primer sequences are listed in Table 1. PCR products were purified using a High Pure PCR Cleanup Micro kit (Roche Diagnostics GmbH, Germany), and products corresponding to the resistance gene fragments were confirmed by sequencing by FASMAC Co. (Japan). The PCR primers were also used to sequence these resistance genes. For sequencing of tetM, a 1862-bp fragment was amplified [Reference Trzcinski27] and sequenced using the same primers and internal primers (Table 1).
Table 1. Primers used in this study
* Used to amplify tetM for sequencing.
† Used as an internal primer for sequencing of tetM.
Molecular characterization
MLST-7 analysis was conducted for all isolates as described previously [Reference Solyman13]. The STs were determined using the S. pseudintermedius MLST database (http://pubmlst.org/spseudintermedius/).
SCCmec typing was performed by PCR amplification of the mec (classes A, B, C) and ccr (types 1, 2, 3, 5) gene regions [Reference Kondo9]. In addition, the structure of SCCmec was determined using Oliveira's strategy [Reference Oliveira and de Lencastre28]. To discriminate SCCmec II–III from SCCmec III, the cadmium resistance gene in the J2 region and the structure of the J1 region were examined by PCR as described previously [Reference Perreten7, Reference Milheiriço, Oliveira and de Lencastre29].
PFGE analysis of SmaI-digested DNA was performed as previously described [Reference Ishihara11, Reference Sasaki30]. PFGE was performed using a CHEF-DR III system (Bio-Rad), as described previously [Reference Sasaki30].
spa genes were amplified using previously described primers and conditions [Reference Perreten7, Reference Moodley8, Reference Ruscher12]. DNA sequences of the spa genes were determined as described above. A S. pseudintermedius spa database, developed by Dr A. Moodley of the University of Copenhagen (personal communication), was used to determine spa types [Reference Moodley8].
RESULTS
Identification
Forty-three isolates, each from a different dog (43/292, 14·7%), were classified as S. intermedius by ID32 STAPH. All of these isolates were further confirmed as S. pseudintermedius by PCR–RFLP, and all were methicillin resistant. These MRSP isolates were obtained from 23 dog patients (23/225, 10·2%; including three dogs with a major complaint of dermatosis), 15 blood donor dogs (15/35, 42·9%), and five dogs owned by veterinary staff (5/13, 38·5%). None of the 19 healthy dogs carried MRSP. The MRSP-positive samples came from 20 different veterinary clinics (20/69, 29·0%).
Molecular characteristics
ST275 (n = 4), ST276 (n = 9), ST323 (n = 2), ST325 (n = 1), ST324 (n = 1) and ST354 (n = 3) were detected in this study and assigned as novel STs in the S. pseudintermedius MLST database. ST71 (n = 11), ST169 (n = 11) and ST121 (n = 1) were also detected (Table 2). The SCCmec types of these 43 isolates are given in Table 2. The predominant SCCmec type was II–III (n = 25), followed by V (n = 13). SCCmec II–III isolates were classified as ST71, ST169, and ST354. The clonal relationships in MRSP SCCmec II–III STs of isolates from this study and others were predicted by BURST analysis using the S. pseudintermedius MLST database (http://pubmlst.org/spseudintermedius/), as shown in Figure 1.
Fig. 1. Clonal relatedness of Staphylococcus pseudintermedius sequence types (STs) as predicted by BURST analysis. STs with ⩾4 loci matching those of ST169 (2–9–1–2–1–1–1) were selected for this analysis. The group including STs of methicillin-resistant Staphylococcus pseudintermedius isolates obtained in this study (* ST71, ST121, ST169, ST354) is shown.
Table 2. Summary of MRSP genotypes
MRSP, Methicillin-resistant Staphylococcus pseudintermedius; ST, sequence type; MLST-7, multilocus sequence typing based on seven housekeeping genes; VCs, veterinary clinics; n.d., not determined.
* spa could not be amplified using any of the primer pairs, therefore spa type was not determined.
† No. of VCs where MRSP isolates were obtained from dogs.
‡ MRSP-ST275 isolates obtained from one VC.
§ Although class A mec complex was determined, ccr was not amplified.
|| Two different STs were obtained for isolates from four VCs.
All 25 SCCmec II–III isolates harboured open reading frames (ORFs) identical to those of SCCmec III-MRSA in the J1 region and dcs in the J3 region, but the cadmium resistance gene was absent from the J2 region. The SCCmec structure confirmed by PCR [Reference Perreten7, Reference Kondo9, Reference Oliveira and de Lencastre28, Reference Milheiriço, Oliveira and de Lencastre29] of these 25 SCCmec II–III MRSP isolates matched the SCCmec II–III DNA sequence available from GenBank (accession no. AM904732). Three SCCmec III isolates harboured both dcs and a cadmium resistance gene, but not the ORF identical to that of SCCmec III-MRSA in the J1 region. All 13 SCCmec V isolates harboured Tn554.
spa types could only be determined for 16 (37·2%) of 43 isolates (Table 2) because the remaining isolates did not yield a product from spa PCR analysis using the various primer pairs. The predominant spa type was t02 (11 isolates), and new spa types t58, t60, and t62 were also identified (Table 2).
PFGE divided MRSP isolates into eight clusters (clusters A–H), with similarities within each cluster of ⩾60% (data not shown). Two SCCmec V isolates did not belong to any cluster. Molecular characteristics of MRSP isolates are given in Table 3. Clusters A (n = 6) and E (n = 4) contained only ST71-SCCmec II–III isolates. Cluster D contained only five ST169-SCCmec II–III isolates. Cluster C contained only SCCmec II–III isolates; however, it included ST71 (n = 1), ST169 (n = 6), and ST354 (n = 3) isolates. Cluster C included two sub-clusters; the first sub-cluster contained ST71-SCCmec II–III and ST169-SCCmec II–III isolates, while the second sub-cluster contained ST354-SCCmec II–III isolates. The similarity between PFGE band patterns of the four clusters containing SCCmec II–III isolates was ⩾50%. SCCmec V isolates were divided between clusters F (n = 8, ST276), G (n = 1, ST121), H (n = 2, ST323), and others (n = 2, ST276 and ST324).
Table 3. Molecular characteristics and antimicrobial resistance of MRSP isolates from dogs in Japan
MRSP, Methicillin-resistant Staphylococcus pseudintermedius; ST, sequence type; PFGE, pulsed-field gel electrophoresis; CIP, ciprofloxacin; VCs, veterinary clinics; UT, untypable; n.d., not determined.
* ST determined by multilocus sequence typing based on seven housekeeping genes.
† Clusters assigned by PFGE.
‡ No. of VCs where MRSP isolates were obtained from dogs.
§ R, resistant; S, susceptible.
|| These isolates were not included in any clusters.
Antimicrobial resistance
Antimicrobial resistance gene profiles and susceptibility to ciprofloxacin results are given in Table 3. All 25 SCCmec II–III isolates contained mecA, blaZ, aac(6′)-Ie-aph(2′)-Ia, aph(3′)-III, ant(6′)-Ia, sat4, ermB, and dfrG, and were resistant to ciprofloxacin (MIC range 4–32 μg/ml). None of the ST71-SCCmec II–III isolates harboured tetM, while nine out of 11 ST169-SCCmec II–III isolates and all ST354-SCCmec II–III isolates (n = 3) contained tetM (Table 3). Nine (69·2%) of the 13 SCCmec V isolates were susceptible to ciprofloxacin (⩽0·125 μg/ml).
The tetM sequences (1671 bp) from 30 MRSP isolates with SCCmec II–III, III, IV or V were determined, and phylogenetic analysis of these sequences revealed three homology groups (Fig. 2). The tetM sequences were identical in isolates belonging to the same ST. Moreover, the tetM sequence of ST354 isolates was identical to that of ST169 isolates.
Fig. 2. Phylogenetic tree based on DNA sequences of tetM genes. tetM DNA sequences (1671 bp) from 30 methicillin-resistant Staphylococcus pseudintermedius isolates obtained from dogs in Japan were determined in this study. The sequence of nine representative isolates (* bold font, sequence type (ST) determined by multilocus sequence typing based on seven genes, along with the SCCmec type is shown in parentheses) were used for this phylogenetic tree. The remaining sequences were obtained from the GenBank database [bacterial species, strain code, and accession number (in parentheses) are shown]. The tree was constructed using the neighbour-joining method of GENETYX-tree (Genetyx Corp., Japan)
All isolates examined in the current study were susceptible to arbekacin (MIC range 0·25–2 μg/ml), minocycline (0·25–4 μg/ml), vancomycin (0·25–1 μg/ml), teicoplanin (⩽0·125–2 μg/ml), quinupristin-dalfopristin (0·25 μg/ml), and linezolid (0·25–1 μg/ml). The antimicrobial resistance patterns determined by phenotypic analysis mainly agreed with patterns of detected antimicrobial resistance genes. Only two SCCmec V isolates without cat pC221 were resistant to chloramphenicol (16–32 μg/ml). Although almost all isolates were resistant to ⩾2 antimicrobials in addition to β-lactam antibiotics, the MRSP SCCmec IV isolate was only resistant to β-lactam antibiotics and tetracycline.
DISCUSSION
ST71-spa t02-SCCmec II–III and ST68-spa t06-SCCmec V have been identified as the genotypes of major MRSP clonal lineages in Europe and North America, respectively [Reference Perreten7]. A previous study classified MRSP isolates by MLST based on four housekeeping genes and the 16S rRNA gene (MLST-4) [Reference Bannoehr3]. In the current study, MRSP isolates were discriminated using a recently developed MLST method based on seven housekeeping genes (MLST-7) [Reference Solyman13]. The ST68 and ST71 designations were maintained to provide continuity between the MLST-4 and MLST-7 techniques [Reference Solyman13].
Because all MRSP SCCmec II–III isolates obtained from European countries [Reference Perreten7, Reference Ruscher12], North China [Reference Wang31], and Japan [Reference Bardiau10] have previously been typed as ST71 by MLST-4, all MRSP SCCmec II–III isolates were thought to belong to one clonal lineage that had spread worldwide. Thereafter, all MRSP SCCmec II–III isolates from Europe [Reference Solyman13] and Brazil [Reference Quitoco32] were also typed as ST71 by MLST-7, supporting the theory of a single clonal lineage. However, in this study we confirmed that in Japan, SCCmec II–III elements are present in three related clonal lineages: ST71, ST169, and ST354. Recently, two MRSP ST169-SCCmec II–III isolates have also been reported in Thailand [Reference Chanchaithong5].
The current analysis showed that the allele sequences of two loci (ack and sar) differ between ST71 and ST169 isolates. These molecular characteristics suggest that ST71 and ST169 clones have been derived from a common ancestral clone, or one clone might be derived from the other through an unknown third clone. The common ancestor or third clone would be arranged in the red zone surrounding ST71 in Figure 1. In either case, neither clone was directly derived from the other. However, ST354 is probably a variant clone of ST169, the predominant ST in Japan, because only one locus (purA) differed between ST169 and ST354.
The majority of ST169 and ST354 isolates harboured tetM, while none of the ST71 isolates contained this gene. Although tetM genes from the MRSP isolates were divided into three groups in the current study, tetM sequences in ST169 (n = 9) and ST354 (n = 3) isolates were identical to each other. Therefore, it is suspected that the ancestor clone of ST169 already contained tetM, or that ST169 isolates acquired tetM prior to dissemination. The antimicrobial resistance gene patterns, other than for tetM, were similar in ST71, ST169, and ST354 isolates in Japan and ST71 isolates in Europe [Reference Perreten7].
The present study detected three novel STs (ST276, ST323, ST324), in MRSP SCCmec V isolates from dogs in Japan, with ST276 being the major type in SCCmec V isolates in this study. Other SCCmec V isolates (ST121, ST323, ST324) were unlikely to be related to the ST276 clone, as allele sequences from these isolates differed from those of ST276 at ⩾5 loci. In 2007, 4507 (61·9%) of 7281 dogs that passed quarantine and gained entry into Japan came from North America (USA, including Hawaii and Canada) (Animal Quarantine Service, http://www.maff.go.jp/aqs/tokei/toukei.html). However, the present study showed that while various MRSP SCCmec V clones were present in dogs in Japan, neither ST68 nor related STs that are more common in North America were found in the tested isolates. A further 1654 (22·7%) dogs were from Asia (Taiwan, 611; Korea, 413; China, 193; Thailand, 160; Singapore, 89; and ‘other’, 188), and 578 (7·9%) dogs were from Europe (UK, 160; Germany, 104; France, 93; Italy, 44, Sweden, 34; and ‘other’, 143). These numbers suggest that the genotypes of MRSP isolates in Japan do not reflect the origins of their canine hosts.
Like the current study, two previous molecular analyses of Japanese MRSP isolates from clinical samples of companion animals revealed that SCCmec II–III and V were predominant types in clinical MRSP isolates in Japan [Reference Bardiau10, Reference Onuma, Tanabe and Sato33]. Moreover, all MRSP SCCmec II–III isolates were classified as ST71 by MLST-4 [Reference Bardiau10]. Therefore, MRSP isolates obtained from buccal samples in the current study are likely to be closely related to these previous clinical MRSP isolates.
In our original study, we evaluated animal patients as a potential source of MRSA contamination of veterinary staff. Therefore, buccal swabs from dog patients were collected to reveal the prevalence of MRSA in dogs which were cared for in veterinary clinics [Reference Ishihara14]. However, MRSA carriage was rare (3/292) in the dogs [Reference Ishihara14]. On the other hand, many mecA-positive non-MRSA staphylococcal isolates were obtained from these animals. Nearly two-thirds of the mecA-positive isolates were identified as S. peudintermedius, and these isolates were examined in the current study.
The prevalence of MRSP in blood donor dogs (42·9%, 15/35; P<0·01) and dogs owned by veterinary staff (38·5%, 5/13; P<0·05) was significantly higher than for dog patients (10·2%, 23/225). Moreover, common STs (ST275, ST276, ST354) of MRSP isolates were detected from ⩾2 blood donor dogs or dogs owned by veterinary staff of the same veterinary clinics. These results suggest that MRSP is transmitted between dogs in veterinary clinics, and that preventative measures should be implemented.
spa typing of S. pseudintermedius divided MRSP ST71-SCCmec II–III isolates obtained in Europe into multiple types [Reference Perreten7, Reference Moodley8, Reference Ruscher12]. The major spa type of ST71 isolates in this study was t02 (10/11), which is also the major type in Europe [Reference Perreten7]. spa genes were not amplified from 27 MRSP isolates by the various primers used in this study. Feng et al. also reported that spa types could only be confirmed for 15 (21·7%) out of 69 MRSP isolates obtained in South China [Reference Feng6].
In conclusion, MRSP SCCmec II–III isolates from Japan were divided into three related STs: ST71, ST169, and the newly registered ST354. Of SCCmec V isolates, various novel STs (ST276, ST323, ST324) and ST121 were observed; however, ST68, which is a major genotype in North America, was not found.
ACKNOWLEDGEMENTS
We thank the staff at the veterinary clinics for providing samples, and Dr T. Nomura of the Sapporo Veterinary Association for assisting us with collection of these samples.
This work was supported by the Japan Society for the Promotion of Science (JSPS KAKENHI Grant Number 26850194), and by the Japanese Ministry of Health, Labour, and Welfare (H24-Shokuhin-Ippan-008).
DECLARATION OF INTEREST
None.
