Detection of virulence-associated genes in Escherichia coli O157 and non-O157 isolates from beef cattle, humans, and chickens

Food-producing animals can be reservoirs of pathogenic Escherichia coli strains that can induce diseases in animals or humans. Contamination of food by E. coli O157:H7 raises immediate concerns about public health, although it is not clear whether all E. coli O157 isolates of animal origin are equally harmful to humans. Inversely, the pathogenic potential of atypical E. coli O157 isolates and several non-O157 serotypes often is ignored. We used a DNA microarray capable of detecting a subset of 346 genes to compare the virulence-associated genes present in eight E. coli O157 isolates from human cases, 14 antibiotic-resistant and/or hypermutable E. coli O157 isolates from beef cattle, and four antibiotic-resistant, sorbitol-negative, non-O157 E. coli isolates from healthy broiler chickens. Hybridization on arrays (HOA) revealed that O157 isolates from beef cattle and humans were genetically distinct, although they possessed most of the same subset of virulence genes. HOA allowed discrimination between hypermutable and antibiotic-resistant O157 isolates from beef cattle based on hybridization results for the stx2 and ycgG genes (hypermutable) or ymfL, stx1, stx2, and hlyE(avian) genes (resistant). However, the absence of hybridization to gene yfdR characterized human isolates. HOA also revealed that an atypical sorbitol-fermenting bovine O157 isolate lacked some genes of the type 3 secretion system, plasmid pO157, and the stx1 and stx2 genes. This isolate had a particular pathotype (eaeA(beta) tir(alpha) espA(alpha) espB(alpha) espD(alpha)) not found in typical E. coli O157:H7. HOA revealed that some non-O157 E. coli isolates from healthy chickens carried genes responsible for salmochelin- and yersiniabactin-mediated iron uptake generally associated with pathogenic strains.     

Isolation and molecular characterization of Escherichia coli O157 isolated from cattle,pigs and chickens at slaughter

From 1999 until 2001, 3625 food samples were examined for the presence of Escherichia coli O157. Samples were from bovine origin (ground beef, n=549; carcasses, n=2452), calves (carcasses, n=147), chicken (breast, n=203; carcasses, n=71) and pigs (carcasses, n=85; trimmings, n=118). Vidas ECO detected 451 (12%) samples positive, but from only 27 (0.74%) samples was E. coli O157 isolated. One strain was isolated from bovine ground beef (0.18%), one from a pig carcass (1.17%) and all others were isolated from bovine carcasses (1.02%). All strains possessed the attaching-and-effacing gene, the enterohemorrhagic plasmid and verotoxin (VT) genes, except the strain isolated from the pig carcass that was therefore eliminated. Six of the strains were urease-positive. Strains were typed by two DNA fingerprinting methods: random amplification of polymorphic DNA (RAPD) and pulsed field gel electrophoresis (PFGE). PFGE revealed a similarity of 71.05%, while RAPD was 77.36% similar. None of the typing methods were able to classify all urease-positive strains to one pattern. Strains in the same PFGE cluster did not belong to one RAPD cluster. This paper highlights that Belgian fresh meat at retail level can be contaminated with E. coli O157 and that two different typing methods divide strains into different types.     

Antimicrobial resistance of Escherichia coli O157 from cattle in Korea

Of 45 Escherichia coli O157 isolates from cattle feces, which were collected between May 2000 and September 2003 in Korea, 32 were resistant to at least 1 antibiotic and 28 were resistant to 4 or more antibiotics, with 32, 30 and 30 of the isolates being resistant to streptomycin, tetracycline and sulfisoxazole, respectively. Two isolates were resistant to fluoroquinolones and to 10 or more of the 22 other antimicrobial agents that were tested. Thirteen antimicrobial resistant patterns were observed. The most frequent resistance type, which was found for 11 isolates, was streptomycin-tetracycline-kanamycin-ampicillin-piperacillin-cephalothin-sulfisoxazole-ticarcillin. Polymerase chain reaction (PCR) analysis of the isolates for E. coli O157 virulence markers revealed that 25 of the resistant E. coli O157 isolates tested positive for stx2 or both stx1 and stx2 genes. These findings suggest that many of the resistant E. coli O157 isolates might cause disease in humans.     

Isolation and characterization of verocytotoxin-producing Escherichia coli O157 from Turkish cattle

The objective of this study was to collect rectal swabs from the cattle in a slaughterhouse located in Hatay (Turkey) immediately after slaughter for the isolation and characterization of verotoxin-producing Escherichia coli O157 in each month during a 1-year period. The rectal swab samples were analyzed for the isolation of E. coli O157 through pre-enrichment, immunomagnetic separation and selective plating on CT-SMAC agar. E. coli O157 was isolated from 77 (13.6%) of the samples. The presence of E. coli O157 changed during a 1-year period, in that the occurrence of E. coli O157 was the highest in July and November and lowest in February. A total of 66 isolates out of 77 were serotype O157:H7 and 11 were serotype O157:NM. PCR analysis of E. coli O157 virulence genes revealed that all O157:H7/NM were positive for rbf(O157), 74 positive for EhlyA, 72 positive for eaeA, 62 positive for vtx2, and 3 positive for both vtx1 and vtx2. It was presented by cytotoxicity tests that many of E. coli O157 isolates showed high cytotoxicity on Vero cells. All of the isolates containing EhlyA showed enterohaemolysin production.     

Typing of Escherichia coli O157 strains isolated from fresh sausage

E. coli O157 is a foodborne pathogen responsible for serious human illnesses, such as hemorrhagic colitic and hemolytic uremic syndrome. Ground beef products are among the foods that are more commonly contaminated, and the strains isolated have been frequently found to carry virulence factors of this pathotype. This paper reports the results of serotyping assays and of investigations performed to screen for virulence factors of 10 E. coli O157 strains isolated from fresh sausages purchased at retail meat outlets in various provinces of Apulia (southern Italy). The presence of verocytotoxins was assessed on VERO cells and ELISA tests. Multiplex PCR assays were performed for the eae, stx₁, stx₂ and hlyA genes. Six of the 10 strains examined presented the H7 antigen and all of them proved to be potentially pathogenic due to the presence of individual or multiple virulence factors.     

Evaluation of molecular typing methods for Escherichia coli O157:H7 isolates from cattle, food, and humans

Escherichia coli O157:H7, a Shiga toxin-producing E. coli, has been the causative agent of many cases of severe, often life-threatening foodborne illness. Because of the importance of E. coli O157:H7 to public health, many molecular typing methods have been developed to determine its transmission routes and source of infection during epidemiological investigations. Pulsed-field gel electrophoresis (PFGE) is currently used by public health organizations to track infections of E. coli O157:H7 and other foodborne pathogens. In this study, we compared the ability of PFGE, multilocus sequence typing (MLST), and repetitive-element PCR (Rep-PCR) to distinguish among 92 E. coli O157:H7 isolates from cattle, food, and infected humans. Several virulence genes, including the intimin gene (eaeA), the hemolysin gene (hlyA), and the H7 fimbrial gene (fliC), and a housekeeping gene for beta-glucuronidase (uidA) were included in MLST. Rep-PCR reactions were performed using a commercially available typing kit (Bacterial Barcodes Inc., Houston, Tex.) with the provided Uprime-RI primer set. Results of the study indicated that PFGE provided the most discrimination among the techniques, identifying 72 distinct PFGE profiles for the isolates; Rep-PCR elucidated 14 different profiles, whereas MLST generated five profiles. Additionally, there did not appear to be any correlation among the typing methods examined in this study. Therefore, to date, PFGE remains the technique of choice for molecular subtyping of E. coli O157:H7.     

Detection and frequency of VT1, VT2 and eaeA genes in Escherichia coli O157 and O157:H7 strains isolated from cattle, cattle carcasses and abattoir environment in Istanbul

The aim of this study was to detect VT1, VT2 and eaeA genes and to determine the frequency of these genes in Escherichia coli O157 and O157:H7 strains isolated from cattle, cattle carcasses and environmental samples of the 5 abattoirs located in Istanbul, Turkey. For this, the presence of VT1, VT2 and eaeA genes in 26 strains of E. coli O157:H7 and 6 strains of O157 was investigated by multiplex-PCR. The results have shown that eaeA gene was detected in all O157 and O157:H7 strains tested. Both VT2 and eaeA genes were detected in 4 (80%) of 5 strains of E. coli O157 and eaeA alone in 1 strain of O157. In 27 strains of O157:H7, 5 (18.5%) strains were found to be positive for VT1, VT2 and eaeA genes, 19 (70.3%) strains for both VT2 and eaeA and, 3 (11.1%) strains for only eaeA gene. Either VT1 alone or VT2 alone was not detected in any strains tested. eaeA gene alone in 2 strains, VT2-eaeA genes in 9 strains and VT1-VT2-eaeA genes in 2 strains were detected in 13 of E. coli O157:H7 strains isolated from cattle. eaeA alone in 1 strain, VT2-eaeA genes in 5 strains and VT1-VT2-eaeA genes in 2 strains were detected in 8 of E. coli O157:H7 strains isolated from carcasses. VT2-eaeA genes in 5 strains (isolated from hands, apron, knife and floor) and VT1-VT2-eaeA genes in 1 strain (isolated from knife) were also detected in 6 of E. coli O157:H7 strains isolated from environmental samples. This study reveals that most of the strains are found to be toxigenic and it is most likely that strains isolated from carcasses and abattoir environment originated from cattle feces. Therefore, HACCP systems are necessary from farm to table especially in the abattoirs to prevent contamination of meat and abattoir environment with intestinal content.     

Isolation and characterization of verocytotoxin-producing Escherichia coli O157 from slaughter pigs and poultry

Rectal contents and tonsils from Dutch slaughter pigs collected immediately after slaughter were examined for the presence of verocytotoxin (VT)-producing Escherichia coli (VTEC) of serogroup O157 (O157 VTEC). In addition, fresh fecal material from poultry layer flocks and turkey flocks collected on poultry farms was examined for the presence of O157 VTEC. E. coli O157 strains were isolated from two (1.4%) of 145 pigs. The strains were isolated from samples of rectal contents, all samples of tonsils being negative. While all 501 fecal samples from chicken flocks were found negative, E. coli O157 strains were isolated from six (1.3%) of 459 pooled fecal samples from turkey flocks. One of the porcine isolates and one of the turkey isolates contained the VT2 gene, the E. coli attaching-and-effacing gene, as well as the enterohemorrhagic E. coli hemolysin gene. Production of VT was confirmed by cytotoxicity tests on Vero cells. Based on these characteristics, the two stains were regarded as potentially pathogenic for humans. The porcine and the turkey isolate were further characterized as being of phage types 4 and 14, respectively. While biochemically typical of E. coli O157, the remaining six isolates were nonverocytotoxigenic and negative for both the E. coli attaching-and-effacing gene and the enterohemorrhagic E. coli hemolysin gene. All eight E. coli O157 isolates did not carry genes that encode E. coli heat-labile and heat-stable enterotoxins. It was concluded that pigs and poultry can be a source of O157 VTEC strains characteristic of those causing illness in man. The extent to which pigs and poultry play a role in the epidemiology of human O157 VTEC infection needs further research.     

Escherichia coli O157 and non-O157 Shiga toxin-producing Escherichia coli in fecal samples of finished pigs at slaughter in Switzerland

Fecal samples from 630 slaughtered finisher pigs were examined by PCR to assess the shedding of Escherichia coli O157 (rfbE) and Shiga toxin-producing E. coli (STEC, stx). The proportion of positive samples was 7.5% for rfbE and 22% for stx. By colony hybridization, 31 E. coli O157 and 45 STEC strains were isolated, and these strains were further characterized by phenotypic and genotypic traits. Among E. coli O157 strains, 30 were sorbitol positive, 30 had an H type other than H7, and none harbored stx genes. Intimin (eae), enterohemolysin (ehxA), EAST1 (astA), and porcine A/E-associated protein (paa) were present in 10, 3, 26, and 6% of strains. Among them, one eae-gamma1-positive O157:H7 strain testing positive for ehxA and astA and two eae-alpha1-positive O157:H45 strains were classified as enteropathogenic E. coli (EPEC). The O157:H45 EPEC harbored the EAF plasmid and the bfpA gene, factors characteristic for typical EPEC. The isolated STEC strains (43 sorbitol positive) belonged to 11 O:H serotypes, including three previously reported in human STEC causing hemolytic uremic syndrome (O9:H-, O26:H-, and O103:H2). All but one strain harbored stx2e. The eae and ehxA genes, which are strongly correlated with human disease, were present in only one O103:H2 strain positive for stx1 and paa, whereas the astA gene was found more frequently (14 strains). High prevalence of STEC was found among finisher pigs, but according to the virulence factors the majority of these strains seem to be of low virulence.     

Escherichia coli O157:H7 in faeces from cattle, sheep and pigs in the southwest part of Norway during 1998 and 1999

During a 2-year period from January 1998 to December 1999, intestinal content from 1541 cattle, 665 sheep and 1976 pigs were analysed for Escherichia coli O157:H7 using the immunomagnetic separation procedure. The animals originated from 848, 605 and 832 herds from the southwest part of Norway, respectively. E. coli O157:H7 was present in three samples from cattle from different herds, giving a herd prevalence of 0.35% and an animal prevalence of 0.19%. From pigs, E. coli O157:H7 was isolated from two pigs from different herds, giving a herd prevalence of 0.24% and an animal prevalence of 0.1%. A follow-up study revealed another positive testing pig from one of these herds. E. coli O157:H7 was not found from any of the 665 investigated sheep. By PCR analysis, all six E. coli O157:H7 isolates were shown to contain the genes encoding Shiga toxin 2 (stx2), the intimin protein (eae) and the H7 flagellum (fliC-H7). One of the cattle isolates also harboured the Shiga toxin 1 encoding (stx1) gene. The six isolates were differentiated into three pulse-field gel electrophoresis profiles. The results indicate that the occurrence of E. coli O157:H7 in cattle, sheep and pigs in the southwest part of Norway is low compared to other European countries.     

Prevalence of Escherichia coli O157 in cattle and swine in central Mexico

Escherichia coli O157:H7 is a foodborne pathogenic bacterium that can reside undetected in the gastrointestinal tract of cattle because colonization by this bacterium is asymptomatic. Recent research has indicated that swine can carry and transmit this pathogen as well. The development of more advanced and sensitive detection techniques has improved the limit of detection and increased sensitivity for this important pathogen. This study was undertaken to determine the prevalence of E. coli O157 in cattle and swine in Mexico with the more sensitive detection technique of immunomagnetic bead separation. Samples (n = 60 per farm) were taken from four cattle and four swine farms (n = 240 cattle samples, n = 240 swine samples) located throughout central Mexico in October 2001. The prevalence of E. coli O157 was found to be only 1.25% on cattle farms and 2.1% on swine farms. The prevalence in cattle in this study is lower than that reported in the United States and could be related to the lower reported prevalence of E. coli O157 in humans in Mexico. However, further research is needed to verify prevalence throughout other regions of Mexico, as well as prevalence during other seasons of the year.     

Tracking verocytotoxigenic Escherichia coli O157, O26, O111, O103 and O145 in Irish cattle

The purpose of this study was to investigate carriage and transfer of verocytotoxigenic Escherichia coli (VTEC) O157, O26, O111, O103 and O145 from faeces and hide to dressed carcasses of Irish cattle as well as establishing the virulence potential of VTEC carried by these cattle. Individual cattle was tracked and faecal samples, hide and carcass (pre-evisceration and post-wash) swabs were analysed for verotoxin (vt1 and vt2) genes using a duplex real-time PCR assay. Positive samples were screened for the five serogroups of interest by real-time PCR. Isolates were recovered from PCR positive samples using immunomagnetic separation and confirmed by latex agglutination and PCR. Isolates were subject to a virulence screen (vt1, vt2, eaeA and hlyA) by PCR. Isolates carrying vt genes were examined by Pulsed-Field Gel Electrophoresis (PFGE). Of the VTEC isolated, E. coli O157 was the most frequently recovered from hide (17.6%), faeces (2.3%) and pre-evisceration/post-wash carcass (0.7%) samples. VTEC O26 was isolated from 0.2% of hide swabs and 1.5% of faeces samples. VTEC O145 was isolated from 0.7% of faeces samples. VTEC O26 and VTEC O145 were not recovered from carcass swabs. Non-VTEC O103 was recovered from all sample types (27.1% hide, 8.5% faeces, 5.5% pre-evisceration carcass, 2.2% post-wash carcass), with 0.2% of hide swabs and 1.0% of faeces samples found to be positive for VTEC O103 isolates. E. coli O111 was not detected in any samples. For the four serogroups recovered, the direct transfer from hide to carcass was not observed. This study shows that while VTEC O157 are being carried by cattle presented for slaughter in Ireland, a number of other verotoxin producing strains are beginning to emerge.     

Prevalence and antimicrobial susceptibility of Escherichia coli O157:H7 in vegetables sold in the Amathole District, Eastern Cape Province of South Africa

Fresh vegetables have been implicated in outbreaks of Escherichia coli O157:H7 in most parts of the world. Microbiological quality of vegetables used as recipes for salads is very crucial. Residents of the Amathole District in the Eastern Cape Province of South Africa consume salads frequently, although the microbial quality of recipe vegetables is questionable. The present study investigated the prevalence and antimicrobial susceptibility of E. coli O157:H7 isolated from selected vegetables sold within the Amathole District. One hundred eighty samples of the vegetables were analyzed. Strains of E. coli O157:H7 were isolated by enrichment culture and by immunomagnetic separation and identified by conventional and molecular techniques. In three settlements in this district, the mean counts of presumptive E. coli O157 for the vegetables ranged between 9 x 10(3) and 1.6 x 10(6) CFU/g for Fort Beaufort, 1.6 x 10(3) and 1.6 x 10(5) CFU/g for Mdantsane, and 1.3 x 10(3) and 4.1 x 10(4) CFU/g for Alice. Four (10.3%) of 39 vegetable samples were confirmed to carry E. coli O157:H7. Four representative E. coli O157:H7 isolates from these vegetables were susceptible to at least one of the eight antimicrobial agents tested against them. Even though the prevalence of E. coli O157:H7 was low and those isolated were susceptible to most of the antimicrobials, there remains a need for E. coli O157:H7 surveillance in vegetables used in salad recipes in urban and rural areas of South Africa.     

Frequency of Escherichia coli O157:H7 in Turkish cattle

In this study, five abattoirs in Istanbul were visited between January 2000 and April 2001. During these visits, 330 cattle were selected by a systematic sampling method. Cattle were examined clinically and breed, age, and sex were recorded. Rectal swabs were taken immediately after slaughter. Immunomagnetic separation was performed, and sorbitol-negative colonies were selected on sorbitol MacConkey agar with cefixime and tellurite (CT-SMAC agar). These colonies were checked for 4-methylenebelliferyl-beta-D-glucuronide, indol, rhamnose, and urease activity and motility. Serotypes of bacteria were determined by using antisera specific for Escherichia coli O157 and H7. All cattle selected were clinically healthy. Of 88 sorbitol-negative colonies selected on CT-SMAC agar, isolates from only 14 (4.2%) cattle reacted with anti-O157, and 13 of these isolates also reacted with anti-H7. E. coli O157:H7 was isolated from all breeds, but the numbers of isolates were largest for Holstein and Swiss Brown cows. E. coli O157:H7 was most frequently isolated from 2-year-old cattle. Similarly, it was most frequently isolated from male cattle. E. coli O157:H7 was isolated from cattle slaughtered in four of the five abattoirs studied.     

UV inactivation, liquid-holding recovery, and photoreactivation of Escherichia coli O157 and other pathogenic Escherichia coli strains in water

Drinking water, water used in food production and for irrigation, water for fish farming, waste water, surface water, and recreational water have been recently recognized as a vector for the transmission of pathogenic Escherichia coli, especially serotype O157:H7. We investigated the UV (253.7 nm) inactivation behavior and the capability of dark repair (liquid-holding recovery) and photoreactivation of seven pathogenic (including three enterohemorrhagic E. coli) strains and one nonpathogenic strain of E. coli (ATCC 11229) with respect to the use of UV light for water disinfection purposes. Because most bacteria and yeast are known to be able to repair UV damage in their nucleic acids, repair mechanisms have to be considered to ensure safe water disinfection. We found a wide divergence in the UV susceptibility within the strains tested. A 6-log reduction of bacteria that fulfills the requirement for safe water disinfection was reached for the very most susceptible strain O157:H7 (CCUG 29199) at a UV fluence of 12 J/m2, whereas for the most resistant strain, O25:K98:NM, a UV fluence of about 125 J/m2 was needed. Except for one strain (O50:H7) liquid-holding recovery did not play an important role in recovery after UV irradiation. By contrast, all strains, particularly strains O25:K98:NM, O78:K80:H12, and O157:H7 (CCUG 29193), demonstrated photorepair ability. For a 6-log reduction of these strains, a UV fluence (253.7 nm) up to 300 J/m2 is required. The results reveal that the minimum fluence of 400 J/m2 demanded in the Austrian standard for water disinfection is sufficient to inactivate pathogenic E. coli. A fluence of 160 J/m2 (recommendation in Norway) or 250 J/m2 (recommendation in Switzerland) cannot be regarded as safe in that respect.     

Determining the prevalence of Escherichia coli O157 in cattle and beef from the feedlot to the cooler

Prevalence of Escherichia coli O157 on cattle entering the slaughter floor may range from 10 to > 70%. This study was conducted to determine the effect of E. coli O157 prevalence in fecal pats collected from feedlot pen floors on subsequent E. coli O157 prevalence on carcasses at various points in the slaughter process. Fecal pats from the feedlot pen floor were collected within 3 days before slaughter. During cattle processing at the slaughter facility, additional samples were collected from the hide, from the colon, and from the carcasses before and after evisceration and after final decontamination. Of 15 lots (a group of cattle from the same pen from a feedlot) sampled, 87% had at least one positive fecal pat from the feedlot floor, 47% had a positive hide sample, 73% had a positive colon/fecal sample, and 47% had a positive carcass sample preevisceration; however, only 8% of lots had a positive carcass sample postevisceration or after final intervention. Of the total samples tested (n = 1,328), 24.7, 14.7, 27.6, 10.1, 1.4, and 0.3% of fecal pats from the feedlot floor, hide, colon, preevisceration, postevisceration, and final intervention samples, respectively, were positive for E. coli O157. Pens with greater than 20% positive fecal pats from the feedlot floor had 25.5% hide, 51.4% colon, and 14.3, 2.9, and 0.7% carcass samples positive at preevisceration, at postevisceration, and after final intervention, respectively. However, fecal pats from feedlot floor samples that contained less than 20% positive fecal samples showed lower pathogen prevalence, with 5.0% hide, 7.5% colon, and 6.3, 0, and 0% carcass positive samples at preevisceration, postevisceration, and post-final intervention, respectively. Data from this study can be used as part of risk assessment processes in order to identify mitigation strategies to minimize prevalence of E. coli O157 on fresh beef carcasses.     

Molecular characterization of Escherichia coli O157 contamination routes in a cattle slaughterhouse

In a cattle slaughterhouse, sampling was performed over a 1-week period to examine the prevalence and possible contamination routes of Escherichia coli O157. Each sampling day, swab samples were collected from the slaughterhouse environment before onset of slaughter, from the slaughterline, and from 20 successively slaughtered animals. Isolation of E. coli O157 consisted of a 6-hour enrichment followed by immunomagnetic separation and selective plating. From the 394 samples taken, 84 (21%) were positive for E. coli O157. Pulsed-field gel electrophoresis (PFGE) of collected isolates produced 26 different profiles, from which 5 PFGE profiles carried two or more Stx genes. The combination of PFGE profiles and Stx types resulted in 32 different E. coli O157 types. E. coli O157 was found in the slaughterhouse environment before the onset of slaughter. The first two sampling days, feces and carcasses were found negative. On the third sampling day, five fecal samples and four carcasses from animals negative in the feces were positive. Hide of the anal region and the shoulder were found positive every sampling day. The shoulder hide was more than twice as contaminated as the anal region hide. Typing of different isolates from a sample showed that frequently different E. coli O157 types were presented. On sampling days 1 and 2, types present in the environment and on the hides of the slaughtered animals differed. On the third sampling day, two dominant types were found in the environment (even before the onset of slaughter), as well as on the hides, feces, and carcasses. Although examined animals originated from different farms, one (two on day 3) dominant E. coli O157 type was present on their hides each sampling day. These data indicated that (i) the progress of contamination can differ from day to day within a slaughterhouse and (ii) contact between animals after the departure from the farm can have a large effect on the spread of E. coli O157 hide contamination.     

Fecal shedding of Escherichia coli O157, Salmonella, and Campylobacter in Swiss cattle at slaughter

Fecal samples from 2,930 slaughtered healthy cattle were examined with the following goals: (i) to monitor the shedding of Escherichia coli O157, Salmonella, and Campylobacter in cattle; and (ii) to further characterize the isolated strains. The percentage of the 2,930 samples that tested positive for E. coli O157 by PCR was 1.6%. Thirty-eight strains from different animals that agglutinated with Wellcolex E. coli O157 were isolated. Of the six sorbitol-negative strains, five tested positive for stx2 genes (two times for stx2c and three times for stx2), and one strain tested positive for stx1 and stx2c genes. All sorbitol-negative strains belonged to the serotypes O157:H7- and O157:H7 and harbored the eae type gamma 1 and ehxA genes. The 32 sorbitol-positive strains tested negative for stx genes and belonged to the serotypes O157:H2, O157:H7, O157:H8, O157:H12, O157:H19, O157:H25, O157:H27, O157:H38, O157:H43, O157:H45, and O157:H-. All O157:H45 strains harbored the eae subtype alpha 1 and therefore seem to be atypical enteropathogenic E. coli strains. Whereas none of 1,000 examined samples was positive for Salmonella, 95 of 935 (10.2%) samples were positive for Campylobacter, and all strains were identified as C. jejuni. Sixteen Campylobacter strains were resistant to tetracycline, five were resistant to nalidixic acid/ciprofloxacin, four were resistant to streptomycin, and one was resistant to nalidixic acid/ciprofloxacin and streptomycin. Fecal shedding of zoonotic pathogens in slaughter animals is strongly correlated with the hazard of carcass contamination. Therefore, the maintenance of slaughter hygiene is of crucial importance.     

Antibiotic resistance and hypermutability of Escherichia coli O157 from feedlot cattle treated with growth-promoting agents

In a longitudinal study (165 days), we investigated the effect of growth-promoting agents (monensin and trenbolone acetate-estradiol) and an antibiotic (oxytetracycline) on the incidence in feedlot steers of Escherichia coli O157, including antibiotic-resistant and hypermutable isolates. Eighty steers in 16 pens were treated with eight combinations of promoters, and each treatment was duplicated. Fecal samples were collected at nine different sampling times for detection of E. coli O157. Overall, 50 E. coli O157 isolates were detected in treated animals, and none were found in untreated animals. Compared with untreated controls, there was a significant association between the utilization of growth-promoting agents or antibiotics and the shedding of E. coli O157 at day 137 (P = 0.03), when a prevalence peak was observed and 50% of the isolates were detected. Multiplex PCR assays were conducted for some virulence genes. PCR results indicated that all except one isolate possessed at least the Shiga toxin gene stx2. MICs for 12 antibiotics were determined, and eight oxytetracycline-resistant E. coli O157 strains were identified. Antibiotic-resistant strains were considered a distinct subpopulation of E. coli O157 by pulsed-field gel electrophoresis typing. Seven of these antibiotic-resistant strains were isolated early in the study (on or before day 25), and among them two were also hypermutable as determined by rifampin mutation frequencies. The proportion of hypermutable strains among E. coli O157 isolates remained relatively constant throughout the study period. These results indicate that the use of growth-promoting agents and antibiotics in beef production may increase the risk of environmental contamination by E. coli O157.     

Selection of recently isolated colicinogenic Escherichia coli strains inhibitory to Escherichia coli O157:H7

Escherichia coli strains were screened for their ability to inhibit E. coli O157:H7. An initial evaluation of 18 strains carrying previously characterized colicins determined that only colicin E7 inhibited all of the E. coli O157:H7 strains tested. A total of 540 strains that had recently been isolated from humans and nine different animal species (cats, cattle, chickens, deer, dogs, ducks, horses, pigs, and sheep) were tested by a flip-plating technique. Approximately 38% of these strains were found to inhibit noncolicinogenic E. coli K12 strains. The percentage of potentially colicinogenic E. coli per animal species ranged from 14% for horse isolates to 64% for sheep strains. Those isolates that inhibited E. coli K12 were screened against E. coli O157:H7, and 42 strains were found to be capable of inhibiting all 22 pathogenic strains tested. None of these 42 strains produced bacteriophages, and only 24 isolates inhibited serotype O157:H7 in liquid culture. The inhibitory activity of these strains was completely eliminated by treatment with proteinase K. When mixtures of these 24 colicinogenic strains were grown in anaerobic continuous culture, the four-strain E. coli O157:H7 population was reduced at a rate of 0.25 log10 cells per ml per h, which was fivefold faster than the washout rate. Two strains originally isolated from cat feces (F16) and human feces (H30) were identified by repetitive sequences polymerase chain reaction as the predominant isolates in continuous cultures. The results of this work indicate that animal species other than cattle can be sources of anti-O157 colicinogenic strains, and these results also lead to the identification of at least two isolates that could potentially be used in preharvest control strategies.