Escherichia coli is a common cause of urinary tract infections in dogs,1–3 and antimicrobial-resistant E coli organisms have been increasingly identified as contributing to treatment failure and the cost of health care for affected dogs.3,4 In particular, an increase in the prevalence of fluoroquinolone resistance in canine urinary E coli isolates has been reported.2,3 Such isolates are often MDR.5–8
An important risk factor for emergence of fluoroquinolone resistance in E coli is the use of fluoroquinolone antimicrobials.6,8 Enrofloxacin is a fluoroquinolone commonly used to treat dogs with urinary tract infection9 and is an important contributor to the emergence of MDR E coli.10,11 Although antimicrobial susceptibility testing of E coli continues to be the reference (gold) standard for the detection of antimicrobial resistance, this technique is laborious and time-consuming, requiring 2 to 5 days from the point of specimen collection to the point results are reported to clinicians. This lag can increase the rate of treatment failure, particularly if treatment is initiated with an antimicrobial to which the infecting organism is already resistant.12 This treatment failure along with the increase in the prevalence of MDR infections emphasizes a need for an alternative method that would allow rapid and sensitive detection of MDR or fluoroquinolone resistance in clinical isolates and specimens.13
Mutations in the QRDR of the genes for DNA gyrase (gyrAB) and topoisomerase IV (parCE) are the most common genetic mechanisms conferring fluoroquinolone resistance in bacterial species.14–16 These single nucleotide mutations can be easily detected by hybridization probes and qPCR assay.17,18 Other molecular techniques with the ability to detect fluoroquinolone resistance–causing mutations, such as mismatch detection, sequencing, and gene expression assays, have all been developed for the detection of fluoroquinolone-resistant clinical E coli isolates.19–21 However, these methods also are time-consuming and used primarily for research.
A qPCR-based assay can achieve precise discrimination when used in combination with FRET technology. By monitoring the temperature-dependent hybridization of sequence-specific probes to single-stranded DNA while performing melting curve analysis,22 one can easily detect the destabilizing nucleotide mismatches characteristic of fluoroquinolone resistance. In this situation, the mutant gyrA results in a melting curve with a peak at a lower Tm than wild-type gyrA, rendering mutations and, hence, fluoroquinolone resistance rapidly detectable. The purpose of the study reported here was to determine the effectiveness of a FRET-qPCR assay for detection of gyrA mutations (fluoroquinolone resistance) in canine urinary E coli isolates and clinical canine urine specimens.
Materials and Methods
Samples
Arbitrarily selected canine urinary E coli isolates (n = 264) that had been included in a previous study8 were used to validate the FRET-qPCR assay technique by use of pure E coli cultures. Results of enrofloxacin susceptibility testing and MICs of enrofloxacin (broth microdilution method) for these isolates were obtained from that previous study.8 For quality control purposes, an E coli standard straina was included in each sample set.
In addition, 357 urine specimens collected from dogs by cystocentesis, catheterization, or free catch and submitted to the Auburn University Small Animal Teaching Hospital through the Bacteriology and Mycology Laboratory or private veterinary diagnostic laboratories were included. On receipt of urine specimens at the Auburn University Clinical Pharmacology Laboratory, an aliquot of each was placed in a 4.5-mL tube and stored at 4°C. Another 10-μL aliquot was transferred to nonselective bacterial culture mediumb and incubated at 37°C overnight for isolation, differentiation, and enumeration of urinary tract pathogens by means of standard techniques. Individual colonies of each recovered bacterial species were then transferred to trypticase soy agarc for enrofloxacin susceptibility testing. Bacterial isolates were also tested for enrofloxacin susceptibility via agar diffusion with a commercial gradient strip test.d The MICs of enrofloxacin were recorded, and results for each isolate were interpreted as susceptible or resistant in accordance with Clinical and Laboratory Standards Institute breakpoint standards.23
DNA sample preparation for FRET-qPCR assay
Canine urinary E coli isolates were revived on trypticase soy agar plates at 37°C overnight. Extraction of DNA was then performed by use of a commercial reagent,e in accordance with the manufacturer's instructions.
To replicate the assay as it would be used as a diagnostic test, DNA was extracted from clinical urine specimens that yielded growth on bacterial culture before performance of the FRET-qPCR assay. For this process, urine specimens were applied to commercial centrifugal microconcentratorsf and centrifuged for 40 minutes at 1,000 × g. Afterward, the filter was washed with 150 μL of microbe-free urine and the wash material was collected for DNA extraction. Extraction of DNA was performed by use of a commercial DNA extraction kit.g The obtained DNA was subsequently eluted to 50 μL and stored at −20°C pending analysis.
FRET-qPCR assay
A PCR assay systemh was used for amplification and melting curve analysis. Frozen DNA samples were thawed at room temperature (approx 22°C) before the FRET-qPCR assay was performed. Primers and probes were designed to be specific for a consensus QRDR wild-type sequence.24 Fluorophores were selected with 3′-labeled 6-carboxyfluorescein for the donor probe and 5′-labeled, 3′-fluorescent oligonucleotide-labeling dyei for the reporter probe (Appendix). The FRET-qPCR assay reactions were performed in 96-well plates. Each well contained 5 μL of sample nucleic acids and 15 μL of genotyping PCR reagentj supplemented with 2.0 U of Taq polymerase.k
The thermocycling program, which was based on that used in a prior study24 with modifications for 96-well plates, consisted of 18 high-stringency step-down cycles succeeded by 35 amplification and fluorescence acquisition cycles with a final melting curve. The high-stringency step-down cycling program involved 1 cycle at 95°C for 5 minutes, 6 cycles at 95°C for 15 seconds, 1 cycle at 72°C for 30 seconds, 9 cycles at 95°C for 15 seconds, 1 cycle at 70°C for 30 seconds, 3 cycles at 95°C for 15 seconds, 1 cycle at 68°C for 30 seconds, and 1 cycle at 72°C for 30 seconds. Amplification was then achieved by 35 cycles of denaturation at 95°C for 15 seconds, annealing at 52°C for 15 seconds, 1 cycle at 66°C for 30 seconds, and extension at 72°C for 30 seconds. Light emittance for the PCR assay systemh was set at 498 nm and absorption at 640 nm. Results of the FRET-qPCR assay were reported as Tm.
Validation of FRET-qPCR results
To validate the FRET-qPCR assay, a subset of 20 isolates from the initial 264 canine urinary E coli isolates that were identified as positive for fluoroquinolone resistance by the assay were assessed to confirm fluoroquinolone resistance by means of Sanger sequencingl of the gyrA locus and comparison with the E coli K12 gyrA sequencem, 25 via a bioinformatics search tool.n
To determine the limits of detection of the FRET-qPCR assay with respect to bacterial cell count for urine specimens, the pure E coli cultures from the previous study8 were inoculated into aerobic bacteria–free canine urine. For this determination, aliquots of canine urine specimens collected via cystocentesis were inoculated onto nonselective bacterial culture mediumb and incubated at 37°C for 48 hours to verify aerobe-free status. After confirmation, 4.5-mL urine aliquots were prepared for bacterial dilutions. Seven E coli isolates representing increasing enrofloxacin susceptibilities were suspended in saline (0.9% NaCl) solution to 0.5 McFarland standardn (approx 109 CFUs). Ten-fold dilutions were made from 101 to 106 CFUs in aerobe-free urine, and DNA was subsequently extracted for FRET-qPCR assay as previously described. The limits of detection were determined by assaying the 6 dilutions of each of the 7 isolates in duplicate and recording the lowest dilution that produced 1 peak during the melting-curve phase of FRET-qPCR assay.
Statistical analysis
Statistical analyses were performed with the aid of statistical software.o To evaluate that usefulness of the FRET-qPCR assay for detection of gyrA mutations in the canine urinary E coli isolates or clinical canine urine specimens, enrofloxacin susceptibility testing of bacterial isolates (ie, E coli isolates obtained from the previous study8 or any bacterial species recovered from the evaluated urine specimens) was considered the reference standard.
To evaluate the ability of the FRET-qPCR assay to discriminate between enrofloxacin-resistant and enrofloxacin-susceptible E coli, an ROC curve was constructed. This ROC curve reflected the relationship between sensitivity and (1 – specificity) at different Tm cutoffs for defining positive and negative results. Results were considered positive for enrofloxacin resistance if the observed Tm was less than or equal to the cutoff value and negative if the Tm was greater than the cutoff value. The AUC and its SE were calculated to provide a measurement of the overall discriminative ability of the assay by means of a nonparametric approach.26 The Youden (J) index value (sensitivity + specificity – 1) for each Tm cutoff was computed from the ROC curve, and the cutoff with the maximum index value was selected for further analysis of FRET-qPCR assay for urine specimens and E coli isolates recovered from those specimens.
Results for FRET-qPCR assay were considered truly positive if the tested sample (E coli isolate or clinical urine specimen) yielded an MIC on susceptibility testing that suggested enrofloxacin resistance with an MIC ≥ 4.23 Results were considered truly negative if the tested sample yielded an MIC on susceptibility testing that suggested enrofloxacin susceptibility with an MIC ≤ 0.5.23 For each type of sample, sensitivity, specificity, PPV, and NPV and exact binomial 95% CIs for these values were computed.
Assessment of agreement between results of the FRET-qPCR assay and enrofloxacin susceptibility testing was performed by calculation of Cohen K values and associated 95% CIs. The K values were interpreted with the Landis and Koch classification system27 as follows: < 0.00, poor agreement; 0.00 to 0.20, slight agreement; 0.21 to 0.40, fair agreement; 0.41 to 0.60, moderate agreement; 0.61 to 0.80, substantial agreement; and 0.81 to 1.00, almost perfect agreement. Values of P < 0.05 were considered significant.
Results
Initial assay validation with canine urinary E coli isolates
Results of enrofloxacin susceptibility testing in the previous study8 indicated that 8.3% (22/264) of canine urinary E coli isolates included in the present study were resistant to enrofloxacin and 91.7% (242/264) were susceptible. The ROC curve for the ability of the FRET-qPCR assay to discriminate between enrofloxacin-resistant and enrofloxacin-susceptible E coli isolates yielded an AUC of 0.92 (95% CI, 0.85 to 1.00) and SE of 0.04 (P = 0.01; Figure 1). A Tm cutoff of ≤ 66°C yielded the highest Youden index value (0.83) for detection of enrofloxacin resistance, suggesting the optimal balance (least difference) between sensitivity and specificity (Table 1). Therefore, results from FRET-qPCR assay were consequently considered positive for enrofloxacin resistance if the Tm was ≤ 66°C or negative for enrofloxacin resistance (ie, positive for enrofloxacin susceptibility) if the Tm was > 66°C.
Summary data for the ability of a FRET-qPCR assay to discriminate between enrofloxacin-resistant (ie, possessing gyrA mutations) canine urinary Escherichia coli isolates (n = 22) and enrofloxacin-susceptible isolates (242) at various Tm cutoffs for designating a positive result (enrofloxacin resistance), with enrofloxacin susceptibility testing used as the reference standard.
| Tm (C°) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Youden index value |
|---|---|---|---|---|---|
| ≤ 56 | 4.6 (0.1–22.8) | 100 (97.0–100) | 100 (NA) | 85.1 (83.9–86.2) | 0.05 |
| ≤ 57 | 13.6 (2.9–34.9) | 100 (97.0–100) | 100 (NA) | 86.3 (84.3–88.2) | 0.14 |
| ≤ 58 | 31.8 (13.9–54.9) | 99.2 (95.4–100) | 87.5 (47.5–98.2) | 88.8 (85.6–91.3) | 0.31 |
| ≤ 59 | 77.3 (54.6–92.2) | 99.2 (95.4–100) | 94.4 (70.4–99.2) | 96.0 (91.7–98.1) | 0.76 |
| ≤ 64 | 77.3 (54.6–92.2) | 97.5 (92.9–99.5) | 85.0 (64.4–94.7) | 95.9 (91.5–98.1) | 0.75 |
| ≤ 65 | 77.3 (54.6–92.2) | 96.7 (91.7–99.1) | 81.0 (61.2–92.0) | 95.9 (91.5–98.0) | 0.74 |
| ≤ 66* | 86.4 (65.1–97.1) | 96.7 (91.7–99.1) | 82.6 (64.1–92.7) | 97.5 (93.1–99.1) | 0.83 |
| ≤ 68 | 86.4 (65.1–97.1) | 95.8 (90.5–98.6) | 79.2 (61.3–90.1) | 97.5 (93.0–99.1) | 0.82 |
| ≤ 70 | 100.0 (84.6–100) | 0.0 (0.0–3.0) | 15.5 (15.5–15.5) | NA | 0.00 |
Values in parentheses represent 95% CIs.
The Tm cutoff value deemed to yield the optimal balance (least difference) between sensitivity and specific and highest Youden index value.
NA = Not applicable.
Nonparametric ROC curve of the ability of a FRET-qPCR assay to discriminate between enrofloxacin-resistant (ie, possessing gyrA mutations) canine urinary Escherichia coli isolates (n = 22) and enrofloxacin-susceptible isolates (242) at various Tm cutoffs for designating a positive or negative result (A), and scatterplot of isolate enrofloxacin susceptibility for resistant and susceptible isolates relative to Tm (B). The optimal Tm cutoff value (66°C; dotted line) yielded a sensitivity of 86.4% and specificity of 96.7%.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.755
Nonparametric ROC curve of the ability of a FRET-qPCR assay to discriminate between enrofloxacin-resistant (ie, possessing gyrA mutations) canine urinary Escherichia coli isolates (n = 22) and enrofloxacin-susceptible isolates (242) at various Tm cutoffs for designating a positive or negative result (A), and scatterplot of isolate enrofloxacin susceptibility for resistant and susceptible isolates relative to Tm (B). The optimal Tm cutoff value (66°C; dotted line) yielded a sensitivity of 86.4% and specificity of 96.7%.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.755
Nonparametric ROC curve of the ability of a FRET-qPCR assay to discriminate between enrofloxacin-resistant (ie, possessing gyrA mutations) canine urinary Escherichia coli isolates (n = 22) and enrofloxacin-susceptible isolates (242) at various Tm cutoffs for designating a positive or negative result (A), and scatterplot of isolate enrofloxacin susceptibility for resistant and susceptible isolates relative to Tm (B). The optimal Tm cutoff value (66°C; dotted line) yielded a sensitivity of 86.4% and specificity of 96.7%.
Citation: American Journal of Veterinary Research 79, 7; 10.2460/ajvr.79.7.755
The 20 canine urinary E coli isolates selected for assay validation via gyrA sequence analysis included 5 isolates from each of the true positive (truly enrofloxacin resistant and Tm ≤ 66°C), true negative (truly enrofloxacin susceptible and Tm > 66°C), false positive (truly enrofloxacin susceptible but Tm ≤ 66°C), and false negative (truly enrofloxacin resistant but Tm > 66°C) results. Among the 5 truly negative isolates, no gyrA mutations were identified. Among all 5 truly positive isolates, 2 missense mutations (Ser83Leu and Asp87Asn) were identified at the position of the probe attachment (Supplementary Table S1, available at: http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.79.7.755). The isolates yielding a false positive reading contained 1 point mutation within the probe attachment site (Ser83 deleted and Ser83Leu). Additionally, within the sequenced region but outside the region targeted by our probes, a set of synonymous mutations (Arg91Arg, Tyr100Tyr, and Ser111Ser) were identified in 14 of the 20 isolates. These mutations were found together and occurred in the +3 position within the codon (Arg91, C273T; Tyr100, T300C; and Ser111, T333C). Tests of the limits of detection for the FRET-qPCR assay (involving inoculation of E coli colonies into aerobic bacteria–free urine) revealed that E coli were detectable at dilutions as low as 101 CFUs/mL.
Assay validation with clinical canine urine specimens
Of the 357 clinical canine urine specimens included in the study, 283 (79.3%) yielded aerobic bacterial growth, 59 of which yielded multiple organisms (Supplementary Table S2, available at: http://avmajournals.avma.org/doi/suppl/10.2460/ajvr.79.7.755). Two-hundred forty-nine urine specimens yielded E coli. Other identified aerobic species included Klebsiella spp (n = 19), Enterococcus spp (16), Staphylococcus spp (9), Proteus spp (7), Streptococcus spp (6), and Pseudomonas spp (2).
Antimicrobial susceptibility testing of bacterial isolates (of any species) from the 283 specimens revealed that isolates from 40 (14.1%) specimens were truly enrofloxacin resistant; however, only 25 of these 40 specimens yielded positive results on FRET-qPCR assay for gyrA mutations (in DNA samples extracted from the urine specimens), resulting in 15 (40 – 25) false-negative results and an assay sensitivity of 62.5% (95% CI, 45.8% to 77.3%). Alternatively, susceptibility testing revealed that isolates from 243 specimens were truly enrofloxacin susceptible; however, only 226 of these 243 specimens yielded negative results on FRET-qPCR assay, resulting in 17 (243 – 226) false-positive results and an assay specificity of 93.0% (95% CI, 89.0% to 95.9%). The corresponding PPV and NPV were 59.5% (25/42; 95% CI, 43.3% to 74.4%) and 93.8% (226/241; 95% CI, 89.9% to 96.5%), respectively. Focusing solely on the 249 E coli isolates recovered from the urine specimens (and not on results for direct testing of urine), sensitivity of the FRET-qPCR assay for detection of gyrA mutations was 77.8% (28/36; 95% CI, 60.9% to 89.9%), specificity was 94.8% (202/213; 95% CI, 91.0% to 97.4%), PPV was 71.8% (28/39; 95% CI, 55.1% to 85.0%), and NPV was 96.2% (202/210; 95% CI, 92.6% to 98.3%).
In comparisons of agreement between results of FRET-qPCR assay and enrofloxacin susceptibility testing, moderate agreement was obtained when the assayed substance was clinical urine specimens containing any bacterial species (κ = 0.54; 95% CI, 0.40 to 0.68), but substantial agreement was obtained when the assayed substance was E coli isolates recovered from those urine specimens (κ = 0.70; 95% CI, 0.58 to 0.83).
Discussion
The widespread use of fluoroquinolones for common infections such as urinary tract infections has raised concern regarding the possibility of accelerated development of MDR E coli,6 and some evidence exists that such resistance is sustained in the fecal flora of dogs even when such use is discontinued.28 The rapid detection of antimicrobial-resistant bacterial pathogens combined with early implementation of appropriate intervention reportedly reduces the prevalence of antimicrobial resistance.29,30 Appropriate, targeted selection of antimicrobials is simplified when a reliable result of culture and antimicrobial susceptibility testing is available; however, the minimum 2-day lag associated with receiving the results has hindered this practice. With use of the FRET-qPCR assay evaluated in the present study for detection of gyrA mutations (conferring fluoroquinolone resistance) in canine urinary E coli isolates or canine urine specimens, the turnaround time for detection of fluoroquinolone resistance could reduce this lag to hours in facilities with or with access to appropriately equipped laboratories.
The mechanism of fluoroquinolone resistance in bacterial species is based on several pathways, including mutation of topoisomerases, expression of efflux pumps, and inactivation of fluoroquinolones. The primary mechanism is gyrase impairment resulting from the mutation of gyrAB. In E coli, the mutation in subpart gyrA is primarily located between Ala67 and Gln1066,31 and in subpart gyrB at Asp426 and Lys447.32 Also in E coli, just a single gyrA mutation can increase the MIC of fluoroquinolones up to 64-fold33 and a double mutation at codons 83 and 87 of gyrA can result in even greater resistance. The substitution of Ser83Leu is considered the first step to the development of fluoroquinolone resistance. In contrast, the gyrB mutation appears less effective in conferring resistance against fluoroquinolones than the gyrA mutation.32 Mutation of topoisomerase IV has also been described; however, this mutation has consistently been identified along with ≥ 1 gyrase mutation. Consequently, the FRET-qPCR assay in the present study was developed to detect the most common fluoroquinolone resistance–conferring mutations.
False-positive results of the FRET-qPCR assay in the study reported here could have been attributable to the methodology used, the step mutation of bacteria, or other factors. The DNA extraction was performed in bulk, directly from urine, allowing the possibility of nucleic acids from other microorganisms interfering with probe specificity, and curious multiple melting peaks for Tm were obtained. After examination of the alignment of the reporter probe and strains of the additional microbial species detected in the canine urine specimens, we noticed that considerable homology existed between the front of the reporter probe and regions within the Staphylococcus and Streptococcus QRDR of gyrA. This homology provided the probe with an anchor near the fluorophore, allowing the fluorophore to become excited and register a low Tm melting curve.
The evaluated FRET-qPCR assay had low sensitivity (high proportion of false-negative results) for detection of the gyrA mutation, particularly considering results achieved for DNA extracted from the urine specimens. This low sensitivity could have been attributable to the presence of unculturable organisms from the urinary bladder microbiome34 within the tested urine specimens or residual reagents from DNA extraction methods, which could have affected the PCR assay reaction. After the study concluded, we further analyzed the urine specimens and found that 7 false-negative specimens contained mixed E coli strains. Although antimicrobial susceptibility testing is the reference standard, this method is not without limitations. First, only 1 colony of a selected isolate is used for antimicrobial susceptibility testing, even though a urinary tract infection might be caused by multiple strains or species. Second, false-negative results might be attributable to fluoroquinolone resistance caused by genetic and physiologic factors other than mutations in the QRDR of E coli gyrA. Indeed, plasmids containing genes from the qnr family, qepA, or aac(6‘)-Ib-cr have also been linked to fluoroquinolone-resistance related MDR phenotypes,35–38 which complicates testing for the various genes.
The high AUC of the ROC curve for discrimination between enrofloxacin-resistant and enrofloxacin-susceptible E coli isolates and high agreement between the FRET-qPCR assay and reference standard suggested that the evaluated FRET-qPCR assay may be useful in clinical settings for detection of fluoroquinolone resistance in uropathogens of dogs. The assay rapidly distinguished between enrofloxacin-resistant and enrofloxacin-susceptible E coli in clinical isolates by detection of gyrA mutations. Although we do not suggest that this assay replace antimicrobial susceptibility testing, we believe it could be clinically useful for identifying urinary tract infections cause by fluoroquinolone-resistant E coli, allowing the choice of other antimicrobials to avoid the unnecessary use of fluoroquinolones and potential resultant antimicrobial resistance.
Acknowledgments
This manuscript represents a portion of a thesis submitted by Dr. Behringer to the Auburn University Department of Anatomy, Physiology, and Pharmacology as partial fulfillment of the requirements for a Master of Science degree.
Supported in part by Morris Animal Foundation grant No. MAF24490.
ABBREVIATIONS
| AUC | Area under the curve |
| CI | Confidence interval |
| FRET | Fluorescence resonance energy transfer |
| MDR | Multidrug resistant |
| MIC | Minimum inhibitory concentration |
| NPV | Negative predictive value |
| PPV | Positive predictive value |
| qPCR | Quantitative PCR |
| QRDR | Quinolone-resistance determining regions |
| ROC | Receiver operating characteristic |
| Tm | Melting temperature |
Footnotes
ATCC 25922, American Tissue Cell Culture, Manassas, Va.
CHROMagar orientation medium, BD Diagnostics, Franklin Lakes, NJ.
BD Diagnostics, Franklin Lakes, NJ.
Epsilometer test, BioMérieux, Marcy l'Etoile, France.
PrepMan ULTRA preparation reagent, Applied Biosystems, Foster City, Calif.
100K Microsep advance centrifugal microconcentrators, Pall Corporation, Port Washington, N Y.
EZNA viral RNA kit, Omega Bio-Tek, Norcross, Ga.
LightCycler 480 PCR system, Roche Diagnostics, Indianapolis, Ind.
LightCycler red 640-N-hydroxysuccinimide ester, Roche Diagnostics, Indianapolis, Ind.
LightCycler 480 genotyping master, Roche Applied Science, Indianapolis, Ind.
Platinum Taq DNA polymerase, Invitrogen, Carlsbad, Calif.
Macrogen, Rockville, Md.
GenBank accession No. X06373, NCBI, Bethesda, Md.
BLAST, National Center for Biotechnology Information, National Institutes of Health, Bethesda, Md. Available at: blast.ncbi.nlm.nih.gov/. Accessed Jul 31, 2016.
Thermo Fisher Scientific, Waltham, Mass.
SAS, version 9.3, SAS Institute Inc, Cary, NC.
References
1. Ling GV, Biberstein EL, Hirsh DC. Bacterial pathogens associated with urinary tract infections. Vet Clin North Am Small Anim Pract 1980;9:617–630.
2. McMeekin CH, Hill K, Gibson I, et al. Antimicrobial resistance patterns of bacteria isolated from canine urinary samples submitted to a New Zealand veterinary diagnostic laboratory between 2005–2012. N Z Vet J 2017;65:99–104.
3. Wong C, Epstein SE, Westropp JL. Antimicrobial susceptibility patterns in urinary tract infections in dogs (2010–2013). J Vet Intern Med 2015;29:1045–1052.
4. Guardabassi L, Prescott JF. Antimicrobial stewardship in small animal veterinary practice: from theory to practice. Vet Clin North Am Small Anim Pract 2015;45:361–376.
5. Shaheen BW, Boothe D, Oyarzabal O, et al. Antimicrobial resistance profiles and clonal relatedness of canine and feline Escherichia coli pathogens expressing multidrug resistance in the United States. J Vet Intern Med 2010;24:323–330.
6. Liu X, Boothe DM, Thungrat K, et al. Mechanisms accounting for fluoroquinolone multidrug resistance Escherichia coli isolated from companion animals. Vet Microbiol 2012;161:159–168.
7. Piras C, Soggiu A, Greco V, et al. Mechanisms of antibiotic resistance to enrofloxacin in uropathogenic Escherichia coli in dog. J Proteomics 2015;127:365–376.
8. Thungrat K, Price SB, Carpenter DM, et al. Antimicrobial susceptibility patterns of clinical Escherichia coli isolates from dogs and cats in the United States: January 2008 through January 2013. Vet Microbiol 2015;179:287–295.
9. Weese JS, Giguère S, Guardabassi L, et al. ACVIM consensus statement on therapeutic antimicrobial use in animals and antimicrobial resistance. J Vet Intern Med 2015;29:487–498.
10. Aly SA, Debavalya N, Suh S-J, et al. Molecular mechanisms of antimicrobial resistance in fecal Escherichia coli of healthy dogs after enrofloxacin or amoxicillin administration. Can J Microbiol 2012;58:1288–1294.
11. Black DM, Rankin SC, King LG. Antimicrobial therapy and aerobic bacteriologic culture patterns in canine intensive care unit patients: 74 dogs (January–June 2006). J Vet Emerg Crit Care (San Antonio) 2009;19:489–495.
12. Bubenik LJ, Hosgood GL, Waldron DR, et al. 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 2007;231:893–899.
13. Vernel-Pauillac F, Hogan TR, Tapsall JW, et al. Quinolone resistance in Neisseria gonorrhoeae: rapid genotyping of quinolone resistance-determining regions in gyrA and parC genes by melting curve analysis predicts susceptibility. Antimicrob Agents Chemother 2009;53:1264–1267.
14. Oram M, Fisher LM. 4-Quinolone resistance mutations in the DNA gyrase of Escherichia coli clinical isolates identified by using the polymerase chain reaction. Antimicrob Agents Chemother 1991;35:387–389.
15. Willmott CJ, Maxwell A. A single point mutation in the DNA gyrase A protein greatly reduces binding of fluoroquinolones to the gyrase-DNA complex. Antimicrob Agents Chemother 1993;37:126–127.
16. Piddock LJ. Mechanisms of fluoroquinolone resistance: an update 1994–1998. Drugs 1999;58:11–18.
17. Sun Y, Lu X, Su F, et al. Real-time fluorescence ligase chain reaction for sensitive detection of single nucleotide polymorphism based on fluorescence resonance energy transfer. Biosens Bioelectron 2015;74:705–710.
18. Klein D. Quantification using real-time PCR technology: applications and limitations. Trends Mol Med 2002;8:257–260.
19. Zirnstein G, Li Y, Swaminathan B, et al. Ciprofloxacin resistance in Campylobacter jejuni isolates: detection of gyrA resistance mutations by mismatch amplification mutation assay PCR and DNA sequence analysis. J Clin Microbiol 1999;37:3276–3280.
20. Johnson JR, Johnston B, Clabots C, et al. Escherichia coli sequence type ST131 as the major cause of serious multidrug-resistant E. coli infections in the United States. Clin Infect Dis 2010;51:286–294.
21. Suzuki S, Horinouchi T, Furusawa C. Prediction of antibiotic resistance by gene expression profiles. Nat Commun 2014;5:5792.
22. Gibson NJ. The use of real-time PCR methods in DNA sequence variation analysis. Clin Chim Acta 2006;363:32–47.
23. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial disk and dilution susceptibility tests for bacterial isolated from animals; approved standard. 4th ed. CLSI document VET01–A4 (standard) and VET01–S2 (supplement). Wayne, Penn: Clinical and Laboratory Standards Institute, 2013.
24. Shaheen BW, Wang C, Johnson CM, et al. Detection of fluoroquinolone resistance level in clinical canine and feline Escherichia coli pathogens using rapid real-time PCR assay. Vet Microbiol 2009;139:379–385.
25. Swanberg SL, Wang JC. Cloning and sequencing of the Escherichia coli gyrA gene coding for the A subunit of DNA gyrase. J Mol Biol 1987;197:729–736.
26. Zou KH, O'Malley AJ, Mauri L. Receiver-operating characteristic analysis for evaluating diagnostic tests and predictive models. Circulation 2007;115:654–657.
27. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–174.
28. Boothe DM, Debavalya N. Impact of routine antimicrobial therapy on canine fecal Escherichia coli antimicrobial resistance: a pilot study. Int J Appl Res Vet Med 2011;9:396–406.
29. Cunningham R, Jenks P, Northwood J, et al. Effect on MRSA transmission of rapid PCR testing of patients admitted to critical care. J Hosp Infect 2007;65:24–28.
30. Kluytmans J. Control of meticillin-resistant Staphylococcus aureus (MRSA) and the value of rapid tests. J Hosp Infect 2007; 65:100–104.
31. Yoshida H, Bogaki M, Nakamura M, et al. Quinolone resistance-determining region in the DNA gyrase gyrA gene of Escherichia coli. Antimicrob Agents Chemother 1990;34:1271–1272.
32. Nakamura S, Nakamura M, Kojima T, et al. gyrA and gyrB mutations in quinolone-resistant strains of Escherichia coli. Antimicrob Agents Chemother 1989;33:254–255.
33. Drlica K, Zhao X. DNA gyrase, topoisomerase I V, and the 4-quinolones. Microbiol Mol Biol Rev 1997;61:377–392.
34. Hilt EE, McKinley K, Pearce MM, et al. Urine is not sterile: use of enhanced urine culture techniques to detect resident bacterial flora in the adult female bladder. J Clin Microbiol 2014;52:871–876.
35. Jacoby GA, Walsh KE, Mills DM, et al. qnrB, another plasmid-mediated gene for quinolone resistance. Antimicrob Agents Chemother 2006;50:1178–1182.
36. Machuca J, Briales A, Labrador G, et al. Interplay between plasmid-mediated and chromosomal-mediated fluoroquinolone resistance and bacterial fitness in Escherichia coli. J Antimicrob Chemother 2014;69:3203–3215.
37. Yamane K, Wachino J, Suzuki S, et al. New plasmid-mediated fluoroquinolone efflux pump, QepA, found in an Escherichia coli clinical isolate. Antimicrob Agents Chemother 2007;51:3354–3360.
38. Vetting MW, Park CH, Hegde SS, et al. Mechanistic and structural analysis of aminoglycoside N-acetyltransferase AAC (6′)-Ib and its bifunctional, fluoroquinolone-active AAC (6′)-Ib-cr variant. Biochemistry 2008;47:9825–9835.
