key: cord-007484-nvhu7blo
authors: Richard Dorn, C.; Francis, David H.; Angrick, Elisabeth J.; Willgohs, JoAnn A.; Wilson, Richard A.; Collins, James E.; Jenke, Bruce H.; Shawd, Stacey J.
title: Characteristics of vero cytotoxin producing Escherichia coli associated with intestinal colonization and diarrhea in calves
date: 2002-11-13
journal: Vet Microbiol
DOI: 10.1016/0378-1135(93)90136-u
sha: 
doc_id: 7484
cord_uid: nvhu7blo

Isolates of Escherichia coli which produce Vero cytotoxin (VTEC) were obtained during 1983–1989 from calves raised in 5 north-central states of the USA. All of the calves experienced intestinal epithelial colonization by VTEC, diarrhea or both; twelve of the calves had bloody diarrhea. Twenty one isolates were serogroup O111 and the others were O103, O69, O45, O26, O5, or non-typable (4 isolates). All but one of the isolates hybridized with the CVD419 probe which identifies most VTEC strains. Thirty two isolates hybridized with the VT1 probe, 3 with both the VT1 and VT2 probes, and one with neither probe. The culture filtrate of the VT probe negative isolate was partially neutralized by SLT I monoclonal antibody. For the other isolates, the results of toxin neutralization by anti-SLT I and anti-SLT II monoclonal antibodies corresponded exactly with the VT1 and VT2 probe hybridization results. Three of the strains adhered in a localized manner to HEp-2 cells and Intestine 407 cells.

first reported that certain E. coli isolated from human diarrhea produced cytotoxins which acted in vitro on Vero cells. Two serologically distinct Vero cytotoxins have been described: VT 1 and VT2, also called SLT I and SLT II because of their similarity to the toxin produced by Shigella dysenteriae 1 (O'Brien and Holmes, 1987). E. coli producing Vero cytotoxin (VTEC) have been isolated from cattle, beef, other meats, milk and milk products (Karmali, 1989; Giffin and Tauxe, 1991 ) . Many of the bovine VTEC have been isolated from cattle with diarrhea. Pathological examination of the intestinal tracts of some of these animals revealed characteristic epithelial lesions described as attaching and effacing with loss of microvilli and pedestal and cup formations of the cell membrane (Hall et al., 1985; Moxley and Francis, 1986 ) .

VTEC of animal and human origin did not hybridize with the enteropathogenic E. coli (EPEC) adherence factor (EAF) DNA probe (Dorn et al., 1988 ) . Many of the VTEC of bovine origin have characteristics in common with those proposed for the enterohemorrhagic E. coli (EHEC) (Levine, 1987 ) . These E. coli of human origin produce Vero cytotoxin; they are associated with hemorrhagic colitis; they lack production of enterotoxins LT and ST characteristic of enterotoxigenic E. coli (ETEC); and they are not invasive. A DNA probe, designated CVD419, has been developed which hybridizes with most E. coli O 157: H7 and other serotypes that fit the above criteria .

The source of the CVD419 probe was serotype O 157:H7 which is associated with human outbreaks and sporadic illnesses. This serotype is a serious human pathogen capable of producing hemorrhagic colitis (HC), hemolytic uremic syndrome (HUS) and thrombotic thrombocytopenic purpura (Karmali, 1989) . The sources of infection and associated human illnesses have been traced to undercooked beef and unpasteurized milk; serotype O 157: H7 E. coli have been isolated from cattle (Karmali, 1989; Giffin and Tauxe, 1991 ) . The isolates have been from healthy cattle, except for four reports of serotype O 157: H7 from diarrheic calves (Kashiwazaki et al., 1981; Orskov et al., 1987; Blanco et al., 1988; Yano et al., 1988) . None of these calves was reported to have blood present in the feces. The purpose of this study was to characterize the VTEC isolates from cattle obtained by two veterinary diagnostic laboratories. This information may be used in the diagnosis and epidemiologic investigation of VTEC infections of cattle, and in understanding their possible role in human infections.

Bacterial strains and tissue specimens VTEC isolates from calves with diarrhea were collected by the South Dakota Animal Disease Research and Diagnostic Laboratory and the Department of Diagnostic Investigation at the University of Minnesota. Entire calves and tissues collected at field necropsies were routinely submitted to the laboratories. Historical and clinical information about each animal was obtained from submission forms and interviews with producers and veterinarians. Intestines were processed for histopathological examination using routine methods. Colonization was based upon observations under light microscopy of short bacilli which were intimately adherent to intestinal epithelial cells exhibiting irregular erosive lesions typical of attaching and effacing infections (Moxley and Francis, 1986; Janke et al., 1990) .

Specimens of intestines were cultured aerobically on blood agar, tergitol-7 agar (Difco Laboratories, Detroit, MI ) and E agar (Francis et al., 1982 ) and on Brilliant green agar plates (Flow Laboratories, McLean, VA). The following biotyping tests were performed: indole, methyl red, Voges-Proskauer, citrate, H2S, urease, phenylalanine deaminase, lysine decarboxylase, arginine decarboxylase, ornithine decarboxylase, motility, gelatin hydrolysis, salicin, malonate, esculin hydrolysis, nitrate, oxidase, and 11 carbohydrate and 6 sugar alcohol fermentations. Isolates from the intestines identified as E. coli were examined for production of enterotoxin by methods previously described (Francis, 1983 ) and serotyped by the E. coli Reference Center, Pennsylvania State University, University Park, Penn.

The VT1 probe was a HinclI fragment of 0.75 kb specific for VT1 sequences cloned from a VT 1 encoding phage carried by H 19, E. coli 026: H 11 (Willshaw et al., 1985) . The VT2 probe was a 0.85 kb AvaI-PstI fragment specific for VT2 sequences cloned from a VT2-encoding phage carried by E. coli strain E32511 of serotype O157:NM . The CVD419 probe was a 3.4 kb HindlII fragment cloned from strain 933, E. coli O 157: H7 . The probe fragments were cut from low melting point agarose gels and labelled with deoxyadenosine 5'-a [35S ] thiotriphosphate (Amersham, Arlington Heights, IL) using random primers (Feinberg and Vogelstein, 1984) . Hybridization with the 35S-labelled probes was performed according to Maniatis et al. ( 1982 ) . In all hybridizations, the following three controls were used: E3787, E. coli 026 :H 11, VT 1 positive, VT2 negative and CVD419 positive; E32511, E. coli O157:NM, VT1 negative, VT2 positive and CVD419 positive; and 14R519, E. coli K12, negative with all of these probes.

The Vero cytotoxin assay used was a modification of the method of Scotland et al. (1985) . Five-fold dilutions of filtrates was tested on Vero cell monolayers. The plates were resealed and incubated at 37°C, 5% CO2 for 2 days. A filtrate was considered positive if a 1 : 10 or greater dilution produced a cytotoxic effect on the monolayer and the corresponding heat-inactivated preparation was negative.

Neutralization tests were performed using two mouse ascitic fluid reagents containing monoclonal antibody (mAb) 13C4 and mAb BC5-BB 12, respec-tively, obtained from Dr. N. A. Strockbine, Centers for Disease Control, Atlanta, Georgia. mAb 13C4 has approximately a 1:1000 neutralizing titer against 25 CDso of SLT I. mAb BC5-BB 12 has approximately a 1 : 50000 neutralizing titer against 25 CDs0 of SLT 11. The stock ascitic fluid preparations were diluted 1 : 1000 and 1 : 100 for use in the neutralization tests. The respective ascitic fluid dilutions (0.025 ml) were mixed with an equal amount of filtrate in 0.2 ml supplemented MEM and incubated in a 96 well plate at 37 °C for 3 hrs. The neutralization mixtures were then transferred to 96 well plates containing Vero cell monolayers and incubated at 37°C, 5% C02 for 2 days. Neutralization of the original filtrate was recorded when the addition of anti-SLT specific ascitic fluid resulted in a decrease in the cytotoxic titer by one or more dilutions from the cytotoxic titer of the original filtrate.

The method of Scotland et al. ( 1985 ) was used to test each strain for ability to adhere to HEp-2 cells and Intestine 407 cells, in the presence of D-mannose, during a 6 h incubation period. Only localized adherence as described by Nataro et al. (1987) was observed and it was scored as the percentage (mean of at least 2 tests) of HEp-2 or Intestine 407 cells with 10 or more adherent bacteria. Strains which adhered to less than 10% of the HEp-2 or Intestine 407 cells were considered non-adhesive for that cell line.

A total of 36 VTEC isolates, each from a different calf between 1983 and 1989, were identified for study (Table 1 ) . They were isolated predominantly from calves designated as dairy or dairy-crossbred; only 4 of the isolates were from beef or beef-crossbred calves. One calf was 3 months old and all the other calves of known age were less than 3 weeks of age. Sixteen of the calves were from Minnesota, 14 from South Dakota, 4 from Iowa, one from North Dakota and one from Nebraska. Thirty two of the 36 calves had bacteria colonizing the epithelium of their intestinal tract and diarrhea, 3 had colonization with no observed diarrhea, and one had diarrhea but colonization was obscured by post-mortem autolysis. Thirteen of these were observed by histopathology to have lesions of the colon. Twelve of the calves were observed to have bloody diarrhea.

All 36 isolates fermented sorbitol within 24 hours. The four 05:NM and two other isolates did not ferment raffinose. None of the isolates were positive for the K99 antigen.

Twenty one of the isolates were serogroup O 111, 4 isolates were non-typable, 4 isolates were 05 and the other 7 isolates were distributed among 6 other 

Except for isolate 84-5406, all of the isolates hybridized with the CVD419 probe (Table 1 ) . Thirty two of the 36 isolates hybridized with only the VT1 probe; two of the serotype O 111 :NM isolates and one of the O5:NM isolates hybridized with both VT 1 and VT2 probes. None of the isolates hybridized with only the VT2 probe.

The highest titer obtained for these VTEC isolates was 31250. Filtrates of the other isolates had lower titers: 8 were 6250 and the other 27 were less than 6250 (Table 1 ) . For the VT1 and VT2 positive isolates, the results of neutralization of their filtrates by anti-SLT I and anti-SLT II ascites fluid matched exactly the VT probe hybridization results. The filtrate of isolate 84-5406, which did not hybridize with either VT 1 or VT2, was partially neutralized by anti-SLT I monoclonal antibody. The anti-SLT I ascites fluid completely neutralized the filtrates of 7 of the 36 SLT I positive isolates. Incomplete neutralization of SLT I was observed with filtrates from the remaining 29 SLT I positive isolates. Anti-SLT II ascites fluid completely neutralized one of the 3 filtrates of SLT II positive isolates.

Three isolates possessed the ability to adhere in a localized manner to greater than 10% of both HEp-2 and Intestine 407 cells; two were serotype O 111 :NM and one was serotype O 103: H2,12. The adherence test results for a particular isolate were similar for both HEp-2 and Intestine 407 cells.

Other pathogenic agents were isolated from 24 of the 36 calves; 8 were infected with two other pathogens besides VTEC. Fifteen co-infections involved cryptosporidia, 7 coronavirus, 9 rotavirus, one enterotoxigenic E. coli (ETEC) and one bredavirus.

The absence of serotype O 157: H7 in this series of VTEC isolates from calves is consistent with previous reports that O157:H7 is not a frequent cause of illness in cattle. There are only 4 reports of this serotype being isolated from ill cattle (Kashiwazaki et al., 1981; Orskov et al., 1987; Blanco et al., 1988; Yano et al., 1988 ) and in none of these was O157:H7 confirmed as the cause of illness.

The most common serogroup in this series, O 111, has been recorded and also identified as verotoxigenic in other studies of bovine isolates (Sherwood et al., 1985; Marques et al., 1986; Mohammad et al., 1986; Mainil et al., 1987; Schoonderwoerd et al., 1988; Smith et al., 1988; Dorn et al., 1989) . Other serogroups represented in this series have also been previously isolated from cattle and shown to be verotoxigenic: O 103 (Mohammad et al., 1986; D orn et al., 1989; Gonzalez et al., 1989) : 026 (Kashiwazaki et al., 1981; Sherwood et al., 1985; Marques et al., 1986; Mainil et al., 1987; Smith et al., 1988; Rietra et al., 1989; Scotland et al., 1990; Woodward et al., 1990) : and 05 (Chanter et al., 1986; Hall et al., 1985; Dorn et al., 1989) . No previous published reports of verotoxigenic isolates belonging to serogroups 069 and 045 were found. None of the VTEC isolates showed delayed sorbitol fermentation, a unique characteristic of serotype O 157 : H7 (Farmer and Davis, 1985 ) .

Twelve of the 36 calves had bloody diarrhea. Seven of these calves yielded cultures of serotype O 11 I:NM, 3 were non-typeable, one was O103:H2,12 and one was 05 : NM. Serotypes which were isolated from calves in this study and which have been reported from HC in humans include O11 I:NM, O45:H2, O26:H11 and O5:NM (Bopp et al., 1987) Other 05 : NM strains of bovine origin have been associated with hemorrhagic enteritis in cattle. A serotype 05 : NM E. coli was isolated from an outbreak of dysentery among 8 to 21 day old calves in England (Chanter et al., 1986 ) . In both natural and experimentally reproduced disease with this strain, the calves passed copious bright red blood in the feces and developed diarrhea (Hall et al., 1985 ) . This O5:NM strain produced a toxin active on Vero cells which was neutralized by anti VT 1 and hybridized with the VT 1 and CVD419 probes (Dorn et al., 1989) . One of the serotype O5:NM E. coli in the series ( 84-5406 in Table 1 ) was previously reported by Moxley and Francis ( 1986 ) to produce bloody diarrhea in calves and lambs. In the present study, isolate 84-5406 produced a toxin active on Vero cells which was incompletely neutralized by anti SLT I ascites fluid. However, it failed to hybridize with the VT 1, VT2 and CVD419 probes. It apparently produces another toxin active on Vero cells, neutralizable by anti-SLT I ascites fluid, but not mediated by the phage genes encoding SLT I and SLT II. A similar observation has been made with porcine edema disease strains which produce a variant of SLT II that is chromosomally mediated and active on Vero cells, but not on HeLa cells Weinstein et al., 1988) .

Our observation that 32 of 36 VTEC isolates from calves in north-central states hybridized with only the VT 1 probe is consistent with reports that most of the non-O157:H7 bovine VTEC isolated by laboratories located in England were positive for only VT1 (Smith et al., 1988; Dorn et al., 1989; Scotland et al., 1990) . However, few of the non-O 157 :H7 VTEC isolated from healthy cattle (usually mature) hybridized with only the VT 1 probe in Germany and in Thailand, 5% and 14%, respectively (Montenegro et al., 1990; Suthienkul et al., 1990) . Serotype O 157:H7 and other serotypes isolated from human HC and HUS cases usually produce SLT II alone or both SLT I and II, but much less frequently SLT I alone Dickie et al., 1989; Ostroff et al., 1989) . The SLT production pattern of VTEC isolates from human cases appears to be similar to that of VTEC isolates from healthy cattle, but dissimilar to that of calves with diarrhea. It is possible that SLT II strains are more pathogenic for humans while SLT I strains are more pathogenic for calves resulting in higher frequencies of these respective toxins among isolates from humans and from calves. It is also possible that SLT I producing E. coli are more common in calves and SLT II E. coli infections are more common among mature cattle. The higher prevalence of SLT IIE. coli than SLT I E. coli infections among mature cattle (Montenegro et al., 1990; Suthienkul et al., 1990 ) might result in more passive transfer of SLT II antibody than SLT I antibody, resulting in lower incidence of clinical illness due to SLT II producing VTEC strains in calves. This hypotheses should be explored in future studies.

The VTEC isolates of this study were from midwestern states with a substantial beef production industry, yet 28 of 32 VTEC isolates came from dairy or dairy cross calves. The predominance of dairy calves as the source of VTEC strains suggests that these animals may be more susceptible to life threatening illness by VTEC than are beef calves. Dairy calves are usually fed a milk replacement starter and many may either not receive or poorly absorb colostrum: they are thus more at risk than beef calves which are usually nursed. This is consistent with the observation of higher incidence of gastroenteritis in breast-fed children infected with E. coli serogroup O 111 compared to that in formula-fed infants, similarly infected (Belnap and O'Donnell, 1955 ) .

All but one of the 36 VTEC isolates in this study hybridized with the CVD419 probe (Table 1 ). In another study of VTEC isolated from healthy cattle, only 71% hybridized with this probe. It is possible that the EHEC probe is more specific for pathogenic VTEC than for VTEC strains associated with subclinical infections; however, additional studies of cattle with different ages and from different geographic areas are needed.

Epidemic gastroenteritis due to Escherichia coli 0111

Production of toxins by Escherichia coli strains isolated from calves with diarrhea in Galicia

Unusual verotoxin-producing Escherichia coli associated with hemorrhagic colitis

Dysentery in calves caused by an atypical strain of Escherichia coli

Purification of an Escherichia coli serogroup O 157 : H7 verotoxin and its detection in North American hemorrhagic colitis isolates

Properties of Veto cytotoxin producing Escherichia coli of human and animal origin belonging to serotypes other than O157:H7

H7 antiserum sorbitol fermentation medium: a single tube screening medium for detecting Escherichia coli O 157:H7 associated with hemorrhagic colitis

A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity

Production of K88, K99 and 987P antigens of Escherichia coli cultured on synthetic and complex media

Use of immunofluorescence, gram's staining, histologic examination, and seroagglutination in the diagnosis of porcine colibacillosis

The epidemiology of infections caused by Escherichia coli O 157:H7, other enterohemorrhagic E. coli, and the associated hemolytic uremic syndrome

Serotypes and antibiotic resistance of Verotoxigenic (VTEC) and Necrotizing (NTEC) Escherichia coli strains isolated from calves with diarrhea

Dysentery caused by Escherichia coli (S 102-9) in calves: natural and experimental disease

Attaching and effacing Escherichia coli infection as a cause of diarrhea in young calves

Infection by Vero cytotoxin-producing Escherichia coli

Veto cytotoxin produced by Escherichia coli strains of animal origin

Vero response to a cytotoxin of Escherichia coli

Escherichia coli that cause diarrhea: enterotoxigenic enteropathogenic, enteroinvasive, enterohemorrhagic, and enteroadherent

A DNA probe to identify enterohemorrhagic Escherichia coli of O 157:H7 and other serotypes that cause hemorrhagic colitis and hemolytic uremic syndrome

Shiga-like toxin production and attaching effacing activity of Escherichia coli associated with calf diarrhea

Molecular Cloning: a Laboratory Manual

Production of Shiga-like toxin by Escherichia coli

Escherichia coli strains isolated from pigs with edema disease produce a variant of Shiga-like toxin II

Serotypes of verocytotoxic Escherichia coli isolated from cattle and buffalo calf diarrhoea

Detection and characterization of fecal verotoxin-producing Escherichia coli from healthy cattle

Natural and experimental infection with an attaching and effacing strain ofEscherichia coli in calves

Patterns of adherence of diarrheagenic Escherichia coli to HEp-2 cells

Shiga and Shiga-like toxins

Cattle as reservoir of verotoxin-producing Escherichia coli O 157 : H7. Lancet, ii

Toxin genotypes and plasmid profiles as determinants of systemic sequelae in Escherichia coliO 157: H7 infections

Comparison of vero-cytotoxin-encoding phages from Escherichia coli of human and bovine origin

Colitis in calves: natural and experimental infection with a verotoxin-producing strain of Escherichia coli O111 :NM. Canad

Laboratory tests for enterotoxin production, enteroinvasion and adhesion in diarrhoeagenic Escherichia coli

Properties of strains of Escherichia coli belonging to serogroup O 157 with special reference to production of Vcro cytotoxins VTI and VT2

Properties of strains of Escherichia coli 026: H 11 in relation to their enteropathogenic or enterohemorrhagic classification

Shiga-like toxin production from Escherichia coli associated with calf diarrhea

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

Shiga-like toxin-producing Escherichia coli in retail meats and cattle in Thailand

Cloning and sequencing of a Shiga-like toxin type II variant from an Escherichia coli strain responsible for edema disease of swine

Cloning of genes determining the production of Vero cytotoxin by Escherichia coli

Heterogeneity of Escherichia coli phages encoding Vero cytotoxins: comparison of cloned sequences determining VTI and VT2 and development of specific gene probes

DNA probes for the detection of toxin genes in Escherichia coli isolated from diarrhoeal disease in cattle and pigs

Fimbria-like adhesive factor (EAF44) from verocytotoxigenic Escherichia coli

The authors wish to thank G.A. Willshaw of the Division of Enteric Pathogens, Central Public Health Laboratory, Colindale, London, UK for the cloned VT 1 and VT2 probes. The cloned CVD419 probe was kindly provided by M.M. Levine, Center for Vaccine Development, University of Maryland, School of Medicine, Baltimore; and the SLT-specific monoclonal antibodies were kindly provided by N.A. Strockbine, Centers for Disease Control, Atlanta, Georgia.