Serotypes, virulence genes, and PFGE patterns of enteropathogenic Escherichia coli isolated from Cuban pigs with diarrhea.

Thirty-six enteropathogenic Escherichia coli strains isolated from Cuban pigs with diarrhea were serotyped and screened by PCR for the presence of virulence genes. The 36 isolates belonged to 11 O serogroups and 14 O:H serotypes, with 53% of the isolates belonging to only two serotypes: O141:H- (13 isolates) and O157:H19 (6 isolates). Genes coding for STb, STa, VT2e, and LT toxins were identified in 69, 61, 53, and 6% of the isolates, respectively. The most prevalent fimbrial adhesin was F18, detected in 22 (61%) isolates. The gene encoding F6 (P987) colonization factor was identified in three (8%) isolates. None of the 36 isolates assayed contained genes encoding F4 (K88), F5 (K99), or F41. The seropathotype O141:H-:STa/STb/VT2e/F18 (13 isolates) was the most frequently detected, followed by O157:H19:VT2e/F18 (5 isolates). A genetic diversity study, carried out by pulsed-field gel electrophoresis (PFGE) of 24 representative isolates, revealed 21 distinct restriction patterns clustered in 18 groups (I-XVIII). Isolates of the same serotype were placed together in a dendrogram, but isolates of serotype O157:H19 showed a high degree of polymorphism. The results of this study demonstrate the presence in Cuba of different clusters among one of the most prevalent serotypes isolated from pigs with diarrhea. Further experiments are needed to determine whether some of these clusters have appeared recently; if so, their evolution, as well as their possible association with pathogenicity in farms should be studied.     

Vero cell toxins in Escherichia coli and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chickens and pigs

Sixty-eight of 519 strains of Escherichia coli and six of 10 strains of Pseudomonas aeruginosa produced toxins acting on Vero cells (VT+); all of 63 Salmonella, Shigella, Klebsiella, Enterobacter and Proteus strains were VT-. Most of the VT+ E. coli strains were from weaned pigs suffering from oedema disease and/or diarrhoea and belonged to serogroups O141:K85,88, O141:K85, O138:K81, and O139:K82; six VT+ E. coli strains were from diarrhoeic human babies, four of serogroup O26 and two of serogroup O128. The VT genes in two of the O26 strains and in the O128 strains were located in the genome of the phages with which they were lysogenized. One O141:K85,88 pig E. coli strain transferred its VT genes, probably by conjugation, to E. coli K12. The VTs of the human E. coli strains, the pig E. coli strains and the P. aeruginosa strains were antigenically different from each other; unlike the others, the P. aeruginosa VT was heat-resistant. Cell-free preparations of cultures of E. coli K12 to which the VT genes of the four human E. coli strains had been transferred caused fluid accumulation in ligated segments of rabbit intestine. Inoculated intravenously, they were lethal for mice and rabbits; similar preparations of E. coli K12 to which the VT genes of the pig E. coli strain had been transferred produced a disease in pigs that clinically and pathologically resembled oedema disease.     

Vero cytotoxin production and presence of VT genes in Escherichia coli strains of animal origin

Vero cytotoxin (VT) production has been studied in 34 Escherichia coli strains isolated from animals with enteric diseases. The strains were tested by DNA hybridization with probes for VT1 and VT2 sequences and also in toxin neutralization experiments with specific antisera. Twenty bovine strains, belonging to nine different O serogroups, produced VT1 or VT2 but not both toxins. In contrast, all 14 porcine strains of four different O serogroups produced VT2 only. Six of these porcine strains, belonging to serogroups O138, O139 and O141, were isolated from cases of oedema disease. In general, the porcine isolates produced toxin at a lower level than the bovine strains.     

The expression of an R3 lipopolysaccharide-core by pathotypes of Escherichia coli

AIMS: To investigate the incidence of an R3 lipopolysaccharide (LPS)-core amplicon in a range of pathotypes of Escherichia coli, including Verocytotoxin-producing E. coli (VTEC), enteroaggregative E. coli (EAggEC) and enteropathogenic E. coli (EPEC). METHODS AND RESULTS: A total of 100 strains of E. coli belonging to a range of pathotypes, including 41 strains of VTEC, were screened for the genes encoding the R3 LPS-core using PCR. Fifty-four per cent produced an amplicon with the R3 primer set. Of the 41 VTEC, 66% had an R3 LPS-core with a PCR product being observed with all strains belonging to serotypes O26:H11, O111ac:H- and O145:H25. However, 46% of enteroaggregative E. coli and 50% of enteropathogenic E. coli were also shown to have an R3 LPS-core structure. CONCLUSIONS: Strains with an R3 LPS-core are widely distributed within the species E. coli. SIGNIFICANCE AND IMPACT OF THE STUDY: Strains of E. coli with an R3 LPS-core structure appear not to be associated with a specific pathotype.     

An eight-month study of a population of verocytotoxigenic Escherichia coli (VTEC) in a Scottish cattle herd

AIMS: Strains of Verocytotoxin-producing Escherichia coli (VTEC) from Scottish beef cattle on the same farm were isolated during four visits over a period of eight months. Characteristics of these strains were examined to allow comparisons with strains of VTEC associated with human infection. METHODS AND RESULTS: Strains were characterized to investigate the relationship between these bovine isolates with respect to serotype, Verocytotoxin (VT) type, intimin-type, and presence or absence of the enterohaemolysin genes. VT genes were detected in 176 of 710 (25%) faecal samples tested using PCR, although only 94 (13%) VTEC strains were isolated using DNA probes on cultures. Forty-five different serotypes were detected. Commonly isolated serotypes included O128ab:H8, O26:H11 and O113:H21. VTEC O26:H11 and O113:H21 have been associated with human disease. Strains harbouring the VT2 genes were most frequently isolated during the first three visits to the farm and those with both VT1 and VT2 genes were the major type during the final visit. Of the 94 strains of non-O157 VTEC isolated, 16 (17%) had the intimin gene; nine had the gene encoding beta-intimin and seven strains had an eta/zeta-intimin gene. Forty-one (44%) of 94 strains carried enterohaemolysin genes. CONCLUSIONS: Different serotypes and certain transmissible characteristics, such as VT-type and the enterohaemolysin phenotype, appeared to be common throughout the VTEC population at different times. SIGNIFICANCE AND IMPACT OF THE STUDY: Detailed typing and subtyping strains of VTEC as described in this study may improve our understanding of the relationship between bovine VTEC and those found in the human population.     

Identification of two new intimin types in atypical enteropathogenic Escherichia coli.

Stool specimens of patients with diarrhea or other gastrointestinal alterations who were admitted to Xeral-Calde Hospital (Lugo, Spain) were analyzed for the prevalence of typical and atypical enteropathogenic Escherichia coli (EPEC). Atypical EPEC strains (eae+ bfp-) were detected in 105 (5.2%) of 2015 patients, whereas typical EPEC strains (eae+ bfp+) were identified in only five (0.2%) patients. Atypical EPEC strains were (after Salmonella) the second most frequently recovered enteropathogenic bacteria. In this study, 110 EPEC strains were characterized. The strains belonged to 43 O serogroups and 69 O:H serotypes, including 44 new serotypes not previously reported among human EPEC. However, 29% were of one of three serogroups (O26, O51, and O145) and 33% belonged to eight serotypes (O10:H-, O26:H11, O26:H-, O51:H49, O123:H19, O128:H2, O145:H28, and O145:H-). Only 14 (13%) could be assigned to classical EPEC serotypes. Fifteen intimin types, namely, alpha1 (6 strains), alpha2 (4 strains), beta1 (34 strains), xiR/b2 (6 strains), gamma1 (13 strains), gamma2/q (16 strains), delta/k (5 strains), epsilon1 (9 strains), nuR/e2 (5 strains), zeta (6 strains), iota1 (1 strain), muR/iota2 (1 strain), nuB (1 strain), xiB (1 strain), and o (2 strains), were detected among the 110 EPEC strains, but none of the strains was positive for intimin types mu1, mu2, lambda, or muB. In addition, in atypical EPEC strains of serotypes O10:H-, O84:H-, and O129:H-, two new intimin genes (eae-nuB and eae-o) were identified. These genes showed less than 95% nucleotide sequence identity with existing intimin types. Phylogenetic analysis revealed six groups of closely related intimin genes: (i) alpha1, alpha2, zeta, nuB, and o; (ii) iota1 and muR/iota2; (iii) beta1, xiR/beta2B, delta/beta2O, and kappa; (iv) epsilon1, xiB, eta1,eta2, and nuR/epsilon2; (v) gamma1, muB, gamma2, and theta; and (vi) lambda. These results indicate that atypical EPEC strains belonging to large number of serotypes and with different intimin types might be frequently isolated from human clinical stool samples in Spain.     

Verotoxin-producing Escherichia coli in Spain: prevalence,serotypes, and virulence genes of O157:H7 and non-O157 VTEC in ruminants,raw beef products,and humans

In Spain, as in many other countries, verotoxin-producing Escherichia coli (VTEC) strains have been frequently isolated from cattle, sheep, and foods. VTEC strains have caused seven outbreaks in Spain (six caused by E. coli O157:H7 and one by E. coli O111:H- [nonmotile]) in recent years. An analysis of the serotypes indicated serological diversity. Among the strains isolated from humans, serotypes O26:H11, O111:H-, and O157:H7 were found to be more prevalent. The most frequently detected serotypes in cattle were O20:H19, O22:H8, O26:H11, O77:H41, O105:H18, O113:H21, O157:H7, O171:H2, and OUT (O untypeable):H19. Different VTEC serotypes (e.g., O5:H-, O6:H10, O91:H-, O117:H-, O128:H-, O128:H2, O146:H8, O146:H21, O156:H-, and OUT:H21) were found more frequently in sheep. These observations suggest a host serotype specificity for some VTEC. Numerous bovine and ovine VTEC serotypes detected in Spain were associated with human illnesses, confirming that ruminants are important reservoirs of pathogenic VTEC. VTEC can produce one or two toxins (VT1 and VT2) that cause human illnesses. These toxins are different proteins encoded by different genes. Another virulence factor expressed by VTEC is the protein intimin that is responsible for intimate attachment of VTEC and effacing lesions in the intestinal mucosa. This virulence factor is encoded by the chromosomal gene eae. The eae gene was found at a much less frequency in bovine (17%) and ovine (5%) than in human (45%) non-O157 VTEC strains. This may support the evidence that the eae gene contributes significantly to the virulence of human VTEC strains and that many animal non-O157 VTEC strains are less pathogenic to humans.     

A newly discovered verotoxin variant, VT2g, produced by bovine verocytotoxigenic Escherichia coli

A new verotoxin (VT) variant, designated vt2g, was identified from a bovine strain of verocytotoxigenic Escherichia coli (VTEC) serotype O2:H25. When vt2g was aligned with published sequences of vt2 and vt variants, it exhibited the highest DNA sequence homology with vt2 and vt2c. However, vt2g was not detected by vt2-specific primers and probes, although it was partially neutralized by an antiserum to the VT2A subunit. VT2g was cytotoxic for Vero and HeLa cells and was not activated by mouse intestinal mucus. The vt2g gene was detected in 3 of 409 (0.7%) bovine VTEC strains, including serotypes O2:H25, O2:H45 and Ont:H-.     

Verotoxin-producing Escherichia coli in culled beef cows grazing rangeland forages

The objective of this study was to assess prevalence of verotoxin-producing Escherichia coli (VTEC) in culled beef cows at the time of shipping to slaughter. Feces were collected from 82 cows on eight Nevada ranches during fall and winter (from September to January) after grazing rangeland forages. A random sample (n = 154) of potential VTEC isolates were tested for verotoxicity and were screened for the presence (polymerase chain reaction) and expression (VTEC-reversed passive latex agglutination assay) of the toxin genes (i.e., VT1 and VT2). Seventeen isolates from four ranches were VTEC. Of these, four had the VT1 gene, five had the VT2 gene, seven had both genes, and one did not have either gene despite its toxicity to Vero cells. Except for one isolate (i.e., untypeable that reacted with VT1-latex beads without having VT1 gene), the genotype and phenotype data of the VTEC isolates matched. Another isolate (O8:H- [nonmotile]) was verotoxic, but neither had nor expressed the toxin genes. Of the 17 isolates, four (from one cow) were O157:H7, 11 (from five cows on three ranches) were non-O157:H7 (two O8:H-, three O105:H-, three O116:H-, and three O141:H-), and two were untypeable. Because some of these VTEC serotypes (i.e., O8:H-, O141:H-, and O157:H7) are known to cause human illnesses, it is beneficial to identify VTEC-positive cows before slaughter. This is a critical step in any pre- or post-harvest strategy to minimize the risk of beef contamination with such pathogens.     

Production of cytolethal distending toxin and other virulence characteristics of Escherichia coli strains of serogroup O86

Genetic and phenotypic virulence markers of different categories of diarrhoeagenic Escherichia coli were investigated in 106 strains of enteropathogenic E. coli (EPEC) serogroup O86. The most frequent serotype found was O86:H34 (86%). Strains of this serotype and the non motile ones behaved as EPEC i.e., carried eae, bfpA and EAF DNA sequences and presented localised adherence to HeLa cells. Serotypes O86:H2, O86:H6, O86:H10, O86:H18, O86:H27 and O86:H non determined, belonged to other categories. The majority of the strains of serotype O86:H34 and non motile strains produced cytolethal-distending toxin (CDT). The ribotyping analysis showed a correlation among ribotypes, virulence markers and serotypes, thus suggesting that CDT production might be a property associated with a universal clone represented by the O86:H34 serotype.     

Genetic relationship of diarrheagenic Escherichia coli pathotypes among the enteropathogenic Escherichia coli O serogroup

The genetic relationship among the Escherichia coli pathotypes was investigated. We used random amplified polymorphic DNA (RAPD) data for constructing a dendrogram of 73 strains of diarrheagenic E. coli. A phylogenetic tree encompassing 15 serotypes from different pathotypes was constructed using multilocus sequence typing data. Phylogram clusters were used for validating RAPD data on the clonality of enteropathogenic E. coli (EPEC) O serogroup strains. Both analyses showed very similar topologies, characterized by the presence of two major groups: group A includes EPEC H6 and H34 strains and group B contains the other EPEC strains plus all serotypes belonging to atypical EPEC, enteroaggregative E. coli (EAEC) and enterohemorrhagic E. coli (EHEC). These results confirm the existence of two evolutionary divergent groups in EPEC: one is genetically and serologically very homogeneous whereas the other harbors EPEC and non-EPEC serotypes. The same situation was found for EAEC and EHEC.     

Molecular evolution and mosaic structure of alpha, beta, and gamma intimins of pathogenic Escherichia coli.

Two types of pathogenic Escherichia coli, enteropathogenic E. coli (EPEC) and enterohemorrhagic E. coli (EHEC), cause diarrheal disease by disrupting the intestinal environment through the intimate attachment of the bacteria to the intestinal epithelium. This process is mediated by intimin, an outer membrane protein that is homologous to the invasins of pathogenic Yersinia. The intimin (eae) gene is part of a pathogenicity island, a 35-kb segment of DNA that has been acquired independently in different groups of pathogens. Nucleotide sequences of eae of three EPEC and four EHEC strains representing distinct clonal lineages revealed an exceptionally high level of divergence (15%) in the amino acid sequences of alpha, beta, and gamma intimin molecules, most of which is concentrated in the C-terminal region. The gamma intimin sequences from E. coli strains with serotypes O157:H7, O55:H7, and O157:H- are virtually identical, supporting the hypothesis that these bacteria belong to a single clonal lineage. Sequences of beta intimin of EPEC strains of serotypes O111:H2 and O128:H2 show substantial differences from alpha and gamma intimins, indicating that these strains have evolved independently. Strong nonrandom clustering of polymorphic sites indicates that the intimin genes are mosaics, suggesting that protein divergence has been accelerated by recombination and diversifying selection.     

Escherichia coli Serogroups Isolated from Streams in Pennsylvania, 1965 to 1972

Of 3,200 cultures of Escherichia coli isolated from streams in Pennsylvania over a 7-year period, 82.46% or 2,639 were O serogrouped. The largest number of cultures (33.4%) belonged to O groups 1 to 26, and the second highest number (16.8%) belonged to O groups 60 to 88. The individual E. coli O groups most frequently isolated were ADO3, 18ac, 2a, 3, 7, 73, 139, and OX13. Practically every known standard E. coli O group was found in the streams. It was not possible to identify the K and H antigen of every E. coli isolate. Serotypes of E. coli O2a:K1:H6, O26:K60:H11, O55:K59:H27, O86:K62:H2, 112ab:K68:H2, 125ab:K70:H21, 128ab:K67:H2, and O138:K81:H14 known to be pathogenic for humans and animals were identified. Cultures having the same K antigen but a different H antigen for enteropathogenic E. coli O groups 6, 18ab, 18ac, 111ab, 126, 127a, 139, 141, and 147 were also isolated.     

Studies on Coli-aerogenes and Other Gram-negative Intestinal Bacteria in Various Animals and Birds

S ummary . Some faecal material from native animals and birds (especially) showed either no coli-aerogenes bacteria or contained types other than Escherichia coli I.
Amongst coli-aerogenes isolates, E. coli I was frequent but the high percentage of other biotypes indicated that the animal host may serve as a pool of intermediate and irregular strains. Paracolons and Proteus spp. were abundant, and Salmonella spp. were isolated from dogs and native animals possibly in contact with man.
Animals were only occasionally carriers of enteropathogenic E. coli serotypes and seldom of types locally important in the etiology of infantile diarrhoea, such as E. coli O26:B6 and O55:B5. In this series O86 serotypes other than the diarrhoeal strain isolated from babies were found. The results of the study indicated some degree of host specificity. In the group of 20 O serotypes (referred to as specific E. coli ), O8 was most commonly listed. Most strains were sensitive to chemotherapeutic agents but resistance of some strains especially to sulphonamide was recorded.
The implication of the presence in the animal gut of coli-aerogenes bacteria other than E. coli type I is discussed in relation to bacterial standards for drinking water.     

Prevalence of enterohaemorrhagic Escherichia coli from serotype O157 and other attaching and effacing Escherichia coli on bovine carcasses in Algeria

AIMS: Bovine meat is the principal source of human contamination of attaching and effacing Escherichia coli, including enterohaemorrhagic E. coli O157. The aim was to study the prevalence of these strains on bovine carcasses in Algeria. METHODS AND RESULTS: Two-hundred and thirty carcasses were swabbed and analysed by classical microbiological methods for total E. coli counts and for the presence of pathogenic E. coli. The E. coli counts were high, with a 75th percentile of 444.75 CFUs cm(-2). For pathogenic E. coli, more than 7% of the tested carcasses were positive for E. coli O157. Eighteen E. coli O157 strains were isolated and typed by multiplex PCR. The main isolated pathotype (78%) was eae+ stx2+ ehxA+. In addition to E. coli O157, other attaching and effacing E. coli (AEEC) were also detected from carcasses by colony hybridization after pre-enrichment and plating on sorbitol MacConkey agar using eae, stx1 and stx2 probes. Thirty carcasses (13%) on the 230 analysed harboured at least one colony positive for one of the tested probes. These positive carcasses were different from those positive for E. coli O157. Sixty-six colonies (2.9%) positive by colony hybridization were isolated. The majority (60.6%) of the positive strains harboured an enteropathogenic E. coli-like pathotype (eae+ stx-). Only three enterohaemorrhagic E. coli (EHEC)-like (eae+ stx1+) colonies were isolated from the same carcass. These strains did not belong to classical EHEC serotypes. CONCLUSIONS: In this study, the global hygiene of the slaughterhouse was low, as indicated by the high level of E. coli count. The prevalence of both E. coli O157 and other AEEC was also high, representing a real hazard for consumers. SIGNIFICANCE AND IMPACT OF THE STUDY: This is the first study of this type in Algeria, which indicates that the general hygiene of the slaughterhouse must be improved.     

Analysis of fimbrial gene clusters and their expression in enterohaemorrhagic Escherichia coli O157:H7

The sequence of two enterohaemorrhagic Escherichia coli (EHEC) O157:H7 strains reveals the possession of at least 16 fimbrial gene clusters, many of the chaperone/usher class. The first part of this study examined the distribution of these clusters in a selection of EHEC/EPEC (enteropathogenic E. coli) serotypes to determine if any were likely to be unique to E. coli O157:H7. Six of the clusters, as determined by the presence of amplified main subunit or usher gene sequences, were detected only in the E. coli O157 and O145 serotypes tested. With the exception of one serotype O103 strain that contained an lpf2 cluster, lpf sequences were only detected in E. coli O157 of the serotypes tested. Expression from each cluster was measured by the construction of chromosomally integrated lacZ promoter fusions and plasmid-based eGFP fusions in E. coli O157:H7. This analysis demonstrated that the majority (11/15) of main fimbrial subunit genes were not expressed under the majority of conditions tested in vitro. One of the clusters showing promoter activity, loc8, has a temperature expression optimum indicating a possible role outside the host. From the presence of pseudogenes in three of the clusters, the lack of FimH-like minor adhesins in the clusters and their limited expression in vitro, it would appear that E. coli O157:H7 has a limited repertoire of expressed functional fimbriae. This restricted selection of fimbriae may be important in bringing about the tropism E. coli O157:H7 demonstrates for the terminal rectum of cattle.     

Isolation of K88 Antigen Variants (ab,ac, ad) from Porcine Enterotoxigenic Escherichia coli Belonging to Different Serotypes

Fimbriae isolation by means of thermal shock was applied to fifteen K88-positive (three K88ab, nine K88ac and three K88ad) Escherichia coli reference strains belonging to serotypes O8:K87, O32, O45, O138:K81, O141:K85, O147:K89, O149:K91, and O157, as well as to ten K88-positive enterotoxigenic strains isolated from porcine diarrhea in Spain, all of them belonging to the O149 serogroup. Fimbriae were removed from the bacterial cells by thermal shock at 60 C and then precipitated using ammonium sulfate. The final amount of K88 antigen and the purification degree were not related to the serogroup of the bacteria or to the antigen variant but were related to the buffer used for isolation. Phosphate buffer containing urea was shown to be more effective than Tris-HCl for isolation of K88 antigen. The molecular weights by SDS-PAGE for K88ab, K88ac, and K88ad were 28.5, 29.2, and 31.0 kDa, respectively. All enterotoxigenic E. coli strains isolated in Spain showed the K88ac variant.     

Genetic Mapping of the Antigenic Determinants of Two Polysaccharide K Antigens, K10 and K54, in Escherichia coli

The genes controlling synthesis of the Escherichia coli acidic polysaccharide capsular antigens K10 and K54 were transferred by conjugation to E. coli strains of other serotypes. The genes concerned with these K antigen determinants showed genetic linkage with the serA locus. We propose to name the K antigen-controlling gene kpsA. The genetic determinants of the two K antigens could also be transferred to enteropathogenic serotypes, even though such strains have never been found in nature with special acidic polysaccharide K antigens. A noncapsulated derivative, K(-), of the K10 strain can transfer the genetic determinant of the K antigen, demonstrating the probable existence of another chromosomal locus involved in the production of such acidic polysaccharide K antigens.     

Identification of EPEC and non-EPEC serotypes in the EPEC O serogroups by PCR-RFLP analysis of the fliC gene

Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis of the flagellin gene (fliC) was performed in 233 strains of enteropathogenic Escherichia coli (EPEC) O serogroups for determining their flagellar antigen (H) status. The serological detection of flagellin is the basis for the H-codes typing system in E. coli. Thus, it is impossible to serotype nonmotile bacteria (i.e. to assign H-codes). Twenty-eight fliC restriction patterns were obtained for motile (H2, H4, H6, H7, H8, H9, H10, H11, H12, H18, H21, H27, H32, H34, H35, H40 and H51) and nonmotile serotypes (H(-)). Each motile serotype was characterized by one or two fliC specific restriction patterns. The only exception was serogroup O128ab, where a common restriction pattern was found for serotypes O128ab:H2 and O128ab:H35, even after digestion with RsaI, AluI and Sau3AI endonucleases. These two serotypes were, however, discriminated by single strand conformation polymorphism (SSCP) analysis of RsaI restriction fragments. Nonmotile strains showed fliC restriction patterns identical to some known H serotypes. The PCR-RFLP analysis of fliC gene proved to be a useful method for identifying the H variants in motile and nonmotile EPEC O serogroups.     

Detection of Escherichia coli serogroups O26, O103, O111 and O145 from bovine faeces using immunomagnetic separation and PCR/DNA probe techniques

AIMS: The aim of this study was to isolate Escherichia coli O26, O103, O111 and O145 from 745 samples of bovine faeces using (i) immunomagnetic separation (IMS) beads coated with antibodies to lipopolysaccharide, and slide agglutination (SA) tests and (ii) PCR and DNA probes for the detection of the Verocytotoxin (VT) genes. METHODS AND RESULTS: IMS-SA tests detected 132 isolates of presumptive E. coli O26, 112 (85%) were confirmed as serogroup O26 and 102 had the VT genes. One hundred and twenty-two strains of presumptive E. coli O103 were isolated by IMS-SA, 45 (37%) were confirmed as serogroup O103 but only one of these strains was identified as Verocytotoxin-producing E. coli (VTEC). Using the PCR/DNA probe method, 40 strains of VTEC O26 and three strains of VTEC O103 were isolated. IMS-SA identified 21 strains of presumptive E. coli O145, of which only four (19%) were confirmed as serogroup O145. VTEC of this serogroup was not detected by either IMS-SA or PCR/DNA probes. E. coli O111 was not isolated by either method. CONCLUSION: IMS beads were 2.5 times more sensitive than PCR/DNA probe methods for the detection of VTEC O26 in bovine faeces. SIGNIFICANCE AND IMPACT OF THE STUDY: IMS-SA is a sensitive method for detecting specific E. coli serogroups. However, the specificity of this method would be enhanced by the introduction of selective media and the use of tube agglutination tests for confirmation of the preliminary SA results.