The role of lipopolysaccharide and Shiga-like toxin in a mouse model of Escherichia coli O157:H7 infection

The role of lipopolysaccharide (LPS) and Shiga-like toxin (SLT) in the pathogenesis of hemolytic uremic syndrome (HUS) was studied in a mouse model. Mice inoculated intragastrically with Escherichia coli O157:H7 developed gastrointestinal, neurologic, and systemic symptoms, necrotic foci in the colon, glomerular and tubular histopathology, and fragmented erythrocytes. LPS-responder (C3H/HeN) mice developed a combination of neurologic and systemic symptoms, whereas LPS-nonresponder (C3H/HeJ) mice had a biphasic course of disease, first developing systemic symptoms and later severe neurologic symptoms. Mice inoculated with SLT-II-positive strains developed severe neurotoxic symptoms and a higher frequency of systemic symptoms and glomerular pathology compared with SLT-II-negative strains. Anti-SLT-II antibodies protected against these symptoms and pathology. These results demonstrate that this model could be used to study aspects of human HUS and that both LPS and SLT are important for disease development.

[Predictive indicators for progression to severe complications(hemolytic-uremic syndrome and encephalopathy) and their prevention in enterohemorrhagic Escherichia coli infection]

The treatment of infection with enterohemorrhagic Escherichia coli(EHEC) aims for early prediction and prevention of severe complications such as hemolytic-uremic syndrome, encephalopathy and/or thrombotic thrombocytopenic purpura. Factors related to the complications are divided into three categories; risk factors or predisposition, predictors, and indicators of severity and outcome. Risk factors for complications include two extreme ages, infection with verotoxin 2 producing E. coli, positive stool culture for EHEC, use of antimotility drug, use of trimethoprim-sulfamethoxazole. Predictors for complications include severe abdominal pain and bloody diarrhea development of high fever, change of consciousness, urinal protein and/or occult blood, abrupt increase of white blood cell count, urinal NAG, alpha 1 microglobulin, beta 2 microglobulin, low osmolar urine, high thrombomodulin level, marked thickening of intestinal wall, increased brightness of kidney in ultrasound sonography. No preventive treatment for these complications is proven except SYNSORB-pk which is expected to effectively aborb verotoxin in the intestine.

Norepinephrine induced growth and expression of virulence associated factors in enterotoxigenic and enterohemorrhagic strains of Escherichia coli

The small intestine is richly innervated by the sympathetic nervous system. High concentrations of monoamines, most notably norepinephrine, are found throughout the various intestinal layers. In order to determine whether norepinephrine is capable of influencing bacterial pathogenesis, the growth and production of virulence factors in ETEC and EHEC were examined in a physiologically relevant medium utilizing very low initial bacterial inoculums to more closely mimie in vivo conditions. The growth of ETEC strain B44 and the production of the K99 pilus adhesin on a protein equivalent basis was greatly increased in the presence of norepinephrine. Growth of EHEC O157:H7 was also increased in norepinephrine containing medium as well as production of SLT-I and SLT-II. The ability of norepinephrine to increase both bacterial growth and expression of virulence factors was shown to be non-nutritional in nature. Given the abundant adrenergic innervation in the small intestine, these in vitro results suggest that the neurohumoral environment of the host may play a role in bacterial growth and expression of virulence factors.

Intracellular transport and processing of protein toxins produced by enteric bacteria

Bacterial toxins are associated with disease in humans and animals. Toxins can either be preformed in food or produced by bacteria in the intestine. There are two types of toxins: heat-labile protein toxins and heat stabile toxins. Heat labile toxins are produced by Bacillus cereus, Clostridium perfringens, Escherichia coli, and Vibrio cholerae, and heat-stabile enterotoxins consisting of relatively few amino acids are produced by Escherichia coli and acts by activation of guanylate cyclase. Similarly, heat-stabile entero-toxins are also produced by Staphylococcus aureus, a common cause of food poisoning in the United States, and Yersenia enterocolitica. Protein toxins produced by enteric bacteria can intoxicate intestinal cells and can also be taken up from the gut and reach other cells in the body. For example the Shiga-like toxins (vero-toxins) can intoxicate endothelial cells in the kidney and cause kidney failure. Intracellular transport and processing of a few of the protein toxins produced by enteric bacteria, namely Clostridium difficile toxin A and B, cholera toxin and the related heat-labile toxin produced by Escherichia coli, and Shiga toxin and Shiga-like toxins are presented.

Variation in virulence in the gnotobiotic pig model of O157:H7 Escherichia coli strains of bovine and human origin

Escherichia coli strains of serotype O157:H7 have been incriminated in outbreaks and sporadic cases of food-borne illness, including diarrhea, hemorrhagic colitis, hemolytic uremic syndrome and thrombotic thrombocytopenic purpura. Food-producing animals, particularly cattle, are believed to be reservoirs of the organism. Whether all strains of bovine origin pose human health risk is unknown and was the impetus for this investigation. We compared the virulence of ten SLT-I, SLT-II, and eae DNA probe-positive O157:H7 strains from cattle to 10 like strains associated with human diarrheal disease outbreaks for virulence in one day-old gnotobiotic pigs. All strains caused diarrhea, and only four pigs inoculated with either of two bovine strains failed to develop that condition. Signs of central nervous system disease, death, debilitation requiring euthanasia before the end of an eight day observation period, and/or encephalomalacia occurred in 32/42 pigs inoculated with the strains isolated from human beings, 13/39 pigs inoculated with strains from cattle, and 7/7 pigs inoculated with a positive control strain. More strains of human origin (9/10) than bovine origin (5/10) caused these effects. The results of this study indicate considerable variability in virulence of O157:H7 strains possessing the same known virulence determinants, and suggest that disease outbreaks tend to be caused by the more virulent of these strains.

Magnetic resonance imaging and histopathological study of brain lesions in rabbits given intravenous verotoxin 2

When rabbits were given intravenously purified verotoxin 2 (VT2) at 5 microg/kg of body weight, they developed hemorrhagic diarrhea, flaccid paresis, an ataxic gait, an opisthotonic posture, and convulsions. To examine the effects of VT2 toxemia on the rabbit central nervous system, magnetic resonance imaging and ultrastructural studies were performed. At 24, 57, and 80 h after injection of VT2 into 12 rabbits, T2-weighted images of the central nervous system were obtained. The initial lesion was noted at 24 h in the hypothalamic areas of all experimental animals. At 57 h, the T2 value increased in the medulla of the cerebral hemisphere or the hippocampus, with a brain stem lesion in six rabbits (50%). The rabbits with the brain stem lesions, in which neurological signs were very severe, died within 6 days. Lesions in the cerebellar hemisphere and/or vermis were noted in four rabbits (33%) that survived more than 1 month. To better understand the pathogenesis of VT2 in these brain lesions, we examined the deterioration of the blood-brain barrier and cerebrospinal fluid-brain barrier by using horseradish peroxidase as a tracer. The tracer was detected by electron microscopy both in the subendothelial layer, including the basal lamina, and throughout the cytoplasm of the ependymal cell layer covering the ventricle after intravenous or intrathecal treatment with horseradish peroxidase. We also determined the localization of VT2 by immunoelectron microscopy and found that it was localized on edematous endothelial cells of capillaries, ependymal cells, and myelin sheaths. The present study suggests that VT2 was conveyed from the endothelial and ependymal cell layers and caused edematous changes in the rabbit brain.

[Investigations of raw milk for Shiga-like toxin producing Escherichia coli by the polymerase chain reaction]

This is a report about the occurrence of Shiga-like toxin (SLT) producing Escherichia coli in milk in southern Sachsen-Anhalt. Samples were taken from tanks of milk at the farms. The samples were investigated by polymerase chain reaction as recommended by the "Bundesinstitut für gesundheitlichen Verbraucherschutz und Veterinärmedizin". Using PCR, in 11% of the tank samples sequences coding for SLT I and/or SLT II were detected. 75% of the as positive detected samples had eae gen. Some cows in these farms spread potentially enterohemorrhagic E. coli. These farms should not sell milk for raw consumption.

[Verocytotoxin-producing Escherichia coli O157 in Dutch veal calves and beef cattle]

In 1994 (August-September) and in 1995 (July-October) faeces of Dutch veal calves and adult cattle was examined for the presence of verocytotoxin-producing Escherichia coli O157 (O157 VTEC). The samples were collected at slaughterhouses. In 1994, O157 VTEC were isolated from three (0.9%) of 365 fecal samples of veal calves. Faeces of adult cattle was not collected. In 1995, O157 VTEC were isolated from one (0.5%) of 18.3 fecal samples of veal calves and from 30 (11.1%) of 270 fecal samples of adult cattle. In 1994, the organisms were isolated by selective plating onto both sorbitol MacConkey agar (SMAC) and SMAC containing tellurite and cefixime (TC-SMAC) following selective enrichment in modified trypton soya broth with acriflavin (mTSB + a). In 1995, the samples were enriched in modified E. coli broth with novobiocin (mEC + n), and in addition to directly plating onto TC-SMAC the enriched cultures were plated after incorporation of an immunomagnetic separation step of E. coli O157. All 30 strains isolated from adult cattle were isolated with the immunomagnetic separation procedure. Four of the 30 samples were determined positive also by directly plating onto TC-SMAC. The sample from the veal calve was determined positive only by directly plating onto TC-SMAC. Plating onto SMAC resulted only in negative results. Characterization of the O157 VTEC isolates showed similarities between isolates of humans and cattle. Additional experiments need to be done concerning this aspect. From this study it appeared that also Dutch cattle are a reservoir of O157 VTEC.

Vero cytotoxigenic Escherichia coli infection in dairy farm families

Fecal samples from 335 dairy farm residents and 1458 cattle on 80 farms were tested for Vero cytotoxin (VT)-producing Escherichia coli (VTEC). Residents were also tested for antibodies to VT1 and O157 lipopolysaccharide (LPS). Residents and cattle on farms with VTEC-positive persons or E. coli O157:H7-positive cattle were retested. Twenty-one persons (6.3%) on 16 farms (20.8%) and 46% of cattle on 100% of the farms had VTEC in fecal samples. Human VTEC isolates included E. coli O157:H7 and 8 other serotypes, 4 of which were present in cattle on the same farms. More persons had antibodies to VT1 (41%) than to O157 LPS (12.5%). Seropositivity to O157 LPS was associated with isolation of E. coli O157:H7 on the farm (P = .022). Human VTEC infection was negatively associated with age (P < .05) and was not associated with clinical illness. Many dairy farm residents experience subclinical immunizing VTEC infections at a young age, which frequently involve non-O157 VTEC found in cattle.

Comparative cytotoxicity of purified Shiga-like toxin-IIe on porcine and bovine aortic endothelial and human colonic adenocarcinoma cells

Porcine and bovine aortic endothelial cells and human colonic adenocarcinoma cells were compared for their susceptibility to the toxic effect of purified Shiga-like toxin IIe (SLT-IIe), measured by the neutral red cytotoxicity assay. Cytotoxicity correlated with toxin binding as indicated by fluorescence activated cell sorter analysis and with the globotriosylceramide (Gb3) and globotetraosylceramide (Gb4) content of cells determined by high pressure liquid chromatography. One line of porcine aortic endothelial cells was 1400-fold more susceptible than the line of bovine aortic endothelial cells that was tested, but a second line of porcine aortic endothelial cells was highly refractory to SLT-IIe. Human colonic adenocarcinoma cells lacked detectable levels of Gb4 and were least susceptible to SLT-IIe.

A standard immunoglobulin preparation produced from bovine colostra shows antibody reactivity and neutralization activity against Shiga-like toxins and EHEC-hemolysin of Escherichia coli O157:H7

Enterohemorrhagic Escherichia coli (EHEC) causes a variety of clinical conditions, the most important being hemorrhagic colitis and hemolytic uremic syndrome. A curative therapy of EHEC diseases is not yet feasible. This study investigates the antibody reactivity of Lactobin, a standardized immunoglobulin (Ig) preparation, obtained from the colostra of non-immunized cows. Three different batches of Lactobin exhibited equally high titers of specific antibodies against Shiga-like toxins (SLTs, verocytotoxins) and EHEC hemolysin (EHEC-Hly) produced by E. coli O157. In addition, Lactobin blocked the cytotoxic effect of SLT-I and SLT-II on Vero cell monolayers and inhibited the cytolytic effects of EHEC-Hly on human erythrocytes. Since Lactobin contains high levels of antibodies and neutralizing activity against important virulence factors of EHEC O157, this drug has potential use in the treatment of diarrhea and the prevention of EHEC-associated hemolytic uremic syndrome.

Evaluation of an enzyme immunoassay for verotoxin detection in Escherichia coli

Verotoxin-producing Escherichia coli strains (VTEC) cause hemorrhagic colitis and hemolytic-uremic syndrome in humans. Laboratory diagnosis by conventional methods is slow and cumbersome. The results of a new rapid enzyme immunoassay (EIA Premier EHEC) for verotoxin detection both in isolated strains and in clinical samples are presented, and they are compared with cell culture (CC) and polymerase chain reaction (PCR) techniques. Fifty-four strains have been analyzed by both EIA and PCR, and 33 by all three methods. The kit has also been evaluated for experimentally infected stool samples directly and after their enrichment on MacConkey broth. Nineteen, out of the 54 strains, were positive by EIA and 20 by PCR. The results of the 33 strains evaluated by the three techniques were coincident with one exception. The latter was uninterpretable by CC, negative by EIA and positive by PCR. The sensitivity of the kit for experimentally infected stool samples was approximately 5 x 10(7) bacteria/ml in the direct test, and 5 x 10(4) bacteria/ml after broth enrichment. EIA sensitivity and specificity were similar to those of CC and PCR. The diagnostic times were 18h for EIA, 3 days for PCR and 5 days for CC. Sensitivity, rapidity and ease of performance make this technique especially valuable for clinical diagnosis.

Detection of enterohaemorrhagic Escherichia coli O157:H7 in minced beef using immunomagnetic separation

Enterohaemorrhagic Escherichia coli (EHEC) O157:H7 has been recently recognized as a human pathogen associated with haemorrhagic colitis and haemolytic uraemic syndrome. Most outbreaks of haemorrhagic colitis resulted from the consumption of undercooked minced beef or raw milk. Dairy cattle have been identified as a reservoir of EHEC O157: H7. In this study E. coli O157 specific antibody, coated on magnetic beads, was used to concentrate and release EHEC O157:H7 from meat samples. A survey of retail fresh minced beef and hamburger samples using this procedure revealed that 3 (5%) of 58 beef samples were positive for EHEC O157:H7. Two of the strains produced both VT1 and VT2 verotoxins, and one produced only VT2. Immunomagnetic separation is a sensitive and simple technique for the isolation of E. coli O157 from food, and could be useful for a further elucidation of the epidemiology of this organism. The relatively high prevalence of EHEC O157:H7 in beef samples may constitute a risk for public health. Thus, a suitable epidemiologic control and effective methods of prevention should be applied.

The enterohemolysin phenotype of bovine Shiga-like toxin-producing Escherichia coli (SLTEC) is encoded by the EHEC-hemolysin gene

Naturally occurring enterohemolysin negative variants were observed during studies on bovine Shiga-like toxin-producing E. coli (SLTEC). Examination of three strains (413/89-1 and 332, 026:H-, and 570/89, O111:H-) and their isogenic variants (413/89-6, 332-I and 570/89-I, respectively) showed, that in each strain loss of the enterohemolytic phenotype correlated with the loss of a large plasmid ranging from 94 to 104 kb in size. The hemolysin determinant present on the 94 kb plasmid of strain 413/89-1 was cloned and discovered by DNA and N-terminal aminoacid sequence analysis to be highly homologous to the recently published EHEC-hemolysin (HlyEHEC; Schmidt et al., 1994; 1995). When a recombinant plasmid harboring this determinant was reintroduced into the enterohemolysin negative isogenic mutant 413/89-6, the enterohemolytic phenotype was restored. Southern blot hybridization analysis was used to demonstrate that the HlyEHEC is plasmid-borne in SLTEC-strains. Our cumulative data suggest that the enterohemolytic phenotype of SLTEC is encoded by the plasmid-borne HlyEHEC. These results further demonstrate the close similarity between SLTEC-isolates from bovine and human.

Prevalence of enterohemorrhagic Escherichia coli in raw and treated municipal sewage

Municipal sewage was screened for DNA encoding Shiga-like Toxin (SLT) II, a key protein involved in the virulence of enterohemorrhagic Escherichia coli. PCR analysis of sewage concentrates showed that DNA encoding SLT II was present in a single sample of untreated sewage and absent in all other samples tested (n = 6). Thermotolerant E. coli cultured from the sewage (n = 1,520) also tested negative for SLT II by colony hybridization.

Typing of bovine attaching and effacing Escherichia coli by multiplex in vitro amplification of virulence-associated genes

Attaching and effacing Escherichia coli is a new causal agent of diarrhea in calves. Its major virulence factors are the intimin protein, encoded by the eaeA gene, and the Shiga-like toxins, encoded by slt genes. Because the sequences of these genes are available, we selected specific primers to amplify each virulence gene so as to develop a new identification test based on multiplex amplification of virulence-associated genes. Of 30 tested strains, 14 were eaeA+, 15 were eaeA+ slt-I+, 1 was eaeA+ slt-I+ slt-II+, and 1 was eaeA+ slt-II+. The method proved in our hands to be fast and specific and in perfect correlation with the hybridization method.

An Escherichia coli hemolysin transport system-based vector for the export of polypeptides: export of Shiga-like toxin IIeB subunit by Salmonella typhimurium aroA

The export of Escherichia coli hemolysin across the cytoplasmic and the outer membranes requires the COOH-terminal signal sequence of HlyA, the two specific translocator proteins HlyB and HlyD, and the outer membrane protein TolC. We have developed an export cloning system that is composed of two vectors: one in which the fusion of the desired gene with the 3'-end of hlyA is generated, and a second in which the sequences containing the fusion are combined with the accessory genes hlyB and hlyD, thereby reconstructing the natural organization of the hly locus. In the second vector the fusion and the accessory genes are flanked by Notl sites, allowing subcloning of the whole cluster into a variety of minitransposons to achieve the stable integration of the constructs into the chromosome of Gram-negative bacteria. Since some applications may require the production of transcriptional fusions, an alternative version of the system provides the efficient translation initiation region of T7 phage gene 10 upstream of the fusion protein coding sequence. The usefulness of the system was assessed by constructing a fusion between the gene encoding the B subunit of Shiga-like toxin lle and the 3'-end of hlyA. An attenuated Salmonella typhimurium vaccine strain harboring the resulting construct, either in multicopy or monocopy, efficiently expressed and exported the chimeric protein. We anticipate that this system will lead to a higher stability of the engineered function and permit a faithful monitoring of the export of the recombinant peptide under physiologic single-copy conditions.

Survival rate of mice after transient colonization with Escherichia coli clones carrying variant Shiga-like toxin type II operons

We have previously employed a streptomycin-treated mouse model to demonstrate that Escherichia coli clones expressing particular variant Shiga-like toxin type II operons differ in oral virulence, as judged by median survival time. Differences in virulence were not seen between all toxin variants, including two which differed significantly in cytotoxicity for Vero cells. In the present study, we have modified the animal model by withdrawing antibiotic selection and reintroducing normal mouse intestinal flora at various times after oral challenge with three different variant SLT-II-producing clones. This resulted in a transient colonization more akin to that seen in natural human infections. This has enabled detection of significant differences in survival rate between mice challenged with E. coli clones producing different SLT-II variants, which were not observed when colonization was maintained at high levels.

Serum antibody responses of cattle following experimental infection with Escherichia coli O157:H7

Oral inoculation of calves and steers with 10(10) CFU of Escherichia coli O157:H7 induced prompt and sustained increases in serum antibodies to O157 lipopolysaccharide. Neutralizing antibodies to verotoxin 1 also increased rapidly in most steers but more gradually in calves. None of the animals developed neutralizing antibodies to verotoxin 2. These serological responses were not correlated with elimination of infection in calves or steers or protection of calves against reinfection with the same strain.

Interaction of verotoxin 2e with pig intestine

In pigs with edema disease, verotoxin 2e (VT2e) is produced in the intestine and transported to tissues, but neither the mechanism by which toxin passes through the intestine nor its failure to induce an enterotoxic reaction is understood. Binding of VT2e to pig intestine was examined by enzyme-linked immunosorbent assay involving microvillus membranes (MVM) and crude mucus; thin-layer chromatographic overlay immunoassay with total lipids extracted from MVM; and indirect immunofluorescence of toxin bound to thin sections of jejunum, ileum, and colon. VT2e bound significantly to MVM from pig jejunum and ileum but not to crude mucus. Verotoxin 2e-binding glycolipids, globotetraosylceramide and globotriaosylceramide, were detected by thin-layer chromatographic overlay immunoassay in extracts of MVM from jejunum and ileum. Indirect immunofluorescence showed that VT2e bound to vessels within the submucosa and muscularis mucosa of the jejunum, ileum, and colon and to enterocytes at the lower portion but not at the tips of villi in the jejunum and ileum. Receptors for VT2e are therefore present in the intestine of the pig, but their role in absorption of VT2e is unclear since intraintestinal inoculation of pigs with large quantities of VT2e does not result in edema disease. Previously reported lack of enterotoxicity of verotoxins in pig intestine may be explained by the absence of toxin receptors in the villus absorptive enterocytes.

Activation of Shiga-like toxins by mouse and human intestinal mucus correlates with virulence of enterohemorrhagic Escherichia coli O91:H21 isolates in orally infected, streptomycin-treated mice

The enterohemorrhagic Escherichia coli (EHEC) O91:H21 isolates B2F1 and H414-36/89 are virulent in an orally infected streptomycin-treated mouse model. Previous studies demonstrated that B2F1 and H414-36/89 grow to high levels in mucus isolated from mouse small intestine and colon and that growth in small-intestine mucus is related to virulence. We measured the levels of Shiga-like toxins (SLTs) SLT-IIvha and SLT-IIvhb produced by B2F1 after growth in Luria-Bertani (LB) broth supplemented with mouse intestinal mucus by assaying the cytotoxicity of culture supernatants on Vero cells. Culture supernatants from B2F1 grown in mouse intestinal mucus, but not EHEC strains that produce SLT-II or SLT-IIc, were approximately 35- to 350-fold more toxic for Vero cells than supernatants from B2F1 grown in LB broth. This increased toxicity was not reflected by a concomitant increase in SLT antigen content. Furthermore, when culture supernatants from B2F1 or K-12 strains carrying plasmids encoding SLTs cloned from H414-36/89 or purified SLT-IIvhb from B2F1 were incubated with mouse intestinal mucus, the samples exhibited greater cytotoxicity than when they were incubated with N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) buffer alone. These toxin preparations also showed increased cytotoxicity after incubation with human colonic mucus. In contrast, culture supernatants from LB-grown EHEC isolates that produced SLT-I, SLT-II, SLT-IIc or SLT-IIe did not show increased cytotoxicity after incubation with mouse or human intestinal mucus. The A subunits of purified SLT-II and SLT-IIvhb that had been treated with mouse intestinal mucus or trypsin were cleaved to A1 fragments by the mucus, but trypsin-mediated cleavage, unlike treatment with mouse intestinal mucus, did not result in increased Vero cell cytotoxicity activity. This finding implies that the increased cytotoxicity of SLT-IIvhb detected after incubation with mucus is probably not due to cleavage of the A subunit into the A1 and A2 fragments. Taken together, these results indicate that mouse or human intestinal mucus directly activates SLT-II-related toxins from B2F1 and H414-36/89 and suggest that toxin activation may explain the low 50% lethal doses of B2F1 and H414-36/89 in streptomycin-treated mice.

Cluster of cases of haemolytic uraemic syndrome due to unpasteurised cheese

A cluster of four patients (1 girl, 3 boys) from a French village (2,000 inhabitants) had acute haemolytic uraemic syndrome (HUS) between March 1992 and May 1993. All had prodromes with fever and diarrhoea, then acute renal failure, anaemia, schistocytosis and thrombocytopenia. Peritoneal dialysis was carried out in three children (duration 3-12 days). The verotoxin VT2 gene was identified by polymerase chain reaction in the stools of two children. Some days prior to the diarrhoea, all children had eaten a cheese made with unpasteurised mixed cows' and goats' milk from the same farm. A case control study showed that the occurrence of HUS was linked to the consumption of this milk product (P = 0.006). The VT 2 gene was isolated from the cheese and from the stools of goats and cows from the farm, but not from the stools of farm employees.

Analysis of the enterohemorrhagic Escherichia coli O157 DNA region containing lambdoid phage gene p and Shiga-like toxin structural genes

In this study, we determined the nucleotide sequence of the p gene contained within a 5-kb EcoRI restriction fragment cloned from Shiga-like toxin II (SLT-II)-converting phage 933W of Escherichia coli O157:H7 strain EDL933. The p gene was 702 bp long and had 95.3% sequence similarity to the p gene of phage lambda. Multiple hybridization patterns were obtained when genomic DNA fragments were hybridized with both p and slt-I, slt-II, or slt-IIc sequences. All O157 isolates also possessed an analog of lambda gene p which was not linked with either slt-I or slt-II. Restriction fragment length polymorphism comparisons of clinical O157 isolates and derivates undergoing genotype turnover during infection were made, and loss of large DNA fragments that hybridized with slt-II and p sequences was observed. To further analyze the DNA region containing the p and slt genes, we amplified fragments by using a PCR with one primer complementary to p and the other complementary to either the slt-I or the slt-II gene. PCR analysis with enterohemorrhagic E. coli O157 and non-O157 strains yielded PCR products that varied in size between 5.1 and 7.8 kb. These results suggest that even within O157 isolates, the genomes of SLT-converting phages differ. The methods described here may assist in further investigation of SLT-encoding phages and their role in the epidemiology of infection with enterohemorrhagic E. coli.