Received 15 November 2005; returned 1 December 2005; revised 8 February 2006; accepted 10 February 2006

Transcription

1 Journal of Antimicrobial Chemotherapy (2006) 57, doi: /jac/dkl057 Advance Access publication 8 arch 2006 Identification of bla CY-7 and associated plasmid-mediated resistance genes in multidrug-resistant Escherichia coli isolated from dogs at a veterinary teaching hospital in Australia Hanna E. Sidjabat 1,2, Kirsty. Townsend 1, Nancy D. Hanson 3, Jan. Bell 4, H. W. Stokes 5, Kari S. Gobius 6, Susan. oss 1 and Darren J. Trott 1 * 1 School of Veterinary Science, The University of Queensland, Brisbane, QLD 4072, Australia; 2 edical Faculty of the Christian University of Indonesia (FK-UKI), Cawang Atas, Jakarta, Indonesia; 3 Center for Research in Anti-Infectives and Biotechnology, Department of edical icrobiology and Immunology, School of edicine, Creighton University, Omaha, Nebraska, USA; 4 icrobiology and Infectious Diseases, Women s and Children s Hospital, North Adelaide, SA 5006, Australia; 5 Department of Chemistry and Biomolecular Sciences, acquarie University, Sydney, NSW 2109, Australia; 6 Food Science Australia, PO Box 3312, Tingalpa DC, QLD 4173, Australia Received 15 November 2005; returned 1 December 2005; revised 8 February 2006; accepted 10 February 2006 Objectives: To determine clonality and identify plasmid-mediated resistance genes in 11 multidrugresistant Escherichia coli (DREC) isolates associated with opportunistic infections in hospitalized dogs in Australia. ethods: Phenotypic (IC determinations, modified double-disc diffusion and isoelectric focusing) and genotypic methods (PFGE, plasmid analysis, PCR, sequencing, Southern hybridization, bacterial conjugation and transformation) were used to characterize, investigate the genetic relatedness of, and identify selected plasmid-mediated antimicrobial resistance genes, in the canine DREC. Results: Canine DRECs were divided into two clonal groups (CG 1 and 2) with distinct restriction endonuclease digestion and plasmid profiles. All isolates possessed bla CY-7 on an 93 kb plasmid. In CG 1 isolates, bla TE, cata1 and class 1 integron-associated dfra17-aada5 genes were located on an 170 kb plasmid. In CG 2 isolates, a second 93 kb plasmid contained bla TE and unidentified class 1 integron genes, although a single CG 2 strain carried dfra5. Antimicrobial susceptibility profiling of E. coli K12 transformed with CG 2 large plasmids confirmed that the bla CY-7 -carrying plasmid did not carry any other antimicrobial resistance genes, whereas the bla TE /class 1 integron-carrying plasmid carried genes conferring resistance to tetracycline and streptomycin also. Conclusions: This is the first report on the detection of plasmid-mediated bla CY-7 in animal isolates in Australia. DREC isolated from extraintestinal infections in dogs may be an important reservoir of plasmid-mediated resistance genes. Keywords: multidrug-resistant E. coli, bla CY-7, bla TE, class 1 integrons Introduction The increasing prevalence of antimicrobial resistance is a serious concern for both human and veterinary medicine. Horizontal gene transfer among bacteria by mobile genetic elements such as plasmids, transposons and integrons has facilitated the widespread distribution of multiple antibiotic resistance genes among Enterobacteriaceae. 1 In human clinical isolates, resistance to extended-spectrum b-lactam antimicrobial agents is generally due to the production of extended-spectrum b-lactamases... *Corresponding author. Tel: ; Fax: ; d.trott@uq.edu.au Present address. Division of Veterinary and Biomedical Sciences, urdoch University, urdoch, WA 6150, Australia Ó The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 Plasmid-mediated resistance genes in canine DREC (ESBLs) which can be variants of the classical TE-1, TE-2 and SHV-1 enzymes. ESBL prevalence has increased dramatically over the past 15 years in human medicine, with many isolates producing multiple combinations of enzymes. 2 Interestingly, ESBLs are rarely reported in veterinary isolates, although cases of opportunistic infections due to ESBL-producing Escherichia coli have been reported in companion animals. 3,4 Instead, the most commonly identified b-lactamases that confer extendedspectrum resistance in veterinary isolates are plasmid-mediated AmpC b-lactamases, in particular the cephalosporinase bla CY-2. In North America, bla CY-2 has been detected with increased frequency in multidrug-resistant E. coli (DREC) and Salmonella enterica serotype Newport (NewportDR-AmpC) isolated from food-producing and companion animals, pet treats and associated human infections. 5 7 Sanchez et al. 6 described the emergence of DREC associated with nosocomial infections in dogs in the United States. The isolates were shown to contain flo and bla CY genes conferring chloramphenicol/florfenicol and extended-spectrum cephalosporin resistance, respectively, as well as to contain class 1 integrons. Our laboratory also identified DREC as a cause of opportunistic infections in dogs in Australia at approximately the same time. 8 The isolates were resistant to all major classes of antimicrobials used in veterinary medicine, including extendedspectrum b-lactams and b-lactamase inhibitors. The aim of the present study was to determine b-lactam, chloramphenicol and class 1 integron-associated resistance mechanisms of 11 DREC isolates obtained from 10 cases of clinical infection in hospitalized dogs, the majority of which occurred at the University of Queensland Veterinary Teaching Hospital (UQVTH) between 1999 and Isolates were also typed using PFGE and plasmid analysis to identify clonal relationships between strains, and selected resistance genes were mapped to specific plasmids using Southern hybridization and transformation experiments. aterials and methods Bacterial strains, culture conditions and phenotypic tests The species assignment of 11 DREC isolates was confirmed biochemically using the icrobact 24E system (edvet Diagnostics, Thebarton, SA, Australia) and using species-specific PCR amplification of E. coli uspa. 9 The collection represented clinical isolates from 10 cases of opportunistic infection in hospitalized dogs (cystitis, post-surgical and miscellaneous wound infections) obtained from clinical specimens submitted to The University of Queensland Veterinary Diagnostic Laboratory for culture and sensitivity test (Table 1). All cases occurred at UQVTH, except for Case 7, which occurred at a private veterinary referral hospital in Brisbane. Isolates obtained from Cases 1 9 have been described previously. 8 The additional isolates were obtained from a surgical wound site (1) and urine (2a). Isolates from Case 8 lost their viability and isolates from Case 10 were subsequently shown to belong to Enterobacter cloacae and were excluded from the study. The only multiple isolates that were included in the study came from Case 2 and were obtained 22 days apart from the same surgical site (osteomyelitis following a fracture repair). The ICs of 20 antimicrobials (Table 2) were determined for the 11 isolates using broth microdilution as described by the CLSI, using E. coli ATCC as a control. 10 Phenotypic examination for ESBLs was performed using the modified double-disc test (DDT) developed by Pitout et al. 11 This test has the ability to demonstrate ESBLs regardless of the conflicting presence of AmpC b-lactamases. Sonic extracts from four of the 11 isolates (, C3, C5a, C7a) were subjected to isoelectric focusing (IEF) as described previously. 7 Following IEF, b-lactamase bands were visualized by staining with nitrocefin Table 1. Restriction endonuclease digestion, and plasmid and antimicrobial resistance gene profiles of 11 clinical DREC isolates b-lactamase d,e (PCR) Isolate a Date isolated REDP b Clonal group PP c TE CY-7 Chloramphenicol resistance gene e inti1 e Integron cassette array f 28/10/1999 1a 1 I + + cata1 + dfra17-aada5 g C2a h 17/04/2000 1b 1 I + + cata1 + dfra17-aada5 h 08/05/2000 2a 2 IIa NP C3 02/05/2000 2a 2 IIb NP C4 12/05/2000 2a 2 IIa NP C5a 06/07/2000 2a 2 IIa dfra5 27/07/2000 1b 1 I + + cata1 + dfra17-aada5 C7b 22/09/2000 2b 2 IIa NP 08/12/2000 1b 1 I + + cata1 + dfra17-aada5 1 04/04/2001 2c 2 IIc NP 2a 22/08/2001 2d 2 IIa NP a C, clinical isolate. b Restriction endonuclease digestion profile. c Plasmid profile. d b-lactamase identification by PCR amplification using known b-lactamase-specific primers. e +/ indicates the presence or absence of the genes. f Class 1 integron cassettes were identified using primers HS317 and HS320. NP, no product was amplified with these primers. g The integron gene cassette array from this isolate was sequenced fully, then specific primers were designed to identify these genes in other isolates. h Samples were from the same animal. 841

3 Sidjabat et al. Table 2. ICs for 11 clinical DREC isolates IC (mg/l) a,b Clonal group Isolate no. CIP STR SPT CHL TZP c FOX CRO CAZ CAZ/CLA d CTX CTX/CLA d AT AT/CLA d > C2a > > > > C C4 64 > C5a 128 > C7b > a 256 > > a All 11 isolates showed resistance to tetracycline (IC >32 mg/l), enrofloxacin (>32 mg/l), ampicillin (>512 mg/l), sulfamethoxazole/trimethoprim (32/608 mg/l), ticarcillin (>512 mg/l), piperacillin (>256 mg/l) and amoxicillin/clavulanic acid (>32/16 mg/l). All 11 isolates were susceptible to amikacin and imipenem. b CIP, ciprofloxacin; STR, streptomycin; SPT, spectinomycin; CHL, chloramphenicol; TZP, piperacillin/tazobactam; FOX, cefoxitin; CRO, ceftriaxone; CAZ, ceftazidime; CAZ/CLA, ceftazidime/clavulanic acid; CTX, cefotaxime; CTX/CLA, cefotaxime/clavulanic acid; AT, aztreonam; AT/CLA, aztreonam/clavulanic acid. c Tested in the presence of a constant concentration of tazobactam (4 mg/l). d Tested in the presence of a constant concentration of clavulanic acid (2 mg/l). (Oxoid). Gels were also overlaid with clavulanate (for detection of non AmpC-type enzymes) and cloxacillin (AmpC-type enzymes). PFGE and plasmid analysis Genomic DNA of the 11 isolates was prepared and digested with the restriction endonuclease XbaI (New England Biolabs, Australia) as described previously. 12 A low range lambda ladder (New England Biolabs) was used as a DNA size marker. The resulting fragments were separated in a GeneNavigator system (Pharmacia) at 200 volts with pulse times of 5 50s and linear ramping at a temperature of 12 C for 25 h. Relatedness between the restriction endonuclease digestion profiles (REDP) was determined by pair-wise comparison as described previously. 13 Plasmid DNA for the 11 isolates was isolated using the alkaline lysis method. 14 Plasmids were visualized by electrophoresis in 0.5% agarose prepared in 0.5 TBE. Plasmid sizes were estimated by comparison with plasmids of known size from the reference strains E. coli R27 (112 Da = 169 kb), R1 (62 Da = 93 kb) and RP4 (36 Da = 54 kb), as well as BAC- Tracker T Supercoiled DNA Ladder (Epicentre, adison, Wisconsin). PCR detection of resistance genes The isolates were tested for the presence of genes encoding bla TE and bla SHV b-lactamases using PCR as described previously. 15 DNA sequencing of bla TE products was also undertaken using the same primers. Plasmid-mediated AmpC b-lactamase gene amplification was carried out using multiplex PCR as described previously. 16 Oligonucleotide primers CY25F1 (nt ) and CYDR1 (nucleotides ) (Table 3) were designed by Hanson et al. 17 to flank the entire structural gene of Citrobacter freundii-origin plasmid-mediated AmpC (bla CY-2 ). The isolates were also tested for the presence of class 1 integron and chloramphenicol resistance genes using primer-specific PCR screening (Table 3). Oligonucleotide primer pairs used for the amplification of integrases and integron-associated gene cassette arrays were as follows: HS464 and HS463A, HS317 and HS320, respectively (Table 3). 18,19 Primers were also designed specifically to detect the presence of dfra5 and dfra17 gene cassettes in the isolates. DREC strains that demonstrated chloramphenicol resistance were tested for the presence of known chloramphenicol resistance genes (i.e. cata1, cata2, cata3, flo and cmla) Template DNA was prepared as described previously. 15 The total volume of the PCR mixture was 25 ml. Amplifications of products were performed using the GeneAmpÔ PCR system 2400 (Perkin-Elmer), in a reaction mixture containing 3.2 pmol of each primer, 200 m of each dntp, 1 PCR buffer, 2 m gcl 2, and 0.25 U of Red Hot Ò DNA Polymerase (ABgene), with an initial denaturation for 5 min at 95 C, followed by 30 cycles of denaturation for 30 s at 94 C, annealing for 30 s at various temperatures and extension for 30 s at 72 C, followed by a final extension for 5 min at 72 C (Table 3). Long-template amplification of all products for sequence analysis was performed using the Expand High Fidelity PCR System (Roche Diagnostic GmbH) as directed by the manufacturer. The amplified products for sequencing templates were purified using the QIAquick PCR purification kit (QIAGEN). Nucleotide sequences were determined using the ABI BigDye Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit, Version 2.0 (Applied Biosystems), and the reactions analysed using an ABI 377 automatic DNA sequencer at the Australian Genome Research Facility, Brisbane, Australia. Southern hybridization analysis DNA probes were prepared from PCR products generated for bla CY (462 bp), bla TE (971 bp), cata1 (581 bp) and class 1 integronassociated dfra17-aada5 (652 bp) using a digoxigenin nucleic acid labelling and detection system (Roche Diagnostic GmbH). Plasmid DNA was transferred onto positively charged nylon membranes (Amersham) on a vacuum blotter (Bio-Rad) at 5 mm Hg. embranebound DNA was hybridized to each probe according to conditions described previously. 14 The hybridization was carried out overnight at 42 C. Hybridization to the four different probes was also carried 842

4 Table 3. List of oligonucleotide primers used in the study Plasmid-mediated resistance genes in canine DREC Primer name Target gene Primer sequence (5 0!3 0 ) Amplicon size (bp) Annealing temperature ( C) Reference or accession number a uspa-f uspa CCGATACGCTGCCAATCAGT uspa-r ACGCAGACCGTAGGCCAGAT CY25F1 bla CY-2 or CAATGTGTGAGAAGCAGTC CYDR1 bla CY-7 CGCATGGGATTTTCCTTGCTG HS463A inti1 CTGGATTTCGATCACGGCACG position (complement) in NC_ HS464 ACATGCGTGTAAATCATCGTCG position in NC_ HS458 gene cassette GTTTGATGTTATGGAGCAGCAACG variable regions of integron class 1 HS459 GCAAAAAGGCAGCAATTATGAGCC HS317 gene cassette regions of GAACCTTGACCGAACGC variable 50 position in J01773 integron class 1 HS320 AGTAAAGCCCTCGCTAG 22 dfra17-f dfra17-aada5 ATATGTCTCTGGCGGGGGTC AF aada5-r CCGAACTGAAGCTCACGCCG this study dfra5-f dfra5 GGTTGCGGTCCACACATAC X12868 dfra5-r CTCCTTCCGGCTCAATATCA this study dfra17-f dfra17 ATGGCTCCTTGTCGGAAGA AF dfra17-r GTTGCGGCTTTGTGGAATAC this study CAT1-FWD cata1 AATGTACCTATAACCAG CAT1-REV TGCCTTAAAAAAATTAC CAT2-FWD cata2 CCTGGAATCGCAGGGAAC CAT2-REV CCTGCTGAAACTTTGCCA CAT3-FWD cata3 ATTGGGTTCGCCGTGAGC CAT3-REV AGTCTATCCCCTTCTTG FLO-FWD flo TATCTCCCTGTCGTTCCAG FLO-REV AGAACTCGCCGATCAATG CLA-FWD cmla CCGCCACGGTGTTGTTGTTATC CLA-REV CACCTTGCCTGCCCATCATTAG a GenBank accession number for the sequence used for primer design. out using plasmid DNA that had been digested with the BamHI restriction endonuclease enzyme (Roche Diagnostic GmbH). Conjugation and transformation experiments Conjugation experiments were carried out using a liquid mating procedure and on filters, as described previously. 7 Donor strains from both CG 1 (, C2a, ) and CG 2 (, 2a) were chosen, and E. coli J53Az R and S. enterica serotype Typhimurium (Salmonella Typhimurium) strain LT2 were the recipients. Transconjugants were selected on acconkey and ueller Hinton agar plates containing sodium azide (150 mg/l) and ampicillin (50 mg/l) for counter-selection of E. coli J53Az R. Possible transconjugants of S. Typhimurium strain LT2 were selected on ueller Hinton agar plates containing ampicillin (50 mg/l). Antimicrobial susceptibility testing of putative transconjugants was conducted for 16 antimicrobials using the CLSI disc diffusion methodology. 10 Transformation of plasmid DNA from the same CG 1 and CG 2 strains that were used for conjugation was performed using electroporation with E. coli K12 as the recipient. 15 Transformants were selected on ueller Hinton agar containing 50 mg/l ampicillin or 16 mg/l cefotaxime. Transfer of canine DREC plasmids into E. coli K12 was verified by plasmid profiling, Southern hybridization, PCR detection of bla CY-7 and bla TE, and disc diffusion antimicrobial susceptibility tests described by the CLSI. 10 Results Antimicrobial resistance profile, DDT and IEF All canine DREC isolates were resistant to tetracycline, enrofloxacin, ciprofloxacin, streptomycin, trimethoprim/ sulfamethoxazole, ampicillin, ticarcillin, piperacillin, amoxicillin/clavulanic acid and ticarcillin/clavulanic acid. In addition, all isolates were resistant or showed intermediate susceptibility to the extended-spectrum cephalosporins ceftriaxone, ceftazidime and cefotaxime and the cephamycin cefoxitin. Inclusion of the b-lactamase inhibitor clavulanic acid (2 mg/l) had no effect on the susceptibility of isolates to ceftazidime, cefotaxime and 843

5 Sidjabat et al. aztreonam. Piperacillin in combination with tazobactam (4 mg/l) increased the susceptibility of all isolates at least 8-fold compared with their susceptibility to piperacillin alone, although one isolate (2a) still remained resistant (piperacillin IC data not shown). Four isolates showed resistance to both chloramphenicol and spectinomycin (Table 2). All the isolates were sensitive to cefepime, imipenem and amikacin (data not shown). All isolates were shown to possess an AmpC b-lactamase using the DDT method, with no indication of ESBL activity. This was confirmed by demonstration of resistance to third-generation cephalosporins and the lack of a keyhole effect between amoxicillin/ clavulanate and/or tazobactam/piperacillin with any third- or fourth-generation cephalosporin. Representative isolates had identical IEF profiles showing a pi >8.2 enzyme inhibited by cloxacillin and a pi 5.4 enzyme inhibited by clavulanate, suggesting that they harboured AmpC and TE-like b-lactamases, respectively. PFGE and plasmid analysis Based on PFGE, the 11 isolates were subdivided into six distinct REDPs located within 2 unrelated E. coli clonal groups (CG 1 and 2). Four isolates were located in CG 1 and were subdivided into two closely related REDPs (1a 1b) (Table 1 and Figure 1). These isolates were resistant to chloramphenicol, had reduced susceptibility to spectinomycin and shared an identical plasmid profile (Figure 2a). The remaining seven isolates in CG 2 were divided into 4 REDPs (2a 2d) (Table 1 and Figure 1). These isolates shared similar plasmid profiles, IIa c (Table 1 and Figure 2a). A large plasmid of similar size to a 93 kb plasmid from E. coli strain R1 was present in both CG 1 and CG 2 isolates. Isolates belonging to CG 1 possessed an additional large plasmid which was slightly larger than a 169 kb plasmid possessed by E. coli strain R27 (Figure 2a and c), whereas CG 2 isolates possessed a second 93 kb plasmid that was slightly larger than the plasmid common to both clonal groups. When compared with CG 1 isolates, CG 2 isolates differed in their susceptibilities not only to chloramphenicol and spectinomycin but also to ciprofloxacin (respective ICs of 16 and 128 mg/l). The first recorded clinical case was a 14-year-old Welsh Pembroke corgi with pyelonephritis that was hospitalized in 1999 for 49 days and treated with multiple classes of antimicrobials. 8 A CG 1 strain was isolated from the urine of this animal. The second clinical case, a 2-year-old St Bernard admitted with chronic osteomyelitis following surgical repair of a fracture, did not develop a DREC infection until 6 months after Case 1. This animal was infected with a CG 1 DREC strain initially, whereas a second culture from an orthopaedic screw removed 22 days after the first admission yielded the first CG 2 strain. 8 CG 1 DRECs were subsequently isolated from Cases 6 and 9. The same or related strains of CG 2 DREC were responsible for the remaining cases of clinical infection (Cases 3, 4, 5, 7, 11, and 12). PCR detection of b-lactam, chloramphenicol and class 1 integron-associated resistance genes Direct PCR amplification generated products for all 11 isolates with primers specific for bla TE. No product was obtained using primers specific for bla SHV. Nucleotide sequence analysis C2a C3 C4 C5a C7a 1 2a 340 kb 291 kb 243 kb 194 kb 145 kb 97 kb 48.5 kb 23.1 kb 9.4 kb Figure 1. Pulsed-field gel electrophoresis restriction endonuclease digestion profiles of 11 clinical DREC isolates. = low range lambda ladder. identified the gene in two of the isolates as bla TE-1B. The multiplex AmpC PCR suggested the presence of a Citrobacterbased plasmid-mediated AmpC gene in all isolates. Following detection of a CIT-like AmpC, the full-length PCR-amplified product from each isolate was amplified using primers designed to flank the entire structural gene for bla CY-2 or bla CY-7. Nucleotide sequencing analysis of both strands identified the genes as bla CY-7. The sequence was identical with the bla CY-7 from S. Typhimurium strain 100 (GenBank accession number: AY324388). PCRs were carried out using primers designed to detect the presence of class 1 integrons and to recover class 1 integron-associated cassette arrays. All isolates produced a product with the primers HS463A and HS464 designed to amplify a region internal to inti1, implying that all isolates possess a class 1 integron. However, long-template PCR amplification using the primers HS317 and HS320 (designed to amplify integron-associated gene cassettes) produced products only for the four CG 1 isolates and one CG 2 isolate (Table 1). Nucleotide sequence analysis of the products amplified using HS317 and HS320 primers indicated that the four CG 1 isolates possessed a dfra17-aada5 gene cassette array. These genes respectively confer resistance to trimethoprim and spectinomycin/streptomycin. The amplicon of the variable region of class 1 integrons from isolate C5a contained dfra5, which encodes a dihydrofolate reductase and confers resistance to 844

6 Plasmid-mediated resistance genes in canine DREC (a) CG 1 CG 2 (b) C2a C3 C4 C5a C7b 1 2a RP4 R1 R27 C2a C3 C4 C5a C7b 1 2a RP4 R1 R kb 120 kb 95 kb 55 kb 38 kb chr 169 kb 93 kb 54 kb bla CY-7 plasmid chr (c) uncut cut with BamHI CG1 CG2 R1, R27 2a pcy-7 pte 1 2a pcy-7 pte λ/hindiii (d) uncut cut with BamHI CG1 CG2 R1, R27 2a pcy-7 pte 1 2a pcy-7 pte λ/hindiii 169 kb 93 kb bla CY-7 plasmid fragments 165 kb 120 kb 95 kb 55 kb 38 kb 23 kb 9.4 kb 6.5 kb chr bla CY-7 plasmid fragments 4.3 kb 2.3 kb 2.0 kb Figure 2. Plasmid profiles, BamHI-digested plasmids and Southern hybridization of 11 clinical DREC isolates and transformants. (a) Plasmid profiles and (b) Southern hybridization of plasmids using the DIG-labelled bla CY-7 probe. (c) Comparison of uncut and BamHI-digested plasmids of clinical isolates and transformants and (d) hybridization using the DIG-labelled bla CY-7 probe. = plasmid marker (BAC-Tracker T Supercoiled DNA Ladder). RP4, R1 and R27 are E. coli reference strains that contain plasmids of known size. chr, chromosomal DNA. trimethoprim. Integron-associated gene cassettes could not be identified in any of the remaining CG 2 isolates using primer pairs HS317 and HS320 or HS458 and HS459. Use of a longer annealing temperature also did not result in production of amplicons for these isolates (data not shown). All four chloramphenicol-resistant CG 1 isolates were positive for cata1 and were negative for PCR amplification using primers designed to amplify other phenicol resistance genes. Southern blot hybridization The bla CY-7 gene was shown to reside on the 93 kb plasmid that was present in both CG 1 and CG 2 (Figure 2b). Following restriction endonuclease digestion of plasmids by BamHI, Southern hybridization revealed that bla CY-7 was present on the same size plasmid fragment (12 kb) possessed by all DREC isolates (Figure 2c and d), confirming that there was 845

7 Sidjabat et al. no detectable difference in the location of this gene on the common 93 kb plasmid shared by CG 1 and CG 2. In CG 1 isolates, bla TE, cata1 and the class 1 integron dfra17-aada5 gene cassette were confirmed to be located on the large 170 kb plasmid (data not shown). Of the two 93 kb plasmids present in CG 2 isolates, the second, slightly larger plasmid was hybridized using the class 1 integron dfra17-aada5 and bla TE probes. It is possible that the remaining isolates within CG 2 may contain a single dfr gene encoding trimethoprim resistance. However, none of the remaining CG 2 isolates was positive in PCR amplifications designed to amplify either dfra5 or dfra17 genes. Conjugation and transformation experiments Despite multiple attempts, bla CY-7 - and bla TE -containing plasmids could not be transferred by conjugation to the sodium azide-resistant E. coli strain and S. Typhimurium strain LT2. The only colonies appearing on selective agar containing ampicillin and sodium azide were confirmed to be CG 1 and CG 2 DREC mutants that had developed spontaneous sodium azide resistance (data not shown). However, transformation of each of the 93 kb plasmids from E. coli CG 2 strain 2a into E. coli K12 using electroporation was achieved successfully. The rate of transformation was very low, since only four isolates were obtained from a total population of 10 8 E. coli K12, and each of the transformants was shown to possess a single large 93 kb plasmid. Two transformants designated E. coli K12:pCY-7HSA and K12:pCY-7HSB were confirmed to contain bla CY-7 by PCR and had a typical AmpC b-lactamase phenotype. They were resistant to all b-lactams and third-generation cephalosporins tested, including their combination with clavulanic acid, although they showed intermediate susceptibility to aztreonam. They were susceptible to all other antimicrobials tested. The two remaining transformants (E. coli K12:pTEHSA and K12:pTEHSB) were confirmed by PCR to contain bla TE, and they showed resistance to ampicillin, tetracycline, streptomycin and sulfamethoxazole/trimethoprim. The presence of bla CY-7 in the E. coli K12:pCY-7HSA transformant and its absence in the other E. coli K12:pTEHS transformants also was confirmed by Southern hybridization (Figure 2c and d). Discussion In this study, we report the identification of plasmid-mediated b-lactamases, chloramphenicol and class 1 integron-associated resistance genes in two clonal groups of DREC isolated from clinical infections in dogs at a veterinary teaching hospital in Australia. Apart from the 93 kb plasmid demonstrated to contain bla CY-7 that appeared to be common to both clonal groups, there was a marked difference in the distribution of plasmid-mediated resistance genes between CG 1 and CG 2. CG 1 strains shared the same plasmid profile, were resistant to chloramphenicol and spectinomycin, and possessed cata1, dfra17-aada5 and bla TE genes on a large 170 kb plasmid. In contrast, CG 2 isolates were sensitive to chloramphenicol and spectinomycin and possessed class 1 integron-associated and bla TE genes on a second 93 kb plasmid that was not present in CG 1 strains. Southern blotting of plasmids digested with BamHI and hybridized using a bla CY probe confirmed that the bla CY-7 gene was present on a 12 kb fragment common to both DREC clonal groups, suggesting that the 93 kb plasmid that carries bla CY-7 is highly conserved between CG 1 and CG 2. These results confirm that opportunistic infections at UQVTH were caused by clonal expansion of two distinct genetic groups of DREC, rather than by the transfer and spread of a single multidrug-resistant plasmid between genetically unrelated isolates. Epidemiological relationships between the 11 canine DRECs characterized in the present study and a large collection of DR coliforms obtained from rectal swabs of hospitalized dogs and the hospital environment over the same period are the subject of a separate study (H. E. Sidjabat, N. D. Hanson and D. J. Trott, unpublished results). The isolates characterized in the present study shared highly similar resistance mechanisms with canine DREC from a previous study in the United States, although the US isolates showed greater genetic diversity. 6 Canine DREC from the United States and Australia possessed a plasmid-mediated Citrobacter-type AmpC and class 1 integron-associated gene cassettes conferring resistance to trimethoprim (dfra17 and dfra5) and streptomycin/spectinomycin (aada5). Interestingly, analysis of the GenBank submission (accession no. AF475279) of bla CY identified in the study of Sanchez et al. 6 showed that this AmpC b-lactamase was also bla CY-7 and not bla CY-2,as reported. In the Australian isolates, chloramphenicol resistance in the CG 1 isolates was mediated by cata1. However, chloramphenicol resistance in the US isolates was in part due to the presence of the florfenicol resistance gene (flo), which mediates resistance to both chloramphenicol and florfenicol, an antimicrobial registered for use in food-producing animals only. There has been only one other report demonstrating the presence of plasmid-mediated bla CY-7 in a clinical isolate from Australia. This concerns an unusual isolate of S. enterica serotype Typhimurium that also produced an ESBL (SHV-9) and OXA The Salmonella strain 100 plasmid encoding bla CY-7 could be transferred into a sodium azide-resistant recipient E. coli strain. 24 However, in the present study, attempts to transfer bla CY-7 -carrying plasmids from CG 1 and 2 into the same sodium azide-resistant recipient E. coli strain or into S. Typhimurium strain LT2 (data not shown) were unsuccessful. Several previous studies have documented difficulties in transferring very large AmpC-containing plasmids by conjugation, though greater success has been obtained by direct transformation of plasmid DNA. 5,25 In our study, the transformation of the two 93 kb plasmids possessed by CG 2 strain 2a into E. coli K12 confirmed the array of resistance genes carried by each plasmid. The plasmid conferring an AmpC b-lactamase phenotype was shown to contain only bla CY-7, conferring the AmpC resistance phenotype, whereas the plasmid carrying bla TE was shown to carry in addition resistance to tetracycline, streptomycin and sulfamethoxazole/trimethoprim and conferred reduced susceptibility to gentamicin (data not shown). In Italy, Carattoli et al. 3 suggested that increased expanded spectrum cephalosporin resistance detected in E. coli isolates from dogs and cats was due to diffuse off-label veterinary use of extended-spectrum cephalosporins. Extended-spectrum cephalosporins are rarely used by veterinarians in companion animal practice in Australia, although first-generation cephalosporins are commonly administered by the oral route for urinary tract and skin infections and by the parenteral route for surgical prophylaxis. ore important, broad-spectrum b-lactam/ clavulanate combinations and fluoroquinolones were heavily 846

8 Plasmid-mediated resistance genes in canine DREC utilized in the UQVTH between 1999 and 2001, when the majority of DREC clinical infections occurred. Other antimicrobials in frequent use included doxycycline and trimethoprim/ sulfamethoxazole. Plasmid transformation experiments confirmed that plasmid co-selection of bla CY-7 was not operating through the clinical use of any of these additional antimicrobials at UQVTH, as the bla CY-7 -carrying plasmid did not appear to contain any other resistance genes. The origin of this bla CY-7 -carrying plasmid and the possibility of dog to human and human to dog horizontal transmission require further investigation. Integrons have been increasingly associated with multidrug resistance in both clinical and environmental bacteria, as they are capable of incorporating, collecting and expressing antibiotic gene-containing cassettes. 18 Sanchez et al. 6 also identified the dfra17-aada5 gene cassette array in canine DREC isolated from the United States, and this integron cassette array appears to be very common in E. coli and other enteric bacteria isolated from a wide variety of sources. 26 IC determinations confirmed that only the four CG 1 isolates that possessed the aada5 gene were resistant to spectinomycin. The remaining seven isolates possessed an inti1 gene and, by inference, a class 1 integron. However, gene cassette array amplified products were not obtained under the conditions used in this study. These isolates were all sensitive to spectinomycin, confirming that they did not possess aada5. The failure to produce a product may be a result of one of the primer target sites being absent in particular, the primer HS320, which is located within the 3 0 -conserved segment, since this segment may carry deletions or be absent entirely. 27 As further evidence of the absence of the 3 0 -conserved segment in the CG 2 strains, suli and sulii were confirmed to be present in CG 1 strains, whereas only sulii was found in CG 2 (H. E. Sidjabat, N. D. Hanson and D. J. Trott, unpublished data) It is clear from this and previous studies 3,6 that the intestinal microbiota of dogs represent a previously overlooked reservoir of antibiotic resistance genes. The emergence of plasmid-mediated AmpC b-lactamase-producing strains in veterinary hospitals has raised concern regarding prudent antimicrobial usage in companion animal medicine. Without antimicrobial selection pressure, a CG 1 canine DREC strain was rapidly eliminated from the gastrointestinal tract of experimentally infected dogs using commensal coliforms. 28 ore important, infection control procedures similar to those instituted in human hospitals were beneficial in identifying carrier animals, eliminating DREC from the hospital environment and reducing the subsequent number of clinical infections (H. E. Sidjabat, N. D. Hanson and D. J. Trott, unpublished results). These practices need to be implemented regularly in veterinary hospitals, particularly in large referral hospitals, to impede the emergence of DREC strains and reduce the public health risk posed by a high prevalence of these organisms in the gastrointestinal tract of dogs. Acknowledgements We would like to acknowledge Dr Narelle Fegan (Food Science Australia) and Kent Wu (Elizabeth acarthur Agriculture Institute, Australia) for technical assistance with pulsed-field gel electrophoresis and Dr Johann Pitout (University of Calgary, Alberta, Canada) for assistance with the interpretation of the modified disc diffusion method. We also would like to acknowledge Dr yat Kyaw-Tanner (School of Veterinary Science, The University of Queensland) for her excellent technical assistance. This work was supported by the Australian Companion Animal Health Foundation, the New South Wales Canine and Veterinary Foundation and a University of Queensland Research Development Grant. H. E. Sidjabat was the recipient of an AusAID postgraduate scholarship. Transparency declarations None to declare. References 1. Kang HY, Jeong YS, Oh JY et al. Characterization of antimicrobial resistance and class 1 integrons found in Escherichia coli isolates from humans and animals in Korea. J Antimicrob Chemother 2005; 55: Paterson DL, Yu VL. Extended-spectrum b-lactamases: a call for improved detection and control. Clin Infect Dis 1999; 29: Carattoli A, Lovari S, Franco A et al. Extended-spectrum b- lactamases in Escherichia coli isolated from dogs and cats in Rome, Italy, from 2001 to Antimicrob Agents Chemother 2005; 49: Féria C, Ferreira E, Correia JD et al. Patterns and mechanisms of resistance to b-lactams and b-lactamase inhibitors in uropathogenic Escherichia coli isolated from dogs in Portugal. J Antimicrob Chemother 2002; 49: Pitout JD, Reisbig D, ulvey et al. Association between handling of pet treats and infection with Salmonella enterica serotype Newport expressing the AmpC b-lactamase, CY-2. J Clin icrobiol 2003; 41: Sanchez S, Stevenson A, Hudson CR et al. Characterization of multidrug-resistant Escherichia coli isolates associated with nosocomial infections in dogs. J Clin icrobiol 2002; 40: Winokur PL, Brueggemann A, DeSalvo DL et al. Animal and human multidrug-resistant, cephalosporin-resistant Salmonella isolates expressing a plasmid-mediated CY-2 AmpC b-lactamase. Antimicrob Agents Chemother 2000; 44: Warren A, Townsend K, King T et al. ulti-drug resistant Escherichia coli with extended-spectrum b-lactamase activity and fluoroquinolone resistance isolated from clinical infections in dogs. Aust Vet J 2001; 79: Chen J, Griffiths W. PCR differentiation of Escherichia coli from other gram-negative bacteria using primers derived from the nucleotide sequences flanking the gene encoding the universal stress protein. Lett Appl icrobiol 1998; 27: National Committee for Clinical Laboratory Standards. Performance Standards for Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Isolated from Animals Second Edition: Approved Standard 31-A2. NCCLS, Wayne, PA, USA, Pitout JD, Reisbig D, Venter EC et al. odification of the doubledisk test for detection of Enterobacteriaceae producing extendedspectrum and AmpC b-lactamases. J Clin icrobiol 2003; 41: Barrett TJ, Lior H, Green JH et al. Laboratory investigation of a multistate food-borne outbreak of Escherichia coli O157:H7 by using pulsed-field gel electrophoresis and phage typing. J Clin icrobiol 1994; 32: Tenover FC, Arbeit RD, Goering RV et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin icrobiol 1995; 33: Sambrook J, Fritsch EF, aniatis T. olecular Cloning: A Laboratory anual. 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press,

9 Sidjabat et al. 15. Pitout JD, Thomson KS, Hanson ND et al. b-lactamases responsible for resistance to expanded-spectrum cephalosporins in Klebsiella pneumoniae, Escherichia coli, and Proteus mirabilis isolates recovered in South Africa. Antimicrob Agents Chemother 1998; 42: Pérez-Pérez FJ, Hanson ND. Detection of plasmid-mediated AmpC b-lactamase genes in clinical isolates by using multiplex PCR. J Clin icrobiol 2002; 40: Hanson ND, oland ES, Hossain A et al. Unusual Salmonella enterica serotype Typhimurium isolate producing CY-7, SHV-9 and OXA-30 b-lactamases. J Antimicrob Chemother 2002; 49: Holmes AJ, Gillings R, Nield BS et al. The gene cassette metagenome is a basic resource for bacterial genome evolution. Environ icrobiol 2003; 5: Nield BS, Holmes AJ, Gillings R et al. Recovery of new integron classes from environmental DNA. FES icrobiol Lett 2001; 195: Alton NK, Vapnek D. Nucleotide sequence analysis of the chloramphenicol resistance transposon Tn9. Nature 1979; 282: urray IA, artinez-suarez JV, Close TJ et al. Nucleotide sequences of genes encoding the type II chloramphenicol acetyltransferases of Escherichia coli and Haemophilus influenzae, which are sensitive to inhibition by thiol-reactive reagents. Biochem J 1990; 272: urray IA, Hawkins AR, Keyte JW et al. Nucleotide sequence analysis and overexpression of the gene encoding a type III chloramphenicol acetyltransferase. Biochem J 1988; 252: Keyes K, Hudson C, aurer JJ et al. Detection of florfenicol resistance genes in Escherichia coli isolated from sick chickens. Antimicrob Agents Chemother 2000; 44: Hossain A, Reisbig D, Hanson ND. Plasmid-encoded functions compensate for the biological cost of AmpC overexpression in a clinical isolate of Salmonella typhimurium. J Antimicrob Chemother 2004; 53: Winokur PL, Vonstein DL, Hoffman LJ et al. Evidence for transfer of CY-2 AmpC b-lactamase plasmids between Escherichia coli and Salmonella isolates from food animals and humans. Antimicrob Agents Chemother 2001; 45: Barlow RS, Pemberton J, Desmarchelier P. Isolation and characterization of integron-containing without antibiotic selection. Antimicrob Agents Chemother 2004; 48: Radstrom P, Swedberg G, Skold O. Genetic analyses of sulfonamide resistance and its dissemination in gram-negative bacteria illustrate new aspects of R plasmid evolution. Antimicrob Agents Chemother 1991; 35: Trott DJ, Filippich LJ, Bensink JC et al. Canine model for investigating the impact of oral enrofloxacin on commensal coliforms and colonization with multidrug-resistant Escherichia coli. J ed icrobiol 2004; 53: