THE ROLE OF GYRA, EFFLUX PUMP AND INTEGRONS IN MEDIATING THE EMERGENCE OF MULTI-DRUG RESISTANCE AMONG CANINE AND FELINE PATHOGENIC CLINICAL ESCHERICHIA COLI Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information. _________________________ Bashar Wajeeh Shaheen Certificate of Approval: _________________________ Omar A. Oyarzabal Associate Professor Poultry Science _________________________ Calvin M. Johnson Professor and Head Pathobiology _________________________ Dawn M. Boothe, Chair Professor Anatomy, Physiology and Pharmacology _________________________ George T. Flowers Dean Graduate School THE ROLE OF GYRA, EFFLUX PUMP AND INTEGRONS IN MEDIATING THE EMERGENCE OF MULTI-DRUG RESISTANCE AMONG CANINE AND FELINE PATHOGENIC CLINICAL ESCHERICHIA COLI Bashar Wajeeh Shaheen A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements of Doctor of Philosophy Auburn, Alabama May 9, 2009 iii THE ROLE OF GYRA, EFFLUX PUMP AND INTEGRONS IN MEDIATING THE EMERGENCE OF MULTI-DRUG RESISTANCE AMONG CANINE AND FELINE PATHOGENIC CLINICAL ESCHERICHIA COLI Bashar Wajeeh Shaheen Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon the request of individuals or institutions and at their expense. The author reserves all publication rights. _________________________ Signature of Author _________________________ Date of Graduation iv VITA Bashar W. Shaheen, son of Wajeeh Raghib Shaheen and Aziza Abdulfatah Naser, was born August 29, 1975, in Kuwait, State of Kuwait. He graduated from comprehensive high school in 1993. He completed his degree of Bachelor of Veterinary Medicine and Surgery (BVMS) equivalent to DVM from Mosul University, Iraq in September, 1998. After working as a veterinarian in Greater Amman Municipality (food safety and inspector in charge) for three years, he immigrated to the United States and entered Graduate School in the Department of Poultry Science at Auburn University, Alabama in January 2004 and graduated with a Master of Science degree in May, 2006. In 2006, he began his Doctoral work in Biomedical Science at Auburn University College of Veterinary Medicine. v DISSERTATION ABSTRACT THE ROLE OF GYRA, EFFLUX PUMP AND INTEGRONS IN MEDIATING THE EMERGENCE OF MULTI-DRUG RESISTANCE AMONG CANINE AND FELINE PATHOGENIC CLINICAL ESCHERICHIA COLI Bashar Wajeeh Shaheen Doctor of Philosophy, May 9, 2009 (M.S., Auburn Univesity, 2006) (BVMS, Mosul University, 1998) 176 Typed Pages Directed by Dawn M. Boothe A series of experiments was performed to elucidate the emergence of fluoroquinolone (FQ)/multi-drug resistance (MDR) and the contribution of selected mechanisms of resistance to MDR (i.e. mutations in gyrA, efflux pump, and the integrons) in canine and feline pathogenic Escherichia coli. E. coli isolates (n= 377) were collected from dogs or cats with spontaneous infection between May and September 2005 for phenotypic (i.e. E-test?) and genotypic (i.e PFGE) characterization. Of isolates expressing resistance to any drug, two isolates expressed single drug resistance (SDR) only to enrofloxacin. The remaining 109 isolates expressing resistance to enrofloxacin also expressed MDR, with resistance to all 7 drugs (Z phenotype) representing the largest proportion (18.3%; 20/109). The drugs most commonly involved in MDR phenotypes (n=109) were amoxicillin (96.3 %; 105/109), amoxicillin-clavulanic acid (85%; 93/109) vi and enrofloxcin (61.5%; 67/109). Genotypically, isolates were extensively diverse, regardless of resistance phenotype, and phenotypes and genotypes were not related. For the first time, a FRET PCR was developed to identify enrofloxacin-resistance in clinical E. coli isolates that carry mutations in codon 83 and 87 of gyrA. The assay identified as few as four genome copies per reaction from culture and 19 genome copies in urine with a very short time (1-2 hrs). For the 70 isolates tested, the sensitivity of the test was 87.5% (95% CI = 75% to 95.3%) (n=42/48) whereas specificity was 100% (95% CI = 87.3% to 100%) (n=22/22). MICs for E. coli isolates (n= 536) were determined for enrofloxacin and five other drug classes using broth micro-dilution. FQ resistance was significantly associated with the MDR phenotype compared to non FQ-R isolates. The results suggest that the double mutation in gyrA confers high level of resistance to the majority of FQ-R isolates. Furthermore, the impact of an efflux pump was studied. Blocking the action of the AcrAB efflux pump (i) decrease the intrinsic level of the MICs to FQ in susceptible isolates, (ii) decrease the MICs below the susceptible break point for FQ-R isolates even with single mutation in gyrA, (iii) decrease the magnitude of resistance in highly resistance isolates in the presence of double mutations in gyrA (IV) decrease the MICs below the susceptible break point for some of the highly FQ-R isolates without gyrA mutations, and (V) decrease the resistance to structurally unrelated drugs, thus reducing the incidence of MDR. Integrons of Class I and 2 were identified in 27% of the isolates; thus only 2.4% of the isolates carried Class 2 integrons. All integron positive isolates were MDR, compared to 56.6% of integron negative isolates. Resistance resolved in the 3 MDR integron positive isolates subjected to plasmid curing (using 10 % SDS). This is the first report of class 2 integrons in E. coli from companion animals in US. vii ACKNOWLEDGMENTS This dissertation could not have been written without Dr. Dawn M. Boothe who not only served as my supervisor but also encouraged me and challenged throughout my graduate program. For everthing you?ve done for me, Dr. Boothe, I thank you. Also I would like to thank my committee members, Dr. Omar A. Oyarzabal, and Dr. Calvin M. Johnson who offered guidance and support. Special thanks are due to Dr. Leonardo De La Fuente for his time and consideration to review this dissertation. I would also like to thank Dr. Bernhard Kaltenboeck and Dr. Chengming Wang for their assistance and guidance in FRET PCR design. I would like to thank my parents, my father Wajeeh Shaheen and my mother Aziza Naser, who taught me the value of hard work by their own experience. I would like to share this moment of happiness with my father and mother-in law ?Hafez and Khawla", who rendered me enormous support during the whole study. Finally, and most importantly, I would like to thank my wife Suhair. Her support, encouragement, quiet patience and unwavering love were undeniably the bedrock upon which the past eight years of my life have been built. Her tolerance of my occasional absence is a testament in itself of her unyielding devotion and love. Without her love, support and understanding I would never have completed my present work. Particularly, I owe to my son Kareem and my daughter Sarah for their silent prayer for my work to be successfully completed. viii Style manual or journal used Journal of Veterinary Microbiology Computer software used Microsoft? Windows? XP Professional Statistical software MINITAB ? 15 package Vector NTI? Advance 10 SWIN? Epidemiology Module ix TABLE OF CONTENTS LIST OF TABLES..................................................................................................... LIST OF FIGURES................................................................................................... I. LITERATURE REVIEW: ESCHERICHIA COLI AND FLUOROQUINOLE RESISTANCE IN COMPANION ANIMALS??????????????.. II. PHENOTYPIC AND GENOTYPIC CHARACTERIZATION OF CLINICAL ESCHERICHIA COLI ISOLATES EXPRESSING MULTIDRUG RESISTANCE RECOVERED FROM CANINE AND FELINE WITH SPONTANEOUS DISEASE................................................................................................................... III. DETECTION OF FLUOROQUINOLONE RESISTANCE LEVEL IN CLINICAL CANINE AND FELINE ESCHERICHIA COLI PATHOGENS USING RAPID REAL-TIME PCR ASSAY??????????????..... IV. THE CONTRIBUTION OF GYRA MUTATION AND EFFLUX PUMPS TO FLUOROQUINOLONE AND MULTI-DRUG RESISTANCE IN CANINE AND FELINE PATHOGENIC ESCHERICHIA COLI ISOLATES????????... V. THE ROLE OF CLASS 1 AND 2 INTEGRONS IN MEDIATING ANTIMICROBIAL RESISTANCE AMONG CANINE AND FELINE CLINICAL E. COLI ISOLATES FROM US??????????????? x xii 1 31 64 98 132 x LIST OF TABLES I. 1. RESISTANCE TO FLUOROQUINOLONES BY DIFFERENT MECHANISMS????????????????????? II. 1. DISTRIBUTION OF RESISTANT PHENOTYPES WITH THEIR GEOGRAPHICAL LOCATIONS ???.??????????.. 2. DISTRIBUTION OF RESISTANT PHENOTYPES WITH THEIR CLINICAL SOURCES??????????????????. III. 1. E. COLI STRAINS AND ISOLATES USED IN THIS STUDIES AND THEIR ANTIMICROBIAL SUSCEPTIBILITY TO ENROFLOXACIN???????????????????? IV. 1. ANTIMICROBIAL DRUGS USED IN THIS STUDY?????? 2. PRIMERS AND PROBES USED IN FRET PCR ASSAY AND IN THE MELTING CURVE ANALYSIS????????????. 3. ANTIMICROBIAL RESISTANCE PATTERN OF FLUOROQUINLONE RESISTANT AND SUSCEPTIBLE ISOLATES??????????????????????? 4. CHARACTERISTICS OF ESCHERICHIA COLI ISOLATED FROM DOGS AND CATS USED IN THIS STUDY?????????.. 5 51 54 90 124 125 126 127 xi V. 1. ANTIMICROBIAL DRUGS USED IN THIS STUDY????... 2. PRIMERS SEQUENCES AND PCR CONDITIONS USED IN THIS STUDY?????????...???????????.. 3. THE ASSOCIATION BETWEEN THE ANTIBIOTIC PROFILE AND CLASS 1 INTEGRONS IN E. COLI?????????? 4. ANTIBIOTIC RESISTANCE PATTERNS OF E. COLI ISOLATES AND THEIR RELATIONSHIP WITH OCCURRENCE OF THE INTEGRONS??????????... 5. A BLAST SEARCH FOR ANTIMICROBIAL GENE CASSETTES????????????????????... 5. ANTIMICROBIAL RESISTANCE PROFILES AND MICS FOR THE ISOLATES BEFORE AND AFTER CURING???... 155 156 158 159 160 162 xii LIST OF FIGURES I. 1. SCHEMATIC REPRESENTATION OF THE GYRA REGION?... 2. FLUORESCENCE RESONANCE ENERGY TRANSFER (FRET)????????????????????????... 3. USING FRET TO DIFFERENTIATE THE MUTANT AND THE WILD-TYPE OF Y.PESTIS????????????????... 4. DIAGRAMMATIC COMPARISON OF THE FIVE FAMILIES OF EFFLUX PUMPS??????????????????? 5. GENERAL STRUCTURE OF CLASS 1 INTEGRON?............ II. 1. RESISTANT ISOLATES OF E.COLI FROM FOUR US REGIONS AND THEIR CLINICAL SOURCES???????. 2. THE NUMBER AND PROPORTION (%) OF EACH SDR (A) AND MDR (B) PHENOTYPE(S) EXPRESSED IN 377 ISOLATES OF E. COLI CAUSING SPONTANEOUS DISEASE IN DOGS AND CATS THROUGHOUT THE US.???????????. 3. THE OVERALL PROPORTION OF SDR, MDR, AND NR PHENOTYPES AMONG E. COLI CLINICAL ISOLATES BY GEOGRAPHICAL REGION ??????????????? 7 12 12 14 16 53 55 56 xiii 4. THE OVERALL PROPORTION OF SDR, MDR, AND NR PHENOTYPESAMONG E. COLI CLINICAL ISOLATES BY TISSUE SOURCE???????????????????.. 5. DENDROGRAM OF SDR, MDR, AND N PHENOTYPES AMONG E. COLI CLINICAL ISOLATES COLLECTED AT THE SAME INSTITUTION (SOUTH)????????.??? 6. THE DENDROGRAM FOR E. COLI PHENOTYPES COLLECTED FROM ONE TISSUE (URINE)???????? 7. FREQUENCY OF THE MDR PHENOTYPE IN ENROFLOXACIN ISOLATES SHOWING RESISTANCE TO NON-FQ ANTIMICROBIAL AGENTS??????????... III. 1. DNA SEQUENCE ALIGNMENTS OF THE GYRA QRDRS OF E. COLI ISOLATES INCLUDING WILD ISOLATES AND MUTANT CLINICAL ISOLATES WITH SER-83 AND ASP-87 MUTATIONS?????????????????????? 2. AMPLIFICATION AND MELTING CURVE ANALYSIS OF PCR AMPLICON CONTAINING THE QRDR OF THE GYRA GENE OF CLINICAL E. COLI ISOLATES?????????... 57 58 60 62 93 94 xiv 3. MELTING TEMPERATURES CORRELATE AND PREDICT THE ENROFLOXACIN MICS IN CLINICAL ISOLATES?........... IV. 1. COMPARISON OF THE FREQUENCY OF RESISTANCE PATTERN OF FQ-R VERSUS FQ-S ISOLATES TO DIFFERENT DRUG CLASSES????????????????????.. 2.EFFECT OF GYRA MUTATIONS AND EPI ON THE IN VITRO MICS OF ENROFLOXACIN FOR 56 CLINICAL E. COLI ISOLATES ISOLATED FROM DOGS AND CATS??????... 3. THE EFFECT OF EPI ON MICS FOR DIFFERENT ANTIMICROBIAL DRUGS????????????????. VI. 1. ASSOCIATION BETWEEN ANTIMICROBIAL RESISTANT PROFILE AND OCCURRENCE OF CLASS 1 AND 2 INTEGRONS IN ESCHERICHIA COLI STAINS ISOLATED FROM DOGS AND CATS????????????????... 2. PFGE PROFILES OF THE EXTRACTED PLASMID DNA OF THE 3 ISOLATES BEFORE AND AFTER CURING????.. 96 129 130 131 157 161 1 CHAPTER I LITERATURE REVIEW: ESCHERICHIA COLI AND FLUOROQUINOLE RESISTANCE IN COMPANION ANIMALS Fluoroquinolone resistance among canine and feline E. coli Escherichia coli is an important pathogen associated with urogenital and other infections in dogs and cats. An increase in the rate of resistance to fluoroquinolones has been reported among E. coli isolates from dogs and cats (Cohn et al., 2003; Cooke et al., 2002; Guardabassi et al., 2004). Cooke et al. (2002) reported an increase in enrofloxacin- resistant Escherichia coli associated with urinary tract infections in dogs at a veterinary medical teaching hospital and from other veterinary hospitals in California. Those enrofloxacin-resistant Escherichia coli isolates were also resistant to other antimicrobials such as ampicillin, trimethoprim-sulphamethoxazole, and cephalexin. The trend of increased resistance to fluoroquinolone among uropathogenic E. coli isolates from dogs between 1992 and 2001 in Missouri was determined for 1,478 bacterial isolates including Escherichia coli (n=547 isolates). The overall resistance of E. coli (n=547) to fluoroquinolones increased significantly over time (P < 0.0001). For each drug, the increased resistance of E. coli to ciprofloxacin was significant (P = 0.0001), but resistance to enrofloxacin failed to reach statistical significance (P = 0.022 at significant P value of 0.01). Boothe et al. (2006) also studied the pharmacodynamic and 2 pharmacokinetic indices of efficacy for 5 fluoroquinolones toward pathogens of dogs and cats. Percent resistance for all Gram-negative and Gram-positive isolates did not differ among drugs or organisms; however E. coli resistance to fluoroquinolones approximated 40%. Resistance contributed to a bimodal distribution of MIC, with the MIC 50 at < 0.25 mcg/ml, but the MIC 90 > 32 ?g/ml. Similarly, studies from human medicine indicated the association of fluoroquinolone with multidrug-resistant phenotypes commonly with the clinical cases of uropathogenic E. coli (Karlowsky et al., 2002, 2003, 2006). previous work of Boothe et al indicated a high incidence of MDR in E. coli uropathogens isolated from canine patients admitted to Auburn University?s Veterinary Teaching Hospital (n=175). Only 50% of uropathogens were E. coli. Overall E. coli resistance to all 17 antimicrobials was 31% and approximated 40% to 50% toward ?first choice? empirical antimicrobials (amoxicillin, amoxicillin-clavulanic acid, cephalexin and sulfonamide-trimethoprim) and toward enrofloxacin (Boothe et al., 2005). Furthermore our pilot data indicated that fluoroquinolone-resistant isolates (i.e. those resistance to enrofloxacin) were more likely to be multi-drug resistant (MDR). Only 2 out 85 (2%) isolates expressed single drug resistant (SDR) phenotype, while (61.5%; n=67/109) of MDR phenotypes were associated with enrofloxacin resistance (Shaheen et al., 2008a, 2008b). An important risk factor associated with the emergence of fluoroquinolones resistance is the approval and the possibility of increased use of broad-spectrum antimicrobials (i.e. fluoroquinolones) at veterinary hospitals as previously described (Sanchez et al., 2002). Furthermore, the enrofloxacin resistant isolates were also 3 associated with resistance to other antimicrobial agents. This association was also observed at a human hospital setting (Richard et al., 1994) especially with patients who had a long term treatment with ciprofloxacin (Muder et al., 1991). As consequence, the emergence of resistance to E. coli and treatment failure has been observed during treatment of E. coli infections with quinolines (Troung et al., 1995). Emergent resistance in E. coli or other bacterial organisms in animals receiving fluoroquinolones might present a public health risk, as has been demonstrated thorough transferring resistant isolates from food producing animals previously treated with FQs (Webber and Piddock, 2001). For dogs and cats, the risk may be increased through the close physical contact between household pets and humans (3). Indeed, this is supported by the fact that resistance mechanisms (i.e. resistance genes) in these isolates are the same as those in the resistant strains found in humans (Guardabassi et al., 2004, Webber and Piddock, 2001). The antimicrobial action of fluoroquinolones Quinolones interact with two type II topoisomerases (bacteria DNA gyrase) and topoisomerase IV. Type II topoisomerases act by breaking both strands of duplex DNA, passing another DNA strand through the break, and resealing the gap. DNA gyrase is a tetrameric enzyme composed of two gyrA subunits and two gyrB subunits. It maintains the DNA structure through catalyzing the negative supercoiling of DNA (Reece and Maxwell, 1991). Negative supercoils are important for initiation of DNA replication and facilitate binding of initiation proteins. During transcription, the replication fork and RNA polymerase tend to introduce positive supercoils which must be removed to avoid stalling of the replication fork (Hooper, 1999). Topoisomerase IV is also a tetrameric 4 enzyme consisting of two ParC and two ParE subunits. Its function is important during the DNA replication by segregation of replicated daughter chromosomes (Drlica and Zhao, 1997). Whereas both DNA gyrase and topoisomerase IV can remove positive supercoils, only DNA gyrase can convert them directly to negative supercoils. The mechanisms of resistance to fluoroquinolones The common mechanism of action of fluoroquinolone is inhibition of topoisomerase. This is demonstrated by point mutations that result in amino acid substitutions within the topoisomerase genes, accompanied with decreased expression of outer membrane porins and overexpression of multidrug efflux pumps. In addition, the discovery of plasmid-mediated quinolone resistance suggests that they which could play a role in horizontal transfer of resistance genes between strains. In addition to that is the discovery of plasmid-mediated quinolone resistance which could play a role in horizontal transfer of resistance genes between strains. Table 1 summarizes the common mechanisms of resistance to quinolone and fluoroquinolone. It is should be noted that mutations in topoisomerase genes cause irreversible changes that will be inherited by all daughter cells of the resistant isolate. This can be a cause of the therapeutic failure due to the emergence of large populations of resistant bacteria. In contrast, plasmid mediated resistance is reversible by removal of the antimicrobial agent, resulting in a loss of positive selective pressure for those bacteria carrying the plasmid. 5 Table 1. Resistance to fluoroquinolones by different mechanisms. a C, chromosomal DNA; P, plasmid; I, integron. Resistance mechanism Resistance gene(s) Gene product(s) Bacteria involved Location of Resistance gene a Mutational modificatio n of the target site Mutation in gyrA resulting in changes of Ser-83 (to Tyr, Phe, or Ala) and Asp-87 (to Gly, Asn, or Tyr) were detected most commonly. Stepwise mutations in gyrA and parC can result in an incremental increase resistance to fluoroquinolones. Mutation in the genes for DNA gyrase and topoisome rase IV Various gram- positive and ? negative bacteria C Mutational modificatio n of regulatory genes that control the efflux pump systems The level of production of the AcrAB-TolC efflux system in E. coli is under the control of several regulatory genes (marRAB, soxR, or acrR genes). Mutation in these regulatory genes leads to overproduction of the AcrAB-TolC efflux pump and the expression of MDR phenotypes. Mutation in the marRAB, soxR, or acrR genes E. coli C Efflux via multidrug transporter s mexB-mexA-oprM, acrA- acrB-TolC Multidrug efflux in combinati on with specific outer membrane proteins Pseudo monas, E. coli, Salmone lla C Decreased drug uptake OmpF is an important porin for the entery of quinolones and fluoroquinolones into the bacterial cell. MAR- mediated down regulation of OmpF porin Gram- nagative bacteria C 6 production . Mutation in different gene loci (cfxB, norB, nfxB, norc, or nalB) are also associated with decreased permeabili ty. DNA gyrase protection Low level quinolone resistance via protection of DNA gyrase from inhibition by quinolones is mediated by the gyr gene. Qnr gene K. pneumo niae, and E. coli P, I Mutations in the topoisomerase genes The mutations in DNA gyrase and topoisomerase IV are the two important mechanisms that confer resistance to the quinolone. Many studies have reported that topoisomerase IV in gram-negative bacteria is the secondary target for quinolones (Khodursky et al., 1995; Kumagai et al., 1996; Heisig, 1996), and the primary target for gram positive bacteria (Ferrero et al., 1994). Fluoroquinolones inhibit the DNA synthesis through stabilizing the breaks in the DNA made by the DNA gyrase or topoisomerase IV. Furthermore, amino acid changes in the quinolone resistance-determining region (QRDR) of GyrA alter the structure of the site of quinolone binding near the interface of the 7 enzyme and DNA and that resistance is then caused by reduced drug affinity for the modified enzyme-DNA complex. Mutation in DNA gyrase Most mutations have been described previously are located in a region called the quinolone resistance-determining region (QRDR). The region (Fig. 1) in E. coli is located between amino acids Ala67- and Gln106 in gyrA (Yoshida et al., 1990) and at amino acid Asp426 and Lys447 in gyrB (Yoshida et al., 1991). Mutations occur mostly at codons 83 and 87, which changes the binding of quinolones to the active site and confers resistance to quinolones (Willmott and Maxwell, 1993). Figure 1. Schematic representation of the gyrA region amplified by primers gyrA6 and gyrA631R (arrows), including the 120-bp QRDR (heavy line) encoding amino acids 67 to 106 of the E. coli GyrA protein (Swanberg and Wang, 1987) DNA gyrase mutations in E. coli The level of resistance to fluoroquinolone is condon specific and can be affected by the number of the mutations involved. For example, a single mutation in gyrA in E. coli conferred high-level resistance to nalidixic acid (Markham and Neyfakh, 1996; Luo 8 et al., 2006). However, further mutations in gyrA and/or the topoisomerase IV genes are essential to confer high-level resistance to fluoroquinolone (Everett et al., 1996; Ozeki et al., 1997; Weigel et al., 1998). Clinically, including the veterinary strains of E. coli, most mutations in gyrA occur at codon 83. In E. coli the mutation at codon 83 involves most commonly the substitution of serine residue for a leucine which confers high-level resistance to nalidixic acid and fluoroquinolones (Ozeki et al., 1997; Weigel et al., 1998; White et al., 2000). Further mutation at codon Asp87 generates slightly higher resistance to fluoroquinolones (Lehn et al., 1996; S?enz et al., 2003). In contrast, the single mutation at Ser83 has its greater impact on the isolate to become more resistant to fluoroquinolones than those with a single mutation at Asp87 or other regions (Lehn et al., 1996). Many studies reported that single mutations in Gly81Asp, Asp82Gly, Asp87Asn, Asp87Gly and Asp87Tyr result in low-level resistance to fluoroquinolones (McDonald et al., 2001). In general the mutations at codons Ser83 and Asp87 confer higher levels of quinolone resistance than mutations in other codons (Yoshida et al., 1990). Mutation in Topoisomerase IV genes Many studies indicated that topoisomerase IV plays a secondary role as a target for quinolones in Gram-negative bacteria (Hoshino et al., 1994; Kato et al., 1992; Khodursky et al., 1995; Peng and Marians, 1993). Other studies found that mutations in parC and parE of E. coli always occurred in conjunction with mutations in gyrA (Everett et al., 1996; Komp Lindgren and Karlsson, 2003; Fendukly et al., 2003; S?enz et al., 2003). These findings suggested that the DNA gyrase mutations should proceed the topoisomerase IV mutations to confer high-level fluoroquinolone resistance in E. coli 9 (Everett et al., 1996; Heisig, 1996). Another in vitro study indicated that single mutation within parC did not confer the strain to be resistant to the quinolone and that expression to level resistance to fluoroquinolone was achieved by a combination of the gyrA double mutation and one parC mutation (Bagel et al., 1999). Another in vitro study revealed that mutation in parC was the second step confering high-level resistance to fluoroquinolone in Escherichia coli (Bachoual et al., 1998). Therefore mutations topoisomerases IV genes impact the level of resistance to fluoroquinolones. The advantages of real-time PCR For our clinical purposes, utilized the real time PCR has major advantages for detection of mutations in codon 83 and 87 of gyrA in E. coli. The LightCycler (Roche, Inc. Indianapolis, IN) System has been designed to reduce the time required to achieve results from PCR reactions and to monitor the amplification of the target sequence simultaneously, in real-time. The system has an advantage to maximize the amount of information that can be obtained from a single reaction, making the system a powerful tool for quantitative PCR and mutation analysis. A major benefit of using this system is the short cycling time; so that a complete PCR run of 20 -30 cycles can be performed in 20 ? 30 minutes. Furthermore, because no gel electrophoresis is required, contamination was minimized as result of maintaing tubes sealed at the end of the amplification. Fluorescence Resonance Energy Transfer (FRET) The hybridization probe format was utilized for maximum sequence-specificity for detection of PCR product. Two designed, sequence-specific oligonucleotide probes, labeled with different dyes, were used. The sequence of the two probes can hybridize to 10 the target sequences on the amplified DNA fragment in a head-to-tail arrangement. The donor dye (Fluoroescein) is excited by the blue light source and emits green fluorescent light at a slightly longer wavelength. When the two dyes are in close proximity, the energy emitted excites the acceptor dye attached to the second Hybridization Probe, which then emits fluorescent light at a different wavelength. This energy transfer, referred to as fluorescence resonance energy transfer (FRET) is highly dependent on the close proximity spacing between the two dye molecules (1-5 nucleotides). The amount of fluorescence is directly proportional to the amount of target DNA generated during the PCR process. Typing of nucleotide sequence polymorphisms and analyzing mutations The LightCycler System can achieve precise discrimination of DNA sequence by monitoring the temperature-dependent hybridization of sequence-specific hybridization probe to single stranded DNA while performing the unique melting curve analysis. The melting temperature (T m ) is dependent on the probe length and GC content, and on the degree of homology between the two DNA strands. Hybridization probes bound perfectly to the matching target DNA which required a higher T m to separate in comparison with those bound to DNA containing destabilizing mismatches (i.e. the oligonucleotide and its perfect complement are in duplex). This is a powerful tool in which no post-PCR processing is required, and thus minimizes the possibility of contamination as the amplification and genotyping will be performed in the same sealed capillary. In the mutation analysis experiment, one hybridization probe was labeled with an LC Red flourophore which hybridized to a part of the target sequence that was not 11 mutated and which will function as an anchor probe. The other hybridization probe, will be labeled with a Fluorescein fluorophore and, in the mutated isolates, spanned the (shorter [mutation probe]) mutation site (Fig 2). After hybridization, the two DNA probes come into close proximity, resulting in FRET between the two fluorophores. During the melting curve analysis, as increasing temperature caused the shorter mutation probe to dissociate first, and the two fluorescent dyes were no longer in close proximity, causing fluorescence to decrease. If the mutation was present, the mismatch of mutation probe with the target (anchor) will destabilize the hybrid, causing T m to be lower. In contrast, mismatches did not occur with wild type phenotype, and the heteroduplex will thus have a higher T m (Fig 4). An example is using the FRET probe assay to differentiate Y. pestis KIM 5 gyrA mutants (Cp r ) from wild-type (Cp s ) organisms (Lindler et al., 2001). As shown in Fig. 8, the melting peak temperature was dependent on the homology between probe 1 and the amplified PCR product. All of the mutant PCR products formed a less- stable hybrid with the designed probe 1 than did the wild-type gyrA sequence. The largest difference in melting peak was seen with the cytosine-to-adenosine mutation (M4 in Fig. 3). The two mutations, which were the most distal to the 3' end of probe 1 (M1 and M2 in Fig. 4), had an intermediate T m . The mutant gyrA-probe1 T m was easily differentiated from those generated with wild-type templates. The minimum T m decrease compared to that for the wild-type sequence was with mutant M3, for which the T m was around 3?C (Fig. 3). The maximum T m shift compared to Y. pestis KIM 5 was around 10 ? C for the M4 mutant. 12 Figure 2. Fluorescence resonance energy transfer (FRET). The mutation probe (Fluorescein) spans the mutation site while the anchor probe hybridizes to a part of the target sequence that is not mutated. FRET occurs when the two probes hybridizing to the target sequence come in close proximity (1-5 nucleotides apart). The energy emitted excites LC640, which then emits red fluorescent light. Figure 3. Using FRET to differentiate the mutant and the wild-type of Y. pestis. Wild- type and mutant templates were used in the PCR amplification followed by melting peak analysis. (Lindler et al., 2001). 13 Efflux pump system Five different families of efflux pump proteins (Fig. 4) have been identified in gram positive and negative bacteria and these include the resistance nodulation division (RND) family, the major facilitator superfamily (MFS), and the staphylococcal multi- resistance (SMR), multidrug and toxic compound extrusion (MATE) families, and ABC (ATP binding cassette). The efflux pump systems of the RND family are organized as tripartite efflux pumps. The pump is very important in Escherichia coli and other gram- negative bacteria (e.g., E. coli acrB/AcrB, P. aeruginosa mexB/MexB, Campylobacter jejuni cmeB/CmeB, and Neisseria gonorrhoeae mtrD/MtrD). The efflux pump has three components: a transporter (efflux) protein in the inner membrane (e.g., AcrB), a periplasmic accessory protein (e.g., AcrA), and an outer membrane protein channel (e.g., TolC) (Koronakis et al., 2004). AcrB captures its substrates within either the phospholipid bilayer of the inner membrane of the cytoplasm (Aires and Nikaido, 2005) and effluxes them into the external medium via TolC. The AcrB and TolC are connected via the periplasmic protein AcrA. AcrA protein provides collaboration between AcrB and TolC. The RND pumps are proton antiporters, using the proton gradient across the membrane to power efflux, exchanging one H + ion for one drug molecule (Paulsen, 2003). 14 Figure 4. Diagrammatic comparison of the five families of efflux pumps (Piddock, 2006). The AcrAB-TolC system as an example of RND system in E. coli has wide range of substrates includes chloramphenicol, lipophilic ?-lactams, fluoroquinolones, tetracycline, rifampin, novobiocin, fusidic acid, nalidixic acid, ethidium bromide, acriflavine, bile salts, short-chain fatty acids, SDS, Triton X-100, and triclosan (Folster and Shefer, 2005; Nikaido, 2003; Olliver et al., 2004; Tegos et al., 2002; Wolter et al., 2004). Overexpression of AcrAB-TolC system alone or mutations in the topoisomerase gene is unlikely, alone, to give rise to clinical levels of resistance to certain drugs such as ciprofloxacin, chloramphenicol, tetracycline, and cotrimoxazole (Piddock, 2006; Tegos et al., 2002). This is possibly because the resistance was caused by systems other than AcrAB (e.g. mutations in gyrase or topoisomerase IV) (Tegos et al., 2002). However, many studies suggest that the combination of a mutation in the drug binding site gene with over expression of an efflux pump give rise to high level of resistance to 15 fluoroquinolones (Mazzariol et al., 2000, 2002; Oethinger et al., 1998). The over expression of efflux pumps may confer the bacterium to become resistant to antimicrobials. In one study, reserpine (Inhibitors of MDR Efflux Pumps) was shown to suppress the in vitro emergence of norfloxacin-resistant S. aureus (Markham and Neyfakh, 1996) and ciprofloxacin-resistant S. pneumoniae (Markham, 1999). The efflux pump inhibitor MC-207-110 also suppressed the emergence of levofloxacin-resistant P. aeruginosa (Lomovskaya et al., 2001). One study indicated that deletion of acrAB in E. coli with two mutations in gyrA resulted in increasing the susceptibility to fluoroquinolones. Inactivation of acrB in MDR S. enterica serovar Typhimurium DT204 (with a mutation[s] in topoisomerase genes) resulted in very low MICs (even below the break points). Similar results observed after deletion of cmeB in C. jejuni containing a substitution in GyrA (Luo et al., 2003). Furthermore, in vitro mutant S. enterica serovar Typhimurium lack tolC were unable to become resistant to ciprofloxacin (Ricci et al., 2006). Integrons Integrons ?are natural genetic engineering systems that scavenge and incorporate circularized open reading frames, called gene cassettes, and convert them to functional genes? (by ensuring their correct expression) (Rowe-Magnus and Mazel, 2002). In the integrons, the incorporation of multiple genes into gene cassette creates the problem of multidrug resistance. The integrons can contain up to 8 resistance gene cassettes and have been found in multiple- resistant clinical isolates. More than 80 different antibiotic resistance genes, covering most classes of antimicrobials presently in use, are identified 16 in class 1 integrons. These include genes conferring resistance to know beta-lactams, all aminoglycosides, chloromphenicol, trimethoprim, streptothricin, fosfomycin, rafampin, erythromycin, lincomycin, and antiseptic of quaternary ammonium compound family. The integron platform consists of an integrase gene and a primary recombination site, called the attI site proximal to the integrase gene (Mazel, 2006). The integrase catalyzes recombination between its cognate attI site and a secondary target called an attC site (or 59 base elements). The attC site is normally found associated with a single open reading frame (ORF).The combination structure of the attC-ORF is termed a gene cassette. Insertion of the gene cassette at the attI site, which is located downstream of the common promoter internal to the intI gene, obliges expression of the down stream gene cassettes (Fig 5) (Fluit and Schmitz, 2004). Figure 5. General structure of class 1 integron. P1, promoter for trancription of gene cassettes; P2, second promoter that is usually inactive; int, integrase gene; attI1, integration site; qacE, partially deleted gene that encodes quaternary ammonium compound resistance; sulI, sulphonamide resistance; orf5, unknown function; P, promoters of the qacED and sulI genes; attC, sequence on the gene cassette recognised by the integrase (Fluit and Schmitz, 2004). 17 Five classes of multi-resistant integrons (MRIs) share the same pool of resistance cassettes (Rowe-Magnus and Mazel, 2002). For example, the dfrA1 cassette has been identified in the Class 1, 2, 4 and 5 integrons of various species. Each of the five classes of mobile integrons was defined based on the homology of their integrase genes, which share between 45 - 58% amino acid identities. The integron platforms are found in a variety of genetic contexts. All 5 classes linked to mobile genetic elements; all of which serve as vehicle for inter and intra microbial species transmission of genetic material. The class 1 integron platform is the most common among multi-drug resistant bacterial populations and is found associated with and can be embedded in the Tn21 transposon family. Class 1 integrons have been reported in many Gram negative genera including Acinetobacter, Aeromonas, Alcaligenes, Burkholderia, Campylobacter, Citrobacter, Enterobacter, Escherichia, Klebsiella, Pseudomonas, Salmonella, Serratia, Shigella and Vibrio. It has also been identified in other bacteria such as Corynebacterium glutamicum and Mycobacterium fortuitum, and Enterococcus faecalis (Mazel, 2006). The other 3 classes of integrons are associated with different Tns. Class 2 integrons are embedded in the Tn7 family of transposons and consist of an integrase gene followed by gene cassettes. Class 2 integrons is not as common as class 1 integrons and found in Acinetobacter, Shigella and Salmonella. One observation is that the intI2 gene of Tn7 contains a stop codon and Tn7 cannot therefore produce an active integrase. Thus may explain the high frequent detection of class 1 integrons rather than class 2 among clinical isolates. Class 3 integrons have been described in Pseudomonas aeruginosa, Serratia marcescens, Alcaligenes xylosoxidans, Pseudomonas putida and Klebsiella pneumoniae 18 isolates from Japan. Class 4 is now named Vibrio cholerae SI and is a component of the SXT element (a conjugative self-transmissible integrating element) in this species (Mazel, 2006). The other (class 5) is located in compound transposon carried on a plasmid of Vibrio salmonicida (Mazel, 2006). It has been found that class 4 and 5 integrons harbor only a single antibiotic resistance cassette followed by several other cassettes with no antibiotic resistance function. Research Objectives The main aim of the present dissertation was to investigate the mechanisms by which muti-drug resistance (MDR) phenotypes emerge in association with enrofloxacin resistance in canine and feline clinical E. coli pathogens associated with spontaneous infections. The study was approached in four phases. Chapter II is a preliminary study to generate baseline data on the prevalence and molecular basis of antimicrobial resistance in the canine and feline clinical Escherichia coli isolates. Isolates were collected from dogs or cats with spontaneous infection between May and September 2005. Isolates were phenotypically characterized based on the susceptibility to 7 antimicrobial agents by E-test? according to CLSI guidelines and interpretive standards. In addition, isolates from each phenotype were randomly selected for genotypical characterization using pulsed-field gel electrophoresis (PFGE). Chapter III is to develop a real time- Fluorescence Resonance Energy Transfer (FRET)-PCR assay that is specifically designed to detect FQ resistant levels associated with the mutations in codon 83 and 87 of gyrA in ENR-resistant E. coli clinical isolates. 19 A muli-sequence alignment was initially performed on clinical E. coli isolates from which probes and primers were designed. Only one probe was designed that hybridizes perfectly gyrA sequence in wild type of clinical E. coli isolates. Melting curve analysis (MCA) was used to differentiate between the wild and the mutant-type template of gyrA. This was based on different melting peaks that emerge as a result of the mismatches between the probe and target sequences. Since this method is sensitive, rapid and specific, it can be used clinically in the future to detect these mutations and potential ENR-MDR phenptypes and can be used as an alternative for other conventional PCR techniques. Chapter IV is to confirm and characterize the association of enrofloxacin (ENR) resistance with MDR in canine and feline pathogens. The resistance was described in terms of incidence (susceptible, intermediate and resistant) and phenotypes. The relationship between ENR resistance and various phenotypes was investigated. Pathogens collected from infected dogs and cats throughout the US were obtained from IDEXX (commercial veterinary microbiologic laboratory). For each isolate, the minimal inhibitory concentrations (MIC) were determined for 14 drugs, representing 6 different antimicrobial classes. Drugs were selected based on their historical use in veterinary medicine and their different mechanisms of action and thus resistance. Methods were based on micro-dilution susceptibility testing (TREK ? Diagnostic Systems, Cleveland, OH); however, in contrast to standard methods, drug concentrations covered a range of over 500 fold differences, including concentrations well below and well above the breakpoint MIC as set by CLSI guidlines. The association between ENR resistant isolates 20 and the emergence of MDR resistant phenotypes was investigated using statistical methods of odds ratio and CI. Comparisons were made in proportion of co-resistance to other drug classes for isolates which do exhibit enrofloxacin resistance and the isolates that do exhibit resistance to all drugs but enrofloxacin. Furthermore, we investigated the contribution of resistance nodulation division (RND) efflux pump (AcrAB-TolC system in E. coli) to the emergence of ENR-MDR phenotypes. The role of the pump was confirmed and characterized by repeating susceptibility testing of all isolates, with the addition of the efflux pump inhibitor (EPI; Phe-Arg-?-naphthylamide) to the culture media (100 ?g/ml). Overexpression phenotype (EPO) was defined as a four fold or more decrease in the MIC for ENR when the inhibitor was present compared to its absence. The outcome measurements were included the statistical comparison of MIC statistics for each drug in the presence and absence of the EPI, and the magnitudes (how many fold decrease in MIC) of MIC decrease in the presence of EPI. Chapter VI is to explore the association of mobile genetic elements class 1 and 2 integrons in the emergence of MDR phenotypes and among different phenotypes. This part of the dissertation is to answer the question about role of mobile genetic elements into the dissemination of antimicrobial resistance genes. Integrons are widely distributed in nature and can play a critical role in the transmission of resistance genes among species. Conventional PCR was used to amplify variable regions (i.e. 5? and 3? conserved segment) of class 1 and 2 integrons. In addition, direct amplification of integrase 1 and 2, RFLP, and sequencing were also utilized to confirm the general structure of class 1 and 2 integrons and to identify their gene cassettes. Outcome measurements include the 21 proportion of class 1 and 2 integrons among single drug resistance (SDR), MDR, and susceptible phenotypes (NR). Furthermore, molecular typing of the gene cassette carried by class 1 and 2 integrons was determined by RFLP. The impact of the gene cassette on antimicrobial resistance, which was carried by class 1 and 2 integrons, was further determined using curing experiments. This study is designed to help us understand the potential role of different factors into the emergence of MDR phenotype and can be more helpful in prevention and treatment of infectious disease. 22 References Aires, J.R., Nikaido, H., 2005. Aminoglycosides are captured from both periplasm and cytoplasm by the AcrD multidrug efflux transporter of Escherichia coli. J. Bacteriol. 187, 1923-1929. Bachoual, R., Tankovic, J., Soussy, C.J., 1998. Analysis of the mutations involved in fluoroquinolone resistance of in vivo and in vitro mutants of Escherichia coli. Microb. Drug Resist. 4, 271?276. Bagel, S., H?llen, V., Wiedemann, B., Heisig, P., 1999. Impact of gyrA and parC mutations on quinolone resistance, doubling time, and supercoiling degree of Escherichia coli. Antimicrob. Agents Chemother. 43, 868?875. Boothe D.M., Smaha T.J., Hathcock, T.L., Bixler B.M., 2005. Escherichia coli resistance in canine urinary tract infections. In Congress Processings of ACVIM Forum, Baltimore, MD, June 4 ? 7. p. 249. Boothe, D.M., Boeckh, A., Simpson, R. B., Dubose, K., 2006. Comparison of pharmacodynamic and pharmacokinetic indices of efficacy for 5 fluoroquinolones toward pathogens of dogs and cats. J. Vet. Intern. Med. 20, 1297-1306. Cohn, L.A., Gary, A.T., Fales, W. H., Madsen, R.W., 2003. Trend in fluoroquinolone resistance of bacteria isolated from canine urinary tracts. J. Vet. Diagn. Investig. 15, 338-343. 23 Cook, C.L., Singer, R.S., Jang, S.S., Hirch, D.C., 2002. Enrofloxacin resistance in Escherichia coli isolated from dogs with urinary tract infections. J. Am. Vet. Med. Assoc. 220: 190-192. Drlica, K., Zhao, X., 1997. DNA gyrase, topoisomerase IV, and the 4-quinolones, Microbiol. Mol. Biol. Rev. 61, 377?392. Everett, M.J., Jin, Y.F, Ricci, V., Piddock, L.J., 1996. Contributions of individual mechanisms to fluoroquinolone resistance in 36 Escherichia coli strains isolated from humans and animals. Antimicrob. Agents Chemother. 40, 2380?2386. Fendukly, F., Karlsson, I., Hanson, H.S., Kronvall, G., Dornbusch, K., 2003. Patterns of mutations in target genes in septicemia isolates of Escherichia coli and Klebsiella pneumoniae with resistance or reduced susceptibility to ciprofloxacin. APMIS. 111, 857?866. Ferrero, L., Cameron, B., Manse, B., Lagneaux, D., Crouzet, J., Famechon, A., Blanche, F., 1994. Cloning and primary structure of Staphylococcus aureus DNA topoisomerase IV: a primary target of fluoroquinolones. Mol. Microbiol. 13, 641? 653. Fluit, A.C., Schmitz, F.J., 2004. Resistance integrons and super-integrons. Clin. Microbiol. Infect. 10, 272-288. Folster, J.P., Shafer, W.M., 2005. Regulation of mtrF expression in Neisseria gonorrhoeae and its role in high-level antimicrobial resistance. J. Bacteriol. 187, 3713-3720. 24 Guardabassi, L., Schwarz, S., Lloyd, D.H., 2004. Pet animals as reservoirs of antimicrobial-resistant bacteria. J. Antimicrob. Chemother. 54, 321-332. Heisig, P., 1996. Genetic evidence for a role of parC mutations in development of high- level fluoroquinolone resistance in Escherichia coli. Antimicrob. Agents Chemother. 40, 879?885. Hooper, D.C., 1999. Mechanisms of fluoroquinolone resistance. Drug Resistance Updates. 2, 38?55. Hoshino, K., Kitamura, A., Morrissey, I., Sato, K., Kato, J., Ikeda, H., 1994. Comparison of inhibition of Escherichia coli topoisomerase IV by quinolones with DNA gyrase inhibition. Antimicrob. Agents Chemother. 38, 2623?2627. Karlowsky, J.A., Hoban, D.J., DeCorby, M.R., Laing, N.M., Zhanel, G.G., 2006. Fluoroquinolone-resistant urinary isolates of Escherichia coli from outpatients are frequently multidrug resistant: results from the North American Urinary Tract Infection Collaborative Alliance-Quinolone Resistance study. Antimicrob. Agents Chemother. 50, 2251?2254. Karlowsky, J.A., Kelly, L.J., Thornsberry, C., Jones, M.E., Sahm, D.F., 2002. Trends in antimicrobial resistance among urinary tract infection isolates of Escherichia coli from female outpatients in the United States. Antimicrob. Agents Chemother. 46, 2540?2545. Karlowsky, J.A., Thornsberry, C., Jones, M.E., Sahm, D.F., 2003. Susceptibility of antimicrobial-resistant urinary Escherichia coli isolates to fluoroquinolones and nitrofurantoin, Clin. Infect. Dis. 36, 183?187. 25 Kato, J., Suzuki, H., Ikeda, H., 1992. Purification and characterization of DNA topoisomerase IV in Escherichia coli. J. Biol. Chem. 267, 25676?25684. Khodursky, A.B., Zechiedrich, E.L., Cozzarelli, N.R., 1995. Topoisomerase IV is a target of quinolones in Escherichia coli. Proc. Natl. Acad. Sci. USA. 92, 11801?11805. Komp Lindgren, P., Karlsson, A., 2003. Hughes D., Mutation rate and evolution of fluoroquinolone resistance in Escherichia coli isolates from patients with urinary tract infections. Antimicrob. Agents Chemother. 47, 3222?3232. Koronakis, V., Eswaran, J., Hughes, C., 2004. Structure and function of TolC: the bacterial exit duct for proteins and drugs. Annu. Rev. Biochem. 73, 467-489. Kumagai, Y., Kato, J.I., Hoshino, K., Akasaka, T., Sato, K., Ikeda, H., 1996. Quinolone- resistant mutants of Escherichia coli DNA topoisomerase IV parC gene, Antimicrob. Agents Chemother. 40, 710?714. Lehn, N., Stower-Hoffmann, J., Kott, T., Strassner, C., Wagner, H., Kronke, M., Schneider-Brachert, W., 1996. Characterization of clinical isolates of Escherichia coli showing high levels of fluoroquinolone resistance. J. Clin. Microbiol. 34, 597?602. Lindler, L.E., Fan, W., Jahan, N., 2001. Detection of ciprofloxacin-resistant Yersinia pestis by fluorogenic PCR using the LightCycler. J. Clin. Microbiol. 39, 3649- 3655. Lomovskaya, O., Warren, M., Lee, A., Galazzo, J., Fronko, R., Lee, M., Blais, J., Cho, D., Chamberland, S., Renau, T., Leger, R., Hecker, S., Watkins, W., Hoshino, K., Ishida, H., Lee, V. J., 2001. Identification and characterization of inhibitors of 26 multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. Agents Chemother. 45, 105-116. Luo, N., Sahin, O., Lin, J., Michel, L. O., Zhang, Q., 2003. In vivo selection of Campylobacter isolates with high levels of fluoroquinolone resistance associated with gyrA mutations and the function of the CmeABC efflux pump. Antimicrob. Agents Chemother. 47, 390-394. Markham, P.N., 1999. Inhibition of the emergence of ciprofloxacin resistance in Streptococcus pneumoniae by the multidrug efflux inhibitor reserpine. Antimicrob. Agents Chemother. 43, 988-989. Markham, P.N., Neyfakh, A., 1996. Inhibition of the multi-drug transporter norA prevents emergence of norfloxacin resistance in Staphylococcus aureus. Antimicrob. Agents Chemother. 40, 2673-2674. Mazel, D., 2006. Integrons: agents of bacterial evolution. Nat. Rev. Microbiol.4, 608- 620. Mazzariol, A., Tokue, Y., Kanegawa, T.M., Cornaglia, G., Nikaido, H., 2000. High-level fluoroquinolone-resistant clinical isolates of Escherichia coli overproduce multidrug efflux protein AcrA. Antimicrob. Agents Chemother. 44, 3441?3443 Mazzariol, A., Zuliani, J., Cornaglia, G., Rossolini, G. M., Fontana, R., 2002. AcrAB efflux system: expression and contribution to fluoroquinolone resistance in Klebsiella spp. Antimicrob. Agents Chemother. 46, 3984-3986. McDonald, L.C., Chen, F.J., Lo, H.J., Yin, H.C., Lu, P.L., Huang, C.H., Chen, P., Lauderdale, T.L., Ho, M., 2001. Emergence of reduced susceptibility and 27 resistance to fluoroquinolones in Escherichia coli in Taiwan and contributions of distinct selective pressures, Antimicrob. Agents Chemother. 45, 3084?3091. Muder, R.R., Brennen, C., Goetz, A.M., Wagener, M.M., Rihs, J.D., 1991. Association with prior fluoroquinolone therapy of widespread ciprofloxacin resistance among Gram-negative isolates in a Veterans Affairs medical center. Antimicrob. Agents Chemother. 35, 256?258. Nikaido, H., 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol. Mol. Biol. Rev. 67, 593-656. Oethinger, M., Podglajen, I., Kern, W.V., Levy, S. B., 1998. Overexpression of the marA or soxS regulatory gene in clinical topoisomerase mutants of Escherichia coli. Antimicrob. Agents Chemother. 42, 2089-2094. Olliver, A., Valle, M., Chaslus-Dancla, E., Cloeckaert, A., 2004. Role of an acrR mutation in multidrug resistance of in vitro selected fluoroquinolone-resistant mutants of Salmonella enterica serovar Typhimurium. FEMS Microbiol. Lett. 238, 267-272. Ozeki, S., Deguchi, T., Yasuda, M., Nakano, M., Kawamura, T., Nishino, Y., Kawada, Y., 1997. Development of a rapid assay for detecting gyrA mutations in Escherichia coli and determination of incidence of gyrA mutations in clinical strains isolated from patients with complicated urinary tract infections. J. Clin. Microbiol. 35, 2315?2319. Paulsen, I.T., 2003. Multidrug efflux pumps and resistance: regulation and evolution. Curr. Opin. Microbiol. 6, 446-451. 28 Peng, H., Marians, K.J., 1993. Escherichia coli topoisomerase IV. Purification, characterization, subunit structure, and subunit interactions, J. Biol. Chem. 268, 24481?24490. Piddock, L.J., 2006. Clinically relevant chromosomally encoded multidrug resistance efflux pumps in bacteria. Clin. Microbiol. Rev. 19, 382-402. Reece, R.J., Maxwell, A., 1991. DNA gyrase: structure and function, Crit. Rev. Biochem. Mol. Biol. 26, 335?375. Ricci, V., Tzakas, P., Buckley, A., Coldham, N., Piddock, L.J.V., 2006. Ciprofloxacin- resistant Salmonella enterica serovar Typhimurium strains are difficult to select in the absence of AcrB and TolC. Antimicrob. Agents. Chemother. 50, 38-42. Richard P., Delangle M.H., Merrien D., Barill? S., Reynaud A., Minozzi C., Richet H., 1994. Fluoroquinolone use and fluoroquinolone resistance: is there an association?. Clin. Infect. Dis. 19, 54?59. Rowe-Magnus, D.A., Mazel, D., 2002. The role of integrons in antibiotic resistance gene capture. Int J. Med. Microbiol. 292, 115-125. S?enz, Y., Zarazaga, M., Bri?as, L., Ruiz-Larrea, F., Torres, C., 2003. Mutations in gyrA and parC genes in nalidixic acid-resistant Escherichia coli strains from food products, humans and animals. J. Antimicrob. Chemother. 51, 1001?1005. Sanchez, S.M., McCrackin Stevenson, A., Hudson, C.R., Maier, M., Buffington, T., Dam, O., Maurer, J.J., 2002. Characterization of multidrug-resistant Escherichia coli isolates associated with nosocomial infections in dogs. J. Clin. Microbiol. 40, 3586?3595. 29 Shaheen, B.W., Boothe, D.M., Oyarzabal, O.A., Smaha, T., 2008a. Characterization of clinical Escherichia coli isolates expressing multidrug resistance recovered from canine and feline with spontaneous disease. In Congress Processings of ACVIM Forum, San Antonio, TX, June, 4 -7. p. 786. Shaheen, B.W., Boothe, D.M., Smaha, T., 2008b. The Phenotypic Expression of Multidrug Resistant Escherichia coli Recovered from Spontaneous Diseased Dogs and Cats. In: Proceedings of the 2008 Annual Conference on Antimicrobial Resistance, Bethesda, MD, June 23 ? 25. p.19. Swanberg, S.L., Wang, J. C., 1987. Cloning and sequencing of the Escherichia coli gyrA gene coding for the A subunit of DNA gyrase. J. Mol. Biol. 197, 729-736 Tegos, G., Stermitz, F. R. Lomovskaya, O., Lewis, K., 2002. Multidrug pump inhibitors uncover remarkable activity of plant antimicrobials. Antimicrob. Agents Chemother. 46, 3133-3141. Truong, Q.C., Ouabdesselam, S., Hooper, D.C., Moreau, N.J., Soussy, C.J., 1995. Sequential mutations of gyrA in Escherichia coli associated with quinolone therapy. J. Antimicrob. Chemother. 36, 1055?1059. Webber, M., Piddock, L.J., 2001. Quinolone resistance in Escherichia coli. Vet. Res. 32, 275?284. Weigel, L.M., Steward, C.D., Tenover, F.C., 1998. gyrA mutations associated with fluoroquinolone resistance in eight species of Enterobacteriaceae. Antimicrob. Agents Chemother. 42, 2661?2667. 30 White, D.G., Piddock, L.J., Maurer, J.J., Zhao, S., Ricci, V., Thayer, S.G., 2000. Characterization of fluoroquinolone resistance among veterinary isolates of avian Escherichia coli. Antimicrob. Agents Chemother. 44, 2897?2899. Willmott, C.J., Maxwell, A., 1993. A single point mutation in the DNA gyrase A protein greatly reduces binding of fluoroquinolones to the gyrase-DNA complex. Antimicrob. Agents Chemother. 37, 126?127. Wolter, D.J., Smith-Moland, E., Goering, R.V., Hanson, N.D., Lister, P.D., 2004. Multidrug resistance associated with mexXY expression in clinical isolates of Pseudomonas aeruginosa from a Texas hospital. Diagn. Microbiol. Infect. Dis. 50, 43-50. Yoshida, H., Bogaki, M., Nakamura, M., Nakamura, S., 1990. Quinolone resistance- determining region in the DNA gyrase gyrA gene of Escherichia coli. Antimicrob. Agents Chemother. 34, 1271?1272. Yoshida, H., Bogaki, M., Nakamura, M., Yamanaka, L.M., Nakamura, S., 1991. Quinolone resistance-determining region in the DNA gyrase gyrB gene of Escherichia coli. Antimicrob. Agents Chemother. 35, 1647?1650. 31 CHAPTER II PHENOTYPIC AND GENOTYPIC CHARACTERIZATION OF CLINICAL ESCHERICHIA COLI ISOLATES EXPRESSING MULTIDRUG RESISTANCE RECOVERED FROM CANINE AND FELINE WITH SPONTANEOUS DISEASE Abstract Background: Antimicrobial resistance is emerging among canine and feline clinical E. coli isolates. Objectives: To generate baseline data on the prevalence and molecular basis of antimicrobial resistance in the canine and feline clinical Escherichia coli isolates. Animals: E. coli pathogens isolated from dogs and cats. Methods: A total of 377 E. coli isolates were collected from dogs or cats with spontaneous infection between May and September 2005. Isolates were phenotypically characterized based on the susceptibility to seven antimicrobial agents by E-test? according to CLSI guidelines. Selected isolates from each phenotype were genotyped using pulsed-field gel electrophoresis (PFGE). Results: Of the 377 isolates, 43 phenotypes were generated. Almost 44% (n= 85/194) of the resistance isolates expressed single drug resistance, with resistance to beta-lactams predominating (83%; n=71/85). Only two isolates expressed single drug resistance to 32 enrofloxacin. Whereas 56% (n= 109/194) of the resistant isolates expressed multidrug resistance, with twenty isolates were resistant to all drugs. The drugs most commonly involved in multidrug resistance were amoxicillin (96.3 %; 105/109), amoxicillin- clavulanic acid (85%; 93/109) and enrofloxacin (61.5%; 67/109). Among all four regions, multidrug resistance tended to be greater in the South region (67.9%; 57/84) and least in the West, (46.6 %, 14/30). Dendrograms for 34 E. coli isolate from the state of Alabama revealed a total of 26 different PFGE profiles (based on ? 90% similarity), representing 12 different phenotypes. Conclusions: These results indicated that phenotypes and genotypes are not related, and that pathogenic E. coli have an extensive genetic diversity. The South had a greater incidence of multidrug resistance E. coli. Introduction The emergence of antibiotic resistance has been documented in E. coli isolates from human, animal and environmental sources (von Baum and Marre, 1995). The MDR E. coli is an emerging condition among clinical samples from small animals (Normand et al., 2000). Several studies have demonstrated an increased in multi-drug resistance E. coli as a cause of infection in dogs and cats both in the United State and Europe (Cohn et al., 2003; Guardabassi et al., 2004; Normand et al., 2000a, 2000b). Previous work by the authors indicated a high incidence of MDR in E. coli isolates associated with first time urinary tract infection (UTI) from canine patients admitted to Auburn University?s Veterinary Teaching Hospital (n=175). Overall E. coli resistance to all 17 antimicrobials 33 was 31% and approximated 40% to 50% toward ?first choice? empirical antimicrobials (amoxicillin, amoxicillin-clavulanic acid, cephalexin and sulfonamide-trimethoprim) and toward enrofloxacin. Those factors associated with a greater percent resistance (p? 0.0005) were gender (males), previous antimicrobial therapy (within 30 days), immune suppressive drugs (within 30 days), and days of hospitalization (? 5days) a . Nosocomial infection associated with multi-resistant E. coli has been reported in dogs in intensive care units, and extensive use of broad-spectrum antimicrobials is a contributing factor for multi-drug resistance (Sanchez et al., 2002). In addition to the impact on animal health, emergent pathogenic multi-resistant E. coli may have significant public health consequences if isolates are transmitted to humans from their pets (Beutin, 1999; Johnson et al., 2001). For instance, methicillin-resistant Staphylococcus aureus (MRSA) isolates recovered from pets closely resembled MRSA human isolates which are widespread in German hospitals (Strommenger et al., 2006). Thus, a possibility exists of cross-transmission of MRSA from pets to humans (Strommenger et al., 2006). A similar concern is now emerging with E. coli clinical isolates. Phylogenetic and pathotypic similarities between E. coli isolates from urinary tract infections in dogs and extraintestinal pathogenic E. coli (ExPEC) infections in humans have been described (Johnson et al., 2001; Starcic, et al., 2002). Over 15% of canine feces were found to contain E. coli strains closely related to human virulent ExPEC clones (Johnson et al., 2001). Molecular characterization of resistant isolates utilizing fingerprinting and hybridization techniques, such as pulsed-field gel electrophoresis (PFGE), polymerase 34 chain reaction (PCR) and genetic sequencing, are important tools to identify the spread of clonal unites. Among these methods, genetic fingerprinting of strains using PFGE is particularly useful for demonstrating close relationships among strains, including those manifesting MDR. Understanding the patterns of antimicrobial resistance in canine and feline E. coli isolates is important not only from a veterinary medicine prospective, but also from a global public health prospective. The multi-drug resistant bacteria are detected with an increasing frequency in populations exposed to antimicrobial gents. Therefore, it will become increasingly important for a veterinarian to understand the impact of antimicrobial agents and the emergence of antimicrobial resistance patterns. In this study, we extend the evaluation and monitoring of the emerging resistances and molecular bases of antimicrobial resistance among E. coli isolates in the U.S. In the present study, canine and feline isolates of E. coli with spontaneous infections were collected from different regions in the US and were characterized phenotypically using the susceptibility testing and genotypically using PFGE standard protocol. Materials and methods Sample collection and handling Isolates (n = 377) of E. coli were collected from dogs or cats with spontaneous infection between May and September 2005. The isolates were collected from the different tissue sites. Those isolates were previously isolated and cultured from samples submitted by veterinary practitioners to commercial or veterinary teaching hospital 35 microbiology laboratories serving veterinary practices throughout the US. Cultured isolates were then submitted by nine microbiology (commercial and academic) laboratories in the US to the Clinical Pharmacology Laboratory at Auburn University. Laboratories were divided, based on proximity, into four geographical areas: South, comprising the states of Alabama (AU), Mississippi (MSU), and North Carolina (NCSU); West, with the states of Washington (WSU), and California (IDC), Midwest, including Indiana (CL), Kansas (KSU), and Wisconsin (MAR), and Northeast, with the state of Massachusetts (IDM). Collaborating laboratories submitted isolates on trypticase agar slants (shipped overnight at room temperature). Upon receipt, isolates were streaked on trypticase agar and incubated at 37 o C for 24 h. For standard susceptibility testing, samples were adjusted to 0.5 McFarland standard turbidity (~10 8 CFU) and a sterile swab was used to inoculate two 150 mm Mueller Hinton plates (MH). Susceptibility testing Susceptibility to 7 drugs, representing 5 classes of antimicrobials, was determined using the E-test? b according to Clinical and Laboratory Standards Institute (CLSI; previously the National Committee on Clinical Laboratory Standards) guidelines and interpretive standards (CLSI, 2008). Drugs tested were three beta lactams, amoxicillin (A), amoxicillin-clavulanic acid (X), and cefpodoxime (P); a tetracycline, doxycycline (D); a fluoroquinolone, enrofloxacin (E); an aminoglycoside, gentamicin (G); and a ?potentiated? sulfonamide, trimethoprim-sulfamethoxazole (T ). The drugs were selected based on different mechanisms of action and their historical use on veterinary medicine. 36 MH plates were incubated at 35 o C for 20-24 h. The MIC results were recorded from the strips. Each isolate was then designated as resistant (R) or susceptible (S) to each drug. The CLSI intermediate designation (I) was treated as R using CLSI guidelines and interpretive standards (Livermore et al., 2002). For control purposes, E. coli ATCC 25922 c (range MIC = 0.016-0.125?/mL) and S. pneumoniae ATCC 49619 c (range MIC = 0.5-2.0?/mL) were used. Phenotypic characterization Phenotypes were generated for each isolate and categorized as no resistance (NR), resistant to a single drug or drug class (SDR) (A, X and P considered a drug class) or resistant to two or more drug class (MDR). The distribution of each phenotype expressed for each region or each tissue source was determined. Statistical analysis The associations between the various categories, including the geographical and tissue factors, and the emergence of MDR versus SDR phenotypes, were assessed using the chi square test (X 2 ) using MINITAB ? 15 package. d P values of < 0.05 were reported as statistically significant. Genotypic characterization For genotyping, at least two isolates of each phenotype were selected from each region for PFGE (for a total 90 isolates). PGFE was performed according to protocols 37 developed by the Centers for Disease Control and Prevention (Atlanta, GA) (Barrett et al., 1994). Briefly, the genomic DNA of each isolate was embedded inagarose plugs and the plugs were transferred to 50-ml tubes e containing 5 ml of cell lysis buffer (50 mM Tris, 50 mM EDTA [pH 8.0], 1% sarcosine, 0.1 mg of proteinase K/ml). Samples were lysed for 120 min at 54?C in an orbital shaker water bath with constant agitation (150 to 200 rpm). A total of six washes (twice with sterile ultrapure water and four times with 0.01 M Tris-EDTA buffer, pH 8.0) were used to remove excess reagents and cell debris from the lysed samples. Samples were restricted with XbaI f (40 to 50 U of enzyme at 37?C for 2 h). The plugs were loaded onto a 1% agarose-Tris buffer gel g , and PFGE was performed with a CHEF Mapper XA apparatus h . DNA was electrophoresed for 18 h at a constant voltage of 200 V (6 V/cm), with a pulse time of 5.0 to 40.0 s, an electric field angle of 120 o , and a temperature of 14?C, before being stained with ethidium bromide. Computer analysis of PFGE patterns Gel images were captured on a Gel Doc imaging system i . The images were converted to a TIFF file using the automated GeneSnap j , and analyzed with BioNumerics software k . The TIFF images were normalized using three standards per gel (Salmonella serotype Braenderup H9812), which were loaded on lanes 1, 7, and 15 (Hunter et al., 2005). Matching and dendrogram (UPGMA) analysis of the PFGE patterns was performed using the Dice correlation coefficient. 38 Results Sample descriptions and antimicrobial susceptibility There were 308 samples collected from dogs and 69 collected from cats. The majority of isolates were submitted by the South region (n=133), while the West region submitted 79 isolates (Table 1). Isolates from the South region had the highest percentage of resistance, while the West region had the least percent of resistance compared with other regions. The ratio of the different phenotypes varied among regions (Table 1). However in the South region, the phenotype Z (resistant to all seven antimicrobials) had the highest prevalence among phenotypes (28%, n=16/57). Resistant isolates were most recovered from urine samples in all regions. For example, urine sample represented 61.4% (n= 51/80) and 54.7% (n=29/53) in the South and Midwest regions respectively (Table 2). Among tissue sites, the soft tissue had the highest percent of resistant isolates (73.3 %, n=11/15) while the ear tissue had the least percent of resistance (29%, 8/27) compared with other tissues (Fig 1). Similarly, all phenotypes were distributed in different proportions among different tissues with no particular pattern observed. Phenotypic characterization The result of antimicrobial susceptibility of E. coli (n = 377) to 7 antimicrobials used in veterinary medicine was evaluated (Fig 2A). Forty-four different phenotypes (including no resistance [NR] and resistance to all tested drugs [Z]) occurred among the 377 E. coli isolates. Of the resistant isolates, 85 were grouped within eight SDR phenotypes, whereas 109 isolates grouped within 35 MDR phenotypes. Among the SDR 39 phenotypes, beta lactams were the most commonly represented class of drugs to which resistance was encountered (83.5%, n=71/85) and this included amoxicillin (30.5%, n=26/85) and amoxicillin-clauvalanic acid (28.2%, n=24/85). Other strains were resistant to combinations of amoxicillin, amoxicillin-clauvalanic acid and cefpodoxime (23.5%, n=20/85). Furthermore, less than 20% (n=14/85) of the isolates were resistant to gentamicin, cefpodoxime, doxycycline, enrofloxacin, or trimethoprim-sulfamethoxazole (Fig 2A). The most common MDR phenotypes was Z and the least common was the phenotype resistant to only 2 drug classes (Fig 2B). The percent resistant to 5, 6 or all 7 drugs were 16.6% (n=16), 21.1% (n=23) and 18.3% (n=20), respectively (Fig 2B). The remaining MDR were represented by 34 different phenotypes. Amoxicillin was the drug that was mostly associated with MDR phenotypes 96.3 %; 105/109), followed by amoxicillin-clauvalanic acid (85%; 93/109), and enrofloxacin (61.5%; 67/109). Effects of regions and tissue sites on microbial susceptibility Among the four different regions, the South region had the highest percentage of MDR phenotype (67.9%, n=57/84), while the West region had the highest percentage of susceptible phenotype (62%, n=49/79) (Fig 3). No relationship was identified (p = 0.066) between regions and the emergence of SDR versus MDR phenotypes. The tissue exhibiting the highest proportion of MDR was skin (75%, n=12/16). In contrast, ear tissue showed the high proportion of susceptible phenotypes (70%, n=19/27) (Fig 4). However, no evidence exists for the association (p = 0.664) between tissues and the 40 presence of SDR and MDR phenotypes. The removal of tissues with low isolate numbers from the analysis did not influence the final results. PFGE and genotypic characterizations A total of 90 E. coli isolated from dogs and cats with different phenotypes yielded a total of 82 different PFGE patterns. Samples with similarity coefficients of 90?100% were considered identical. In the South, the largest proportion of resistant isolates were submitted by Auburn University (n=32) compared with Mississippi (n=26) and North Carolina (n=27). Among the selected isolates (n=34; 12 phenotypes), a total of 26 different PFGE patterns were identified (Fig 5). Thus phenotypes were not genotypically related as indicated by the different PFGE profiles of the SDR phenotypes: (AX) and (AXP). One large profile of six isolates and three different phenotypes were genetically related. In addition, three pairs with identical profiles were identified. Within each pair or cluster only isolates with distinct phenotype were found. For the tissue samples, the majority of clinical E .coli isolates were from urine. Among the 46 uropathogens, a total 43 different PFGE patterns were identified (based on 90% similarity level). This represents a total of 13 different phenotypes. Only two profiles from Auburn University, one of 3 isolates and the other with 2 isolates, were related, each phenotypically distinct (Fig 6). Discussion Our current study was confined with restricted time frame (summer 2005), limited sample size, retrospective design, and potential antimicrobials selection bias (limited to 41 seven antimicrobial agents). However, we studied a broad geographic collection of isolates and a different representation of tissue samples. Furthermore, we utilized E test as method of antimicrobial susceptibility testing because at it can provide a wide range of MIC lower and higher than MIC break point. The emergence of antimicrobial resistance among pathogens in human or in animal medicine draws attention towards the increase in the risk of therapeutic failure (Costa et al., 2007). In the present study, 56.2% (n=109/194) of E. coli were phenotypically resistant to multiple antimicrobials (Fig 2B). The majority of E. coli isolates were distinct based on the analysis of PFGE fingerprint patterns. Our results indicated that there is no correlation between the phenotype and the genotypes within the same region (Fig 5), or within the same tissue (i.e. urine sample) (Fig 6). This observation is in agreement with studies showing that the antibiotic resistance patterns differed significantly within PFGE strain types among uropathogenic E. coli isolated from dogs (Cooke et al., 2002; Drazenovich et al., 2004; Mentula et al., 2006). The possibility of lateral horizontal transfer of resistance mobile genetic elements, such as plasmids, integrons or phage-mediated exchange, could play a role in the difference of PFGE patterns in our study as previously reported (Cooke et al., 2002; Drazenovich et al., 2004; Mentula et al., 2006). The different profiles on PFGE of E. coli can be also explained by the large genomic diversity among E. coli. This diversity has also been demonstrated by early studies using multilocus enzyme electrophoresis (Milkman, 1973, 1997), and we know now that multiple isolates considered identical clones and groups of isolates that have similar clonal complexes are frequently recovered from diverse 42 geographical and temporal origins (Feil and Spratt, 2001). In the case of medically important but numerically small number of pathogenic lineages of E. coli, such as enteropathogenic or enterohaemorrhagic strains, recombination appears to play a limited role on the genetic diversity. But the contribution of recombination to the genetic variability of E. coli isolates that produce animal disease has not been addressed (Feil and Spratt, 2001). Many antimicrobials used to treat infections in dogs and cats tend to increase the prevalence of antimicrobial resistance (von Baum and Marre, 1995). The present study revealed a high percentage of resistance to the beta lactams class (i.e. amoxicillin, amoxicillin-clauvalanic acid), a class that is also associated with MDR phenotypes (Fig 2A, 2B). Similarly, the trend of increasing resistance among canine and feline E. coli isolates in the UK during the years 1989?1997 was toward amoxicillin- clavulanic acid and streptomycin, with many isolates showing multi-drug resistance phenotypes (Normand et al., 2000a). The same observations were found among nonenteric E. coli isolates from dogs in the US between 1990 and 1998 , in which isolates exhibited resistance to amoxicillin, carbenicillin, and cephalothin (Oluoch et al., 2001). Furthermore, Feria et al. (2002) characterized the pattern and mechanism of ?-lactams resistance among 72 uropathogenic E. coli isolated from dogs, and illustrated that 36% of the isolates exhibited resistance to amoxicillin and 19% of isolates were resistance to amoxicillin-clavulanic acid. The mechanism of hyperproduction of the plasmid encoded TEM-1 ?-lactamase or chromosomally encoded AmpC ?-lactamase could contribute to the emergence SDR phenotype as previously described (Bradford et al., 1999; Feria et al., 2002). On the other 43 hand, the third generation cephalosporins (i.e. cefpodoxime) have a broad spectrum activity and were approved for treating acute and chronic skin infections in dogs (Cherni et al., 2006). Only one out of 194 E. coli isolates from dogs in this study was resistant to cefpodoxime. This observation is consistent with the observations by Feria et al. (2002) who found that resistance was more often to cephalothin (25%) than to either ceftazidime or ceftriaxone (1.4%). Our observations indicated that fluoroquinolone-resistant isolates (i.e. enrofloxacin) were more likely to be multi-drug resistant. Only 2 out 85 isolates expressed SDR phenotype, while 69 out 71 isolates were MDR. Enrofloxacin-MDR isolates showing resistance towards a ?-lactam antimicrobial agent have the highest prevalence of the association compared to other drug classes (Fig 7). Specifically, enrofloxacin isolates which are MDR phenotype have the highest prevalence of resistance toward A (97%), followed by AX (94%). The enrofloxacin-MDR phenotype was present in about half of the isolates with resistance to T (63%), P (60%), G (53%), and D (53%). This observation is in agreement with previous reported studies which also suggest that the selection of fluoroquinolone-resistant mutants of E. coli is more frequent. Cooke et al. (2002) indicated that the enrofloxacin resistant isolates were also MDR to at least three other antimicrobials that are commonly used to treat UTI. Similarly, studies from human medicine indicated the association of fluoroquinolone with multidrug- resistant phenotypes commonly with the clinical cases of uropathogenic E. coli (Karlowsky et al., 2002, 2003, 2006). The appearance of multidrug-resistant phenotypes maybe correlated with the higher use of fluoroquinolones in veterinary medicine as 44 reported in human medicine (Karlowsky et al., 2006), along with inadequate infection hygiene, and sanitation control measures (Johnson et al., 2006). The present study showed an increased emergence of multidrug-resistant isolates which is consistent with previously published studies. For example, Sanchez et al. (2002) characterized 34 of multidrug-Resistant E. coli isolates associated with nosocomial infections in dogs. The isolates were resistant to most cephalosporins, beta-lactams, and the beta-lactamase inhibitor clavulanic acid as well as resistance to tetracycline, spectinomycin, sulfonamides, chloramphenicol, and gentamicin. In our study, the MDR phenotypes were associated with the South region (Fig 3). In contrast, the West region had the highest percent of susceptible phenotype with lowest percent of MDR phenotype. These results probably highlight the different prescribing practices by the veterinarians among different geographic regions. In other words, increasing prescription of the broad spectrum antimicrobials (i.e. fluoroquinolones) at veterinary hospitals and clinics may be contributing to this observation, as previously reported (Sanchez et al., 2002). There is a possibility that more antimicrobial agents were prescribed based on the veterinarian?s first choice and neither was based on antimicrobial susceptibility testing or considered the previous antimicrobial case history. Focusing on the previous antimicrobial drug used and their resistance mechanisms of the pathogen will be determined for our next studies. The dissemination of MDR phynotypes among E. coli isolates from skin tissue may occur as a consequence of horizontal gene transfer (Osborn and B?ltner, 2002). Recent evidence indicates that aerobic Gram-positive corynebacteria can act as reservoir of class 1 integrons in poultry litter (Nandi et al., 2004). This indicates that the 45 environment could play critical role in genes cassette exchange between different mobile integrons. With increase sequence information, many resistance genes have been found in many bacterial species, and homologous genes share high similarities among non-related bacteria (Aarestrup, 2006). The situation becomes very important in the exposed environment, such as the skin, in which the spread of resistance gene can be facilitated among heterogeneous bacterial communities through some mobile elements (Dionisio et al., 2002). In conclusion, the antimicrobial-susceptible and multidrug-resistant phenotypes identified in this study were genetically not related by PFGE within the same tissue or among the same geographic location collected over the same period. Furthermore, this study indicated large genomic diversity among clinical E. coli isolated from dogs and cats from different regions and/or tissues. MDR phenotypes were associated with South regions and enrofloxacin resistance was associated with MDR phenotypes. Therefore, a large surveillance study in which more E. coli isolates from different regions with previous antimicrobial history will be conducted in our laboratory. This will help a veterinarian to understand the impact of previous antimicrobials use on the emergence of E. coli MDR phenotypes. Additionally, class 1 integrons were established to be more common in clinical E. coli isolates from animals and further work is going on in our laboratory to characterize these integrons. 46 References Aarestrup, F.M., 2006. The origin, evolution, and local and global dissemination of antimicrobial resistance. In: Aarestrup, F.M. (Eds.), Antimicrobial resistance in bacteria of animal origin, ASM Press, Washington DC, pp. 339-360. Barrett, T.J., Lior, H., Green, J.H., Khakhria, R., Wells, J.G., Bell, B.P., Greene, K.D., Lewis, J., Griffin, P.M.., 1994. Laboratory investigation of a multistate food- borne outbreak of Escherichia coli O157:H7 by using pulsed-field gel electrophoresis and phage typing. J. Clin. Microbiol. 32, 3013?3017. Beutin, L., 1999. Escherichia coli as a pathogen in dogs and cats. Vet. Res. 30, 285 298. Bradford, P.A., Petersen, P.J., Fingerman, I.M., White, D.G., 1999. Characterization of expanded-spectrum cephalosporin resistance in E. coli isolates associated with bovine calf diarrhoeal disease. J. Antimicrob. Chemother. 44, 607?610. Cherni, J.A., Boucher, J.F., Skogerboe, T.L., Tarnacki, S., Gajewski, K.D., Lindeman, C.J., 2006. Comparison of the efficacy of cefpodoxime proxetil and cephalexin in treating bacterial pyoderma in dogs. Intern. J. Appl. Res. Vet. Med. 4, 85?93. Clinical and Laboratory Standards Institute, 2008. Performance standards for antimicrobial susceptibility testing; 18th informational supplement. CLSI document M100-S18 Wayne, PA. Cohn, L.A., Gary, A.T., Fales, W.H., 2003. Trends in fluoroquinolone resistance of bacteria isolated from canine urinary tracts. J. Vet. Diagn. Invest. 15, 338?343. Cooke, C.L., Singer, R.S., Jang, S.S., Hirsh, D.C., 2002. Enrofloxacin resistance in Escherichia coli isolated from dogs with urinary tract infections. J. Am. Vet. 47 Med. Assoc. 220, 190?192. Costa, D., Poeta, P., S?enz, Y., Coelho, A.C., Matos, M., Vinu?, L., Rodrigues, J., Torres, C., 2007. Prevalence of antimicrobial resistance and resistance genes in faecal Escherichia coli isolates recovered from healthy pets. Vet. Microbiol.127,97?105. Dionisio, F., Matic, I., Radman, M., Rodrigues, O.R., Taddei, F., 2002. Plasmids spread very fast in heterogeneous bacterial communities. Genetics. 162, 1525?1532. Drazenovich, N., Ling, G.V., Foley, J., 2004. Molecular investigation of Escherichia coli strains associated with aarently persistent urinary tract infection in dogs. J. Vet. Intern. Med. 18, 301?306. Feil, E.J., Spratt, B.G., 2001. Recombination and the population structures of bacterial pathogens. Annu. Rev. Microbiol. 55, 561?590. Feria, C.E., Ferreira, J.D., Correia, J., Gon?alves, J., Cani?a, M., 2002. Patterns and mechanisms of resistance to ?-lactams and ?-lactemase inhibitors in uropathogenic Escherichia coli isolated from dogs in Portugal. J. Antimicrob. Chemother. 49, 77?85. Guardabassi, L., Schewarz, S., Lloyd, D.H., 2004. Pet animals as reservoirs of antimicrobial-resistant bacteria. J. Antimicrob. Chemother. 54, 321?332. Hunter, S.B., Vauterin, P., Lambert-Fair, M.A., Van Duyne, M.S., Kubota, K., Graves, L., Wrigley, D., Barrett, T., Ribot, E., 2005. Establishment of a universal size standard strain for use with the PulseNet standardized pulsed-field gel electrophoresis protocols: converting the national databases to the new size standard. J. Clin. Microbiol. 43, 1045?1050. 48 Johnson, J.R., Delavari, P., Stell, A.L., Whittam, T. S., Carlino, U., Russo, T.A., 2001. Molecular comparison of extraintestinal Escherichia coli isolates of the same electrophoretic lineages from humans and domestic animals. J. Infect. Dis. 183, 154 159. Johnson, J.R., Kuskowski, M.A., Menard, M., Gajewski, A., Xercavins, M., Garau, J., 2006. Similarity between human and chicken Escherichia coli isolates in relation to ciprofloxacin resistance status. J. Infect. Dis. 194, 8?10. Johnson, J.R., Stell, A.L., Delavari, P., 2001. Canine feces as a reservoir of extraintestinal pathogenic Escherichia coli. Infect. Immun.69, 1306?1314. Johnson, J.R., Stell, A.L., Delavari, P., Murray, A.C., Kuskowski, M., Gaastra, W., 2001. Phylogenetic and pathotypic similarities between Escherichia coli isolates from urinary tract infections in dogs and extraintestinal infections in humans. J. Infect. Dis. 183, 897?906. Karlowsky, J.A., Hoban, D.J., DeCorby, M.R., Laing, N.M., Zhanel, G.G., 2006. Fluoroquinolone-resistant urinary isolates of Escherichia coli from outpatients are frequently multidrug resistant: results from the North American Urinary Tract Infection Collaborative Alliance-Quinolone Resistance study. Antimicrob. Agents Chemother. 50, 2251?2254. Karlowsky, J.A., Thornsberry, C., Jones, M.E., Sahm, D.F., 2003. Susceptibility of antimicrobial-resistant urinary Escherichia coli isolates to fluoroquinolones and nitrofurantoin. Clin. Infect. Dis. 36, 183?187. Karlowsky, L.J., Kelly, C., Thornsberry, M.E., Jones, M.E., Sahm, D.F., 2002. Trends in 49 antimicrobial resistance among urinary tract infection isolates of Escherichia coli from female outpatients in the United States. Antimicrob. Agents. Chemother. 46, 2540?2545. Livermore, D.M., James, D., Reacher, M., Graham, C., Nichols, T., Stephens, P., Johnson, A.P., George, R.C., 2002. Trends in fluoroquinolone (ciprofloxacin) resistance in Enterobacteriaceae from bacteremias, England and Wales, 1990? 1999. Emerg. Infect. Dis. 8, 473?478. Mentula, S., Virtanen, T., Kanervo-Nordstr?m, A., Harmoinen, J., Westermarck, E., Rautio M, Huovinen, P., K?n?nen, E., 2006. Relatedness of Escherichia coli strains with different susceptibility patterns isolated from beagle dogs during ampicillin treatment. Int. J. Antimicrob. Agents. 27, 46?50. Milkman, R., 1973. Electrophoretic variation in Escherichia coli from natural sources. Science. 182, 1024?1026. Milkman, R., 1997. Recombination and population structure in Escherichia coli. Genetics. 146, 745?750. Nandi, S., Maurer, J.J., Hofacre, C., Summers, A.O., 2004. Gram-positive bacteria are a major reservoir of Class 1 antibiotic resistance integrons in poultry litter. Proc. Natl. Acad. Sci. USA. 18, 7118?7122. Normand, E.H., Gibson, N.R., Reid, S.W., Carmichael, S., Taylor, D.J., 2000a. Antimicrobial-resistance trends in bacterial isolates from companion-animal community practice in the UK. Prev. Vet. Med. 46, 267?278. Normand, E.H., Gibson, R.N., Taylor, D.J., Carmichael, S., Reid, S.W., 2000b. Trends of 50 antimicrobial resistance in bacterial isolates from a small animal referral hospital. Vet. Rec.146, 151?155. Oluoch, A.O., Kim, C.H., Weisiger, R.M., Koo, H.Y., Siegel, A.M., Campbell, K.L., Burke, T.J., McKiernan, B.C., Kakoma, I., 2001. Nonenteric Escherichia coli isolates from dogs: 674 cases (1990-1998). J. Am. Vet. Med. Assoc.218,381?384. Osborn, A.M., B?ltner, D., 2002. When phage, plasmids, and transposons collide: genomic islands, and conjugative- and mobilizable-transposons as a mosaic continuum. Plasmid. 48, 202?212. Sanchez, S., McCrackin Stevenson, M.A., Hudson, C.R., Maier, M., Buffington, T., Dam, Q., Maurer, J.J., 2002. Characterization of multidrug-resistant Escherichia coli isolates associated with nosocomial infections in dogs. J. Clin. Microbiol.40, 3586?3595. Starcic, M., Johnson, J.R., Stell, A.L., van der Goot, J., Hendriks, H.G., van Vorstenbosch, C., van Dijk, L., Gaastra, W., 2002. Haemolytic Escherichia coli isolated from dogs with diarrhea have characteristics of both uropathogenic and necrotoxigenic strains. Vet. Microbiol. 85, 361?377. Strommenger, B., Kehrenberg, C., Kettlitz, C., Cuny, C., Verspohl, J., Witte, W., Schwarz, S., 2006. Molecular characterization of methicillin-resistant Staphylococcus aureus strains from pet animals and their relationship to human isolates. J. Antimicrob. Chemother. 57, 461?465. von Baum, H., Marre, R., 1995. Antimicrobial resistance of Escherichia coli and therapeutic implications. Int. J. Med. Microbiol.295, 503?514. 51 Regions a No. of resistant phenotype(s) b Antimicrobial phenotype(s) c Midwest (108) 53 (49 %) A (10); AD, AXDEGT, AXET, AXP, AXPD, AXPDET, E (2); ADG, AT, AXD, AXDE, AXDET, AXE, AXEGT, AXPDEG, AXPDT, AXPEGT, AXPET, AXPGT, AXT, DE, DEG, P, T, Z (1); AX (6); D (5) Northeast (57) 27 (47.3 %) A, AX (4); ADEGT, AXDGT, AXDT, AXE, AXPDEG, AXT, E (1); AXDEGT, AXPD (2); AXP (5); Z (3) South (133) 83 (62.4 %) A, AXPE (5); AD, ADET, ADGT, AP, AXD, AXDET, AXDGT, AXDT, AXET, AXPDGT, AXPDT, AXPEG, AXPEGT, AXPT, AXT, T (1); AT, AXE, AXDEG, AXEGT, AXEGT, AXPD, AXPDEG, D (2); AX, AXP (9); AXPDET (8); Z (16) West (79) 30 (38 %) A (6); AD, AT, AXPDE, AXPDEG, AXPEGT, AXT, D, DT, EGT (1); AX (5); AXD, AXPD (3); AXP (4) 52 Table 1. Distribution of resistant phenotypes and their geographical locations. A = amoxicillin; X = amoxicillin-clavulanic acid; P = cefpodoxime; D = doxycycline; E = enrofloxacin; G = gentamicin; T = trimethoprim-sulfamethoxazole. a Number in parentheses is the total number of isolates collected from that region. b Number in parentheses is the percent of resistant isolates collected from that region. c Number in parentheses is the total number of phenotype(s) collected from that region. 53 29 7 51 3 3 3 3 1 4 3 7 2 4 3 1 2 4 5 4 2 2 1 2 4 5 4 1 1 3 2 20 2 1 1 3 0% 20% 40% 60% 80% 100% Midwest (53) Northeast (27) South (83) West (30) Urine Feces Skin Ear Lung Nose Abdomen Genital organs Liver NA Mouth Bone Blood Lymph node Figure 1. Resistant isolates of E. coli from four US regions and their clinical sources. NA = the source was not determined. a Number in parentheses is the total number of isolates collected from that region. b Number in parentheses is the percent of resistant isolates collected from that region. c Number in parentheses is the total number of resistant isolates collected from the tissue site. 54 Type of infection a No. of phenotype(s) b Antimicrobial phenotype(s) c Skin (27) 16 (59.2 %) A, AX, AXE, AD, AXPD, AXDGT, AXPDET, ADET, AXPEGT (1); Z (3); AXDEGT, AXP (2) Ear (27) 8 (29.6 %) A, AX, (2); AXPDEG, AXT, AXE, Z (1) Lung (22) 15 (68.2 %) A, P, AX, AXDET, AXPT, AXPE, AXPD (1); Z, AXP (3); AXT (2) Soft Tissue (15) 11 (73.3 %) A, AXPDET, T, AXPT, AXD, AXDT, AXP, AXPDET, AX (1); Z (2) Urine (174) 91 (52.3 %) A (12); AXDET, AT, AXPEGT, AXPDT, AXET, AXDET, AXPDEG (2); AXPDET (7); AXP (8); AXPE, AD, E, AXPD, AXEGT (3); AXDT, T, AXDGT, AXT, ADGT, AP, DE, DEG, AXPDE, AXD, AXE (1); D (6); AX, Z (9) Table 2. Distribution of resistant phenotypes with their clinical sources. A = amoxicillin; X = amoxicillin-clavulanic acid; P = cefpodoxime; D = doxycycline; E = enrofloxacin; G = gentamicin; T = trimethoprim-sulfamethoxazole. a Number in parentheses is the total number of isolates collected from that tissue. b Number in parentheses is the percent of resistant isolates collected from that tissue. c Number in parentheses is the total number of R phenotype(s) collected from that tissue. 55 26 1 24 20 8 3 1 2 0 5 10 15 20 25 30 35 A AP AX AXP D E P T Percent Resistant Phenotype (%) A 20 10 9 5 5 3 3 4 4 4 2 2 2 2 4 2 5 4 4 3 0 5 10 15 20 AD AT AXD AXDEG AXDEGT AXDET AXDGT AXDT AXE AXEGT AXPDET AXET AXPDT AXPD A XPDEG AXT AXPE AXPEGT AXT Z Percent Resistant Phenotype (%) B Figure 2. The number and proportion (%) of each SDR (A) and MDR (B) phenotype(s) expressed in 377 isolates of E. coli causing spontaneous disease in dogs and cats throughout the US. Numbers of isolates are indicated above each bar. SDR = single drug resistance; MDR = multiple drug resistance; N = susceptible phenotype. Z = resistance to all 7 drugs. A = amoxicillin; X = amoxicillin-clavulanic acid; P = cefpodoxime; D = doxycycline; E = enrofloxacin; G = gentamicin; T = trimethoprim-sulfamethoxazole. 56 16 27 1427 1456 1326 49 50 3054 0% 20% 40% 60% 80% 100% Midw est (53) Northeast (27) South (83) West (30) Regions P e r cen t of I s o l a t es SDR MRR N Figure 3. The overall proportion of SDR, MDR, and NR phenotypes among E. coli clinical isolates by geographical region. For SDR and MDR, the proportion is of total resistant isolates (n=194); for NR, the proportion is out of all 377 isolates. Numbers of isolates for each phenotype are indicated within each bar. SDR = single drug resistance; MDR = multiple drug resistance; N = susceptible phenotype. 57 40464 4 51 7 9 412 83 4 7 19 11 0% 20% 40% 60% 80% 100% Skin (16) Ear (8) Lung (15) Soft T.(11) Urine (91) Tissues P e r cen t o f I s ol at es SDR MRR N Figure 4. The overall proportion of SDR, MDR, and NR phenotypes among E. coli clinical isolates by tissue source. For SDR and MDR, the proportion is of total resistant isolates (n=194); for NR, the proportion is out of the total isolates in that tissue. Numbers of isolates for each phenotype are indicated within each bar. SDR = single drug resistance; MDR = multiple drug resistance; N = susceptible phenotype. 58 59 Figure 5. Dendrogram of SDR, MDR, and N phenotypes among E. coli clinical isolates collected at the same institution (South). N = susceptible phenotype; Z = resistance to all 7 drugs. A = amoxicillin; X = amoxicillin-clavulanic acid; P = cefpodoxime; D = doxycycline; E = enrofloxacin; G = gentamicin; T = trimethoprim-sulfamethoxazole. a Isolates? identification. b The phenotype for each isolate. c Isolates? tissue source. 60 61 Figure 6. The dendrogram for E. coli phenotypes collected from one tissue (urine). N = susceptible phenotype; Z = resistance to all 7 drugs. A = amoxicillin; X = amoxicillin- clavulanic acid; P = cefpodoxime; D = doxycycline; E = enrofloxacin; G = gentamicin; T = trimethoprim-sulfamethoxazole. a The phenotype for each isolate b Isolates? identification. c The regional source of isolates. 62 48 4040 46 65 67 0 10 20 30 40 50 60 70 80 90 100 AAXP G D T Antimicrobial Drugs ( % ) o f r esi st a n ce t o o t h e r a n t i m i cr o b i al dr u g s MDR-FQ (n=69) Figure 7. Frequency of the MDR phenotype in enrofloxacin isolates showing resistance to non-fluroquinolone antimicrobial agents. Numbers of isolates are indicated above each bar. A = amoxicillin; X = amoxicillin-clavulanic acid; P = cefpodoxime; D = doxycycline; E = enrofloxacin; G = gentamicin; T = trimethoprim-sulfamethoxazole. 63 Footnotes a Boothe DM, Smaha T, Hathcock T, et al. Escherichia coli Resistance in the Canine Urinary Tract. Am Coll Vet Intern Med Ann Forum, Baltimore, MD, June 2005 (abstract). b Epsilon; AB Biodisk, Solna, Sweden. c American Tissue Cell Culture, Manassas, VA. d Minitab 15, statistical software, State College, PA, USA. e polypropylene tubes; 30 by 115 mm; Becton Dickinson, Franklin Lakes, NJ. f XbaI, restriction enzyme; Roche Applied Science, Indianapolis, IN. g SeaKem Gold Agarose; BioWhittaker Molecular Applications, Rockland, ME. h CHEF Mapper XA; Bio-Rad Laboratories, Hercules, CA. i Gel Doc imaging system; Syngene, Frederick, MA. j GeneSnap; Syngene, Frederick, MA. k BioNumerics software; Applied Maths, Sint-Martens-Latem, Belgium. 64 CHAPTER III DETECTION OF FLUOROQUINOLONE RESISTANCE LEVEL IN CLINICAL CANINE AND FELINE ESCHERICHIA COLI PATHOGENS USING RAPID REAL-TIME PCR ASSAY Abstract Fluoroquinolones are used to treat infections caused by E. coli in canine and feline veterinary patients, particularly those infecting the urinary tract. The gyrA gene is a primary target causing fluoroquinolone resistance in Gram negative coliforms, with mutations in codon 83 and 87 generally associated with high level of resistance E. coli clinical isolates. We have developed a fluorescence resonance energy transfer (FRET) quantitative PCR to identify enrofloxacin-resistance in clinical E. coli isolates that carry mutations in codon 83 and 87 of gyrA. This real-time quantitative PCR assay is rapid, economical, and sensitive compared with cultured antimicrobial susceptibility testing. The assay identified as few as four genome copies per reaction from culture and 19 genome copies in urine. For the 70 isolates tested, the sensitivity was 87.5% (95% CI = 75% to 95.3%) (n=42/48) whereas specificity was 100% (95% CI = 87.3% to 100%) (n=22/22). Furthermore, we were able to accurately differentiate between the wild type and mutants E. coli directly from infected canine urine samples (n=5) within 1-2 hrs. These results were confirmed by sequence alignments of the PCR products and 65 comparison with the susceptibility testing. The FRET-PCR assay appears to have promising clinical application as an early diagnostic tool for rapid and sensitivedetection and differentiation of the level of fluoroquinolone resistance among clinical E. coliisolates. Introduction Escherichia coli is the major cause of urinary tract infections, pyometra and other infections in dogs and cats (Hagman and K?hn, 2002; Chen et al., 2003). Fluoroquinolones historically have been used in veterinary medicine to treat infections caused by E. coli. In recent years, a substantive increase in quinolone resistance in animal E. coli isolates has been reported (Cooke et al., 2002; Cohn et al., 2003; Guardabassi et al., 2004; Boothe et al., 2006). Included are studies that report an increase in the incidence of antimicrobial-resistance, including multi-drug resistant E. coli, particularly from isolates associated with urinary tract infections in canine patients (Hirsh, 1973; Guardabassi et al., 2004; Shaheen et al., 2008a, 2008b). The most common mechanism of resistance to quinolones in E. coli are mutations in genes that encode subunits of the quinolone target topoisomerases, and specifically, DNA gyrase (topoisomerase II, in gyrA and gyrB genes) and topoisomerase IV (in parC and parE) (Everett et al., 1996; Vila et al., 1996; Piddock, 1999). These mutations are primarily located in the quinolone resistance-determining region (QRDR) of the gyrA gene and its homologous region of the parC gene (Oram and Fisher, 1991; Piddock, 1999). Mutations in gyrB and parE genes are less significant and rarely contribute to quinolone resistance (Giraud et al., 2001; Vila et al., 1996). Other contributing 66 mechanisms of resistance include an efflux pump or modifications of porins which preclude effective intracellular drug concentrations, thus decreasing susceptibility to quinolones (Everett et al., 1996; Giraud et al., 2001). These latter mechanisms generally confer multidrug resistance (Piddock, 1999; Mazzariol et al., 2000). Previously it has been reported that the predominant mechanism of high level fluoroquinolone resistance in E. coli appears to be a C to T transition in codon 83 and G to A, A to G, or G to T transitions in codon 87 in the QRDR of gyrA. The codon 83 mutation results in a Ser-83 to Leu-83 substitution whereas the codon 87 mutations results in Asp-87 to Asn-87, Asp-87 to Gly-87 or Asp-87 to Tyr-87 substitution in the functional protein (Oram and Fisher, 1991). Each of these mutations precludes binding of the drug to the target site, rendering the isolate resistant to the drug (Hooper, 1999). Clinical and Laboratory Standards Institute (CLSI) guidelines for routine fluoroquinolone susceptibility testing include agar gel or dilution or micro-dilution methods (CLSI, 2008). For the clinician, delays of 3 to 5 days are often realized between sample acquisition and susceptibility reporting. This delay between sample collection and receipt of results is often untenable, leading the clinician to empirically treat the infection. For urinary tract infections, the infecting organisms is often (correctly) assumed to be E. coli (Bubenik et al., 2007). Drugs presumed (often incorrectly) to be effective include fluoroquinolones. In the face of increasing E. coli resistance to fluoroquinolones, the result is inappropriate therapy (Sanchez et al., 2002). Thus, there is a need for a rapid method of detecting infections associated with fluroquinolone resistant E coli. A rapid PCR-based FRET method has been developed that allows detection of 67 ciprofloxacin resistant Yersinia pestis isolates that carry a mutation in codon 81 or codon 83 of gyrA (Lindler et al., 2001) and for detection of quinolone-resistant Neisseria gonorrhoeae in urine samples (Siedner et al., 2007). The basic principle of the technique is the use of two standard DNA primers. For FRET PCR, two fluorescently labeled hybridization probes bind to target DNA between the flanking primers. Hybridization probe 1 is labeled with fluorescein (the donor dye) at the 3?-end and hybridization probe 2 is labeled with LightCycler-Red 460 (LC Red 640; acceptor dye) at the 5? end. The probes hybridize to the amplified DNA fragment in a head to tail arrangement, thereby bringing the two fluorescent dyes into close proximity. The energy emitted as a result of the proximity excites the acceptor dye attached to the second hybridization probe, which then emits fluorescent light at a different wavelength. Because the stability of hybridization varies with the phenotype (i.e. wild versus mutation, and type of mutation), melting temperatures consequently vary. Discriminating among the different phenotypes that carry mutation(s) is indicated by melting curve analysis (MCA). Most studies that focus on the mechanisms of quinolone resistance in E. coli have been conducted in human clinical strains. Those few reports involving animal strains generally focus on those of public health significance, that is, animals intended for food consumption (Everett et al., 1996; Giraud et al., 2001). Among these reports, most involve conventional PCR techniques, followed by sequence analysis of the amplicon. However, for our purposes, a rapid, sensitive, and specific test is needed for early identification of pathogens causing infections in clinical patients (Ko and Grant, 2006; Abdelbaqi et al., 2007; Siedner et al., 2007); if the technique can also detect the mutation level (that is, level of resistance) of the clinical E. coli isolates, the clinician will be 68 further empowered in the design of dosing regimens that support judicious antimicrobial use (Boothe, 2006). The purpose of this report is to report the detection of canine and feline clinical E. coli isolates that carry mutations in gyrA were targeted utilizing a rapid PCR-based FRET method that detects mutations in codon 83 and codon 87 in the QRDR of gyrA. Materials and methods Bacterial isolates and culture conditions E. coli isolates used for FRET PCR assay development are listed in Table 1. E. coli strains ATCC 11775 (Weigel et al., 1998) and E. coli K12 (Swanberg and Wang; 1987) were used for multi-sequence alignment analysis for gyrA sequence. The other 70 susceptible (n= 22), intermediate (n= 5) and resistant E. coli isolates (n= 43) studied were acquired from clinical veterinary microbiology laboratories between May and September of 2008. In each case, the isolate had previously been cultured from a sample collected from a dog or cat with spontaneous disease by a veterinarian and submitted to the laboratory for identification and susceptibility testing. The veterinary clinical microbiology laboratory identified each isolate as E. coli prior to submission to the Clinical Pharmacology Laboratory (Auburn University, Auburn, AL, USA). Susceptibility testing was repeated in the Clinical Pharmacology Laboratory using a custom microtube dilution plate. Upon receipt in our laboratory, isolates were plated on CHROMagar? Orientation (Becton, Dickinson, Franklin Lakes, NJ) for rapid confirmation and E. coli. isolates were streaked on trypticase agar (Becton Dickinson, Franklin Lakes, NJ) and then incubated at 35?C for 18-24 h. For standard susceptibility testing, cells were harvested and suspended in 4.5 ml of 0.9% normal saline and were 69 adjusted to 0.5 McFarland standard turbidity (~10 8 CFU) using SENSITITER ? Nephelometer (TREK Diagnostic Systems, Cleveland, OH) before testing. Susceptibility testing Susceptibility to fluoroquinolone was determined for enrofloxacin (ENR) using a custom-made microdilution susceptibility testing (TREK Diagnostic Systems, Cleveland, OH) according to CLSI guidelines and interpretive standards (CLSI, 2008). Microdilution plates were incubated at 35 o C for 18-24 h. The results were recorded as true value MIC using the SENSITITER ? VIZION system (TREK Diagnostic Systems, Cleveland, OH). For quality control purposes E. coli ATCC ? 25922 (American Tissue Cell Culture, Manassas, VA; [range MIC = 0.008-0.03?/mL]) and Enterococcus fecalis ATCC ? 29212 (American Tissue Cell Culture, Manassas, VA; [range MIC = 0.12-1.0?/mL]) were used. Each isolate was then designated as resistant (R), intermediate (I) and susceptible (S) to enrofloxacin using CLSI guidelines and interpretive standards (CLSI, 2008). The enrofloxacin break point (BP) was defined according to CLSI guidelines as any isolates exhibition MIC = 4 ?g/mL, while MIC = 0.5 ?g/mL was considered to be the enrofloxacin susceptible BP (CLSI, 2008). DNA extraction and designing of the real-time PCR Bacterial DNA for PCR reaction were prepared by boiling a bacterial culture in 200 ?l of PreMan? Ultra Sample Preparation Reagent (Applied Biosystem, Foster City, CA) for 10 min, followed by centrifugation. Initially primers were designed to amplify the quinolone resistance-determining region (QRDR) (Yoshida et al., 1990) of gyrA. 70 The gyrase A sequences of E. coli was obtained from the GenBank (GenBank accession number: X06373) (Swanberg and Wang, 1987). Primers and probes were designed on the conserved region using Vector NTI ? 10.1 software (Invitrogen, Carlsbad, CA), such that the mutant E. coli strain has two one-bp mutations in the region for the bodipy 630/650 probe as shown in Fig 1. The length of amplification product was 310 bp with the two primers and probes designed specific for gyrA corresponding to bases 152 to 427 of the gyrA gene that included all point mutations within the QRDR (Fig 1). The carboxyfluorescein (6-FAM) probe, which was the unpurified 3? labeled FRET energy donor, was excited by 488 nm light. The bodipy 630/650 probe, which was HPLC- purified 5?-labeled, and 3?-phosphorylated, was used as an acceptor probe (von Ahsen et al., 2000). Fluorescence emitted from bodipy 630/650 probes served for detection of gyr A at 640 nm. We designed only one single probe based on the sequence of the wild type so that the bodipy 630/650 probe bridged the two potential point mutations in codons 83 and 87 of gyrA. This enabled discrimination of the wild-type and mutated isolates based on the MCA. Quantitative real-time FRET-PCR Quantitative determination of E. coli was performed by real-time FRET-PCR in a LightCycler (Roche Diagnostics), and reaction chemistry was performed as described (Wang et al., 2004). Reactions were performed in glass capillaries with 5 ?l of sample nucleic acids or standards and 15 ?l of 1.33 x master mixture. Each 20 ?l reaction contained 2.0 U Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA), and 0.2 U heat-labile uracil-N-glycosylase (Roche Molecular Biochemicals, Indianapolis, IN). 71 Thermal cycling consisted of 18 high-stringency step-down cycles followed by 25 fluorescence acquisition cycles. Eighteen step-down thermal cycles included the following steps: 6 cycles at 95?C for 15 s, 72?C for 30 s; 9 cycles at 95?C for 15 s, 70?C for 30 s; 3 cycles at 95?C for 15 s, 68?C for 30 s, 72?C for 30 s. The fluorescence acquisition cycling consisted of following steps: denaturation at 95?C 15 s, annealing at 56?C for 8 s, 66?C for 30 s, and extension at 72?C 30 s. PCR products were purified for automated DNA sequencing with a QIAquick PCR Purification Kit (Qiagen, Valencia, CA). Nucleotide sequence determination was performed using an ABI automatic DNA sequencer (Model 377; Perkin-Elmer) at the Genomic Sequencing Laboratory, (Auburn University, Auburn, AL, USA) by using the described forward and antisense primers in Figure 1. DNA sequences were analyzed by using Vector NTI ? 10.1 software (Invitrogen, Carlsbad, CA). Quantitative standards were produced by PCR amplification with dTTP, 4 % MetaPhor agarose gel-purification, and quantification of the fragment by Pico-Green assay (Invitrogen, Carlsbad, CA). Standards were verified by DNA sequencing, and used at 10 4 , 10 3 , 10 2 , 10, and 0 copies per 5 ?l in T 10 E 0.1 . The melting curve for the annealing of the PCR product with the FRET probes was determined by monitoring the fluorescence of channel F2/F1 from 55 to 80?C with a temperature transition rate of 0.1?C per second. Specificity of the PCRs was confirmed by a lack of amplification of extracted DNAs from Salmonella, Bartonella, and Yesinia. The size of amplification products was verified by electrophoresis of 1 ?l of the PCR product on 2% agarose gels. 72 PCR discrimination of fluroquinolone resistance in canine urine specimens We examined the sensitivity of our PCR assay in the detection of E. coli in urine samples in 3 resistant isolates selected among the 70 clinical samples known to have mutations in codon 83 and 87. In this experiment, we performed the assay on dilutions of original urine samples such that the bacterial counts/ml for each sample was, in colony forming units (CFU): 10 6 , 10 5 , 10 4 , 10 3 , 10 2 , 10, and 0 CFU/ml. Furthermore, we performed the experiment in two clinical urine samples that had the original bacterial counts of >10 5 CFU/ml and were fluoroquinolone-susceptible. For real-time PCR analysis of the gyrA gene, DNA was extracted directly from urine samples without previous culturing and purification. The DNA from urine specimens (140 ?l) was extracted by using QIAamp Viral RNA Mini Kit (Qiagen, Valencia, CA). The extracted DNA (5 ?l) was used for real-time PCR assay. The FRET PCR amplification was performed as described above. Results Genetic analysis of the gyrA QRDR The DNA sequences of the gyrA gene fragments for the wild type and gyrA mutatnt clinical isolates; including the QRDR (nucleotides 199 to 319 of the corresponding E. coli sequence) and excluding the primer sequences are illustrated in Figure 1. The gyrA QRDR sequences of the clinical E. coli isolates (IDM 15, AU 26, MS 1, MS 10) were aligned and compared to the E. coli type strain (ATCC 11775; GenBank accession number: AF052254) reported by Weigel et al. (1998) and with the E. coli K12 gyrA sequence reported by Swanberg and Wang (1987) (GenBank accession number: 73 X06373). Four nucleotide differences were identified at the following positions: 255, C to T; 267, T to C; 273, and C to T; and 300, T to C; all of these substitutions were nonsense. Mutations were found in isolates MS 1 and MS 10 at the following positions: 248, C to T; 259, G to A of the corresponding codons 83 and 87, respectively (Fig 1). The amino acid sequences of the QRDRs for the six E. coli isolates were highly conserved. The QRDR amino acid sequences of E. coli (ATCC 11775, IDM 15, AU 26) were identical to the E. coli QRDR (E. coli K12). Sequence analysis of E. coli (MS 1 and MS 10) QRDRs revealed that two conserved substitutions were detected: Ser-83 to Leu- 83, and Asp-87 to Asn-87. Furthermore these amino acid substitutions were associated with a high level of resistance to enrofloxacin (MIC >32 ?g/ml). Real-time FRET-PCR for detection of mutations in gyrA sequences In this experiment we were able to differentiate E. coli gyrA mutants (ENR r ) from wild-type susceptible isolates. The quantitative interpretation and the melting curve analysis are illustrated in Fig 2. These results were compared with susceptibility profile results for those isolates. PCR amplification progressed efficiently with different starting DNA concentrations. This can be observed by parallel slopes of the linear phase of the amplification curve corresponding to different starting template numbers (Fig. 2A). Our initial results indicated that the amplicons of the resistant-type gyrA isolates formed a less-stable hybrid with bodipy 630/650 probe than did the susceptible ones. The ENR r isolates MS-1 and MS-10 exhibited T m of 61?C, as compared to the susceptible isolates IDM-15 and AU-15 which exhibited a T m of 71?C (Fig. 2B). In this experiment the ?T m (defined as the T m of the wild type ? the T m of the mutant) was 10?C. Agarose gel 74 electrophoresis of the amplified products confirmed the presence of the 310-bp amplicon for the four positive clinical isolates. Furthermore, the presence of E. coli gyrA in these four clinical samples was also confirmed by sequencing the PCR products and comparing these to published wild-type sequences (Fig 2C). The results of MCA were consistent with the sequencing results which revealed the presence of two transitions C/T and G/A mismatch of the corresponding codons 83 and 87, respectively. These results are in agreement with those obtained with the FRET-PCR assay which can directly detect the single nucleotide polymorphisms (SNP) in E. coli gyrA from the clinical samples. Robustness of the assay in discrimination of mutations in the gyrA gene in 66 clinical E. coli samples After the DNA and amino acid sequences of the QRDRs from the quinolone- susceptible type isolates were determined, the PCR assay was applied to clinical samples of 66 E. coli isolates obtained from canine and feline clinical specimens. Clinical E. coli isolates were selected to represent a wide range of ENR MICs (Table 1). An excellent agreement was obtained between the MCA, the MIC results, and gene sequencing performed on the majority of the isolates (Fig 2 and 3). Most of the resistant isolates had T m that were easily differentiated from those generated with ENR susceptible isolates (Fig 2B). Based on different melting temperatures, the results could be categorized into 3 major groups. In group I the T m was 61?C, while in group III the T m was consistently 71?C (Fig 2B). Those isolates in group I expressed double mutations at codons 83 and 87 and exhibited high MIC range (8- >128 ?g/ml) (Fig 3). In contrast, the isolates in group III had no mutation, and were associated with low MIC range (0.06- 0.5 ?g/ml). The 75 majority of the isolates were grouped into group I or III. Interestingly, a new emerging group II, which consisted of 9 isolates, had an intermediate T m that it was easily differentiated from those generated with resistant or susceptible-type templates of group I and group III, respectively. In group II, T m decreased and formed less stable hybrid compared to group III. We performed multi-sequence alignment analysis on QRDR of gyrA and included the codons alterations in ENR r resistant and -susceptible clinical isolates (Fig 2C). Six isolates in group II exhibited intermediate T m of 68?C (Fig 3). This result had a match with the single mutation that involved C to T transition in codon 83 of gyrA (Fig 2C) and was associated with decreased susceptibility or low levels of resistance to ENR r (MIC range= 0.25-4 ?g/ml; ENR BP = 4 ?g /ml [CLSI, 2008]) (Fig 3). This substitution changed the amino acid sequence from Ser-83 to Leu-83. One isolate has been identified to have T m of 69?C (plotted as T m of 68?C in Fig 2B). This isolate approached the break point of susceptibility to enrofloxacin (MIC = 2 ?g/ml) (Fig 3) (CLSI, 2008) and a substitution that changed the amino acid sequence from Asp-87 to 87-to-Gly (Fig 2C). Furthermore, 2 isolates displayed T m of 66.5?C (Fig 2B, 3). For those isolates, multi sequence analysis revealed one nucleotide mismatch at codon 85 that involved T to C (Fig 2C). All of these substitutions were nonsense. This corresponded well with the susceptibility results (MIC = 0.06 ?g/ml) for those isolates (Fig 3). The real-time PCR assay amplified products in 70 of 70 samples (100%). The assay precisely detected intermediate or high level resistance among 70 different clinical E. coli isolates in 42 of 48 for a sensitivity (correctly predicting true positives, that is either intermediate or high level of resistance) of 87.5% (95% confidence interval [CI] = 75% to 95.3%) corresponding resistant cultures and correctly identified 22 of 22 for a 76 specificity (correctly identifying true negatives or susceptibe) of 100% (95% CI = 87.3% to 100%) susceptible clinical cultures. The assay discriminated between clinical levels of resistance (i.e. intermediate or low level of resistance, reflecting a single mutation, versus high level of fluoroquinolone resistance, reflecting double mutations in gyrA) (Fig 3). The specificity of the PCR primers for E. coli was tested using other bacterial species such as Salmonella, Bartonella, and Yesinia, and the FRET signals were not observed in those species. These results confirm the excellent specificity of the primers, discriminatory power of the hybridization probe assay and the utility of this assay to be used clinically. Sensitivity of PCR amplification in detection of flouroquinlone resistance in 5 clinical urine samples After extraction of DNA, PCR amplifications were performed successfully in randomly selected 5 (susceptible, n=2; mutants, n=3) clinical urine specimens. MCA revealed that the Tm in the wild and mutant type of E. coli found in the urine samples produced a similar pattern of results as those in the 70 clinical samples: 61?C and 71?C Tm for the wild and the double mutant isolates, respectively. PCR successfully detected bacterial dilution counts ranging from as high as 10 6 CFU/ml to as few as 10 CFU/ml, the latter corresponding to the template DNA concentrations of 19 genome copies. Discussion E. coli is emerging as a pathogen that causes a variety of diseases in veterinary medicine (Hagman and K?hn, 2002; Chen et al., 2003). The trend of increasing resistance to fluoroquinolones has been described by many investigators (Cooke et al., 77 2002; Cohn et al., 2003), and its association with resistance to other antimicrobial agents (Cooke et al., 2002; Karlowsky et al., 2003, 2006) has stimulated the need for a rapid method for the detection of fluoroquinolones. In general, the mutations in DNA gyr A confer resistance to fluoroquinolone in both gram-negative and gram-positive organisms (Hooper, 1999). In E. coli the mutations in gyrA leading to fluoroquinolone resistance have been well characterized (Yoshida et al., 1990; Heisig et al., 1993; Conrad et al., 1996); thus a single point mutation in gyrA results in increased resistance to fluoroquinolones (Willmott and Maxwell, 1993), and a higher-level of resistance is associated with double mutations in codon 83 and 87 of gyrA (Heisig et al., 1993; Vila et al., 1994; Conrad et al., 1996). The substitution in the amino acid Ser-83 was considered to be the first step, for an organism, of becoming resistant to fluoroquinolones (Heisig and Tschorny, 1994). The contribution of mutations in the ParC subunit of topoisomerase IV and decreased intracellular drug accumulation play a secondary role and cause an increase in the level of resistance (Oram and Fisher, 1991; Livermore et al., 2002). Accordingly, we have developed a specific and rapid PCR assay to detect ENR r isolates of E. coli using real-time FRET-PCR. The assay detects mutations within E. coli gyrA using a single bi-probe in as low as 4 copies of purified target sequence. We were able to determine if the E. coli were ENR r within 1-2 hours using MCA. Our assay has promising value, exhibiting a high sensitivity in the early determination of antimicrobial susceptibility to fluoroquinolones of non-isolated clinical urine samples (i.e. directly from an infected urine specimens). Although we studied only enrofloxacin as model to study fluoroquinolone resistance, a previous study indicated that this would result in resistance to other second generation fluoroquinolones (Sanchez et al., 2002). In a clinical 78 perspective, this is an important finding because the assay is rapid and specific for gyrA, thus reducing time consumption and the possibility of cross contamination that may occur with other methods. We were able to design E. coli-specific PCR primers and FRET probes based on nucleotide sequence alignment of the QRDRs of gyrA from a clinical E. coli culture collection. The FRET assay identified all the E. coli isolates tested, while failing to amplify the chromosomal DNA of other bacterial species. The real-time FRET-PCR assay discriminated between wild-type E. coli and other E. coli isolates that carry a C to T transition in codon 83 and G to A in codon 87 of gyrA. Furthermore, we used touch- down thermal cycling for our amplification process in which several gradual high- stringency annealing temperatures preceded the fluorescence acquisition of the real-time PCR (Marchand et al., 2003). This increased the specificity of the PCR amplification. The classical standard methods for determining fluoroquinolone resistance based on CLSI guideline (CLSI, 2008) are tedious, time consuming, and generally result in a 2 to 5 day lapse between collection and reporting. Other molecular techniques such as mismatch amplification mutation assay (MAMA) combined with DNA sequencing (Qiang et al., 2002) have been developed for the detection of ciprofloxacin-resistant in clinical E. coli isolates. This method required using the conventional PCR based technique that required post amplification process such as gel electrophoresis and special expensive equipment for sequencing. Furthermore, the cross contamination is another drawback and may occur as a result of reopening the tubes at the end of the PCR amplification cycling (Kaltenboeck and Wang, 2005). In contrast, the real-time FRET- PCR format reported here in can detect SNP in gyrA in clinical E. coli isolates. A 79 potential additional advantage of the technique is that it allows quantitation of E. coli because of the linear relationship observed between initial DNA template and the PCR threshold cycle in real time. This may allow differentiation of infection (which generally requires > 10 5 CFU/ml [Boothe, 2006]) from non-infection. Other real-time PCR methods have been described previously to detect emerging pathogens (Ko and Grant, 2006; Abdelbaqi et al., 2007; Siedner et al., 2007) and to differentiate bacterial species as in Campylobacter (M?nard et al., 2005). For example, SYBR Green?, hybridization probes, TaqMan probes, Molecular Beacons and Scorpion probes are common real-time detection formats currently used (Kaltenboeck and Wang, 2005). However, SYBR Green is not recommended for tests in which specificity is important as no probe is used and several primers are needed. Other probes used in Taqman , Molecular Beacons and Scorpion require many probes; with two labels needed, design and cost would be greater compared to the need for biprobe labeled once as with the FRET hybridization probe. Thus, rather than four different probes need by TaqMan to detect point mutations (one for each individual point mutation in gyrA plus one for the wild type), only one was needed for our method. Taken together, FRET-PCR technique was chosen and proven to be the sensitive and specific method for detection and differentiation of mutations in gyrA in clinical E. coli samples. The initial decision to focus on codons 83 and 87 was based on an initial sequence analysis of the QRDR of gyrA of 4 clinical isolates, for which gyrA mutations were limited to these sites (Fig. 1). Four nonsense mutations have been identified in those isolates in correspondence to the first gyrA sequence reported by Swanberg and Wang (1987). This result is in agreement with a previous report by Weigel et al. (1998) in 80 which gyrA QRDR sequence of the E. coli type strain (ATCC11775) also revealed four nucleotide differences at the same positions. Two of the mutant clinical isolates in this study had C to T transitions at the second position of codon 83 (nucleotide position 248) in gyrA and G to A transitions at the first position of codon 87 (nucleotide position 259) in which the MICs were >32 ?g/ml. We found only two amino acid positions in GyrA that changed in association with reduced ENR susceptibility (MIC range= 0.25- >128 ?g/ml) Ser-83 to Ile-83, Asp-87 to Asn-87 and Asp-87 to Gly-87 (Fig 2C). Thus the double mutations that involve both codons 83 and 87 were associated with high level of resistance (MIC= ?8 ?g/ml) than those isolates with single mutation (Fig 3). In contrast, a single mutation has been identified in our study that changed Ser-83 to Leu-83 and Asp- 87 to Gly-87 which was associated with intermediate to low level of resistance (MIC range = 0.25-4 ?g/ml). These data were consistent with previous studies (Heisig et al., 1993; Willmott and Maxwell, 1993; Vila et al., 1994) where the low-level fluoroquinolone resistance in E. coli was associated with identical single mutations in codon 83 of gyrA and high-level resistance required double mutations in codons 83 and 87. The predominance of the transitions in codon 83 and 87 of gyrA among fluoroquinolone resistant E. coli indicates a key role for gyrA in conferring high-level enrofloxacin resistance. Generally the mutations at codons Ser83 and Asp87 confer higher levels of quinolone resistance than mutations in other codons (Yoshida et al., 1990). The ability to form hydrogen bonds and the negative charge of amino acid at these positions seems to be important for quinolone interactions with the DNA gyrase?DNA complex (Vila et al., 1996; S?enz et al., 2003). 81 The range of ENR resistance in the 43 E. coli clinical isolates tested for this study was (MIC 4- >128 ?g/ml). Those isolates with similar levels of ENR resistance showed similar melting peaks in MCA (Fig 3). Similarly, 27 of the wild-type E. coli clinical isolates had low MIC value range (0.06-0.5 ?g/ml). We have also sequenced the gyrA QRDRs of 5 exceptional E. coli clinical isolates that showed in-agreement between MCA and the MIC results. Interestingly, they were enrofloxacin resistant with MIC range (1- 4 ?g/ml) in 2 isolates and (16 - >128 ?g/ml) in the other 4 isolates (Fig 3). All 5 possessed a sequence identical to the wild isolates. These results emphasize on the importance of other mechanisms beside mutations that could play critical role in the emergence of fluoroquinolone resistance from clinical isolates. In clinical E. coli, the AcrAB-TolC system was reported by many investigators to cause MDR including fluoroquinolone via overexpression (Mazzariol et al., 2000). These studies indicated that fluoroquinolone- resistant clinical isolates of E. coli expressed high levels of acrB and AcrA of the AcrAB- TolC system (Everett et al., 1996, Mazzariol et al., 2000). Furthermore, many studies indicated that the combination of efflux pump overexpression and a mutation(s) in a topoisomerase gene give rise to clinical levels of resistance to fluourquinolone (Mazzariol et al., 2000). However, the overexpression of AcrAB-TolC system in E. coli may confer high level of resistance to fluoroquinolone regardless of mutations topoisomerase genes. This part can be investigated further in future utilizing the efflux pump inhibitor MC- 207,110 (Phe-Arg-?-naphthylamide) that was previously identified (Lomovskaya et al., 2001). In a very short time (1-2 hours) we were able to detect the presence of E. coli in the urine samples and the presence or absence of point mutations in gyrA gene from 82 clinical specimens without culturing. In all nonisolated clinical samples, E. coli was detected in all diluted urine specimens and the detection limit of the test was determined to be 19 genome copies per reaction (10 CFU/ml urine). In cultured samples the assay can detect as low as 4 genome copies per reaction. There is consistent association between mutations in the gyrA gene and resistance to fluoroquinolone. In general, our assay exhibited 93% (65 of 70 samples) sensitivity to detect the correct susceptibility genotype for fluoroquinolone resistance among in clinical samples. This high proportion for detection sensitivity can justify using this assay for surveillance programs especially when we are dealing with a large number of clinical samples (? 1000 sample/year). Conclusion To our knowledge we are the first to develop and report a FRET-PCR assay that can specifically predict the flouroquinolone resistance-associated gyrA mutations in E. coli using MCA. Our assay proved to be sensitive, reproducible, and quantitative for detection of ENR r from canine and feline clinical E. coli isolates and from nonisolated urine specimens. Furthermore, the FRET-PCR assay is rapid and sensitive for detecting mutations in gyrA and this is critical when designing dosing regimens in a clinical veterinary setting. We have proven the specificity of this assay by using different canine and feline E. coli clinical isolates (including the nonisolated specimens) and unrelated bacterial species. However, a perfect correlation is not predicted between antibiotic susceptibility profiles and mutations in QRDR using the molecular techniques. Therefore, more experiments are needed to identify the contribution of other mechanisms, most importunately the efflux pump, in the emergence of fluoroquinolone resistance among 83 clinical E. coli isolates. Our development of a real-time PCR assay will help define the mechanisms of fluoroquinolone resistance in E. coli. This rapid detection method will provide helpful information when formulating the antimicrobial dosing regimen for clinical cases of E. coli infection and will help to reduce therapeutic failure and the emergence of fluoroquinolone and multi-drug resistance from clinical isolates. Furthermore, because of its applicability to uncultured the urine samples, this assay can be used successfully in surveillance and the emergence of fluoroquinolone resistance among clinical samples of E. coli over the time. 84 References Abdelbaqi, K., Buissonni?re, A., Prouzet-Mauleon, V., Gresser, J., Wesley, I., M?graud, F., M?nard, A., 2007. Development of a real-time fluorescence resonance energy transfer PCR to detect arcobacter species. J. Clin. Microbiol. 45, 3015?3021. Boothe, D.M., 2006. Principles of antimicrobial therapy. Vet. Clin. North Am. Small Anim. Pract. 36, 1003?1047. Boothe, D.M., Boeckh, A., Simpson, R.B., Dubose, K., 2006. Comparison of pharmacodynamic and pharmacokinetic indices of efficacy for 5 fluoroquinolones toward pathogens of dogs and cats. J. Vet. Intern. Med. 20, 1297?1306. Bubenik, L.J., Hosgood, G.L., Waldron, D. R., Snow, L.A., 2007. Frequency of urinary tract infection in catheterized dogs and comparison of bacterial culture and susceptibility testing results for catheterized and noncatheterized dogs with urinary tract infections. J. Am. Vet. Med. Assoc. 231, 893?902. Caplin, B.E., Rasmussen, R.P., Bernard, P.S., Wittwer, P.S., 1999. LightCycler hybridization probes: the most direct way to monitor PCR amplification for quantitation and mutation detection. Biochemica. 1, 5?9. Chen, Y.M., Wright, P.J., Lee, C.S., Browning, G.F., 2003. Uropathogenic virulence factors in isolates of Escherichia coli from clinical cases of canine pyometra and feces of healthy bitches. Vet. Microbiol. 94, 57?69. Clinical and Laboratory Standards Institute, 2008. Performance standards for antimicrobial susceptibility testing; 18th informational supplement. CLSI document M100-S18 Wayne, PA. Cohn, L.A., Gary, A.T., Fales, W.H., Madsen, R.W., 2003. Trends in fluoroquinolone 85 resistance of bacteria isolated from canine urinary tracts. J. Vet. Diagn. Invest. 15, 338?343. Conrad, S., Oethinger, M., Kaifel, K., Klotz, G., Marre, R., Kern, W.V., 1996. gyrA mutations in high-level fluoroquinolone-resistant clinical isolates of Escherichia coli. J. Antimicrob. Chemother. 38, 443?455. Cooke, C.L., Singer, R.S., Jang, S.S., Hirsh, D.C., 2002. Enrofloxacin resistance in Escherichia coli isolated from dogs with urinary tract infections. J. Am. Vet. Med. Assoc. 15, 190?192. Everett, M.J., Jin, Y.-F., Ricci, V., Piddock, L.J.V., 1996. Contribution of individual mechanisms to fluoroquinolone resistance in 36 Escherichia coli isolates from humans and animals. Antimicrob. Agents Chemother. 40, 2380?2386. Giraud, E., Leroy-S?trin, S., Flaujac, G., Cloeckaert, A., Dho-Moulin, M., Chaslus- Dancla, E., 2001. Characterization of high-level fluoroquinolone resistance in Escherichia coli O78:K80 isolated from turkeys. J. Antimicrob. Chemother. 47, 341?343. Guardabassi, L., Schwarz, S., Lloyd, D.H., 2004. Pet animals as reservoirs of antimicrobial-resistant bacteria. J. Antimicrob. Chemother. 54, 321?332. Hagman, R., K?hn, I., 2002. Escherichia coli strains isolated from the uterus and urinary bladder of bitches suffering from pyometra: comparison by restriction enzyme digestion and pulsed-field gel electrophoresis. Vet. Microbiol. 84,143?53. Heisig, P., Schedletzky, H., Falkenstein-Paul, H., 1993. Mutations in the gyrA gene of a highly fluoroquinolone-resistant clinical isolate of Escherichia coli. Antimicrob. Agents Chemother. 37, 696?701. 86 Heisig, P., Tschorny, R., 1994. Characterization of fluoroquinolone resistant mutants of Escherichia coli selected in vitro. Antimicrob. Agents Chemother. 38, 1284? 1291. Hirsh, D.C., 1973. Multiple antimicrobial resistance in Escherichia coli isolated from the urine of dogs and cats with cystitis. J. Am. Vet. Med. Assoc. 162, 885?887. Hooper, D.C., 1999. Mechanisms of fluoroquinolone resistance. Drug Resist. Updat. 2, 38?55. Kaltenboeck, B., Wang, C., 2005. Advances in real-time PCR: application to clinical laboratory diagnostics. Adv. Clin. Chem. 40, 219?259. Karlowsky, J.A., Thornsberry, C., Jones, M.E., Sahm, D.F., 2003. Susceptibility of antimicrobial-resistant urinary Escherichia coli isolates to fluoroquinolones and nitrofurantoin. Clin. Infect. Dis. 36, 183-187. Karlowsky, J.A., Hoban, D.J., DeCorby, M.R., Laing, N.M., Zhanel, G.G., 2006. Fluoroquinolone-resistant urinary isolates of Escherichia coli from outpatients are frequently multidrug resistant: results from the North American Urinary Tract Infection Collaborative Alliance-Quinolone Resistance study. Antimicrob. Agents Chemother. 50, 2251?2254. Ko, S., Grant, S.A., 2006. A novel FRET-based optical fiber biosensor for rapid detection of Salmonella typhimurium. Biosens. Bioelectron. 21, 1283-1290. Lindler, L.E., Fan, W., Jahan, N., 2001. Detection of Ciprofloxacin-Resistant Yersinia pestis by Fluorogenic PCR Using the LightCycler. J. Clin. Microbiol. 39, 3649? 3655. Lomovskaya, O., Warren, M., Lee, A., Galazzo, J., Fronko, R., Lee, M., Blais, J., Cho, 87 D., Chamberland, S., Renau, T., Leger, R., Hecker, S., Watkins, W., Hoshino, K., Ishida, H., Lee, V.J., 2001. Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. Agents Chemother. 45, 105?116. Marchand, S., Hajdari, P., Hackman, P., Udd, B., Richard, I., 2003. Touch-down method for high performance sequencing of polymerase chain reaction products. Anal. Biochem. 315, 270?272. Mazzariol, A., Tokue, Y., Kanegawa, T.M., Cornaglia, G., Nikaido, H., 2000. High-level fluoroquinolone-resistant clinical isolates of Escherichia coli overproduce multidrug efflux protein AcrA. Antimicrob. Agents Chemother. 44, 3441?3443. M?nard, A., Dachet, F., Prouzet-Maul?on, V., Oleastro, M., M?graud, F., 2005. Development of a real-time fluorescence resonance energy transfer PCR to identify the main pathogenic Campylobacter spp. Clin. Microbiol. Infect. 11, 281?287. Oram, M., Fisher, L.M., 1991. 4-Quinolone resistance mutations in the DNA gyrase of Escherichia coli clinical isolates identified by using the polymerase chain reaction. Antimicrob. Agents. Chemother. 35, 387?389. Piddock, L.J.V., 1999. Mechanisms of fluoroquinolone resistance: an update 1994?1998. Drugs. 58, 11?18. Qiang, Y.Z., Qin, T., Fu, W., Cheng, W.P., Li, Y.S., Yi, G., 2002. Use of a rapid mismatch PCR method to detect gyrA and parC mutations in ciprofloxacin- resistant clinical isolates of Escherichia coli. J. Antimicrob. Chemother. 49, 549? 552. 88 S?enz, Y., Zarazaga, M., Bri?as, L., Ruiz-Larrea, F., Torres, C., 2003. Mutations in gyrA and parC genes in nalidixic acid-resistant Escherichia coli strains from food products, humans and animals. J. Antimicrob. Chemother. 51, 1001?1005. Sanchez, S.M., McCrackin Stevenson, A., Hudson, C.R., Maier, M., Buffington, T., Dam, O., Maurer, J.J., 2002. Characterization of multidrug-resistant Escherichia coli isolates associated with nosocomial infections in dogs. J. Clin. Microbiol. 40, 3586?3595. Shaheen, B.W., Boothe, D.M., Oyarzabal, O.A., Smaha, T., 2008a. Characterization of clinical Escherichia coli isolates expressing multidrug resistance recovered from canine and feline with spontaneous disease. In: Congress Proceedings of ACVIM Forum, San Antonio, TX, June 4 ? 7. p.786. Shaheen, B.W., Boothe, D.M., Smaha, T., 2008b. The Phenotypic Expression of Multidrug Resistant Escherichia coli Recovered from Spontaneous Diseased Dogs and Cats. In: Proceedings of the 2008 Annual Conference on Antimicrobial Resistance, Bethesda, MD, June 23 ? 25. p.19. Siedner, M.J., Pandori, M., Castro, L., Barry, P., Whittington, W.L., Liska, S., Klausner, J.D., 2007. Real-time PCR assay for detection of quinolone-resistant Neisseria gonorrhoeae in urine samples. J. Clin. Microbiol. 45, 1250?1254. Swanberg, S.L., Wang, J.C., 1987. Cloning and sequencing of the Escherichia coli gyrA gene coding for the A subunit of DNA gyrase. J. Mol. Biol. 197, 729?736. Vila, J., Ruiz, J., Marco, F., Barcelo, A., Go?i, P., Giralt, E., Jimenez De Anta, T., 1994. Association between double mutation in gyrA gene of ciprofloxacin- resistant clinical isolates of Escherichia coli and MICs. Antimicrob. Agents Chemother. 89 38, 2477?2479. Vila, J., Ruiz, J., Go?i, P., Jim?nez de Anta, T., 1996. Detection of mutations in parC in quinolone-resistant clinical isolates of Escherichia coli. Antimicrob. Agents. Chemother. 40, 491?493. von Ahsen, N., Oellerich, M., Schutz, E., 2000. A method for homogeneous color- compensated genotyping of factor V (G1691A) and methylenetetrahydrofolate reductase (C677T) mutations using real-time multiplex fluorescence PCR. Clin. Biochem. 33, 535?539. Wang, C., Gao, D., Vaglenov, A., Kaltenboeck, B., 2004. One-step real-time duplex reverse transcription PCRs simultaneously quantify analyte and housekeeping gene mRNAs. Biotechniques. 36, 508?519. Weigel, L.M., Steward, C.D., Tenover, F.C., 1998. gyrA mutations associated with fluoroquinolone resistance in eight species of Enterobacteriaceae. Antimicrob. Agents Chemother. 42, 2661?2667. Willmott, C.J.R., Maxwell, A., 1993. A single point mutation in the DNA gyrase A protein greatly reduces binding of fluoroquinolones to the gyrase-DNA complex. Antimicrob. Agents Chemother. 37, 126?127. Yoshida, H., Bogaki, M., Nakamura, M., Nakamura, S., 1990. Quinolone resistance- determining region in the DNA gyrase gyrA gene of Escherichia coli. Antimicrob. Agents Chemother. 34, 1271?1272. 90 TABLE 1. E. coli strains and isolates used in this studies and their antimicrobial susceptibility to enrofloxacin. Antimicrobial Susceptibility d Strain/Isoltes ID Tissue Source Strain donor c or reference Origin MIC (?g/ml) Interpretation E. coli K12 a GenBank X06373 (Swanberg and Wang, 1987) S ATCC 11775 a Urine ATCC/GenBank AF052254 (Weigel et al., 1998) ?0.12 e S N7190284 Urine IDEXX-MA Canine 0.06 S M0846371 Urine IDEXX-NC Canine 64 R R5107645 Urine IDEXX-MA Feline 0.06 S L8059588 Urine IDEXX-CA Feline 64 R M0851881 Urine IDEXX-NC Feline 32 R M0854758 Skin IDEXX-NC Canine 64 R M0858621 Urine IDEXX-NC Canine >128 R M0825317 Urine IDEXX-NC Canine 0.06 S M0821846 Urine IDEXX-NC Canine 0.06 S M0824507 Ear IDEXX-NC Canine >128 R I2132223 Urine IDEXX-IL Canine 16 R B5299277 Urine IDEXX-CA Feline 4 R L8070100 Urine IDEXX-CA Canine 32 R L8063713 Urine IDEXX-CA Canine >128 R L8055239 Urine IDEXX-CA Feline 32 R F5093961 Urine IDEXX-CA Feline 2 I K5262419 Urine IDEXX-CA Canine 0.06 S W6316944 Trachea IDEXX-IL Canine >128 R D1056271 Wound IDEXX-IL Canine 32 R W6319937 Nose IDEXX-IL Feline >128 R I3211251 Urine IDEXX-IL Canine 0.06 S C6394833 Urine IDEXX-IL Canine 0.06 S I2473760 Urine IDEXX-IL Canine >128 R I3195059 Urine IDEXX-IL Feline 128 R M0846371 Urine IDEXX-NC Canine 64 R 91 MS-1 b Wound MS Canine >32 f R MS-10 b Ear MS Canine >32 f R AU-26 b Urine AU Canine 0.064 f S IDM-15 b Urine IDEXX-MA Canine 0.064 f S W7528341 Skin IDEXX-OH Canine 64 R W7534250 Urine IDEXX-OH Canine 32 R D8632999 Urine IDEXX-OH Canine 8 R D1083401 Urine IDEXX-OH Canine 64 R W7538455 Urine IDEXX-OH Canine 16 R L9254925 Urine IDEXX-CA Canine 128 R L9245953 Urine IDEXX-CA Canine 2 I K5693300 Urine IDEXX-CA Canine 1 I I2850490 Fracture IDEXX-IL Canine 64 R I2846835 Urine IDEXX-IL Canine 1 I I2888998 Nose IDEXX-IL Feline 64 R M1213579 Urine IDEXX-NC Canine 4 R L9473547 Urine IDEXX-CA Canine 128 R L9470993 Urine IDEXX-CA Canine >128 R I4097858 Urine IDEXX-IL Canine 64 R Y5826621 Urine IDEXX-MA Canine 128 R J7261205 Urine IDEXX-MA Canine 64 R W7538320 Vagina IDEXX-OH Canine 0.5 S F5889381 Urine IDEXX-CA Feline 0.12 S L9254729 Urine IDEXX-CA Canine 0.12 S Y5645394 Vagina IDEXX-MA Canine 0.25 S R6559423 Urine IDEXX-MA Canine 0.5 S Y5643881 Pleural fuid IDEXX-MA Canine 0.5 S B5259772 Urine IDEXX-CA Canine 0.25 S L8063901 Urine IDEXX-CA Canine 0.5 S W6319347 Urine IDEXX-OH Canine 0.25 S D1056065 Urine IDEXX-OH Canine 0.25 S I3634766 Urine IDEXX-IL Canine 2 I B6276689 Urine IDEXX-CA Feline 0.5 S D1079284 Urine IDEXX-OH Feline 0.5 S D1079346 Urine IDEXX-OH Canine 0.12 S I3621025 Abdomenal fluid IDEXX-IL Canine 64 R I3620609 Urine IDEXX-IL Canine 128 R B6276670 Urine IDEXX-CA Canine 32 R 92 Q2100506 Abdomenal fluid IDEXX-MA Canine 64 R D1056271 Tracheal wash IDEXX-OH Canine 32 R B6276651 Urine IDEXX-CA Canine 64 R S3590475 Throat IDEXX-CA Canine 16 R R5271946 Urine IDEXX-MA Feline >128 R D8474601 Urine IDEXX-OH Feline 64 R W7529296 Urine IDEXX-OH Canine 64 R a The DNA sequences of these strains were acquired from GenBank. b Isolates were collected between May and September 2005. c ATCC, American Type Culture Collection, Manassass, VA; IDEXX Laboratories, Inc.,Westbrook, Maine; CA, California; NC, North Carolina; OH, Ohio; MA, Massachusetts; IL, Illinois; MS, Mississippi; AU, Auburn University. d The MIC was determined for enrofloxacin according to CLSI guidlines (CLSI, 2008); R, resistant; I, intermediate; S, susceptible. e The MIC for this strain was determined for ciprofloxacin (Weigel et al., 1998). f The MIC was detemined using the E-test? (Epsilon; AB Biodisk, Solna, Sweden) according to CLSI guidlines and interpretive standard (CLSI, 2008). 93 Figure 1. DNA sequence alignments of the gyrA QRDRs of E. coli isolates including wild isolates and mutant clinical isolates with Ser-83 and Asp-87 mutations. Consensus indicates nucleotide positions identical to the corresponding E. coli gyrA sequence. PCR primers are labeled. Probe 1 (Fluorescein) and probe 2 (Bodipy 630/650) are shown between the upstream and down primer. Probe 1 and probe 2 are identical to with the wild-type E. coli gyrA sequence. The positions of recurring mutations in Ser-83 and Asp-87 codons are indicated (boldfaced letters) in the mutant-type DNA sequence below probe 2. 94 Figure 2. Amplification and melting curve analysis of PCR amplicon containing the QRDR of the gyrA gene of clinical E. coli isolates. A, efficient PCR amplification is indicative by parallel slopes of linear phase corresponded to different starting DNA templates. B, four different melting peaks represented 3 major groups: I, II, and III were generated from the dissociation of the probes from the wild-type and mutant-type templates. The y axis represents the change in fluorescence as a function of time (dF/dT) while x axis represents the temperature. Different peaks in the MCA corresponded to different mutations are indicated. All mutant isolates produced less T m than the wild-type template. C, multi-sequence alignments of the gyrA QRDRs of the clinical E. coli isolates and location of the FRET probes used for 95 real-time PCR amplification. The donor probe (Fluoresceine) and the anchor probe (Bodipy 630/650) are indicated. Nucleotides mismatches are indicated (boldfaced letters) within the anchor probe, and dots indicate nucleotides identical to the corresponding wild E. coli sequence. 96 Fig 3. Melting temperatures correlate and predict the enrofloxacin MICs in clinical isolates. Each dot represents one clinical sample of E. coli from dogs and cats. Isolates were carried double mutations in codons 83 and 87 (filled black circle), or single mutation in codon 83 or 87 (filled grey circle), or nonsense mutation in codon 85 (opened triangle), or wild type genotype of gyrA (opened circle). Group I includes the mutant (83 and 87) isolates; Group II, includes the mutant (83 or 87) and the mutant (85) isolates; Group III, includes the wild type isolates. Mutant (83 & 87), samples (n=37) with a melting-curve temperature of 61?C corresponding to a mutant genotype in codon 83 and 87 of gyrA and had high MICs range (8 - >128 ?g/ml); mutant (85), samples (n=2) with a melting-curve temperature of 66.5?C corresponding to a mutant genotype in codon 85 of gyrA and had 97 low MIC values (0.06 ?g/ml); mutant (83), samples (n=6) with a melting-curve temperature of 68?C corresponding to a mutant genotype in codons 83 of gyrA and had intermediate to low level of resistance (MICs range= 0.06 - 4 ?g/ml); mutant (87), samples (n=1) with a melting-curve temperature of 69?C corresponding to a mutant genotype in codon 87 of gyrA and had susceptible MIC value (0.5 ?g/ml); wild type, samples (n=24) with a melting-curve temperature of 71?C corresponding to a wild-type genotype of gyrA and had a wide MICs range (i.e. 0.06 - >128 ?g/ml). Interestingly, among the wild type genotype of gyrA isolates, 6 out of 24 were exceeding the break point of susceptibility to enrofloxacin (i.e. MIC= 0.5 ?g/ml). The MIC results were interpreted according to CLSI guidelines and interpretive standards (CLSI, 2008). 98 CHAPTER IV THE CONTRIBUTION OF GYRA MUTATION AND EFFLUX PUMPS TO FLUOROQUINOLONE AND MULTI-DRUG RESISTANCE IN CANINE AND FELINE PATHOGENIC ESCHERICHIA COLI ISOLATES FROM THE US Abstract Objectives: The aim of this study was to investigate the contribution of gyrA and efflux pump and the emergence of fluoroquinlone/MDR resistance among the clinical sample of E. coli. Methods: Minimun inhibitory concentrations (MIC) were determined for enrofloxacin and six other drug classes (i.e. beta lactams, aminoglycosides, phenicols, potentiated sulfa, tetracycline, and macrolides) using broth micro-dilution for 536 of E. coli clinical isolates with spontaneous infections. Fifty-six of the isolates were further studied for the presence of the broad substrate AcrAB efflux pump system, gyrA mutations and their relationship to co-resistance between fluoroquinolone and other drug classes. Phe-Arg-?-naphthylamide, an efflux pump inhibitor, was used to examine the contribution of efflux pump overexpression. Results: Fluoroquinolone resistance was significantly associated with a multidrug resistance phenotype compared with fluoroquinolone susceptible isolates. The results suggest that double mutation in gyrA can confer high level of resistance to the majority of 99 fluoroquinolone resistant isolates, and that inhibition by EPI to the efflux pump, presumably AcrAB efflux pump would (i) decrease the intrinsic level of the MICs to FQ in susceptible isolates, (ii) decrease the MICs below the susceptible break point for FQ-R isolates even with single mutation in gyrA, (iii) decrease the magnitude of resistance in highly resistance isolates in the presence of double mutations in gyrA (IV) decrease the MICs below the susceptible break point for some of the highly FQ-R isolates without gyrA mutations, and (V) decrease the resistance to structurally unrelated drugs, thus reducing the incidence of MDR. Conclusion: Our data provide further evidence that fluoroquinolone use and subsequent resistance can adversely affect the susceptibility of E. coli to multiple drug classes. Introduction Escherichia coli is a major cause of urinary tract infections and has a significant role in infections of other tissues in dogs and cats (Chen et al., 2003; Hagman and K?hn , 2002; Wernicki et al., 2002). Fluoroquinolones (FQ) are among the drugs most commonly used to treat infections caused by E coli in dogs and cats. This frequent use may have contributed to the substantial increase in E. coli resistance to FQ from canine and feline clinical isolates that has emerged in the last decade (Cohn et al., 2003; Cooke et al., 2002; Guardabassi et al., 2004). Because FQs are used both by veterinarians and physicians as first line therapy in USA and Europe (Cohn et al., 2003; Goettsch et al., 2000), and E. coli is a common cause of urinary tract and other infections in humans, its resistance is both a medical concern (therapeutic failure in the veterinary patient), and a public health concern (transmission of resistant E. coli from pets to people and back) 100 (Guardabassi et al., 2004; Lanz et al., 2003; Sanchez et al., 2002; Talan et al., 2008). Furthermore, FQ resistance to E. coli is often associated with multi-drug resistance (MDR) as it has been demonstrated for clinical E. coli isolates collected from dogs and cats with spontaneous infections (i.e. non experimentaly induced) (Cooke et al., 2002; Karlowsky et al., 2006; Sanchez et al., 2002; Shaheen et al., 2008). The most common mechanism of resistance to quinolones in E. coli are mutations in the genes that encode the subunits of the quinolone target topoisomerases, specifically DNA gyrase (topoisomerase II, in gyrA), and topoisomerase IV (in parC) (Everett et al., 1996; Piddock, 1999; Vila et al., 1996). These mutations are located in the quinolone resistance-determining region (QRDR) of the gyrA gene and its homologous region of the parC gene (Heisig, 1996; Oram and Fisher, 1991; Piddock, 1999). Mutations in gyrB and parE genes are less prevelant and rarely contribute to quinolone resistance (Giraud et al, 2001; Vila et al., 1996). Double mutations in gyrA gene are generally required for high- level of resistance (Conrad et al., 1996; Heisig et al., 1993; Vila et al., 1994), whereas mutation(s) in parC are less frequent and are associated with lower level resistance (Bagel et al., 1999; Everett et al., 1996; Heisig, 1996). Beside mutations in the QRDR, resistance towards quinolones is now known to be mediated by transfersable resistance genes (i.e. qnrA, B, C, S), enzymatic modification (i.e. aac(6?) Ib-cr) or specific efflux pumps (qepA) (Cattoir and Nordmann, 2009; Park et al., 2006; Robicsek et al., 2006a, 2006b; Yamane et al., 2007). Although this mechanism confers low level of resistance alone, their presence was found to enhance the frequency of selection of chromosomal mutatants upon exposure to FQs (Park et al., 2006; Robicsek et al., 2006a, 2006b). Although a common cause of FQ resistance, mutations in gyrA and parC should 101 not directly cause MDR. As such, mechanisms of resistance that involve other classes of antimicrobials must be involved in the emergence of MDR. Among the other mechanisms by which organisms become resistance to FQ is a decreased intracellular, reduced drug concentration by increasing efflux pump activities is a major mechanism for gram negative bacteria (Everett et al., 1996; Giraud et al., 2001; Kern et al., 2000). Among the different efflux transport proteins, the AcrAB-TolC system is most active for E. coli. This system belongs to the Resistance Nodulation Division (RND) family of transporters. This system transports a broad range of substrates, including drugs, endogenous substrates and toxins (Piddock, 2006). Among the drug substrates for this system are chloramphenicol, lipophilic ?-lactams, FQs, tetracycline, rifampin, novobiocin, fusidic acid, and nalidixic acid (Piddock, 2006; Sulavik et al., 2001). It is assumed that the evolutionary role of MDR efflux pump is to protect the bacteria against hostile environments (i.e. through the transport of harmful substance out of the bacterial cell). Enteric pathogens (i.g. E. coli) constitutively express efflux pump activities that enable them to survive in environments rich in bile salts and fatty acids, which are substrates for E. coli AcrAB-TolC (MA et al., 1997; Thanassi, et al., 1997). However, over expression of the AcrAB-TolC system in E. coli has been demonstrated to be insufficient to cause resistance to ciprofloxacin, chloramphenicol, tetracycline, and cotrimoxazole (Piddock, 2006; Sulavik et al., 2001). It is the combination of a mutation with over expression of an efflux pump that gives rise to high level of resistance to FQs (Mazzariol et al., 2000; Mazzariol et al., 2002). Thus the induction of AcrAB-TolC efflux pump by efflux pump substrate, such as FQs, can cause an increase in the level of resistance to FQs and the emergence of MDR phenotypes (Kriengkauykiat et al., 2005). 102 The increase of the pump expression in the presence of mutations in the QRDR probably explains the high level resistance associated with MDR. The AcrAB-TolC efflux pump can be inhibited by a variety of substrates. Efflux pump inhibitors (EPI) may enhance the activity of some antimicrobials, particularly for gram negative bacteria, including the Enterobacteriaceae (Piddock, 1999). Among the inhibitors of the AcrAB-TolC system, EPI Phe-Arg-?-naphthylamide reduces resistance caused by the RND efflux systems and can reverse the resistance for antimicrobials that are substrate of AcrAB-TolC efflux pump. The association of FQ and MDR is generally not addressed in veterinary medicine especially from pet animals (Cohn et al., 2003; Cooke et al., 2002; Lanz et al., 2003; Meunier et al., 2004). Studies addressing these mechanisms have been largely limited to human strains, or isolates of public health concern, such as those collected from food animals (Everett et al., 1996; Giraud et al., 2001). The purpose of this study was to describe the combined role of topoisomerase mutation and efflux pump activity in pathogenic canine and feline E. coli collected from animals with disease. Isolates were phenotyped using standard micro-broth dilution susceptibility testing. The level of resistance to FQ was determined based on MIC. In isolates resistant to FQ, target mutation(s) in gyrA gene were characterized using fluorescence resonance energy transfer (FRET) PCR. The contribution of efflux pumps to FQ resistance and MDR phenotypes was then tested based on the differences in MIC for each drug in the absence or presence of the EPI. 103 Materials and methods Bacterial isolates and culture conditions Canine and feline pathogenic E. coli isolates (n = 536) were acquired from clinical veterinary microbiology laboratories between May and December 2008. All isolates were submitted by veterinary practitioners treating canine or feline patients with infections associated with the culture. The isolates were collected throughout the US by contributing laboratories. These laboratories identified and performed the susceptibility testing on the isolates. At the Clinical Pharmacology Laboratory (CPL) of Auburn University, isolates were plated on CHROMagar? Orientation (Becton, Dickinson, Franklin Lakes, NJ) for differentiation and confirmation. The antimicrobial susceptibility testing with and without EPI Antimicrobial susceptibility testing was performed using custom made micro- dilution susceptibility plates (TREK Diagnostic Systems, Cleveland, OH) according to Clinical and Laboratory Standards Institute (CLSI) guidelines and interpretive standards (CLSI, 2008). Two panels of 13 antimicrobial agents were tested including ampicillin, amoxicillin/clavulanic acid, cephalothin, cefoxitin, cefpodoxime, cefotaxime, ceftazidime, meropenem, enrofloxacin, gentamicin, doxycycline, chloramphenicol , and trimethoprim/sulfamethoxazole (Table 1). For standard susceptibility testing, cells were harvested and suspended in 4.5 ml of 0.9% normal saline and adjusted to 0.5 McFarland standard turbidity (~10 8 CFU) using SENSITITER ? Nephelometer (TREK Diagnostic Systems, Cleveland, OH). Microdilution plates were incubated at 35 o C for 20-24 hours. The results were recorded as true value MIC using the SENSITITER ? VIZION system 104 (TREK Diagnostic Systems, Cleveland, OH). For quality control purposes, E. coli ATCC ? 25922 (American Tissue Cell Culture, Manassas, VA) and Enterococcus fecalis ATCC ? 29212 (Manassas, VA) were used. Each isolate was then designated as resistant (R), intermediate (I) and susceptible (S) to enrofloxacin (ENR) and other drugs using CLSI guidelines and interpretive standards (CLSI, 2008). Multidrug resistance (MDR) was defined as resistance to two or more classes of antimicrobial agents. Class designations and MIC ranges for custom made micro-plates are indicated in Table 1. In preliminary studies, EPI (Phe-Arg-?-naphthylamide; Thermofisher, Pittsburgh, PA) was prepared first in 11 ml of Mueller-Hinton broth (MHB), then added in a volume of 0.5 ?l into custom made microdilution plates (each well containing 1 ?l final volume), such that the final concentrations were 10, 25, 50, 100 and 120 mg/L. Inhibition at 100 mg/L of Phe-Arg-?-naphthylamide did not appear to impact growth conditions; accordingly, this concentration of the EPI was used for further studies of the effect of the EPI on overexpression phenotypes among the clinical isolates. Fifty-six clinical E. coli isolates were randomly selected that represent a wide range of enrofloxacin MICs for further studies to characterize the roles of mutation and pump overexpression in FQ resistance. Most (43 of the 56 isolates of E. coli) of these clinical isolates were multidrug resistant, and fourteen fully susceptible isolatess were also included in this work. Mutations were determined by FRET PCR assay (As described in Chapter III), and and pump over expression was studied using EPI. Based on MIC and CLSI guidelines (CLSI, 2008), E. coli isolates were designated as susceptible (S, n= 14), intermediate (I, n= 5) or resistant (R, n= 37). 105 An efflux pump-overexpressed phenotype was considered evident for any isolate for which the MIC for ENR decreased at least four fold (two tube dilutions) when the MIC was determined in media containing the EPI compared to media not containing EPI. In addition to ENR, the impact of the EPI on changes in MIC was determined for 13 other drugs. Detection of gyrA gene mutations by FRET-PCR Bacterial DNA for PCR reaction were extracted using PreMan? Ultra Sample Preparation Reagent (Applied Biosystem, Foster City, CA) following the manufacturer?s instructions. The forward and reverse primers and probes designed specific for gyrA are listed in Table 2. The length of the amplified product was 310 bp. The copy number of E. coli gyrA was determined by real-time FRET PCR in a LightCycler (Roche Diagnostics) (Wang et al., 2004). Each 20 ?l reaction contained 2.0 U Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA), and 0.2 U heat-labile uracil-N-glycosylase (Roche Molecular Biochemicals, Indianapolis, IN), 15 ?l of 1.33 X master mix, and 5 ?l of sample nucleic acids. The primers and probes were designed using by using Vector NTI ? 10.1 software (Invitrogen, Carlsbad, CA) showed in Table 2. The execution protocol for stringent PCR and melting curve analysis followed the previous description (Described in details in Chapter III; Lin et al., 2008). 106 Statistical analysis MIC 50 and MIC 90 were calculated for each drug class and for all 56 isolates subjected to EPI. Furthermore, a cumulative percent of susceptibility was used to compare the MIC of isolates before and after EPI. The odds ratio of resistance in FQ-R versus FQ-S isolates (i.e. resistant to the antimicrobials but FQs) associated with each of the drug classes was determined. The association of resistance of each drug class to resistance with ENR was investigated. The proportions of resistance to each drug class in the ENR-S versus ENR-R was calculated using the Fisher Exact Test (P < 0.05 were considered to be statistically significant). Analyses were performed using the statistical software MINITAB ? 15 package (Minitab Inc., PA, USA). The cumulative percent of susceptibility was calculated using SWIN? Epidemiology Module (TREK Diagnostic Systems, Cleveland, OH). Results Fluoroquinolone susceptibility patterns Seventy one percent (381/536) of the isolates expressed resistance to one or more drugs. Of the 381 isolates expressing a resistant phenotype, 20% (n=76) included FQ-R resistance. The FQ-R was significantly associated with MDR, and 98.7% of FQ-R expressed MDR phenotype compared to only 32% of the FQ-S (OR 158.4, CI 27.3? 915.4) (Table 3). Among all antimicrobial drug classes, FQ-R was most frequently associated to resistance to beta lactams (95%). Comparison between FQ-R and FQ-S isolates revealed that FQ-R isolates were significantly associated with doxyxycline resistance (68%; OR 19.1, CI 10.6-35.3), gentamicin resistance (42%; OR 9.8, CI 5.2- 107 18.5), chloramphenicol resistance (39%; OR 2.4, CI 1.4-4.0), and trimethoprim/ sulfamethoxazole resistance (34%; OR 5.8, CI 3.1-11) (Figure 2). Characterization of fluoroquinolone susceptibilities of canine and feline E. coli isolates A total of 56 (10% of the total) clinical E. coli isolates were randomly elected for further characterization of resistance. Isolates represented different MIC and thus resistance levels (susceptible, intermediate or high) for ENR, as well as resistance to other drug classes. Levels of resistance were based on CLSI MIC breakpoints (CLSI, 2008): susceptible at 0.12?0.5 mg/L (n = 14), intermediate at 1?2 mg/L (n=5) or resistant at more than mg/L (n=37) (Table 4). Target mutations in the gyrA gene of the canine and feline isolates of E. coli For susceptible isolates (n=14), the majority (n=12) were free of gyrA mutation(s) in the tested region. For intermediate isolates (n=7, MIC= 0.25-2 mg/L), five had a single mutation in gyrA (S83L), one had a single mutation in gyrA (D87G); while the remaining isolate had no gyrA mutation. All but three resistant isolates (n= 33) expressed double mutations in gyrA (i.e. S83L and D87N) (Table 4). The three exceptions carried no mutations in gyrA despite the high level of resistance (MIC =16, 32 and 128 mg/L) (Table 4). Similarly, one isolate with ENR resistance of 4 mg/L exhibited no mutation in gyrA. This can be attributed to other resistant deteminats such as the mutations in the regulatory genes and their effect on the efflux pump and porins (Piddock, 2006). 108 Effect of the EPI Phe-Arg-?-naphthylamide Inhibition of the efflux pump resulted in a reduction in the MIC below the susceptible break point (0.5 mg/L) for isolates with an MIC of 4 mg/L or less, regardless of the presence of a mutation. Additionally, susceptibility was returned for the wild isolates expressing high level resistance by the addition of EPI. For those isolates with two mutations, EPI resulted in a median decrease in MIC of 16 fold (Figure 2). Interestingly, the median decrease in MIC for single mutations was ?32 fold. It is important to note, however, that although the fold decrease in MIC was still high for isolates with double mutations, the median MICs of ENR for those isolates was still above the resistant (> 4 mg/L) breakpoint for ENR after the treatment with EPI. This is in contrast to the EPI-induced decrease in MIC for low or intermediate isolates with single mutation in gyrA. For these isolates, the median MIC decreased to ? 0.06 mg/L, which is below the susceptible breakpoint. However, in terms of magnitude of decrease in MIC, EPI had the greatest impact on ENR MICs (median decrease ?267 fold) for the isolates with no mutations (n=5; MIC=1->128 mg/L), compared with those with mutation(s) (Figure 2). The treatment with EPI also influenced the susceptibility to other selected antimicrobials, including azithromycin, doxycyline, and chloramphenicol. Based on changes in the cumulative susceptibility of azithromycin, the percent of suscebtible isolates below the susceptible and resistant breakpoints was 0% and 46% prior to treatment with EPI compared to 68% and 96% with active efflux pumps (Figure 3a). For doxycyline, the percent of suscebtible isolates below the susceptible and resistant breakpoints was 32% and 57%, without EPI, compared to 98% and 98%, with EPI 109 (Figure 3b). For chloramphenicol (susceptible and resistant break point of ?8 mg/L and <32 mg/L, respectively), 57% and 71%, of the isolates were susceptible without EPI, compared to 84% and 93%, with EPI (Figure 3c). The presence of EPI had no impact on the susceptibility of the isolates to gentamicin (Figure 3d). Discussion The resistance to FQs has increased since their approval in the late 1980?s, with the resistance increasing particularly in the last decade for pathogenic E. coli and other gram negative bacteria (Boothe et al., 2006; Cohn et al., 2003; Cooke et al., 2002; Guardabassi et al., 2004). This resistance limits therapeutic options, may lead to therapeutic failure, appears to be associated with MDR, and is expected to become more prevalent in many veterinary hospitals and clinics (Shaheen et al., 2008). The purpose of this study was to define the underlying causes of E. coli resistance, including MDR and its association with FQ resistance. Specifically, we were interested in the individual or combined effect of target gene modifications and efflux mechanism on the emergence of the FQ/MDR phenotypes of E. coli from dogs and cats. Data from our previous study (Cooke et al., 2002) indicated an increased prevalence of ENR resistance among uropathogenic E. coli collected from dogs during 1994 to 1998. These isolates had been collected from patients at veterinary teaching hospitals and private practices throughout the United States. Cohn et al. (2003) also documented an increase of FQ resistance in bacteria from canine urinary tract infection between 1992 and 2001 in Missouri. A significant increase in the overall proportion of resistant bacteria (including E. coli) to either ciprofloxacin or ENR was demonstrated 110 separately. These authors concluded that the overall efficacy of FQ remained high (> 80% susceptible), but the data suggested an increase in resistant bacteria isolated from the urinary tract of dogs. Furthermore, Boothe et al. (2006) demonstrated that even those isoalates designated as susceptible based on CLSI guidelines were characterized by MIC sufficiently high that only the highest recommended dose would acehive targeted pharmacodynamic indices, that is Cmax/MIC of 10 or more, or AUC/MIC of 125 or more. Similarly, our results indicated a similar pattern of FQ resistance among E. coli isolates collected between May and December 2008 in the US. These findings may be expected following the extensive use of these drugs since their approval; FQs were often the first drugs of choice used empirically (without the benefit of culture and susceptibility testing) to treat many infections in dogs and cats (Guardabassi et al., 2004; Lanz et al., 2003; Sanchez et al., 2002). Several surveillance studies have tracked changes in E. coli susceptibilities to FQ and other drug classes over time in humans, but few studies have evaluated the mechanisms that account for this increase in resistance in companion animals (Cohn et al., 2003; Cooke et al., 2002; Lanz et al., 2003; Meunier et al., 2004). Our study suggestes the mechanisms by which this resistance is emerging. First, we investigated the target gene modification as one of the common mechanism that confers resistance to FQ among selected resistant and susceptible clinical E. coli isolates. The mutations in DNA gyrase and topoisomerase IV are the most important mechanisms that confer resistance to quinolone (Everett et al., 1996; Piddock, 1999; Vila et al., 1996). In our study, the majority of FQ-R isolates were found to carry double mutations in gyrA, whereas intermediate to low level of resistance were found to be associated with single mutation; 111 a finding consistent with previous reports (Heisig et al., 1993; Vila et al., 1994; Willmott et al., 1993). It is apparent that mutations at codons Ser83 or Asp87 confer higher levels of quinolone resistance than mutations in other codons (Yoshida et al., 1990). The differences in levels of resistance may be attributed to the ability to form hydrogen bonds and the negative charge of amino acid at these positions, which seems to be important for quinolone interactions with the DNA gyrase?DNA complex (S?enz et al., 2003; Vila et al., 1996). Surprisingly, approximately 9% of the isolates exhibited no mutation(s) in gyrA; yet, one isolate expressed intermediate and four expressed high levels of FQ resistance. These results provide a second evidence of the importance of the efflux system and the emergence of FQ resistance/MDR phenotype among clinical isolates of E. coli. It is noteworthy that our study demonstrated, as has others (Cooke et al., 2002; Karlowsky et al., 2006) that many FQ-R isolates exhibit resistance to structurally unrelated antimicrobial agents. Our MDR phenotypes were more likely to be associated with FQ-R (158 times) compared with FQ-S, and that MDR phenotype was expressed to different drug classes. This strong association between FQ-R and MDR suggests that the over expression of efflux pumps contribute to the emergence of MDR phenotype in FQ-R isolates. This might be expected if mutations induce overexpression of the pump and thus confer resistance to a wide range of different drug classes (Piddock, 2006; Sulavik et al., 2001). Therefore, selection for isolates expressing MDR is expected to be induced with previous FQs treatmentbecause these agents are one of the many efflux pump substrates (Piddock, 2006; Sulavik et al., 2001). We screened for the efficacy of EPI (i.e. Phe-Arg-?-naphthylamide) to decrease 112 the magnitude of MICs toward ENR and other drugs based on its ability to potentiate levofloxacin activity (Lomovskaya et al., 2001; Piddock et al., 2006; Renau et al., 1999). Indeed, EPI (i.e Phe-Arg-bnaphthylamide) has been used as to determine the prevalence of efflux pump overexpression phentoypes in P. aeruginosa (Kriengkauykiat et al., 2005). We found that the EPI at low concentration (i.e. 25 mg/L) did not effectively decrease MICs to ENR in E. coli (data no shown), a finding consistent with previous reports (Kern et al., 2006; Mazzariol et al., 2000; S?enz et al., 2004). For ENR, at lease 4-fold reduction in the MICs was observed in almost all of isolates, with exception of one isolates that had a MIC= 0.12 mg/L. These results highlight the important of intrinsic resistance of gram negative bacteria to many antimicrobial drugs. This ressitance has been attributed to a constitutive expression of the efflux pump alone, or together with decreased porins expression that could confer a basal level of resistance for those isolates (Everett et al., 1996; Giraud et al., 2001; Kern et al., 2000; Piddock, 2006). The efficacy of EPI to reduce the MICs was affected by the target mutations in gyrA. This effect was evident among isolates carrying one single mutation in gyrA, in which the MIC of ENR decreased from intermediate level of resistant to susceptible levels. In contrast, EPI was found to decrease the MICs of ENR in 33 out 35 strains that carried double mutation in gyrA. Interestingly, EPI failed to restore the susceptibility to ENR MICs in the presence of double substitutions in the target gene. The presence of efflux pump over expression can contribute to further increase the level of resistance (? 8 fold) in the presence of double mutation in gyrA. This result is consistent with reports on the importance of combined mechanisms that confer high level of resistance among clinical samples (Fernandes et al., 2003; Mazzariol et al., 2000; Mazzariol et al., 2002). 113 These data suggest that double mutations provide the mechanism for clinical isolates to become resistant to FQ and have an enhance efflux system that act additively. Interestingly, EPI decreased the MIC most (? 67 fold) for FQ-R isolates carrying no mutations in gyrA; the levels of resistance of ENR were restored to wild-type levels. Therefore, it appears that over expression of the efflux pump can have an impact on intracellular drug concentrations, dramatically decreasing the concentration such that over expression of the pump alone is sufficient to cause of FQ-R and MDR in E. coli. Evidence to support this result has been documented with E. coli in which inactivation of the acrAB rendered all strains, including those with target gene mutations, hyper- susceptible to FQs and other drugs (Oethinger et al., 2000). Similar findings have been demonstrated with other organisms such as P. aeruginosa, with deletion of MexAB- OprM efflux pump (Lomovskaya et al., 1999), Salmonella enterica serovar Typhimurium DT204, with inactivation of acrB (Baucheron et al., 2002) and in Campylobacter jejuni with cmeB deletion (Luo et al., 2003). We also demonstrated that EPI reduced the MICs of several antimicrobial drugs. EPI appears to have the strongest inhibitor effect on macrolides (i.e. azithormycin), drugs not normally considered effective toward gram negative organisms. The pump also markedly impacted doxycyline and chloromphenicol. Remarkably, MIC 90 s for the isolates resistant to azithromycin decreased by > 64-fold in the presence of an EPI compared to 32 for doxycyline resistant isolates and 16 fold for chloromphenicol resistant isolates (Figure 3). The EPI appears to have not impact on the MIC of aminoglycoside (i.e. gentamicin) when tested over the wide MIC range. This is not surprising, as the AcrAB-TolC efflux system in E. coli has no effect on aminoglycosides (Piddock, 2006) 114 compared to that other efflux system (i.e. AcrD), which does impact efflux of aminoglycosides (Nishino and Yamaguchi, 2001; Rosenberg et al., 2000). Furthermore, the EPI did not decrease the MICs of any beta lactams studied here. It may be important to note that the Phe-Arg-?-naphthylamide inhibits efflux of some but not all the substrates of RND pumps, including AcrAB-TolC, because the inhibitor binding site is specific (Lomovskaya and Watkins, 2001). Both target gene modification and efflux pump-mediated resistance are underestimated mechanisms by which resistance is mediated in clinical isolates of E. coli from dogs and cats. A double mutation in gyrA has proven to be a major mechanism that contributes significantly to emergence of high level of resistance to FQ. A single target- based mutation provides only intermediate level of resistance to ENR. Furthermore, we have provided evidence that isolates resistant to FQ were also resistant to other drug classes simultaneously with enhanced efflux pump over expression. We have demonstrated that inhibition by EPI to the efflux pump, presumably AcrAB efflux pump, at the concentrations used in this study would (i) decrease the intrinsic level of the MICs to FQ in susceptible isolates, (ii) decrease the MICs below the susceptible break point for FQ-R isolates even with single mutation in gyrA, (iii) decrease the magnitude of resistance in highly resistance isolates in the presence of double mutations in gyrA (IV) decrease the MICs below the susceptible break point for some of the highly FQ-R isolates without gyrA mutations, and (V) decrease the resistance to structurally unrelated drugs, thus reducing the incidence of MDR. Thus, the increase in the emergence of FQ-resistant and MDR E. coli calls for a rational usage of these drugs by veterinarians. 115 References Bagel, S., H?llen, V., Wiedemann, B., Heisig, P., 1999. Impact of gyrA and parC mutations on quinolone resistance, doubling time, and supercoiling degree of Escherichia coli. Antimicrob. Agents Chemother. 43, 868-875. Baucheron, S., Imberechts, H., Chaslus-Dancla, E., Cloeckaert, A., 2002. The AcrB multidrug transporter plays a major role in high-level fluoroquinolone resistance in Salmonella enterica serovar Typhimurium phage type DT204. Microb. Drug Resist. 8, 281-289. Boothe, D.M., Boeckh, A., Simpson, R.B., Dubose, K., 2006. Comparison of pharmacodynamic and pharmacokinetic indices of efficacy for 5 fluoroquinolones toward pathogens of dogs and cats. J. Vet. Intern. Med. 20, 1297-306. Cattoir, V., Nordmann, P. 2009. Plasmid-mediated quinolone resistance in gram-negative bacterial species: an update. Curr. Med. Chem. 16, 1028-1046. Chen, Y.M., Wright, P.J., Lee, C.S., Browning, G.F., 2003. Uropathogenic virulence factors in isolates of Escherichia coli from clinical cases of canine pyometra and feces of healthy bitches. Vet. Microbiol. 94, 57-69. Clinical and Laboratory Standards Institute, 2008. Performance standards for antimicrobial susceptibility testing; 18th informational supplement. CLSI document M100-S18 Wayne, PA. Cohn, L.A., Gary, A.T., Fales, W.H., Madsen, R.W., 2003. Trends in fluoroquinolone resistance of bacteria isolated from canine urinary tracts. J. Vet. Diagn. Invest. 15, 338-343. 116 Conrad, S., Oethinger, M., Kaifel, K., Klotz, G., Marre, R., Kern, W.V., 1996. gyrA mutations in high-level fluoroquinolone-resistant clinical isolates of Escherichia coli. J. Antimicrob. Chemother. 38, 443-455. Cooke, C.L., Singer, R.S., Jang, S.S. Hirsh, D.C., 2002. Enrofloxacin resistance in Escherichia coli isolated from dogs with urinary tract infections. J. Am. Vet. Med. Assoc. 220, 190-192. Everett, M.J., Jin, Y.F., Ricci, V., Piddock, L.J., 1996. Contributions of individual mechanisms to fluoroquinolone resistance in 36 Escherichia coli strains isolated from humans and animals. Antimicrob. Agents Chemother. 40, 2380-2386. Fernandes, P., Ferreira, B.S., Cabral, J.M., 2003. Solvent tolerance in bacteria: role of efflux pumps and cross-resistance with antibiotics. Int. J. Antimicrob. Agents. 22, 211-216. Giraud, E., Leroy-S?trin, S., Flaujac, G., Cloeckaert, A., Dho-Moulin, M., Chaslus- Dancla, E., 2001. Characterization of high-level fluoroquinolone resistance in Escherichia coli O78:K80 isolated from turkeys. J. Antimicrob. Chemother. 47, 341-343. Goettsch, W., van Pelt, W., Nagelkerke, N., Hendrix, M.G., Buiting, A.G., Petit, P.L., Sabbe, L.J., van Griethuysen, A.J., de Neeling, A.J. 2000. Increasing resistance to fluoroquinolones in Escherichia coli from urinary tract infections in the Netherlands. J. Antimicrob. Chemother. 46, 223?228. Guardabassi, L., Schwarz, S., Lloyd, D.H., 2004. Pet animals as reservoirs of antimicrobial-resistant bacteria. J. Antimicrob. Chemother. 54, 321-332. 117 Hagman, R., K?hn, I., 2002. Escherichia coli strains isolated from the uterus and urinary bladder of bitches suffering from pyometra: comparison by restriction enzyme digestion and pulsed-field gel electrophoresis. Vet. Microbiol. 84, 143-153. Heisig, P., 1996. Genetic evidence for a role of parC mutations in development of high- level fluoroquinolone resistance in Escherichia coli. Antimicrob. Agents Chemother. 40, 879-885. Heisig, P., Schedletzky, H., Falkenstein-Paul, H., 1993. Mutations in the gyrA gene of a highly fluoroquinolone-resistant clinical isolate of Escherichia coli. Antimicrob. Agents Chemother. 37, 696-701. Karlowsky, J.A., Hoban, D.J., Decorby, M.R., Laing, N.M., Zhanel, G.G., 2006. Fluoroquinolone-resistant urinary isolates of Escherichia coli from outpatients are frequently multidrug resistant: results from the North American Urinary Tract Infection Collaborative Alliance-Quinolone Resistance study. Antimicrob. Agents Chemother. 50, 2251-2254. Kern, W.V., Oethinger, M., Jellen-Ritter, A.S., Levy, S.B., 2000. Non-target gene mutations in the development of fluoroquinolone resistance in Escherichia coli. Antimicrob. Agents Chemother. 44, 814-820. Kern, W.V., Steinke, P., Schumacher, A., Schuster, S., von Baum, H., Bohnert, J.A., 2006. Effect of 1-(1-naphthylmethyl)-piperazine, a novel putative efflux pump inhibitor, on antimicrobial drug susceptibility in clinical isolates of Escherichia coli. J. Antimicrob. Chemother. 57, 339-343. Kriengkauykiat, J., Porter, E., Lomovskaya, O., Wong-Beringer, A., 2005. Uses of an efflux pump inhibitor to determine the prevalence of efflux pump-mediated 118 fluoroquinolone resistance and multidrug resistance in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 49, 565-570. Lanz, R., Kuhnert, P., Boerlin, P., 2003. Antimicrobial resistance and resistance gene determinants in clinical Escherichia coli from different animal species in Switzerland. Vet. Microbiol. 91, 73-84. Li, Y., Wang, C., Allen, K.E., Little, S.E., Ahluwalia, S.K., Gao, D., Macintire, D.K., Blagburn, B.L., Kaltenboeck, B., 2008. Diagnosis of canine Hepatozoon spp. infection by quantitative PCR. Vet. Parasitol. 157, 50-58. Lomovskaya, O., Lee, A., Hoshino, K., Ishida, H., Mistry, A., Warren, M.S., Boyer, E., Chamberland, S., Lee, V.J., 1999. Use of a genetic approach to evaluate the consequences of inhibition of efflux pumps in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 43, 1340-1346. Lomovskaya, O., Warren, M.S., Lee, A., Galazzo, J., Fronko, R., Lee, M., Blais, J., Cho, D., Chamberland, S., Renau, T., Leger, R., Hecker, S., Watkins, W., Hoshino, K., Ishida, H., Lee, V.J., 2001. Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. Agents Chemother. 45, 105-116. Lomovskaya, O., Watkins, W., 2001. Inhibition of efflux pumps as a novel approach to combat drug resistance in bacteria. J. Mol. Microbiol. Biotechnol. 3, 225-236. Luo, N., Sahin, O., Lin, J., Michel, L.O., Zhang, Q., 2003. In vivo selection of Campylobacter isolates with high levels of fluoroquinolone resistance associated with gyrA mutations and the function of the CmeABC efflux pump. Antimicrob. Agents Chemother. 47, 390-4. 119 Ma, D., Cook, D.N., Alberti, M., Pon, N.G., Nikaido, H., Hearst, J.E., 1995. Genes acrA and acrB encode a stress-induced efflux system of Escherichia coli. Mol. Microbiol. 16, 45-55. Mazzariol, A., Tokue, Y., Kanegawa, T.M., Cornaglia, G., Nikaido, H., 2000. High-level fluoroquinolone-resistant clinical isolates of Escherichia coli overproduce multidrug efflux protein AcrA. Antimicrob. Agents Chemother. 44, 3441-3443. Mazzariol, A., Zuliani, J., Cornaglia, G., Rossolini, G.M., Fontana, R., 2002. AcrAB efflux system: expression and contribution to fluoroquinolone resistance in Klebsiella spp. Antimicrob. Agents Chemother. 46, 3984-3986. Meunier, D., Acar, J.F., Martel, J.L., Kroemer, S., Valle, M., 2004. A seven-year survey of susceptibility to marbofloxacin of pathogenic strains isolated from pets. Int. J. Antimicrob. Agents. 24, 592-598. Nishino, K., Yamaguchi, A., 2001. Analysis of the complete library of putative drug transporter genes in Escherichia coli. J. Bacteriol. 183, 5803-5812. Oethinger, M., Kern, W.V., Jellen-Ritter, A.S., McMurry, L.M., Levy, S.B., 2000. Ineffectiveness of topoisomerase mutations in mediating clinically significant fluoroquinolone resistance in Escherichia coli in the absence of the AcrAB efflux pump. Antimicrob. Agents Chemother. 44, 10-13. Oram, M., Fisher, L.M., 1991. 4-Quinolone resistance mutations in the DNA gyrase of Escherichia coli clinical isolates identified by using the polymerase chain reaction. Antimicrob. Agents Chemother. 35, 387-389. 120 Park, C.H., Robicsek, A., Jacoby, G.A., Sahm, D., Hooper, D.C. 2006. Prevalence in the United States of aac(6')-Ib-cr encoding a ciprofloxacin-modifying enzyme. Antimicrob. Agents. Chemother. 50:3953-3955. Piddock, L.J., 1999. Mechanisms of fluoroquinolone resistance: an update 1994?1998. Drugs. 58 Suppl 2, 11-18. Piddock, L.J., 2006. Clinically relevant chromosomally encoded multidrug resistance efflux pumps in bacteria. Clin. Microbiol. Rev. 19, 382-402. Renau, T.E., L?ger, R., Flamme, E.M., Sangalang, J., She, M.W., Yen, R., Gannon, C.L., Griffith, D., Chamberland, S., Lomovskaya, O., Hecker, S.J., Lee, V.J., Ohta, T., Nakayama, K., 1999. Inhibitors of efflux pumps in Pseudomonas aeruginosa potentiate the activity of the fluoroquinolone antibacterial levofloxacin. J. Med. Chem. 42, 4928-4931. Robicsek, A., Jacoby, G.A., Hooper, D.C. 2006a. The worldwide emergence of plasmid- mediated quinolone resistance. Lancet. Infect. Dis. 6, 629-640. Robicsek, A., Strahilevitz, J., Jacoby, G.A., Macielag, M., Abbanat, D., Park, C.H., Bush, K., Hooper, D.C. 2006b. Fluoroquinolone-modifying enzyme: a new adaptation of a common aminoglycoside acetyltransferase. Nat. Med. 12, 83-88. Rosenberg, E.Y., Ma, D., Nikaido, H., 2000. AcrD of Escherichia coli is an aminoglycoside efflux pump. J. Bacteriol. 182, 1754-1756. S?enz, Y., Ruiz, J., Zarazaga, M., Teixid?, M., Torres, C., Vila, J., 2004. Effect of the efflux pump inhibitor Phe-Arg-beta-naphthylamide on the MIC values of the quinolones tetracycline and chloramphenicol in Escherichia coli isolates of different origin. J. Antimicrob. Chemother. 53, 544-545. 121 S?enz, Y., Zarazaga, M., Bri?as, L., Ruiz-Larrea, F., Torres, C., 2003. Mutations in gyrA and parC genes in nalidixic acid-resistant Escherichia coli strains from food products, humans and animals. J. Antimicrob. Chemother. 51, 1001-1005. Sanchez, S., McCrackin Stevenson, M.A., Hudson, C.R., Maier, M., Buffington, T., Dam, Q., Maurer, J.J., 2002. Characterization of multidrug-resistant Escherichia coli isolates associated with nosocomial infections in dogs. J. Clin. Microbiol. 40, 3586-3595. Shaheen, B.W., Boothe, D.M., Oyarzabal, O.A., Smaha, T., 2008. Characterization of clinical Escherichia coli isolates expressing multidrug resistance recovered from canine and feline with spontaneous disease. In: Congress Proceedings of ACVIM Forum, San Antonio, TX, June 4 ? 7. p.786. Sulavik, M.C., Houseweart, C., Cramer, C., Jiwani, N., Murgolo, N., Greene, J., DiDomenico, B., Shaw, K.J., Miller, G.H., Hare, R., Shimer, G., 2001. Antibiotic susceptibility profiles of Escherichia coli strains lacking multidrug efflux pump genes. Antimicrob. Agents Chemother. 45, 1126-1136. Talan, D.A., Krishnadasan, A., Abrahamian, F.M., Stamm, W.E., Moran, G.J.; EMERGEncy ID NET Study Group. 2008. Prevalence and risk factor analysis of trimethoprim-sulfamethoxazole- and fluoroquinolone-resistant Escherichia coli infection among emergency department patients with pyelonephritis. Clin. Infect. Dis. 47, 1150-1158. Thanassi, D.G., Cheng, L.W., Nikaido, H., 1997. Active efflux of bile salts by Escherichia coli. J. Bacteriol. 179, 2512-2518. 122 Vila, J., Ruiz, J., Go?i, P., De Anta, M.T., 1996. Detection of mutations in parC in quinolone-resistant clinical isolates of Escherichia coli. Antimicrob. Agents Chemother. 40, 491-493. Vila, J., Ruiz, J., Go?i, P., De Anta, M.T., 1996. Detection of mutations in parC in quinolone-resistant clinical isolates of Escherichia coli. Antimicrob. Agents Chemother. 40, 491-493. Vila, J., Ruiz, J., Marco, F., Barcelo, A., Go?i, P., Giralt, E., Jimenez de Anta, T., 1994. Association between double mutation in gyrA gene of ciprofloxacin- resistant clinical isolates of Escherichia coli and MICs. Antimicrob. Agents Chemother. 38, 2477-2479. Wang, C., Gao, D., Vaglenov, A., Kaltenboeck, B., 2004. One-step real-time duplex reverse transcription PCRs simultaneously quantify analyte and housekeeping gene mRNAs. Biotechniques. 36, 508-516, 518-519. Wernicki, A., Krzyzanowski, J., Puchalski, A., 2002. Characterization of Escherichia coli strains associated with canine pyometra. Pol. J. Vet. Sci. 5, 51-56 Willmott, C.J., Maxwell, A., 1993. A single point mutation in the DNA gyrase A protein greatly reduces binding of fluoroquinolones to the gyrase-DNA complex. Antimicrob. Agents Chemother. 37, 126-127. Yamane, K., Wachino, J., Suzuki, S., Kimura, K., Shibata, N., Kato, H., Shibayama, K., Konda, T., Arakawa, Y. 2007. New plasmid-mediated fluoroquinolone efflux pump, QepA, found in an Escherichia coli clinical isolate. Antimicrob. Agents Chemother. 51, 3354-3360. Yoshida, H., Bogaki, M., Nakamura, M., Nakamura, S., 1990. Quinolone resistance- 123 determining region in the DNA gyrase gyrA gene of Escherichia coli. Antimicrob. Agents Chemother. 34, 1271-1272. 124 Table 1. Antimicrobial drugs used in this study Drug class Antimicrobial drugs MIC BP (S/R) a MIC range b Beta lactams Ampicillin ?8, ?32 0.5?256 Penicillins Amoxicillin/clavulanic acid ?8/4, ?32/16 0.25?512 Meropenem ?4, ?16 0.25?512 Ticarcillin/ clavulanic acid ?16/2, ?128/2 2?2048 Cephalosporins Cefpodoxime ?2, ?8 0.12?128 Cefotaxime ?8, ?64 1?1024 Ceftazidime ?8, ?32 0.5?512 Fluoroquinolones Enrofloxacin ?0.5, ?4 0.06?128 Tetracycline Doxycycline ?4, ?16 0.25?512 Phenicols Chloramphenicol ?8, ?32 0.5?1024 Macrolides Azithromycin ?0.5 , ?8 0.12?256 Aminoglycosides Gentamicin ?4, ?8 0.12?256 Potentiated sulfa Trimethoprim/ sulfamethoxazole ?2, ?8 0.06?128 MIC, minimum inhibitory concentration; BP, break point; S, susceptible break point; R, resistant break point. a The MIC beak points were designated according to CLSI guidelines and interpretive standards (CLSI, 2008). b The MIC ranges for custom-made micodilution plates. 125 Table 2. Primers and probes used in FRET PCR assay and in the melting curve analysis a The positions of the oliginucleotides correspond to those in GenBank accession number X06373. b Tm, Theoretical melting temperature. c 6-FAM, 6-Carboxyfluorescein. Primer or probe Oligonucleotide sequence (5? ? 3?) Position a Tm (?C) b Upstream primer CCATGAACGTACTAGGCAATGACTG 152-177 57 Down primer TTTTCCGTGCCGTCATAGTTATCAA 256-427 58 Fluorescein GTTGGTGACGTAATCGGTAAATACCA TCCCC-(6-FAM) c 208-238 66 Bodipy 630/650 (Bodipy-630/650)- TGGTGACTCGGCGGTTTATGACACGA -(Phosphate) 240-265 66.8 126 Table 3. Antimicrobial resistance pattern of fluoroquinlone resistant and susceptible isolates Isolates with resistance to a R Cases (FQ-R); n=76 (%) Controls (Non FQ-R); n=305 (%) One class only 1 (1.3) 207 (68) Two classes 13 (17) 63 (20.6) Three classes 12 (15.8) 27 (8.8) Four classes 25 (33) 5 (1.6) Five classes 18 (23.7) 3 (1) Six classes 7 (9.2) 0 (0) MDR phenotype 75 (98.7) 98 (32) FQ-R, fluoroquinolone-resistant isolates; FQ-S, isolates resistant to all but FQ; MDR, multidrug resistance (resistance to two or more classes of antimicrobial agents). a Each of the following represents a ?unique? class for the purpose of study analysis: enrofloxacin/ciprofloxacin, ampicillin/amoxicillin-clavulanic acid/cephalothin/cepfoxitin/cefpodoxime/cefotaxime/ceftazidime/meropenem/tricarcillin- clavulanic acid, doxycycline, chloramphenicol, gentamicin, Trimethoprim/sulfamethoxazole. 127 Table 4. Characteristics of Escherichia coli isolated from dogs and cats used in this study QRDR Phenotypes of Enrofloxacin MIC a Magnitude of Species Source of gyrA Isolates Without inhibitor With inhibitor MIC decrease Urine D87G MDR13 0.5 ? 0.06 ?8 Vagina S83L MDR26 0.25 ? 0.06 ?4 Urine ENR 1 ? 0.06 ?16 ENR, MDR1234 2 ? 0.06 ?32 MDR12346 4 ? 0.06 ?67 S83L, D87N MDR12456 8 1 8 MDR1245 16 2 8 MDR124 16 1 16 MDR12346 32 4 8 MDR146 32 8 4 Tracheal wash MDR13456 32 4 8 Urine MDR1245, MDR12346, MDR13456, MDR12345 64 4 16 MDR12456 64 8 8 Abdomin al fluid MDR12456 64 4 16 MDR2345 64 8 8 Fracture MDR14 64 8 8 Urine MDR12456 64 4 16 MDR1246 128 8 16 MDR124 128 16 8 MDR123456, MDR124 128 4 32 MDR123456 >128 8 ?32 MDR12346, MDR12345, MDR123456 >128 8 ?16 Canine MDR1234 >128 16 ?8 128 Tracheal wash MDR123456 >128 8 ?16 Urine Wild NR 0.12 ? 0.06 ?2 BL 0.12 0.12 0 BL, MDR123 0.25 ? 0.06 ?4 MDR126, MDR12356 0.5 0.12 4 Pleural fluid BL 0.5 ? 0.06 ?8 Vagina MDR125 0.5 ? 0.06 ?8 Urine ENR 1 ? 0.06 ?16 MDR1234 16 ? 0.06 ?267 MDR12345 32 ? 0.06 ?530 Urine S83L MDR1234 2 ? 0.06 ?32 S83L, D87N MDR12456, MDR1246 32 4 8 MDR124, MDR1456 64 8 8 MDR1246 64 4 16 MDR1234 >128 8 16 MDR1245 >128 16 ?8 Wild NR 0.12 ? 0.06 ?2 BL, NR 0.5 ? 0.06 ?8 MDR1234 4 ? 0.06 ?67 Feline Nasal cavity MDR123456 >128 ? 0.06 >2130 QRDR, quinolone resistance-determining region; MIC, minimum inhibitory concentration; MDR, multidrug resistance (resistance to two or more classes of antimicrobial agents: 1, beta lactams; 2, tetracycline; 3, phenicols; 4, fluoroquinolone; 5, potentiated sulfa; 6, aminoglycosides). a The enrofloxacin MIC was determined in the absence and presence of 100 mg of the efflux inhibitor Phe-Arg-?-naphthylamide per L. 129 Figure 1. Comparison of the frequency of resistance pattern of FQ-R versus FQ-S isolates to different drug classes. P < 0.05 for all comparisons except for beta lactam drugs. FQ- R, fluoroquinolone-resistant isolates; FQ-S, isolates resistant to all but FQ; MDR, multidrug resistance (resistance to two or more classes of antimicrobial agents: BL, beta lactam drugs (i.e. ampicillin/amoxicillin-clavulanic acid/cephalothin/cepfoxitin/cefpodoxime/cefotaxime/ceftazidime/meropenem/ tricarcillin-clavulanic acid; DOC, doxycyline; CHL, chloramphenicol; GEN, gentamicin; SMX/TMP, trimethoprim/sulfamethoxazole. 130 Figure 2. Effect of gyrA mutations and EPI on the in vitro MICs of enrofloxacin for 56 clinical E. coli isolates isolated from dogs and cats. With EPI, the median MIC decreased from 0.25 and 2 to ? 0.06 mg/L for isolates with wild type and single mutation in gyrA (i.e. S83L or D87G), respectively. EPI decreased the median MIC for isolates with double mutation in gyrA (i.e. S83 and D87N) by 16 fold, however the MIC is still above the resistant breakpoint (i.e. MIC= 4 mg/L). Interestingly, MIC restored the susceptibility (i.e. ? 0.06 mg/L) for isolates (n =4) that had high MICs (i.e. 4- >128 mg/L) and carried no mutations in gyrA. EPI, efflux pump inhibitor; MIC, minimum inhibitory concentration. 131 Figure 3. The effect of EPI on the susceptibilities of different antimicrobial agents. Cumulative susceptibility to azithromycin (a), doxycyline (b), chloramphenicol (c), and gentamicin (d) alone or in combination with EPI for clinical E. coli isolates from dogs and cats. MIC 90 and MIC 50 are represented by thick dashed line (for isolates treated with EPI) or dotted lines (no EPI treatment). MICs in square and triangle represent the susceptible and resistance breakpoints, respectively, and were designated according to CLSI guidelines and interpretive standards (CLSI, 2008). EPI, efflux pump inhibitor; MIC, minimum inhibitory concentration. 132 CHAPTER V THE ROLE OF CLASS 1 AND 2 INTEGRONS IN MEDIATING ANTIMICROBIAL RESISTANCE AMONG CANINE AND FELINE CLINICAL E. COLI ISOLATES FROM THE US Abstract Integrons are mobile genetic elements that incorporate an open reading frame or gene cassette and have an important role in the acquisition and dissemination of antimicrobial resistance genes. Yet their role in the transmission of resistance in bacterial pathogens of pet animals is seldom addressed. The purpose of this study was to describe the incidence of class I and 2 integrons in Escherichia coli pathogens (n = 82) isolated from dogs and cats with spontaneous infections. The relationship of integrons to antimicrobial resistance was described. Integrons were detected using PCR and then further characterized by restriction fragment length polymorphism (RFLP) analysis and amplicon sequencing. Class 1 was detected in 27% of the total isolates, while only 2.4% (n = 2) of the isolates were positive for the presence of class 2 integrons. Integrons negative isolates represent 72% (n= 59) among selected isolates. Eleven gene cassettes were found either alone or in combination with other gene cassettes, which encoded resistance to aminoglycosides (aadA1, aadA2, aadA5, aacA4, and aadB), trimethoprim (dfrA1, dhfrA17, and dfrA12), chloramphenicol (catB3 and cmlA6), and streptothricin 133 (sat1), respectively. The gene cassette aadA1 (65.2%, n= 15/23) was detected most frequently in class 1 integrons. All integron positive isolates were characterized by resistant to least two drug classes. The role of integrons in mediating resistance was confirmed by curing experiments for three isolates for which resistance was resolved once large plasmids (size range 97-169 kb) carrying the class I integron were lost. To our knowledge, this is the first report of class 2 integrons and antibiotic resistance gene cassettes found in E. coli from companion animals in US. Introduction E. coli is a common cause of canine and feline urinary tract infection and pyometra (Hagman and K?hn, 2002; Chen et al., 2003). Several antimicrobial agents have been used historically to treat these infections in veterinary medicine, including penicillins, cephalosporins, macrolides, lincosamides, fusidic acid, tetracyclines, chloramphenicol, potentiated sulphonamides, aminoglycosides and fluoroquinolones. The use of antimicrobial drugs has been associated with an increase in the trend of antimicrobial resistance among canine and feline clinical E. coli isolates in the last decade (Normand et al., 2000b). In addition to the incidence of resistance, the type of resistance expressed by E. coli also is changing, with more isolates expressing multi-drug resistance (MDR) in the US and Europe (Normand et al., 2000a; Cohn et al., 2003; Guardabassi et al., 2004; Shaheen et al., 2008). The dissemination of MDR phenotypes among E. coli isolates may occur as a consequence of horizontal gene transfer mediated by mobile genetic elements, which 134 include plasmids, transposons, genomic islands, integrons, and natural transformation (Osborn and B?ltner, 2002). Integrons play an essential role in facilitating the transit of the resistance genes and contributing to the creation of MDR-phenotypes (Hall and Collis, 1998; Rowe-Magnus and Mazel, 2002). Integrons incorporate gene cassettes by site specific recombination events mediated by integrase genes, which flank the 5? and 3? conserved segments (Collis et al., 1993). In addition to the gene (intI), the integron structure has a primary recombination site (attI site) at the 5?CS and qacE?I and sulI at the 3?CS region (Recchia and Hall, 1997; Stokes et al., 1997). Gene (intI) encodes for the integrase enzyme that catalyzes the recombination event that involves the primary recombination site (attI) as well as a secondary target site (attC site or 59 base element), which is eventually found associated with gene cassette. To date, five classes of integrons (Classes I-V) have been identified based on the character of the integrase gene associated with the cassette. Class 1 integrons are most commonly found among multi-drug resistance in gram positive and negative bacteria (Rowe-Magnus and Mazel, 2002; Fluit and Schmitz, 2004; Mazel, 2006) and are found associated with the Tn21 transposon family (Rowe-Magnus and Mazel, 2002; Mazel, 2006). Additionally, integrons have been also identified and characterized in bacteria from environmental samples (Rosser and Young, 1999; Petersen et al., 2000). More than 80 gene cassettes are associated with class 1 integrons (Mazel, 2006). Several studies have been conducted in many countries to document and characterize the emergence of class 1 and 2 integrons among E. coli clinical samples isolated from dogs and cats (Lanz et al., 2003; Sidjabat et al., 2006; Cocchi et al., 2007; 135 Kadlec and Schwarz, 2008). Only 22% and 15% of the isolates have been documented to carry class 1 integrons among clinical sample of E. coli from Switzerland and Australia, respectively (Sidjabat et al., 2006; Cocchi et al., 2007). However, there is a little information regarding the acquisition and dissemination of antimicrobial resistance genes among canine and feline clinical E. coli isolates in the US. One study indicated that 30 of 34 nosocomial isolates harbored the intI1 integrase gene and those isolates were resistant to more than one drug (Sanchez et al., 2002). The dissemination of integrons has been attributed to the extensive use of antimicrobials, particularly in production animals (Cocchi et al., 2007). The judicious use of antimicrobials in dogs and cats is incumbent upon understanding the epidemiology of emerging E. coli resistance. The purpose of this study is to define the role of integrons and their associated gene cassettes in mediating antimicrobial resistance in canine and feline pathogenic E. coli. In this study, we build on previous investigators? findings (Grape et al., 2005; van Essen-Zandbergen et al., 2007) by describing the distribution of class 1 and 2 integrons among clinical E. coli isolates associated with disease in dogs and cats, and associating the different the types of integrons with their antimicrobial resistant gene cassettes Materinals and methods Bacterial isolates and culture conditions E. coli isolates were collected from dogs or cats with spontaneous infection between May and September 2008. All isolates had been submitted to the contributing 136 laboratory by veterinary practitioners. The isolates were collected throughout the US. Each isolate was identified as E. coli by the submitting laboratory prior to submission to the Clinical Pharmacology Laboratory (CPL) at Auburn University. Upon receipt by the CPL, isolates were plated on CHROMagar? Orientation (Becton, Dickinson, Franklin Lakes, NJ) to allow for rapid confirmation. Susceptibility testing Isolates also were streaked on trypticase agar (Becton Dickinson, Franklin Lakes, NJ) and then incubated at 35?C for 18-24 h. For standard susceptibility testing, cells were harvested and suspended in 4.5 ml of 0.9% normal saline and adjusted to 0.5 McFarland standard turbidity (~10 8 CFU) using SENSITITER ? Nephelometer (TREK Diagnostic Systems, Cleveland, OH) before testing. Antimicrobial susceptibility testing was performed using custom made broth micro-dilution susceptibility plates (TREK Diagnostic Systems, Cleveland, OH) according to Clinical and Laboratory Standards Institute (CLSI) guidelines and interpretive standards (CLSI, 2008). Panels of 13 antimicrobial agents were tested (Table 1) including ampicillin, amoxicillin/clavulanic acid, cephalothin, cepfoxitin, cefpodoxime, cefotaxime, ceftazidime, meropenem, enrofloxacin, gentamicin, doxycycline, chloramphenicol, and trimethoprim/sulfamethoxazole. Microdilution plates were incubated at 35 o C for 20-24 hours. The MIC values were recorded using the SENSITITER ? VIZION system (TREK Diagnostic Systems, Cleveland, OH). For quality control purposes, E. coli ATCC ? 25922 (American Tissue Cell Culture, Manassas, VA) and Enterococcus fecalis ATCC ? 29212 (Manassas, VA) were used. Eighty-two of E. coli isolates were randomly selected to 137 represent different phenotypes: single drug resistance (SDR) Phenotype to beta lactams (n=16); SDR phenotype to chloramphenicol (n=3); susceptible or non resistance (NR) phenotypes (n=14); and MDR phenotypes (n=47). Bacterial DNA preparation, PCR assays and DNA sequencing Bacterial DNA for PCR reaction was prepared by boiling bacterial cultures in 200 ?l of PreMan? Ultra Sample Preparation Reagent (Applied Biosystem, Foster City, CA) for 10 min, followed by centrifugation. The variable regions (i.e. 5?CS and 3?CS) and integrase genes (i.e. Int1 and IntII) of class 1 and 2 integrons were amplified with a PCR reaction using primers listed in Table 2. All primers were synthesized by Operon, Inc. (Huntsville, AL). Amplification reactions were carried out in 25 ?l reaction of Omnimix (Cepheid) including 3U TaKaRa hot start Taq polymerase, 200 ?M dNTP 4 mM MgCl2 25 mM HEPES buffer pH 8.0? 0.1, 1 ul of DNA preparation, and 1 ul of the forward and reverse primers. Distilled water was added to bring the final volume to 25 ?l. The PCR conditions for class 1 and 2 integrons are delineated in Table 2. The reaction products were analyzed by electrophoresis in agarose gels stained with ethidium bromide, and visualized under UV light. The image was recorded using a gel imaging system (Syngene, Frederick, MA, USA). For each set of PCR reactions, serovar Typhimurium DT104 was included as a positive control for class 1 integrons. 138 Characterization of class 1 and 2 integrons by sequencing and restriction fragment? length polymorphism (RFLP) RFLP was used to differentiate different integrons fragment similar in size and based on restriction patterns generated by restriction enzymes. One representative from each RFLP type (i.e. having similar restriction patterns on the RFLP) was sequenced, while unique integrons with different size were sequenced as well. The PCR amplicons were purified with a QIAquick PCR Purification Kit (Qiagen, Valencia, Calif.) and were sequenced using ABI automatic DNA sequencer (Model 377; Perkin-Elmer) at the Genomic Sequencing Laboratory (Auburn University, Auburn, AL, USA) using the above-described forward and reverse primers, plus the internal, designed primers. DNA sequences were analyzed and compared to published sequences by searching the GenBank database of the National Center for Biotechnology Information via the BLAST network service (http://www.ncbi.nlm.nih.gov/blast/). Typing of class 1 and class 2 integrons were performed by RFLP. For RFLP analysis, the purified PCR amplicons of variable region of class 1 and 2 integrons were digested using AluI and HaeIII restriction endonucleases, respectively, as described (Machado et al., 2005; Machado et al., 2008). Digestions were performed according to the manufacturer's instructions. Plasmid extraction was carried out on integrons positive isolates using the mini- scale alkaline lysis method using the Qiagen miniprep with some modification. Pulsed- field gel electrophoresis (PFGE) was performed to determine DNA fingerprinting profiles of plasmids for the cured and non cured isolates and to estimate the size of the plasmids 139 that may carried by class 1 integrons. Plasmid samples (17 ?l) were loaded onto a 1% agarose-Tris buffer gel (SeaKem Gold Agarose; BioWhittaker Molecular Applications, Rockland, ME), and PFGE was performed with a CHEF Mapper XA apparatus (CHEF Mapper XA; Bio-Rad Laboratories, Hercules, CA). DNA was electrophoresed for 9 hours at a constant voltage of 200 V (6 V/cm), with a pulse time of 6.75 to 21.7 s, an electric field angle of 120 o , and a temperature of 14?C, before being stained with ethidium bromide. Plasmid-curing experiments Curing experiments were performed for all integron positive isolates (n=23) using natural curing. Curing was carried out by overnight growth at lower temperature (i.e. 10?C) in trypticase soy agar (TSA) slant without antibiotics. Subcultures were done on TSA slant every 24 hours and the colonies were tested for antimicrobial susceptibility as previously described. Because curing was unsuccessful despite several weeks of subculture, the sodium dodecyl sulfate (SDS) method of curing was implemented on 11 isolates representing both class I and 2 integrons (Tomoeda et al., 1968). An overnight culture of E. coli was diluted in Brucella broth (with 10% SDS) and was shaken at 37 C or 40 C for 72 h. Cells were checked for their susceptibility to all drug classes. Only 3 out of the 11 isolates loss their plasmid by curing. The MICs for the cured cells were compared with the MIC before curing to all five drug classes to investigate the contribution of the integrons into the emergence of MDR phenotypes. 140 Statistical analysis Susceptibility data were compared from integron-positive and -negative E. coli. A statistical comparison of the frequencies and the association of integron presence in E. coli was conducted by using odds ratios and 95% confidence intervals. Significance (P- value), in terms of the number of resistant and susceptible integron-positive and -negative isolates, was calculated by the Pearson ? 2 -square test (P < 0.05). Analyses were performed using the statistical software MINITAB ? 15 package (Minitab Inc., PA, USA). Results Presence of integrons and the association between integrons and antimicrobial resistance Of the 82 E. coli isolates tested, 27% (n= 22) carried class 1 or 2 integrons. Class 2 integrons where found only in 2.4% (n= 2) isolates. One of these isolates also carried a class 1 integrons. Of the resistant phenotypes (n=65), 34% (n= 22) were positive for either a class 1 and 2 integrons. The association between antimicrobial resistant profile and occurrence of integrons is shown in Figure 1 and Table 3. All integron positive isolates were characterized by resistance to two or more of antimicrobial agents compared with 56.6% (34/60) for integron negative isolates. All four isolates resistant to all 13 antimicrobial agents carried a class 1 integron (Figure 1). A significant association was found between the presence of integrons and resistance to cefotaxime, ceftazidime, meropenem, and trimethoprim-sulfamethoxazole (Table 3). 141 The association between integrons positive isolates and MIC Among beta lactams, the percentages of resistance were high for the integron positive group; 62%, 77%, 87%, and 83% were resistant to amoxicillin/clavulanic acid, cefotaxime, ceftazidime, and meropenem, respectively, compared with 38%, 23%, 12.5% and 17%, respectively for integron-negative group (Table 3). No differences were detected in the proportion of integron-positive and integron-negative isolates expressing resistance to ampicillin. However, integron-positive isolates were associated with a lower proportion of resistance for cephalothin compared to integron-negative isolates (i.e. 35% versus 65%), respectively. At least eight fold differences in the MIC 90s were observed for beta lactams between integron-positive and integron-negative isolates except for ampicillin. The MIC 90s for integron-positive isolates exceeded the MIC breakpoint for all other drug classes (data not shown). In contrast, for integron negative isolates, the MICs were below the susceptible break points (MIC 90 = 4 ?g/ml) for both gentamicin and trimethoprim/sulfamethoxazole. Furthermore, the percentage of isolates resistant to gentamicin and trimethoprim/sulfamethoxazole among integron-positive isolates was significantly high with 65% and 68%, respectively, compared with 35% and 32%, respectively for integron-negative group. Characterization of gene cassettes associated with class 1 and 2 itegrons A total of 11 gene cassettes (Table 4) were identified by PCR amplification of the conserved regions of class 1 integron. The size of PCR product varied from 1 kb to 142 3.2 kb. The integrons were found to contain one to three gene cassettes and to differ in the orders of these gene cassettes. RFLP characterization of class 1 revealed 7 distinct profiles of gene cassette arrays. These were aadA1, dfrA1-aadA1, dfrA17-aadA5, dfrA12- aadA2, aadB-aadA1d, aacA4-catB3-dfrA1and aadB-aadA1-cmlA6. Among these profiles, the gene cassette aadA1 was found most frequently (65.2%, n= 15/23) in class 1 integrons either alone or in combination with other gene cassettes. These gene cassettes encoded for resistance to certain antimicrobial agents, including aadA1, aadA2, and aadA5 (which confers resistance to streptomycin and spectinomycin), aacA4, aadB (which confers resistance to aminoglycosides), the dihydrofolate reductase gene cassettes dfrA1 and dhfrA17, dfrA12, dhfr1 (which confer resistance to trimethoprim), catB3, cmlA6 (which confers resistance to chloramphenicol) and sat1 (which confers resistance to streptothricin). The antimicrobial resistance gene cassettes found in class 1 and 2 integrons correlates with the corresponding antimicrobial phenotypes found in those isolates (Table 4). A BLAST search for the published sequence in GenBank revealed the following high similarities for the most prevalent gene cassettes found in E. coli isolates from different geographical locations and sources (Table 5). The two isolates with class 2 integrons contained the cassette array dfrA1-sat1- aadA1; the class 1 gene cassette array found in one of the isolates was aacA4-catB3- dfrA1. 143 Curing experiments for isolates carrying plasmids harboring class 1 integrons. Curing experiments of plasmids carrying class 1 integrons occurred successfully in the following 3 isolates (gene cassettes): L8055239 (aadA1), I3195059 (aacA4-catB3- dfrA1), and B5729897 (aadB-aadA1-cmlA6). Class 1 integrons were detected in the extracted plasmid DNAs among those isolates before but not after curing using PCR protocol (Figure 2). PFGE revealed that several cured isolates lost 2.3-kb, 23-kb, 97-kb, 145.5-kb, 169-kb, and 194-kb bands after curing (Figure 2). All E. coli cured isolates become susceptible to the antimicrobials for which the integron?s gene cassette array encoded resistance (Table 6). Interestingly, some isolates become susceptible to some beta lactams including ampicillin, cefotaxime, ceftazidime, and ticarcillin/clavulanic acid. None of the gene cassettes identified within the integrons from those isolates confer resistance to beta lactams. Discussion The increase in the incidence of antimicrobial resistance among canine and feline E. coli isolates has been reported in many countries. However the mechanism of acquisition and dissemination of antibiotic resistance genes in clinical E. coli isolates from dogs and cats is not well understood, particular in the US. In our study, we documented the prevalence of class 1 and 2 integrons among clinical samples of E. coli collected from dogs and cats with spontaneous infectious in the US. Additionally, different types of class 1 integrons were characterized which carried different gene cassettes. Furthermore, class 2 integrons were reported for the first time among canine 144 and feline clinical E. coli isolates in the US. To our knowledge, this is the first report that investigates the distribution and to characterization of class 1 and 2 integrons among clinical specimens of E. coli from dogs and cats throughout the US. In the present study, the incidence of class 1 integrons among canine and feline clinical samples (27%)was similar to the data found in Switzerland, in which 22% of the isolates from companion animals carried class 1 integrons (Cocchi et al., 2007). However, many studies have documented a higher proportion of class 1 integrons among E. coli clinical samples from farm animals. It was found that 52% of the strains isolated from farm animals carried class 1 integrons (Cocchi et al., 2007). Other studies indicated an even higher prevalence of class 1 integrons in E. coli from farm animals: 59% reported in calf diarrhea cases (Du et al., 2005), 63% in isolates from chicken (Bass et al., 1999), 64.4% reported in swine with diarrhea (Kang et al., 2005b), and 82% in isolates from chicken (Keyes et al., 2000). The differences between our results and those found in the other reports may reflect the variations with selection pressure in different environments (Rosser and Yound, 1999). Thus, the E. coli strains food production animals (raised for economic purposes) appear to be more frequently exposed to antimicrobial pressure compared with companion animals (Cocchi et al., 2007). In this study we have also reported a very low percentage of isolates that carried a class 2 integron (only two isolates). Although other studies have reported a higher proportions of class 2 integrons in other species, the proportion of class 2 integrons appears to be lower in comparison to class 1 integrons: 14% versus 92% (Goldstein et al., 2001), 9.5% versus 55.2 (Grape et al., 2005), and 6% versus 12% (Sunde, 2005) of the 145 isolates that were found to carry class 2 and 1, respectively. Furthermore, in contrast to class 1 integrons, for which 11 different gene cassettes have been identified, only 3 different gene cassettes have identified (i.e. dhfr1- sat1- aadA1) in class 2 integrons. Similarly the gene cassette array dhfr1- sat1- aadA1 was detected most frequently in E. coli isolates from dogs and cats from Germany (Kadlec and Schwarz, 2008) and in E. coli from animal and human sources (Kang et al., 2005a; van Essen-Zandbergen et al., 2007). Only six gene cassettes have been associated with class 2 integrons (Ram?rez et al., 2005; Mazel, 2006). Thus this reduction in the diversity and prevalence of class 2 integrons compared to class I integrons may suggest that the gene encoding integrase (IntI2), which catalyzes the recombination events between primary recombination site (attI) and a secondary target site (attC site), contains nonsense mutations in codon 179, producing a non functional pseudogene (Hansson et al., 2002; Fluit and Schmitz, 2004; Mazel, 2006). Class 1 integrons were significantly associated with resistance to all the antimicrobials tested in this study. However, integron positive strains were significantly higher among isolates with resistant to beta lactams (e.g. cefotaxime, meropenem), aminoglycosides (e.g. gentamicin) and potentiated sulfanomide (e.g. trimethoprim/sulfamethoxazole). More than 80 gene cassettes, which cover resistance to different drug classes in gram negative and positive bacteria are associated with class 1 integrons (Mazel, 2006). Our study characterized gene cassettes that confer resistance to aminoglycosides, including aadA1, aadA2 aadA5, aacA4 and aadB, and genes that confer resistance to trimethoprim, including dfrA1, dhfrA17 and dfrA12. The most common 146 detectable gene cassette is the resistant gene (Kang et al., 2005a), which encodes for an aminoglycoside adenyltransferase that confers resistance to streptomycin-spectinomycin. A previous study indicated that the presence of a single antibiotic resistance gene, aadA1, was most prevalent among E. coli isolates from veterinary sources (Goldstein et al., 2001). This antibiotic resistance gen was detected most frequently, either alone or in combination with other gene cassettes, in the E. coli isolates from dogs and cats in Germany (Kadlec and Schwarz, 2008). Other studies indicated that aadA1 was detected more frequently in the E. coli isolates from human and animal sources in Netherland (van Essen-Zandbergen et al., 2007) and from environmental sources (Moura et al., 2007). Furthermore, our study is the first to observe the presence of aacA4-catB3-dfrA1in US E. coli from companion animals. These resistance genes confer resistance to chloramphenicol The gene cassettes characterized in the present study showed high identity to those gene cassettes found in same or different bacterial species, either from animals or humans, and distributed among different geographical location. This data suggests that class 1 integrons may contribute significantly to the horizontal transfer of antimicrobial resistance genes among bacterial species from different sources or geographical locations (Chang et al., 2000; Yu et al., 2003). However, the gene cassettes observed in class 1 and class 2 integrons did not correlate with the phenotypes that were observed in canine and feline clinical E. coli isolates. It is likely that other mobile genetic elements (and mechanisms such as efflux pump or modifications of porins contribute to the emergence of multi-drug resistance in these clinical isolates (Bass et al., 1999; Piddock, 1999). 147 The process of plasmid curing was more effective when using 10% SDS compared with natural curing at low temperature. The mechanism of SDS action is not completely understood. However, their effectiveness as anionic detergents can cause disruption of cell membranes - completely or partially - for which plasmid replicons were more closely associated (attached) with it, therefore damaging the plasmid by SDS (El- Mansi et al., 2000). This leads to imperfect replication or unsuccessful segregation of plasmid. Consistent with a previous study, the curing was not completely efficient for these E. coli isolates (Tomoeda et al., 1968). It is possible that plasmid may survive by integration into the genome, even with the absence of selective pressure (El-Mansi et al., 2000). Furthermore, we found that some isolates lost other resistant determinants after curing (i.e. beta lactam drugs) which is not readily carried by class 1 integrons (Table 6). This may be attributed to losing more than the gene cassette that is different from that found in class 1 integrons which is either carried in the same or in different plasmids. The data highlights the importance of other genetic elements, besides integrons, in horizontal gene transfer of resistance genes. The present study demonstrates the emergence of class 1 and 2 integrons among clinical isolates of E. coli from dogs and cats throughout the US. To our knowledge, this is the first report of the presence of class 2 integrons among clinical E. coli isolated from canine and feline in the US. The proportion of isolates carrying class 1 and 2 integrons is not significant as found in food production animals and humans, which may be attributed to different antimicrobial pressures for isolates from companion animals. The gene cassettes characterized in the class 1 and class 2 integrons might reflect the selective 148 pressures resulting from different antimicrobial therapy, drugs used, and the regions from which the isolates originated. Therefore, this study can provide initial information for a surveillance program aimed at the prudent use of antimicrobial agents by veterinarians in the US. 149 References Bass, L., Liebert, C.A., Lee, M.D., Summers, A.O., White, D.G., Thayer, S.G., Maurer, J.J., 1999. Incidence and characterization of integrons, genetic elements mediating multiple-drug resistance, in Avian Escherichia coli. Antimicrob. Agents Chemother. 43, 2925?2929. Chang, C.Y., Chang, L.L., Chang, Y.H., Lee, T.M., Chang, S.F., 2000. Characterization of drug resistance gene cassettes associated with class 1 integrons in clinical isolates of Escherichia coli from Taiwan, ROC. J. Med. Microbiol. 49, 1097? 1102. Chen, Y.M., Wright, P.J., Lee, C.S., Browning, G.F., 2003. Uropathogenic virulence factors in isolates of Escherichia coli from clinical cases of canine pyometra and feces of healthy bitches. Vet. Microbiol. 94, 57?69. Clinical and Laboratory Standards Institute, 2008. Performance standards for antimicrobial susceptibility testing; 18th informational supplement. CLSI document M100-S18 Wayne, PA. Cocchi, S., Grasselli, E., Gutacker, M., Benagli, C., Convert, M., Piffaretti, J. C., 2007. Distribution and characterization of integrons in Escherichia coli strains of animal and human origin. FEMS. Immunol. Med. Microbiol. 50, 126?132. Cohn, L.A., Gary, A.T., Fales, W.H., Madsen R.W., 2003. Trends in fluoroquinolone resistance of bacteria isolated from canine urinary tracts. J. Vet. Diagn. Invest. 15, 338?343. Collis, C.M., Grammaticopoulos, G., Briton, J., Stokes, H.W., Hall, R.M., 1993. Site- 150 specific insertion of gene cassettes into integrons. Mol. Microbiol. 9, 41?52. Du, X., Shen, Z., Wu, B., Xia, S., Shen, J., 2005. Characterization of class 1 integrons- mediated antibiotic resistance among calf pathogenic Escherichia coli. FEMS Microbiol. Lett. 245, 295?298. El-Mansi, M., Anderson, K.J., Inche, C.A., Knowles, L.K., Platt, D.J., 2000. Isolation and curing of the Klebsiella pneumoniae large indigenous plasmid using sodium dodecyl sulphate. Res. Microbiol. 151, 201?208. Fluit, A.C., Schmitz, F.J., 2004. Resistance integrons and super-integrons. Clin. Microbiol. Infect. 10, 272?288. Goldstein, C., Lee, M.D., Sanchez, S., Hudson, C., Phillips, B., Register, B., Grady, M., Liebert, C., Summers, A.O., White, D.G., Maurer, J.J., 2001. Incidence of class 1 and 2 integrases in clinical and commensal bacteria from livestock, companion animals, and exotics. Antimicrobial. Agents. Chemother. 45, 723?726. Grape, M., Farra, A., Kronvall, G., Sundstr?om, L., 2005. Integrons and gene cassettes in clinical isolates of co-trimoxazoleresistant Gram-negative bacteria. Clin. Microbiol. Infect. 11, 185?192. Guardabassi, L., Schewarz, S., Lloyd, D.H., 2004. Pet animals as reservoirs of antimicrobial-resistant bacteria. J. Antimicrob. Chemother. 54, 321?332. Hagman, R., K?hn, I., 2002. Escherichia coli strains isolated from the uterus and urinary bladder of bitches suffering from pyometra: comparison by restriction enzyme digestion and pulsed-field gel electrophoresis. Vet. Microbiol. 84, 143?53. Hall, R.M., Collis, C.M., 1998. Antibiotic resistance in gramnegative bacteria: the role of 151 gene cassettes and integrons. Drug Resist. Updates. 1, 109?119. Hansson, K., Sundstr?m, L., Pelletier, A., Roy, P.H., 2002. IntI2 integron integrase in Tn7. J. Bacteriol. 184, 1712?1721. Kadlec, K., Schwarz, S., 2008. Analysis and distribution of class 1 and class 2 integrons and associated gene cassettes among Escherichia coli isolates from swine, horses, cats and dogs collected in the BfT-GermVet monitoring study. J. Antimicrob. Chemother. 62, 469?473. Kang H.Y., Jeong Y.S., Oh J.Y., Tae S.H., Choi C.H., Moon D.C., Lee W.K., Lee Y.C., Seol S.Y., Cho D.T., Lee J.C. 2005a. Characterization of antimicrobial resistance and class 1 integrons found in Escherichia coli isolates from humans and animals in Korea. J. Antimicrob. Chemother. 55, 639-464. Kang, S.G., Lee, D.Y., Shin, S.J., Ahn, J.M., Yoo, H.S., 2005b. Changes in patterns of antimicrobial susceptibility and class 1 integron carriage among Escherichia coli isolates. J. Vet. Sci. 6, 201?205. Keyes, K., Hudson, C., Maurer, J.J., Thayer, S., White, D.G., Lee, M.D., 2000. Detection of florfenicol resistance genes in Escherichia coli from sick chickens, Antimicrob. Agents Chemother. 44, 421?424. Lanz, R., Kuhnert, P., Boerlin, P., 2003. Antimicrobial resistance and resistance gene determinants in clinical Escherichia coli from different animal species in Switzerland. Vet. Microbiol. 91, 73?84. Levesque, C., Piche, L., Larose, C., Roy, P.H., 1995. PCR mapping of integrons reveals several novel combinations of resistance genes. Antimicrob. Agents Chemother. 152 39, 185-195. Machado, E., Cant?n, R., Baquero, F., Gal?n, J.C., Roll?n, A., Peixe, L., Coque, T.M., 2005. Integron content of extended-spectrum-?-lactamase-producing Escherichia coli strains over 12 years in a single hospital in Madrid, Spain. Antimicrob. Agents Chemother. 49, 1823?1829. Machado, E., Coque, T.M., Cant?n, R., Sousa, J.C., Peixe, L., 2008. Antibiotic resistance integrons and extended-spectrum {beta}-lactamases among Enterobacteriaceae isolates recovered from chickens and swine in Portugal. J. Antimicrob. Chemother. 62, 296-302. Mazel, D., 2006. Integrons: agents of bacterial evolution. Nat. Rev. Microbiol. 4, 608? 620. Moura, A., Henriques, I., Ribeiro, R., Correia, A., 2007. Prevalence and characterization of integrons from bacteria isolated from a slaughterhouse wastewater treatment plant. J. Antimicrob. Chemother. 60, 1243?1250. Normand, E.H., Gibson, N.R., Reid, S.W., Carmichael, S., Taylor, D.J., 2000a. Antimicrobial-resistance trends in bacterial isolates from companion-animal community practice in the UK. Prev. Vet. Med. 46, 267?278. Normand, E.H., Gibson, R.N., Taylor, D.J., Carmichael, S., Reid, S.W., 2000b. Trends of antimicrobial resistance in bacterial isolates from a small animal referral hospital. Vet. Rec. 146, 151?155. Osborn, A.M., B?ltner, D., 2002. When phage, plasmids, and transposons collide: genomic islands, and conjugative- and mobilizable-transposons as a mosaic 153 continuum. Plasmid. 48, 202?212. Petersen, A., Guardabassi, L., Dalsgaard, A., Olsen J.E., 2000. Class 1 integrons containing dhfrI trimethoprim resistance gene cassette in aquatic Acinetobacter spp. FEMS Microbiol. Lett. 182, 73?76. Piddock, L.J.V., 1999. Mechanisms of fluoroquinolone resistance: an update 1994?1998. Drugs. 58, 11?18. Ram?rez, M.S., Vargas, L.J., Cagnoni, V., Tokumoto, M., Centr?n, D., 2005. Class 2 integron with a novel cassette array in a Burkholderia cenocepacia isolate. Antimicrob. Agents. Chemother. 49, 4418?4420. Recchia, D., Hall, R.M., 1997. Origins of the mobile gene cassettes found in integrons, Trends Microbiol. 5, 389?394. Rosser, S.J., Yound, H.K., 1999. Identification and characterization of class l integrons in bacteria from an aquatic environment. J. Antimicrob. Chemother. 44, 11?18. Rowe-Magnus, D.A., Mazel, D., 2002. The role of integrons in antibiotic resistance gene capture. Int. J. Med. Microbiol. 292, 115?125. Sanchez, S., McCrackin Stevenson, M.A., Hudson, C.R., Maier, M., Buffington, T., Dam, Q., Maurer, J. J., 2002. Characterization of multidrug-resistant Escherichia coli isolates associated with nosocomial infections in dogs. J. Clin. Microbiol. 40, 3586?3595. Shaheen, B.W., Boothe, D.M., Oyarzabal, O.A., Smaha, T., 2008. Characterization of clinical Escherichia coli isolates expressing multidrug resistance recovered from canine and feline with spontaneous disease. In: Congress Proceedings of ACVIM 154 Forum, San Antonio, TX, June 4 ? 7. p.786. Sidjabat, H.E., Townsend, K.M., Lorentzen, M., Gobius, K.S., Fegan, N., Chin, J. J., Bettelheim, K.A., Hanson, N.D., Bensink, J.C., Trott, D.J., 2006. Emergence and spread of two distinct clonal groups of multidrug-resistant Escherichia coli in a veterinary teaching hospital in Australia. J. Med. Microbiol. 55, 1125?1134. Stokes, H.W., O'Gorman, D.B., Recchia, G.D., Parsekhian, M., Hall, R.M., 1997. Structure and function of 59- base element recombination sites associated with mobile gene cassettes. Mol. Microbiol. 26, 731?745. Sunde, M., 2005. Prevalence and characterization of class 1 and class 2 integrons in Escherichia coli isolated from meat and meat products of Norwegian origin. J. Antimicrob. Chemother. 56, 1019?1024. Tomoeda, M., Inuzuka, M., Kubo, N., Nakamura, S., 1968. Effective elimination of drug resistance and sex factors in Escherichia coli by sodium dodecyl sulfate. J. Bacteriol. 95, 1078-1089. van Essen-Zandbergen, A., Smith, H., Veldman, K., Mevius, D., 2007. Occurrence and characteristics of class 1, 2 and 3 integrons in Escherichia coli, Salmonella and Campylobacter spp. in the Netherlands. J. Antimicrob. Chemother. 59, 746?50. Yu, H.S., Lee, J.C., Kang, H.Y., Ro, D.W., Chung, J.Y., Jeong, Y.S., Tae, S.H., Choi, C.H., Lee, E.Y., Seol, S.Y., Lee, Y.C., Cho, D.T., 2003. Changes in gene cassettes of class 1 integrons among Escherichia coli isolates from urine specimens collected in Korea during the last two decades. J. Clin. Microbiol. 41, 5429?5433. 155 Table 1. Antimicrobial drugs used in this study Drug class Antimicrobial drugs MIC BP (S/R, mg/L) a MIC range (mg/L) b Beta lactams Ampicillin ?8, ?32 0.5?256 Penicillins Amoxicillin/ clavulanic acid ?8/4, ?32/16 0.25?512 Meropenem ?4, ?16 0.25?512 Ticarcillin/ clavulanic acid ?16/2, ?128/2 2?2048 Cephalosporins Cefpodoxime ?2, ?8 0.12?128 Cefotaxime ?8, ?64 1?1024 Ceftazidime ?8, ?32 0.5?512 Fluoroquinolones Enrofloxacin ?0.5, ?4 0.06?128 Tetracycline Doxycycline ?4, ?16 0.25?512 Phenicols Chloramphenicol ?8, ?32 0.5?1024 Macrolides Azithromycin ?0.5 , ?8 0.12?256 Aminoglycosides Gentamicin ?4, ?8 0.12?256 Potentiated sulfa Trimethoprim/ sulfamethoxazole ?2, ?8 0.06?128 MIC, minimum inhibitory concentration; BP, break point; S, susceptible break point; R, resistant break point. a The MIC beakpoints were designated according to CLSI guidelines and interpretive standards (CLSI, 2008). b The MIC ranges for custom-made micodilution plates. 156 Table 2. Primers sequences and PCR conditions used in this study Primer Oligonucleotide sequence (5' to 3') PCR conditions Reference IntI1-F GGT CAA GGA TCT GGA TTT CG 1 cycle of 4 min at 94?C; 35 cycles of 1 min at 94?C, 1 min at 62?C, 1 min at 72?C; 1 cycle of 8 min at 72?C (Machado et al., 2005) IntI1-R ACA TGC GTG TAA ATC ATC GTC (Machado et al., 2005) 5'CS GGC ATC CAA GCA GCA AG 1 cycle of 4 min at 94?C; 35 cycles of 1 min at 94?C, 1 min at 66?C, 2 min at 72?C; 1 cycle of 10 min at 72?C (Levesque et al., 1995 3'CS AAG CAG ACT TGA CCT GA (Levesque et al., 1995 IntI2-F CAC GGA TAT GCG ACA AAA AGG T Same as for int1 (Machado et al., 2005) IntI2-R GTA GCA AAC GAG TGA CGA AAT G (Machado et al., 2005) attI2-F GAC GGC ATG CAC GAT TTG TA Same as for 5'CS (Machado et al., 2005) orfX-R GAT GCC ATC GCA AGT ACG AG (Machado et al., 2005) 157 Figure 1. Association between antimicrobial resistant profile and occurrence of class 1 and 2 integrons in Escherichia coli stains isolated from dogs and cats. 0 2 4 6 8 10 12 14 16 01234567891011213 No. of resistant antimicrobial drug N o . o f r esit a n t i s o l at es No. of isolates carrying integrons No. of isolates carrying no integrons 158 Table 3. The association between antibiotic profile and class 1 integrons in E. coli Antimicrobial drug Integrons-positive isolates (n=23) Integrons-negative isolates (n=59) %R (n) b %S (n) %R (n) %S (n) OR (95% CI) Ampicillin 50(19) 7(3) 50(19) 93(40) 13.3 (3.51-50.64) Amoxicillin/ clavulanic acid 62(13) 14(8) 38(8) 86(51) 10.3 (3.26-32.84) Cephalothin 35(20) 8(2) 65(37) 92(22) 5.9 (1.27-27.91) Cepfoxitin 59(13) 15(9) 41(9) 85(50) 8.0 (2.65-24.28) Cefpodoxime 58(11) 18(11) 42(8) 82(51) 6.4 (2.08-19.54) Cefotaxime 77(10) 18(12) 23(3) 82(56) 15.5 (3.71-65.19) Ceftazidime 87.5(7) 20.5(15) 12.5(1) 79.5(58) 27 (3.08-237.25) Meropenem 83(5) 23(17) 17(1) 77(58) 17 (1.86-156.13) Enrofloxacin 58(11) 18(11) 42(8) 82(51) 6.4 (2.08-19.53) Doxycycline 50(17) 11(5) 50(17) 89(42) 8.4 (2.67-26.40) Chloramphenicol 48(14) 15(8) 52(15) 85(44) 5.1 (1.80-14.64) Gentamicin 65(11) 28(11) 35(6) 72(53) 8.8 (2.69-28.97) Trimethoprim/ sulfamethoxazole 68(15) 11.5(7) 32(7) 88.5(52) 16 (4.82-52.57) a MICs were determined using microdilution methods according to CLSI standards (CLSI, 2008). b R, resistant; S, susceptible. 159 Table 4. Antibiotic resistance patterns of E. coli isolates and their relationship with occurrence of the integrons RFLP type Length of variable region (bp) Gene cassettes and order Resistance phenotype a No. of isolate s Species Tissue source US State (n) b Class 1 integrons I 1 1000 aadA1 XACDEG R 1 Feline Urine CA (1) II 1 1500 dfrA1- aadA1 XACS 1 Canine Urine NC (1) VI 1 1500 dfrA17- aadA5 ACDGRER S 2 Canine Urine MA (1), NC (1), CA (1) III 1 1800 dfrA12- aadA2 HDS 1 Canine Urine CA (1) 1500 aadB- aadA1d ACDG 1 Canine Urine CA (1) 2500 aacA4- catB3- dfrA1 XATOPZC HDERSG M 4 Canine Urine, Trache al wash OH (1), NC (2), IL (2) 3200 aadB- aadA1- cmlA6 AHDGSXT OPZCER 2 Canine Urine CA (2) 2000 aadA1- unknown XATOPCH DEMGRSZ 9 Canine, Feline Urine, Nasal cavity, Vagina IL (2), CA (5), OH (1), NC (1) Class 2 integrons II 2 2200 dhfr1- sat1- aadA1 XATOPZC HDEGRS 2 Feline, Canine Urine NC (2) a Abbreviation of antimicrobial drugs; A, ampicillin; X, amoxicillin/clavulanic; C, cephalothin; O, cepfoxitin; P, cefpodoxime; T, cefotaxime; Z, ceftazidime; M, meropenem; E, enrofloxacin; D, doxycycline; H, chloramphenicol; G, gentamicin; S, trimethoprim/sulfamethoxazole. b Abbreviation of the origin of the isolates; CA, California; NC, North Carolina; IL, Illinois; MA, Massachusetts; OH, Ohio. 160 Table 5. A BLAST search for the antimicrobial gene cassettes The gene cassette array Source Accession number % of identity dhfr1- sat1- aadA1 Japan (AB188272) 100% Germany (AM990340) 100% Portugal (EU089672) 99% Turkey (EU339237, EF543147) 99% China (AB211124) 99% Ireland (FJ155843) 99% dfrA17-aadA5 Turkey (FJ001849, FJ001848, FJ001847) 99% Netherlands (DQ663488) 99% Japan (AB188265, AB188260) 99% Germany (AM932673) 99% aacA4-catB3-dfrA1 China (DQ836057, EF488369) 99% Germany (AM939644) 99% aadB-aadA1-cmlA6 China (DQ836058, DQ018384) 99% 161 Figure 2. PFGE profiles of the extracted plasmid DNA of the 3 isolates before and after curing. PCR amplification, A, and PFGE profiles, B, for the isolates before and after curing. Lanes 2, 4, 6, PCR positive result for the variable regions of class 1 integrons for isolates L8055239, I3195059, and B5729897 with corresponding amplicons size 1-kb, 2.5-kb, and 3.2-kb, respectively before curing. Lanes 3, 5, 7, negative results of class1 integrons for isolates L8055239, I3195059, and B5729897, respectively. Lanes 8 and 9 represent isolates salmonella serovar Typhimurium DT104 and E. coli ATCC ? 25922, respectively used as positive and negative control for class 1 integrons. Lanes 12, 14, 16, isolates L8055239, I3195059, and B5729897, respectively before the curing. Lanes 13, isolate L8055239 lost a 97-kb band; lanes 15, isolate I3195059 lost a 169-kb and 2.3-kb bands; lanes 17, isolate B5729897 lost a 194-kb, 145.5-kb and 23-kb bands after curing. 162 Table 6. Antimicrobial resistance profiles and MICs for the isolates before and after curing. Bacterial isolate(s) Antimicrobial drugs L80552 39 (B) L80552 39 (A) I31950 59 (B) I31950 59 (A) B57298 97 (B) B57298 97 (A) amoxicillin/ clavulanic acid 16 8 64 16 8 1 ampicillin > 256 4 >256 4 >256 ?0.5 cefotaxime ?1 ?1 16 ?1 ?1 ?1 cepfoxitin 8 8 128 16 2 2 ceftazidime ?0.5 ?0.5 64 2 ?0.5 ?0.5 cephalothin 256 32 >1024 32 16 2 chloramphenicol 4 8 1024 8 32 8 doxycycline 32 32 64 32 16 8 enrofloxacin 64 64 >128 64 1 1 gentamicin 128 1 8 1 128 2 meropenem ?0.25 ?0.25 ?0.25 ?0.25 ?0.25 ?0.25 tricarcillin/ clavulanic acid 256 4 256 8 16 ?2 trimethoprim/ sulfamethoxazole 0.25 0.25 >128 ?0.06 >128 0.12 a MICs were determined using broth microdilution methods according to CLSI standards (CLSI, 2008). Numbers in boldface represent resistant phenotypes that becomes susceptible after curing. Only the curing of isolate B5729897 resulted in susceptible phenotype