Skip to main content
NCBI home page
Journal of Clinical Microbiology logoLink to Journal of Clinical Microbiology
. 2013 Jul;51(7):2082–2088. doi: 10.1128/JCM.00307-13

Molecular and Antimicrobial Susceptibility Analyses Distinguish Clinical from Bovine Escherichia coli O157 Strains

Sinisa Vidovic a, Sarah Tsoi b, Prabhakara Medihala b, Juxin Liu c, John L Wylie d,e, Paul N Levett f, Darren R Korber b,✉
PMCID: PMC3697693  PMID: 23616449

Abstract

A population-based study combining (i) antimicrobial, (ii) genetic, and (iii) virulence analyses with molecular evolutionary analyses revealed segregative characteristics distinguishing human clinical and bovine Escherichia coli O157 strains from western Canada. Human (n = 50) and bovine (n = 50) strains of E. coli O157 were collected from Saskatchewan and Manitoba in 2006 and were analyzed by using the six-marker lineage-specific polymorphism assay (LSPA6), antimicrobial susceptibility analysis, the colicin assay, plasmid and virulence profiling including the eae, ehxA, espA, iha, stx1, stx2, stx2c, stx2d, stx2d-activatable, stx2e, and stx2f virulence-associated genes, and structure analyses. Multivariate logistic regression and Fisher's exact test strongly suggested that antimicrobial susceptibility was the most distinctive characteristic (P = 0.00487) associated with human strains. Among all genetic, virulence, and antimicrobial determinants, resistance to tetracycline (P < 0.000) and to sulfisoxazole (P < 0.009) were the most strongly associated segregative characteristics of bovine E. coli O157 strains. Among 11 virulence-associated genes, stx2c showed the strongest association with E. coli O157 strains of bovine origin. LSPA6 genotyping showed the dominance of the lineage I genotype among clinical (90%) and bovine (70%) strains, indicating the importance of lineage I in O157 epidemiology and ecology. Population structure analysis revealed that the more-diverse bovine strains came from a unique group of strains characterized by a high degree of antimicrobial resistance and high frequencies of lineage II genotypes and stx2c variants. These findings imply that antimicrobial resistance generated among bovine strains of E. coli O157 has a large impact on the population of this human pathogen.

INTRODUCTION

Since the discovery of Escherichia coli O157:H7 as the etiologic agent of hemorrhagic colitis in 1982 (1, 2), the clinical importance of O157 has grown rapidly. For example, the Public Health Agency of Canada reported that the majority (95%) of human cases of E. coli-related illnesses in 1995 involved serovar O157 (3). While the frequency of human illness due to E. coli O157 isolates has fallen in some regions of Canada (4), O157 remains a significant concern due to its demonstrated ability to cause both sporadic and large outbreaks. In July 1996, the city of Sakai, Japan, experienced the largest outbreak of O157 ever recorded, which was part of a series of outbreaks that summer that caused an estimated 8,000 cases and 6 deaths (5). The most potent virulence features of E. coli O157:H7 are the Shiga toxins (Stx1 and Stx2), a family of unique heterodimeric protein toxins (6) that are responsible for a wide spectrum of clinical symptoms, including life-threatening complications such as hemolytic-uremic syndrome and thrombotic thrombocytopenic purpura (7, 8).

The ruminant gastrointestinal tract is the primary reservoir of E. coli O157 (9, 10) and the presence of this human pathogen in nature, in either water or soil, is derived from the primary habitat, usually by fecal contamination. Human infection with E. coli O157:H7 has been associated with various transmission routes, including direct exposure to infected animals (11, 12), person-to-person passage (13), and, primarily, consumption of contaminated foods and water (14, 15). From the public health perspective, it is important to gain insight into the molecular epidemiology of Shiga toxin-producing E. coli (STEC) O157 isolates, which pose a high risk to humans. An increasing number of studies (16–18) indicate the existence of certain lineages of E. coli O157 that are more closely associated with human diseases than are others, and these studies further point to apparent differences in virulence and host ecology within the STEC O157 population. For instance, Ziebell et al. (19) applied the six-marker lineage-specific polymorphism assay (LSPA6), which is based on six selectively neutral markers (20), to populations of E. coli O157 from Canada and found that most E. coli O157:H7 strains isolated from cattle and humans in Canada (82.6%) belonged to lineage I. Recently, Franz et al. (16), performing a similar study with E. coli O157 isolates from the Netherlands, found a distinct fraction of STEC O157 strains in the bovine reservoir being associated with disease in humans. In contrast, human clinical E. coli O157 isolates originating from the Netherlands had a low frequency of LSPA6 lineage I strains and a high frequency of LSPA6 lineage I/II strains, indicating different epidemiological features for E. coli O157 isolates from Canada and the Netherlands. Together with the genetic and virulence differences, it is important to examine the antimicrobial resistance of these two groups of pathogens, as the emergence of antimicrobial resistance plays an important role in the epidemiology of pathogens of public health concern.

The present study genetically and phenotypically characterized and compared two pathogen populations, i.e., bovine and human clinical E. coli O157 strains (n = 100), that were isolated from the Saskatchewan/Manitoba geographic region of western Canada during the same period of time. The ultimate goal of this project was to gain insights into the molecular epidemiology and host ecology of these two groups of E. coli O157 isolates. We combined several molecular typing methods involving 11 important virulence factors (i.e., stx1, stx2, stx2c, stx2d, stx2d-activatable, stx2e, stx2f, eae, espA, ehxA, and iha), LSPA6, and plasmid profiling, together with antimicrobial susceptibility testing and the colicin assay. Further, the pathogen population structure, including all virulence, genotyping, and antimicrobial characteristics, was evaluated using molecular evolutionary analysis with the program Structure.

MATERIALS AND METHODS

Strains of Escherichia coli O157.

In total, 100 E. coli O157 strains from western Canada were analyzed. During the summer of 2006, 50 strains of bovine origin were obtained from various feedlots located across the province of Saskatchewan (21). During the same period, another 50 strains were isolated from humans in the provinces of Saskatchewan and Manitoba, a region in western Canada. Of these 50 human strains, 25 were obtained from the Cadham Provincial Laboratory (Manitoba, Canada) and 25 were provided by the Saskatchewan Disease Control Laboratory (Canada). Each of these 100 strains represented a unique genotype, as determined with either pulsed-field gel electrophoresis or randomly amplified polymorphic DNA methods (21; J. L. Wylie and P. N. Levett, unpublished data), and most likely the studied strains of E. coli O157 are not epidemiologically connected. All strains were stored at −70°C in Luria-Bertani (LB) broth (Difco) with 10% glycerol. For each experiment in this study, fresh cultures derived from the frozen stocks were used.

Extraction of DNA.

Escherichia coli O157 strains were plated from frozen stocks on LB agar plates (Difco) and incubated overnight at 37°C. Bacterial growth from approximately one-quarter of each plate containing isolated colonies was collected using a sterile cotton swab and resuspended in 1 ml of 0.9% saline. After centrifugation at 10,000 × g for 1 min, the supernatant was removed and genomic DNA was extracted by using the Qiagen DNeasy tissue kit (Qiagen Inc., Valencia, CA), according to the manufacturer's instructions.

LSPA6 typing.

All E. coli O157 strains were characterized by the LSPA6 typing method, as described previously (20). Briefly, two multiplex PCRs were performed; the first PCR mixture contained primers for the folD, Z5935, rbsB, and arp-iclR genes and the second PCR mixture contained primers specific for the yhcG and rtcB genes. The amplicons were separated by capillary electrophoresis using an Applied Biosystems 3130 genetic analyzer (Applied Biosystems, Foster City, CA) and the GeneScan 600 LIZ size standard (Applied Biosystems). The size of the amplicons was determined by BioNumerics 6.1 (Applied Maths, Sint-Martens-Latem, Belgium). LSPA6 alleles and genotypes were defined by using a binary character table, as described previously (19). LSPA6 genotype 211111 was considered lineage I/II (19), whereas strains characterized as having LSPA6 genotype 111111 were defined as lineage I. All other allele combinations were considered lineage II. The alleles were placed in the order folD, Z5935, yhcG, rtcB, rbsB, and arp-iclR.

Virulence gene profiling.

The E. coli O157 strains were screened by PCR for the presence of 11 virulence genes, including enterohemorrhagic E. coli attachment-effacement gene (eae), enterohemolysin A gene (ehxA), E. coli secreted protein A gene (espA), IrgA homologue adhesion gene (iha), and various Shiga toxin genes (stx1, stx2, stx2c, stx2d, stx2d-activatable, stx2e, and stx2f). A multiplex PCR was used for the detection of stx1, eae, espA, and iha (22). Identification of the remaining seven virulence genes, ehxA, stx2, stx2c, stx2d, stx2d-activatable, stx2e, and stx2f, was performed individually by employing conventional PCR methods (22). Since the method used for profiling stx2 does not differentiate between stx2, stx2c, and stx2d-activatable (22), we employed restriction analysis using HaeIII, FokI (23), and PstI (22). Two strains, E. coli O157:H7 strain ATCC 43894 and E. coli O157:H7 strain EDL933, were used as positive controls, as they possessed all of the aforementioned virulence determinants. Pseudomonas fluorescens strain cc840406E was used as a negative control. PCRs were carried out in 50-μl reaction volumes, using an Eppendorf Mastercycler thermocycler (Eppendorf AG; Brinkman Instruments, Westbury, NY).

Antibiotic susceptibility testing.

Escherichia coli O157 strains obtained from both bovine and human sources were examined for antibiotic resistance determinants using Sensititre CMV1AGNF plates (TREK Diagnostic Systems, Cleveland, OH). The Sensititre plates each contained 17 antimicrobial agents, i.e., amikacin (0.5 to 32 μg/ml), ampicillin (1 to 64 μg/ml), amoxicillin-clavulanic acid (1/0.5 to 32/16 μg/ml), ceftriaxone (0.25 to 64 μg/ml), ciprofloxacin (0.015 to 4 μg/ml), cefoxitin (0.5 to 32 μg/ml), gentamicin (0.25 to 16 μg/ml), kanamycin (8 to 64 μg/ml), nalidixic acid (0.5 to 32 μg/ml), trimethoprim-sulfamethoxazole (0.12/2.38 to 4/76 μg/ml), ceftiofur (0.12 to 8 μg/ml), sulfisoxazole (16 to 256 μg/ml), streptomycin (32 to 64 μg/ml), tetracycline (4 to 32 μg/ml), and chloramphenicol (2 to 32 μg/ml), dosed in 96 wells at appropriate dilutions, as specified by the National Antimicrobial Resistance Monitoring System of the CDC. Each well of the Sensititre microtiter plate was inoculated according to the instructions of the manufacturer, followed by incubation at 37°C for 24 h. The MIC was manually determined for each isolate as being the lowest concentration of each antibiotic that inhibited visible growth. The MIC breakpoints were determined according to National Committee for Clinical Laboratory Standards performance standards M100-S12 (24) and M31-A2 (25).

Plasmid profiling.

Plasmid DNA was isolated as described elsewhere (26). After extraction, the plasmid DNA was analyzed by electrophoresis (55 min at 100 W) on a 2% (wt/vol) agarose gel containing 0.125 μg ml−1 ethidium bromide. A 1-kb DNA extension ladder (Invitrogen, Burlington, Canada) was used as a molecular weight marker. The bands were visualized under UV light and photographed using Alpha Imager software.

Assay for colicin production.

Colicin production by the E. coli O157 strains was determined using the method of Pugsley and Oudega (27). Briefly, a loopful of inoculum of freshly prepared E. coli O157 culture was placed on a circle (r = 1 cm) of a freshly prepared lawn of E. coli C600 indicator organism grown on a Trypticase soy agar plate. After 24 h of incubation at 37°C, the strains were examined for zones of clearing surrounding the spots of E. coli O157 inoculation. Colicin-positive E. coli O157 strains resulted in a clear zone around the point of inoculation due to colicin inhibition of the indicator organism's growth, whereas colicin-deficient E. coli O157 strains showed no zones of clearing. In this assay, E. coli O157:H7 strain EDL933 was used as a colicin-positive control, whereas E. coli DH5α was used as a negative control.

Population structure analysis.

The program Structure (28) was used to investigate the population structure of E. coli O157 strains, using six loci from the LSPA6 assay (i.e., folD, Z5935, yhcG, rtcB, rbsB, and arp-iclR), 11 virulence-associated genes (i.e., eae, ehxA, espA, iha, stx1, stx2, stx2c, stx2d, stx2d-activatable, stx2e, and stx2f), and antimicrobial resistance phenotypes. A Bayesian model approach was used to infer population structure (K) within the studied collection of E. coli O157 strains and to assign strains individually to the best-fitting population. K values were evaluated by a Markov chain Monte Carlo algorithm with a 10-model run. Each of the 10 runs consisted of 10,000 burn-in steps, followed by 100,000 iterations using the admixture model. The programs CLUMPP (29) and DISTRUCT (30) were used to generate the initial graphs depicting the genetic, virulence, and antimicrobial resistance phenotype distributions across the clinical and bovine strains of E. coli O157.

Statistical analysis.

In order to determine statistically significant differences in the proportions of genes and phenotypes for the human and bovine strains, we used Fisher's exact test (31). With the Bonferroni correction for multiple comparisons, we considered P values of ≤0.05/m statistically significant (m is the total number of members for a given family). Here, our study focuses mainly on three families, LSPA6 (m = 6), virulence factor (m = 11), and antibiotic susceptibility (m = 7). Also, a multivariate logistic regression model was used to illustrate the associations between the genetic, virulence, and antimicrobial resistance variables and the origins of STEC O157. We implemented both Fisher's exact test and the logistic regression model in R (32).

RESULTS

Lineage-specific polymorphism assay.

The majority (80%) of E. coli O157 isolates were typed as LSPA6 lineage I, followed by lineage II (15%) and lineage I/II (5%) (Table 1). Multivariate logistic regression analysis indicated that no significant (P > 0.0083) association existed between different LSPA6 genotypes and the origins of the strains. However, there was a significant difference (P < 0.006) in E. coli O157 LSPA6 genotype distributions for the clinical and bovine isolates. The most common human disease-associated genotype was LSPA6 lineage I (90%), followed by LSPA6 lineage I/II (6%) and LSPA6 lineage II (4%). Among the bovine E. coli O157 strains, LSPA6 lineage I was found at 70% frequency, whereas LSPA6 lineages II and I/II were detected with frequencies of 26% and 4%, respectively. The most significant human disease-associated alleles were LSPA6 lineage I rtcB (P < 0.000) and arp (P < 0.000), being found in 100% of the clinical E. coli O157 isolates. Three other alleles, i.e., LSPA6 lineage I Z5935, yhcG, and rbsB, also showed significant association (98%; P < 0.002) with human E. coli O157 strains. The folD allele showed a nonsignificant frequency (92%; P < 0.017) within this group of E. coli O157 strains. LSPA6 lineage I alleles also showed high frequencies among the bovine strains, ranging from 84% for rtcB to 76% for Z5935 and rbsB.

Table 1.

Distributions of LSPA6 genotypes and lineages within the geographically and temporally well-defined populations of human and bovine strains of E. coli O157

Genotype Lineage No. (%) of strains with the indicated LSPA6 lineage
Bovine strains (n = 50) Human strains (n = 50) Total (n = 100)
111111 I 35 (70) 45 (90) 80 (80)
211111 I/II 2 (4) 3 (6) 5 (10)
222222 II 2 (4) 0 2 (4)
222121 II 2 (4) 0 2 (4)
222122 II 4 (8) 0 4 (8)
111112 II 1 (2) 0 1 (2)
221221 II 3 (6) 0 3 (6)
221222 II 1 (2) 0 1 (2)
112111 II 0 1 (2) 1 (2)
231121a II 0 1 (2) 1 (2)
Open in a new tab
a

The second allele, Z5935, had a unique size and was classified as allele 3.

The bovine E. coli O157 strains showed a higher degree of diversity than the clinical E. coli O157 strains. Besides LSPA6 lineages I and I/II, the bovine strains possessed six genotypes within lineage II, including LSPA6 222121, 222122, 111112, 222222, 221222, and 221221, whereas the clinical strains had two LSPA6 genotypes, 112111 and 231121, within lineage II.

Virulence gene profiling.

The distribution of 11 virulence-associated genes, including eae, ehxA, espA, iha, stx1, stx2, stx2c, stx2d, stx2d-activatable, stx2e, and stx2f, among the 100 E. coli O157 strains is shown in Table 2. All E. coli O157 strains contained the stx2e gene. All bovine strains and virtually all (98%) clinical strains were positive for the attachment-effacement gene (eae) and the enterohemolysin A gene (ehxA). A similar pattern of distribution was observed for the homologue adhesion gene (iha) and the Shiga toxin gene stx2, with all bovine strains and the great majority (96%) of clinical strains being positive for both of these genes. The stx1 gene was found in 86% of the bovine strains and 78% of the clinical strains; however, this difference was not significant (P = 0.43). A variant of the stx2 gene, stx2f, had a much lower frequency of occurrence in both the bovine (34%) and clinical (32%) strains than did either the stx2 genes or the stx1 genes. Most noteworthy is the finding that the distribution of the stx2c gene showed the greatest difference (P = 0.023) between the bovine (24%) and clinical (6%) strains. The secreted protein A gene (espA) was found in only two bovine strains. All of the strains were negative for stx2d-activatable.

Table 2.

Distributions of virulence-associated genes within the collection of human clinical and bovine strains of E. coli O157 obtained from the region of western Canada in 2006

Virulence factor No. (%) of strains positive for the indicated virulence factor
 P
Total (n = 100) Human strains (n = 50) Bovine strains (n = 50)
eae 99 (99) 49 (98) 50 (100) 1.000
espA 2 (2) 0 2 (2) 0.495
ehxA 99 (99) 49 (98) 50 (100) 1.000
iha 98 (98) 48 (96) 50 (100) 0.495
stx1 82 (82) 39 (78) 43 (86) 0.436
stx2 98 (98) 48 (96) 50 (100) 0.495
stx2c 15 (15) 3 (6) 12 (24) 0.023
stx2d 91 (91) 47 (94) 44 (88) 0.487
stx2d-activatable 0 0 0 1.000
stx2e 100 (100) 50 (100) 47 (94) 0.242
stx2f 33 (33) 16 (32) 17 (34) 1.000
Open in a new tab

The most common combination of the various Shiga toxin genes in the examined collection of E. coli O157 isolates was the group with stx2 and stx2e, which accounted for 100% (n = 50) of the bovine strains and 96% (n = 48) of the clinical strains. The combination of three Shiga toxin genes, including stx2, stx2e, and stx2d, was another common feature seen in this collection, where it was shown to occur in 94% of both the bovine (n = 47) and clinical (n = 47) strains. Another common combination, involving four Shiga toxin variant genes, was the combination of stx1, stx2, stx2d, and stx2e, which was present in 76% (n = 38) of the clinical strains and 74% (n = 37) of the bovine strains. One clinical strain, which was isolated in Winnipeg, possessed only two Shiga toxin genes, stx2e and stx2f, of the 11 virulence genes tested.

Antibiotic susceptibility testing.

The antibiotic resistance patterns of the bovine (n = 50) and clinical (n = 50) strains of E. coli O157 are shown in Table 3. Approximately one-half of the E. coli O157 strains (54%; n = 54) were resistant to at least one antibiotic, whereas 46% (n = 46) of the strains were susceptible to all 17 antibiotics tested. Resistance was observed for sulfisoxazole, gentamicin, chloramphenicol, tetracycline, ampicillin, streptomycin, and cefoxitin. Multivariate logistic regression analysis revealed a significant (P = 0.00487) association of the antimicrobial resistance phenotypes with the origins of the strains. The clinical strains of E. coli O157 showed a higher rate of susceptibility (62%); of 50 strains, 31 were susceptible to all tested antibiotics, compared with the bovine strains (30%), of which 15 strains showed susceptibility to the same set of antibiotics. The most profound difference between the clinical and bovine strains was observed for the antibiotic tetracycline (P = 0.000), with 36% of the bovine strains showing resistance to this antibiotic but only 4% of the clinical strains being resistant. Also, strong differences between the clinical and bovine strains were found for sulfisoxazole (P = 0.009) and streptomycin (P = 0.01).

Table 3.

Antimicrobial susceptibility among the human clinical and bovine strains of E. coli O157 from western Canada

Antimicrobial No. (%) of strains
Human
 Bovine
Susceptible Resistant MDRa Susceptible Resistant MDR
Amikacin 50 (100) 0 50 (100) 0
Ampicillin 49 (98) 1 (2) 1 50 (100) 0
Amoxicillin-clavulanic acid 50 (100) 0 50 (100) 0
Ceftriaxone 50 (100) 0 50 (100) 0
Ciprofloxacin 50 (100) 0 50 (100) 0
Cefoxitin 49 (98) 1 (2) 0 50 (100) 0
Gentamicin 49 (98) 1 (2) 0 50 (100) 0
Kanamycin 50 (100) 0 50 (100) 0
Nalidixic acid 50 (100) 0 50 (100) 0
Trimethoprim-sulfamethoxazole 50 (100) 0 50 (100) 0
Ceftiofur 50 (100) 0 50 (100) 0
Sulfisoxazole 35 (70) 15 (30) 1 (2) 20 (40) 30 (60) 8 (16)
Streptomycin 48 (96) 2 (4) 2 (4) 42 (84) 8 (16) 8 (16)
Tetracycline 47 (94) 3 (6) 2 (4) 32 (64) 18 (36) 8 (16)
Chloramphenicol 49 (98) 1 (2) 0 46 (92) 4 (8) 3 (6)
Open in a new tab
a

Appears in the multidrug resistance (MDR) phenotype.

One pattern of multidrug resistance, involving tetracycline, streptomycin, and sulfisoxazole, was observed for both clinical (2%) and bovine (10%) strains. Another two patterns of multidrug resistance, including (i) tetracycline, ampicillin, and streptomycin and (ii) chloramphenicol, tetracycline, streptomycin, and sulfisoxazole, were observed for clinical (2%) and bovine (6%) strains, respectively.

Plasmid profiling.

The entire collection of clinical and bovine strains had eight different plasmid profiles. The great majority (99%) of E. coli O157 strains carried multiple plasmids, whereas only one clinical isolate harbored a single small plasmid (∼15 kb). There was no significant difference (P = 0.1) between the plasmid profiles of the clinical strains and the bovine strains. The most frequent plasmid profile, consisting of pO157 and a small plasmid (∼15 kb), was found in 36 (72%) of the clinical strains and 43 (86%) of the bovine strains. Another four multiple-plasmid profiles were found in seven bovine and nine clinical strains and involved pO157 and the small ∼15-kb plasmid in combination with one of the 20-kb, 35-kb, 40-kb, or 72-kb plasmids. Two more multiple-plasmid profiles, each involving four plasmids, were found in four clinical strains.

Colicin assay.

In total, 16 strains (16%) were found to produce colicin. Colicin production was observed in six clinical strains (12%) and 10 bovine strains (20%). The majority of colicinogenic strains (87.5%) were observed among LSPA6 lineage I, whereas one strain (6.25%) showing the colicinogenic phenotype was found within LSPA6 lineage II and another strain (6.25%) with the same phenotype was found within LSPA6 lineage I/II. No significant difference was found between the frequencies of clinical and bovine colicinogenic strains.

Population structure analysis.

The program Structure, which is based on estimated logarithmic probabilities, strongly suggested that the studied collection of E. coli O157 strains is composed of three distinct populations (K = 3) (Fig. 1). Furthermore, this assumption is confirmed by the fact that the majority of E. coli O157 strains were assigned to one of the three existing populations, creating asymmetric groups, which is a strong indication of the real population structure. The most notable observation on the studied collection of E. coli O157 isolates was that significant associations (88%; P = 0.005) existed between the human strains and population A and/or B (Fig. 2), whereas the strains of bovine origin did not show significant association with any specific population. Rather, the bovine isolates were equally distributed among the three populations (population A, 32.7%; population B, 32.8%; population C, 34.5%) (Fig. 2). The means of the pairwise distances indicated that population C is the most heterogeneous population (mean pairwise distance, 0.2349), whereas populations A (mean pairwise distance, 0.0508) and B (mean pairwise distance, 0.0541) are more homogeneous. Furthermore, triangle plots of the Structure analysis data (Fig. 1) revealed that population B is most likely a subgroup of population A, which indicates their close genetic, virulence, and antimicrobial phenotype profiles. The major characteristics of the strains that belong to populations A and B, which include the great majority of human strains, include (i) antibiotic susceptibility, i.e., to streptomycin (99.5%), tetracycline (91.8%), and sulfisoxazole (64%); (ii) a distinct virulence profile, i.e., stx1 positive (92.8%) and stx2c negative (99.3%); and (iii) a unique genotype profile, lineage I (98.1%). In contrast, the strains of population C, which predominantly includes the strains of bovine origin, are characterized by (i) increased antibiotic resistance, i.e., to sulfisoxazole (66.2%), tetracycline (54.8%), and streptomycin (40.2%); (ii) a distinct virulence profile, i.e., stx2c positive (54.4%) and stx1 negative (46.6%); and (iii) predominantly a different genotype profile, lineage II (66.5%).

Fig 1.

Fig 1

Open in a new tab

Triangle plot of the Structure analysis, a color-coded image showing the distributions of human (green circles) and bovine (red circles) E. coli O157 strains within the 3 distinct populations in the examined collection of E. coli O157 strains, as revealed by Bayesian cluster analysis. Population B is a subpopulation of population A, as they shared similar genetic, virulence, and antimicrobial phenotypes. Human strains of E. coli O157 were almost entirely absent from population C, which was characterized by a combination of different antimicrobial resistance phenotypes, lineage II, and stx2c-positive variants.

Fig 2.

Fig 2

Open in a new tab

Population structure of human and bovine strains of E. coli O157 obtained from western Canada during 2006. The entire collection of E. coli O157 strains consists of three clusters (K = 3). The three clusters are represented by different colors; cluster A is indicated by red, cluster B by green, and cluster C by blue; the values along the y axis represent the membership probabilities for the three clusters. Each strain is represented by a single vertical line partitioned into three segments, representing the genetic, virulence, and antimicrobial phenotype admixture that reflects the overall membership in each of these three clusters. The image was produced using the CLUMPP (42) and DISTRUCT (3) programs.

DISCUSSION

The present population-based study examined the (i) antimicrobial, (ii) genetic, and (iii) virulence profiles of a geographically and temporally well-defined collection of E. coli O157 isolates and revealed that antimicrobial susceptibility was the most distinctive feature (P = 0.00487) distinguishing clinical strains from bovine strains. Both the multivariate logistic regression analysis and Fisher's exact test strongly suggested that, among all of the tested antimicrobial, genetic, and virulence variables, resistance to tetracycline (P < 0.000) and resistance to sulfisoxazole (P < 0.009) were the characteristics most strongly associated with the bovine strains of E. coli O157. In contrast to the bovine strains, these antimicrobial phenotypes occurred at low frequencies among the human E. coli O157 strains. These statistically significant differences between the bovine and human strains of E. coli O157 most likely can be explained by the mass application of antimicrobial growth promoters (AGPs), including oxytetracycline and chlortetracycline, in intensively managed cattle populations in North America (33, 34). Exposure of the bovine strains of E. coli O157 to a constant selective pressure imposed by AGPs may result in the generation of antimicrobial resistance (35) that can be easily transmitted to other pathogens within the same population. We showed in our previous study (21) that the prevalence of this food-borne pathogen increased proportionally with cattle density, which further indicates that any emerging antimicrobial phenotype in these cattle population-dense environments can be easily disseminated throughout the population of E. coli O157 of bovine origin. In addition, O'Connor et al. (36) reported that the use of injectable oxytetracycline in cattle receiving in-feed chlortetracycline was associated with an increase in the prevalence of resistance to sulfisoxazole, another antimicrobial agent with resistance significantly associated with the bovine strains in the present study. The high prevalence of sulfisoxazole and streptomycin resistance among the bovine strains might be explained by profound effects of tetracycline on the transfer, mobilization, and transposition of mobile genetic elements that might contain a variety of antibiotic resistance gene cassettes (37–39). It was shown that a >100-fold increase in gene transfer occurred among bacteria harboring transposons that were exposed to low concentrations of tetracycline (38, 39). Our data support this explanation with the fact that the most common multidrug-resistant phenotype observed in the collection of E. coli O157 strains included resistance to tetracycline, sulfisoxazole, and streptomycin. This observation indicates that these two antibiotic-resistant phenotypes, involving sulfisoxazole and streptomycin, were most likely introduced into the bovine strains via tetracycline-induced transposition of a mobile DNA element that already contained all three resistant determinants. The high rate of antimicrobial susceptibility among the clinical strains of E. coli O157 was most likely associated with the absence of selective pressure within the population of this human pathogen. Another factor contributing to the increased antimicrobial susceptibility among the clinical E. coli O157 strains might be associated with the short-lived nature of tetracycline resistance (40) and spontaneous or stress-induced loss of mobile DNA elements that harbor multidrug resistance genes. A high level of transposon loss was noted when bacteria were exposed to UV light or other stress-inducing treatments (41, 42). It might be assumed that clinical strains of E. coli O157, changing their natural habitat and entering into the human food chain, would be exposed to significant stress conditions for extended periods prior to human infection, which might result in rapid loss of the mobile DNA elements that harbor antibiotic resistance genes. Taken together, these results support the hypothesis that the mass application of AGPs to food animals housed in population-dense environments most likely has a significant effect on the emergence and dissemination of antimicrobial resistance phenotypes (34, 43).

During the past decade, population genetics-based studies have increasingly shaped our understanding of the diversity and complexity of E. coli O157 populations. Several studies using LSPA6 genotyping showed distinct geographically dependent distributions of E. coli O157 isolates with clinical and bovine origins (16, 17, 19, 20). For instance, Yang et al. (20), using a large collection of 1,429 human and bovine isolates from the United States and Australia, showed that lineage I was the LSPA6 genotype most frequently identified in both human and bovine isolates, including dairy herds and feedlots. Another study from Canada also pointed out that lineage I was the predominant LSPA6 genotype among human and bovine isolates (19). A population-based study from Japan (17) revealed that lineages II and I/II were the most common LSPA6 genotypes among bovine isolates, whereas lineage I was the predominant (69.2%) LSPA6 genotype among human clinical isolates. Recently, Franz and colleagues (16) showed that the distributions of LSPA6 genotypes among human and bovine O157 isolates in the Netherlands were quite different from those in the United States, Australia, and Japan. The most notable differences were the low rate of occurrence of lineage I and the high rate of occurrence of lineage I/II in human isolates of E. coli O157 from the Netherlands. The present study showed a predominance of lineage I among human (90%) and bovine (70%) isolates from a vast area of western Canada and, together with the previously published studies (19, 20), indicates the importance of lineage I in the ecology and epidemiology of E. coli O157 isolates from North America.

Evaluation of the E. coli O157 strains for 11 virulence-associated genes revealed that intimin, the enterohemorrhagic E. coli hemolysin, the IrgA homologue adhesin encoded by iha (44), and the Shiga toxins encoded by stx2 and a stx2 variant were distributed virtually across all clinical and bovine strains, indicating the ubiquitous nature of these virulence-associated determinants regardless of the pathogen's origin. A study examining the population genetics of the Shiga toxin-producing E. coli (STEC) in Montana in the United States, targeting the same 11 virulence-associated genes, also showed the high prevalence of intimin (100%), enterohemorrhagic E. coli hemolysin (100%), IrgA homologue adhesin (100%), and stx2 variant (91%) genes within the E. coli O157 population. Notably, these four virulence determinants were less frequent among non-O157 STEC strains, with occurrence rates ranging from 36% for the stx2 variant to 88% for enterohemorrhagic E. coli hemolysin (22), indicating a lower virulence potency of non-O157 STEC than of E. coli O157. In sharp contrast to these four virulence-associated genes, which were evenly distributed among the human and bovine strains, we found that the stx2c Shiga toxin variant was the most segregative virulence determinant (P = 0.02) and was associated with E. coli O157 strains of bovine origin. Our data imply that the association between the stx2c variant and the bovine strains was due to a very strong correlation (93.3%) between lineage II and stx2c. A majority of lineage II genotypes (86.7%) were identified among bovine strains, whereas only a small fraction (13.7%) of these LSPA6 genotypes were observed among clinical strains. Several other studies also reported a strong association between stx2c and bovine isolates (16, 17, 19).

Population structure analysis of human clinical and bovine O157 strains, obtained over the same period of time from a confined geographical region, confirmed the earlier hypothesis that bovine strains represent a more heterogeneous group of organisms than human clinical isolates (16, 17, 45, 46). More importantly, we have shown that the greater diversity of bovine strains is attributable to a distinct population of pathogen that is rarely found among clinical strains (Fig. 1). This unique population of E. coli O157 strains, termed population C, was the most heterogeneous pathogen group, as indicated by the mean of the pairwise distances, with unique antimicrobial, genetic, and virulence profiles. The most important characteristics of population C of E. coli O157 strains were a high degree of antimicrobial resistance, a relatively high prevalence of different LSPA6 genotypes of lineage II, and a high frequency of stx2c variants. In contrast, the other two populations (populations A and B) encompassed a great majority of the human clinical strains (88%) and were characterized by a high frequency of antimicrobial susceptibility, a high prevalence of the lineage I genotype, and a common virulence profile that included a combination of stx1-positive and stx2c-negative variants.

In summary, the present study, which tested a temporally and geographically well-defined population of E. coli O157 strains, clearly indicated that, among different antimicrobial, genetic, and virulence determinants, antimicrobial susceptibility was the most distinctive characteristic that distinguished human strains from bovine strains of E. coli O157 in the region of western Canada. Furthermore, this finding implies that the selective pressure that generates antimicrobial resistance (use of antimicrobials to promote animal growth) in intensively managed livestock operations has a large impact on the population of this human pathogen. In addition, by implementing a broad spectrum of antimicrobial, genetic, and virulence analyses in combination with molecular evolutionary analyses, we were able to identify a unique E. coli O157 group rarely found among clinical strains and commonly identified among bovine strains.

ACKNOWLEDGMENTS

We thank Victor Gannon from the Public Health Agency of Canada for the kind provision of DNA from three E. coli O157:H7 strains that were used as positive controls for the LSPA6 typing. We also thank Daniela Vidovic for her technical support.

The Alberta Beef Cattle Research Council, NSERC, and NSERC-CRD are acknowledged for financial support.

Footnotes

Published ahead of print 24 April 2013

REFERENCES

  • 1. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, Hebert RJ, Olcott ES, Johnson LM, Hargrett NT, Blake PA, Cohen ML. 1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N. Engl. J. Med. 308:681–685 [DOI] [PubMed] [Google Scholar]
  • 2. Stewart PJ, Desormeaux W, Chene J. 1983. Hemorrhagic colitis in a home for the aged: Ontario. Can. Dis. Wkly. Rep. 9:29–32 [Google Scholar]
  • 3. Public Health Agency of Canada, Division of Foodborne and Enteric Diseases, Health Protection Branch 1998. Canadian integrated surveillance report for 1995 on Salmonella, Campylobacter and pathogenic Escherichia coli. Can. Commun. Dis. Rep. 24(Suppl 5):1 [Google Scholar]
  • 4. Wylie JL, van Caeseele P, Gilmour MW, Sitter D, Guttek C, Giercke S. 2013. Evaluation of a new chromogenic agar medium for detection of Shiga toxin-producing Escherichia coli (STEC) and relative prevalences of O157 and non-O157 STEC in Manitoba, Canada. J. Clin. Microbiol. 51:466–471 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Infectious Disease Surveillance Center, National Institute of Infectious Diseases and Infectious Diseases Control Division, Ministry of Health and Welfare of Japan 1997. Verocytotoxin-producing Escherichia coli (enterohemorrhagic E. coli) infection, Japan, 1996-June 1997. Infect. Agents Surveill. Rep. 18:153–154 [Google Scholar]
  • 6. O'Brien AD, Tesh VL, Donohue-Rolfe A, Jackson MP, Olsnes S, Sandvig K, Lindberg AA, Keusch GT. 1992. Shiga toxin: biochemistry, genetics, mode of action, and role in pathogenesis. Curr. Top. Microbiol. Immunol. 180:65–94 [DOI] [PubMed] [Google Scholar]
  • 7. Griffin PM, Ostroff SM, Tauxe RV, Greene KD, Wells JG, Lewis JH, Blake PA. 1988. Illnesses associated with Escherichia coli O157:H7 infections: a broad clinical spectrum. Ann. Intern. Med. 109:705–712 [DOI] [PubMed] [Google Scholar]
  • 8. Su CY, Brandt LJ. 1995. Escherichia coli O157:H7 infection in humans. Ann. Intern. Med. 123:698–714 [DOI] [PubMed] [Google Scholar]
  • 9. Elder RO, Keen JE, Siragusa GR, Barkocy-Gallagher GA, Koohmaraie M, Laegreid WW. 2000. Correlation of enterohemorrhagic Escherichia coli O157 prevalence in feces, hides, and carcasses of beef cattle during processing. Proc. Natl. Acad. Sci. U. S. A. 97:2999–3003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Laven RA, Ashmore A, Stewart CS. 2003. Escherichia coli in the rumen and colon of slaughter cattle, with particular reference to E. coli O157. Vet. J. 165:78–83 [DOI] [PubMed] [Google Scholar]
  • 11. Crump JA, Sulka AC, Langer AJ, Schaben C, Crielly AS, Gage R, Baysinger M, Moll M, Withers G, Toney DM, Hunter SB, Hoekstra RM, Wong SK, Griffin PM, van Gilder TJ. 2002. An outbreak of Escherichia coli O157:H7 infections among visitors to a dairy farm. N. Engl. J. Med. 347:555–560 [DOI] [PubMed] [Google Scholar]
  • 12. Durso LM, Reynolds K, Bauer N, Jr, Keen JE. 2005. Shiga-toxigenic Escherichia coli O157:H7 infections among livestock exhibitors and visitors at a Texas county fair. Vector Borne Zoonotic Dis. 5:193–201 [DOI] [PubMed] [Google Scholar]
  • 13. Gilbert M, Monk C, Wang HL, Diplock K, Landry L. 2008. Screening policies for daycare attendees: lessons learned from an outbreak of E. coli O157:H7 in a daycare in Waterloo, Ontario. Can. J. Public Health 99:281–285 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Karmali MA. 1989. Infection by verotoxin-producing Escherichia coli. Clin. Microbiol. Rev. 2:15–38 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, Widdowson MA, Roy SL, Jones JL, Griffin PM. 2011. Foodborne illness acquired in the United States: major pathogens. Emerg. Infect. Dis. 17:7–15 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Franz E, van Hoek AHAM, van der Wal FJ, de Boer A, Zwartkruis-Nahuis A, van der Zwaluw K, Aarts HJM, Heuvelink AE. 2012. Genetic features differentiating bovine, food, and human isolates of Shiga toxin-producing Escherichia coli O157 in the Netherlands. J. Clin. Microbiol. 50:772–780 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Lee K, French NP, Hara-Kudo Y, Iyoda S, Kobayashi H, Sugita-Konishi Y, Tsubone H, Kumagai S. 2011. Multivariate analyses revealed distinctive features differentiating human and cattle isolates of Shiga toxin-producing Escherichia coli O157 in Japan. J. Clin. Microbiol. 49:1495–1500 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Whitworth J, Zhang Y, Bono J, Pleydell E, French N, Besser T. 2010. Diverse genetic markers concordantly identify bovine origin Escherichia coli O157 genotypes underrepresented in human disease. Appl. Environ. Microbiol. 76:361–365 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Ziebell K, Steele M, Zhang Y, Benson A, Taboada EN, Laing C, McEwen S, Ciebin B, Johnson R, Gannon V. 2008. Genotypic characterization and prevalence of virulence factors among Canadian Escherichia coli O157:H7 strains. Appl. Environ. Microbiol. 74:4314–4323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Yang Z, Kovar J, Kim J, Nietfeldt J, Smith DR, Moxley RA, Olson ME, Fey PD, Benson AK. 2004. Identification of common sub-populations of non-sorbitol-fermenting, beta-glucuronidase-negative Escherichia coli O157:H7 from bovine production environments and human clinical samples. Appl. Environ. Microbiol. 70:6846–6854 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Vidovic S, Korber DR. 2006. Prevalence of Escherichia coli O157 in Saskatchewan cattle: characterization of isolates by using random amplified polymorphic DNA PCR, antibiotic resistance profiles, and pathogenicity determinants. Appl. Environ. Microbiol. 72:4347–4355 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Jelacic JK, Damrow T, Chen GS, Jelacic S, Bielaszewska M, Ciol M, Carvalho HM, Melton-Celsa AR, O'Brien AD, Tarr PI. 2003. Shiga toxin-producing Escherichia coli in Montana: bacterial genotypes and clinical profiles. J. Infect. Dis. 188:719–729 [DOI] [PubMed] [Google Scholar]
  • 23. Russmann H, Schmidt H, Heesemann J, Caprioli A, Karch H. 1994. Variants of Shiga-like toxin II constitute a major toxin component in Escherichia coli O157 strains from patients with haemolytic uremic syndrome. J. Med. Microbiol. 40:338–343 [DOI] [PubMed] [Google Scholar]
  • 24. National Committee for Clinical Laboratory Standards 2002. Performance standards for antimicrobial susceptibility testing; 12th informational supplement. M100-S12 National Committee for Clinical Laboratory Standards, Wayne, PA [Google Scholar]
  • 25. National Committee for Clinical Laboratory Standards 2002. Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standard. M31-A2 National Committee for Clinical Laboratory Standards, Wayne, PA [Google Scholar]
  • 26. Sambrook J, Fritsch EF, Maniatis T. 1999. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY [Google Scholar]
  • 27. Pugsley AP, Oudega B. 1987. Methods of studying colicins and their plasmids, p 105–161 In Hardy KG. (ed), Plasmids: a practical approach. IRL Press, Oxford, England [Google Scholar]
  • 28. Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945–959 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Jakobsson M, Rosenberg NA. 2007. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23:1801–1806 [DOI] [PubMed] [Google Scholar]
  • 30. Rosenberg NA. 2004. DISTRUCT: a program for the graphical display of population structure. Mol. Ecol. Notes 4:137–138 [Google Scholar]
  • 31. Agresti A. 2002. Categorical data analysis. John Wiley & Sons, Inc., Hoboken, NJ [Google Scholar]
  • 32. Development Core Team 2011. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria [Google Scholar]
  • 33. Dibner JJ, Richards JD. 2005. Antibiotic growth promoters in agriculture: history and mode of action. Poult. Sci. 84:634–643 [DOI] [PubMed] [Google Scholar]
  • 34. Marshall BM, Levy SB. 2011. Food animals and antimicrobials: impacts on human health. Clin. Microbiol. Rev. 24:718–733 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Food and Drug Administration 2012. The judicious use of medically important antimicrobial drugs in food-producing animals. Guidance for industry no. 209. Food and Drug Administration, Center for Veterinary Medicine, Washington, DC [Google Scholar]
  • 36. O'Connor AM, Poppe C, McEwen SA. 2002. Changes in the prevalence of resistant Escherichia coli in cattle receiving subcutaneously injectable oxytetracycline in addition to in-feed chlortetracycline compared with cattle receiving only in-feed chlortetracycline. Can. J. Vet. Res. 66:145–150 [PMC free article] [PubMed] [Google Scholar]
  • 37. Davies J. 1994. Inactivation of antibiotics and the dissemination of resistance genes. Science 264:375–382 [DOI] [PubMed] [Google Scholar]
  • 38. Stevens AM, Shoemaker NB, Li LY, Salyers AA. 1993. Tetracycline regulation of genes on Bacteroides conjugative transposons. J. Bacteriol. 175:6134–6141 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Torres OR, Korman RZ, Zahler SA, Dunny GM. 1991. The conjugative transposon Tn925: enhancement of conjugal transfer by tetracycline in Enterococcus faecalis and mobilization of chromosomal genes in Bacillus subtilis and E. faecalis. Mol. Gen. Genet. 225:395–400 [DOI] [PubMed] [Google Scholar]
  • 40. Stabler SL, Fagerberg DJ, Quarles CL. 1982. Effects of oral and injectable tetracyclines on bacterial drug resistance in feedlot cattle. Am. J. Vet. Res. 43:1763–1766 [PubMed] [Google Scholar]
  • 41. Eichenbaum Z, Livneh Z. 1998. UV light induces IS10 transposition in Escherichia coli. Genetics 149:1173–1181 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Godoy VG, Fox MS. 2000. Transposon stability and a role for conjugational transfer in adaptive mutability. Proc. Natl. Acad. Sci. U. S. A. 97:7393–7398 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Wittum TE. 2012. The challenge of regulating agricultural ceftiofur use to slow the emergence of resistance to extended-spectrum cephalosporins. Appl. Environ. Microbiol. 78:7819–7821 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Tarr PI, Bilge SS, Vary JC, Jelacic S, Habeeb RL, Ward TR, Baylor MR, Besser TE. 2000. Iha: a novel Escherichia coli O157:H7 adherence-conferring molecule encoded on a recently acquired chromosomal island of conserved structure. Infect. Immun. 68:1400–1407 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Besser TE, Shaikh N, Holt NJ, Tarr PI, Konkel ME, Malik-Kale P, Walsh CW, Whittam TS, Bono JL. 2007. Greater diversity of Shiga toxin-encoding bacteriophage insertion sites among Escherichia coli O157:H7 isolates from cattle than in those from humans. Appl. Environ. Microbiol. 73:671–679 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Leopold SR, Magrini V, Holt NJ, Shaikh N, Mardis ER, Cagno J, Ogura Y, Iguchi A, Hayashi T, Mellmann A, Karch H, Besser TE, Sawyer SA, Whittam TS, Tarr PI. 2009. A precise reconstruction of the emergence and constrained radiations of Escherichia coli O157 portrayed by backbone concatenomic analysis. Proc. Natl. Acad. Sci. U. S. A. 106:8713–8718 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Clinical Microbiology are provided here courtesy of American Society for Microbiology (ASM)

RESOURCES