Zbl. Bakt. 276,243-253 (1992) © Gustav Fischer Verlag, StuttgartlNew York

Characterization of Shiga-like Toxin Producing Escherichia coli (SLTEe) Isolated from Calves with and without Diarrhoea * LOTHAR H. WIELER, ROLF BAUERFEIND, and GEORG BALJER Institut rur Hygiene und Infektionskrankheiten der Tiere, Justus-Liebig-Universitat GieRen (Gf. Direktor: Prof. Dr. Dr. G. Baljer), D-W6300 GieRen

Received June 5, 1991 . Revision received and accepted August 1, 1991

Summary To determine if shiga-like toxin producing Escherichia coli (SLTEC) are involved in neonatal calf diarrhoea, isolated E. coli strains from diarrhoeic and non-diarrhoeic calves were characterized for shiga-like toxin (SLT) by colony blot hybridization and cytotoxicity assays. None of 150 E. coli strains isolated from diarrhoeic calves in 1985-1988 was positive for SLT, while 7/232 (3.0%) isolated in 1989 were positive for SLT. In contrast, samples collected during 1989 and 1990 from diarrhoeic calves were 21.9% SLTEC positive, and samples from non-diarrhoeic calves were 12.9% SLTEC positive. SLT I positive E. coli strains were isolated more often from diseased (17.8%) than from healthy animals (5.0%), while SLT II positive E. coli were more often detected in non-diarrhoeic (8.9%) than in diarrhoeic calves (4.1 %). The mean percentage of SLT I positive E. coli in the whole E. coli flora of the samples was significantly higher in diarrhoeic than in healthy animals, implying a pathogenic role of SLT I producing E. coli in neonatal calf diarrhoea. Enterohemolysin was produced by 70.8% of the SLT I producing E. coli strains examined. Determination of 0- and K-antigens of SLT positive E. coli revealed a highly diverse spectrum of SLTEC O-groups in calves. While no E. coli isolate belonged to serotype 0157: H7, classical human enteropathogenic E. coli O-groups (026, 0111, 0128) were detected. These results support the theory that cattle serve as a resevoir for human SL TEC infection.

Zusammenfassung

E. coli Stamme von durchfallkranken und gesunden Kalbern wurden mittels Kolonie Blot-Hybridisierung und Cytotoxizitatstests auf das Vorhandensein von Shiga-like Toxinproduzierenden Escherichia coli (SLTEC) untersucht. Wahrend in 150 Kotproben durchfallkranker Kalber aus den Jahren 1985 bis 1988 keine SLTEC nachgewiesen wurden, wiesen 7 von 232 Kotproben (3,0%) aus dem Jahr 1989 SLT-positive E. coli auf. 1m

* Dedicated to Dr. Ida 0rskov and Dr. Frits 0rskov on the occasion of their 70th birthday.

244

L. H. Wieler, R. Bauerfeind, and G. Baljer

Gegensatz dazu lag der Anteil von SLTEC in den Jahren 1989/90 wesentlich hoher. Hier wurden bei 21,9% der durchfallkranken Kalber und bei 12,9% klinisch gesunder Kalber SLTEC nachgewiesen. SLT I-positive E. coli wurden haufiger von durchfallkranken Kalbern isoliert (17,8%) als von klinisch gesunden (5,0%), wahrend SLT II-positive E. coli haufiger bei gesunden (8,9%) als bei durchfallkranken gefunden wurden (4,1%). Der signifikant hOhere Prozentsatz SLT I-positiver E. coli an der E. coli-Gesamtflora durchfallkranker Tiere macht eine Beteiligung der SLTEC I bei der Pathogenese der neonatalen Kalberdiarrhoe wahrscheinlich. SLT I-positive E. coli produzierten zu 70,8 % gleichzeitig Enterohamolysin. Die 36 isolierten SLTEC wiesen 17 verschiedene O-Antigene auf. Wahrend das Serovar 0157: H7 nicht nachgewiesen werden konnte, wiesen einige bovine SLTEC O-Antigene klassischer humaner enteropathogener E. coli auf (026,0111,0128). Diese Ergebnisse unterstreichen die These, dag Rinder ein Reservoir fur humane Infektionen mit SL TEC darstellen. Introduction Shiga-like toxin (SLT) produced by E. coli is so named because it is closely related to Shigella dysenteriae toxin (30). SLT was first described as verotoxin (19). SLT producing E. coli (SLTEC) have been identified as the causative agent in human haemorrhagic colitis (HC), haemolytic uraemic syndrome (HUS), and thrombotic thrombocytopenic purpura (TIP) (18,27). Two separate toxins, SLT I and SLT II, have been identified and cloned, and a variant of SLT II (SLT IIv) has also been described (9,15). While SLT I and SLT II producing E. coli are important pathogens for humans, SLT IIv producing E. coli cause edema disease in piglets (1, 9). SLTEC have been isolated from both healthy cattle and cattle suffering from diarrhoea in Sri Lanka and the United Kingdom (7, 24, 25, 34). In North America, cattle have also been found to be a source of SLTEC (6, 21). In a recent report from the Federal Republic of Germany, 10.8% of the healthy cattle examined were found to be SLTEC positive (28/259) (26). Taken together, these reports suggest that cattle may be an important reservoir for SL TEC involved in human disease. Little is known about a putative pathogenic role for SL TEC in infectious diseases of cattle. One of these diseases, neonatal calf diarrhoea, is a common problem in calf rearing. Recently, we .isolated SLTEC from calves with diarrhoea (4). In the present study,:we characterized these SLTEC isolates by colony blot hybridization, 0- and Kantigen typing and analysis of culture supernatants and celllysates. We compared these results with those obtained from SLTEC isolated from healthy calves. Our results suggest that SLT I producing E. coli may be involved in neonatal calf diarrhoea.

Materials and Methods E. coli strains. E. coli strains from diarrhoeic calves (age < 10 weeks) were received from the Institut fur Medizinische Mikrobiologie, Infektions- und Seuchenmedizin, Universitat Munchen, the Tiergesundheitsdienst Bayern, and the Landesuntersuchungsamt fur das Gesundheitswesen Sudbayern. These strains were isolated from routine diagnostic samples collected from 1985 to 1989. E. coli reference strains were kindly provided by Dr. Montenegro, Institut fur Veterinarmedizin des Bundesgesundheitsamtes and Dr. Karch, Institut fur Medizinische Mikrobiologie und Immunologie der Universitat Hamburg [2324 (SLT II, 2403 (SLT I), SP2389 (SLT IIv), TIP-l (SLT I1II), C600 J1 (SLT I), C600 W34 (SLT II), C600, K87 (STILT), and B41 (LT)].

Characterization of Shiga-like Toxin Producing Escherichia coli

245

Stool samples and faecal swabs. Stool samples from non-diarrhoeic calves were collected in southern Bavaria during 1990, plated on MacConkey agar-, Gassner agar- and 5% sheep blood agar-plates, and incubated 24 to 48 h at Faecal swabs of diarrhoeic calves were obtained from the Tiergesundheitsdienst Bayern, and the Landesuntersuchungsamt ftir das Gesundheitswesen Nordbayern. Samples were stored at 4°C or lyophilized prior to testing. Colony blot hybridization. E. coli strains and E. coli colonies from faecal swabs (10-35 E. coli colonies per stool sample) were plated on nitrocellulose filters overlaying Luria-broth (LB) agar-plates, and grown for 16-24 h at Colony blots were prepared for hybridization as previously described (28). Briefly, the E. coli colonies were lysed with NaOH, neutralized with a TRISIHCI solution, and heated (1 h, 80°C). Two oligonucleotides specific for the SLT I and the SLT II gene were synthesized using the published sequence of probes 772 (GATGATCTCAGTGGGCGTTC), and 849 (TCTGAAACTGCTCCTGTGTA), respectively (16) (DNA synthesizer Modell 380, Applied Biosystems, Weiterstadt, FRG). These 20-mers were endlabelled with gamma- 32 P-ATP (Amersham, Braunschweig, FRG) by T4 polynucleotide kinase (Boehringer, Mannheim, FRG) and purified by chromatography on Sephadex G-50. Colony blots were prehybridized and hybridized as previously described (16). Analysjs of SLTEC colonies. Isolated colonies positive for SLT genes were assayed for production of SLT, heat-labile (LT) and heat-stable (STA) E. coli enterotoxin, a-hemolysin (Hly) and enterohemolysin (EHly) and typed for 0- and K-antigens and biotyped. When multiple colonies isolated from the same stool sample were positive, three colonies were randomly picked and analysed as described below. Assay for SLT production. Vero cells were grown in RPMI 1640 and HeLa cells were cultured in MEM, both media supplemented with 10% heat-inactivated fetal calf serum, 2 mM glutamine, 100 U penicillin and 100 I1g streptomycin/ml (Gibco BRL, Berlin, FRG). These cell lines were cultured at in a humidified, 5% CO 2 incubator. E. coli strains were tested for SLT production in cytotoxicity assays using Vero cells and HeLa cells (10, 19). Briefly, E. coli strains were grown in tryptic soy broth for 16 h at 37°C with aeration, after which they were centrifuged (9000 x g, 20 min, 4°C) and the resulting supernatants harvested and stored at 4 0c. The bacterial pellets were resuspended in polymyxin B (100 11g/ ml) to lyse the cells and to release SLT I (18). After filter sterilization (0.22 11m, Millipore) aliquots of the supernatants and celllysates (log 10 dilutions) were assayed in quadruplicate in 96- well microtiter plates. Each well received 100 111 of the dilution to be tested and 100 111 of the cell suspension (1 x 106/ml Vero cells or 5 X lOs/ml HeLa cells). Cytotoxic effects were evaluated at 24 h, 48 hand 72 h by microscopic evaluation. The reciprocal of the highest dilution causing cytotoxic effects after 4& h in 50% of the wells was multiplied by 10 and recorded as the CDso/ml of supernatants or lysates, respectively. Samples yielding more than 10 CDsolml were considered to be cytotoxic. In each assay, positive controls included stipernatants and celllysates of strains 2324, 2403 and SP2389. Strain C600 was used as a negative control. Assay for heat-labile E. coli enterotoxin (LT) production. The production of LT was assayed by use of a latex agglutination kit (VET-RPLA; Oxoid, Wesel, FRG). E. coli were grown with agitation for 18 h at 37°C in Mundell medium (29), followed by a 1 h incubation at 37°C with polymyxin B (0.1 mg/ml) to lyse cells. The supernatant was harvested by centrifugation and filter sterilized as described above. Serial dilutions were added to 96-well microtiterplates (25 111/well). Latex beads coupled with anti-LT antibodies (25 I1Vwell) were added, the plate was shaken (Flow titertek, FRG) and incubated for 24 h at room temperature. Controls included lysates from the known positive E. coli strains K87 and B41 and the known negative C600 strain. Assay for heat-stable E. coli enterotoxin (STN production. STA-production of the SLTEC strains was evaluated by competitive enzyme immunoassay (EIA; Oxoid) of celllysates and by DNA-hybridization (Snapr hybridization kit; DuPont, Bad Homburg, FRG) of colony blots following the suppliers recommendations. For DNA hybridization to the STKspecific probe, colony blots of SL TEC were pretreated as described above. Controls included strains described above for LT-testing.

3rc.

3rc.

3rC

o

L. H. Wieler, R. Bauerfeind, and G. Balier

246

Assay for a-hemolysin and enterohemolysin production. E. coli strains were grown at 37"C on tryptose blood agar base (Difco, Detroit, USA), supplemented with 10 mM CaCl2 and 5% defribrinated sheep blood (previously washed three times in PBS, pH 7.2). Plates were monitored for hemolysis at 4 hand 24 h (5). 0- and K-antigen typing. E. coli strains were typed for 0- and K-antigens under the kind supervision of Mr. Jansen at the National Institute of Public Health and Environmental Protection, Bilthoven, The Netherlands (11). Biotyping. Biotyping was performed with the Api-bioMerieux 20-system (ApibioMerieux), following the suppliers recommendations. Sorbitol fermentation was detected on Sorbitol-MacConkey agar plates (23). Statistics. Results were analyzed by Fisher-t-test and Wilcoxon-Mann-Whitney-test. Unless mentioned otherwise the level of significance was p< 0.01).

Results

Detection of SLTEC in stool samples from diarrhoeic calves E. coli strains isolated from faecal swabs of diarrhoeic calves were tested for SLT I and SLT II by colony blot hybridization. In initial studies, only one colonylstrain was tested. Strains isolated from 1985-1988 were all negative for SLT genes (01150), but 2.6% of E. coli strains isolated in 1989 (6/232) were positive for SLT I and 0.4% (11 232) for SLT II (Table 1).

Table 1. SLTEC identified in calf faecal swabs by colony blot hybridization Source

a

No. of calves

% of calves with SLTEC

No. of coloniesl calf

SLT-I

SLT-II

Total

Calves with diarrhoea 150 1985-1988 232 1989 73 1989-1990

1 1 10-35

0 2.6 17.8

0 0.4 4.1

0 3.0 21.9

Calves without diarrhoea 1990 101

10-35

5.0

8.9

12.9a

One swab yielded both SLT-I and SLT-II producing E. coli strains

To increase the accuracy of our analysis, multiple colonies (10-35) from each faecal swab of stool samples taken from diarrhoeic calves in 1989 and 1990 were tested by colony blot hybridization. Under these conditions, a total of 21.9% of the calves examined (16173) were found to harbour SLTEC (Table 1). In this population, SLT I positive E. coli were more frequently found, and were detected in 17.8% of the calves (13/73), while only 4.1 % of these calves (3/73) harboured SLT II positive E. coli. Isolated SLT positive E. coli colonies were positive for either SLT I or SLT II, but not for both genes. Moreover, SLT I and SLT II positive strains were never found in the same stool sample.

Characterization of Shiga-like Toxin Producing Escherichia coli

247

Detection of SLTEC in stool samples from non-diarrhoeic calves To evaluate the distribution of SLT I and SLT II producing E. coli, stool samples from non-diarrhoeic calves were collected in 1990 in southern Germany and tested by colony blot hybridization. SLTEC were detected in 12.9% (13/101) of the samples (Table 1). However, in contrast to the results obtained from diarrhoeic calves, SLT II producing E. coli occurred at twice the frequency of SLT I producing strains. SLT II positive E. coli were detected in 8.9% of the samples (9/101), while only 5.0% of the samples contained SLT I producing E. coli (5/101) (Table 1). One sample contained both SLT I and SLT II positive E. coli strains.

Percentage of SLT positive E. coli colonies in SL TEC positive samples from diarrhoeic and non-diarrhoeic calves The percentages of individual E. coli colonies positive for SLT I or SLT II in samples obtained from diarrhoeic and non-diarrhoeic calves in 1989/1990 are shown in Table 2. A high variance was observed in the percentage of the 10-35 E. coli colonies/sample which were positive for SLT production. However, a significant difference was found between the percentages of SLT positive E. coli colonies isolated from healthy and diarrhoeic calves (p < 0.05). The diarrhoeic calves that were positive for SLT I producing E. coli yielded a mean percentage of 47.3% SLT I positive colonies. In 2 faecal swabs 100% of the E. coli colonies were SLT I positive. In marked contrast, healthy calves had a mean percentage of only 9.0% of the colonies which hybridized with the SLT I specific gene probe. No significant difference in the mean percentage of SLT II positive E. coli colonieslfaecal swab from diarrhoeic (27.6%) and healthy (33.0%) calves was found. Table, 2. Percentage of SLT positive E. coli colonies in SLTEC positive samples from 1989-1990 determined by colony blot hybridization SLT genotype

II

a

Isolated from calves

SLTEC+ samples

Percentage of SL TECpositivea colonies mean range

with diarrhoea without diarrhoea

13 5

47.3% 9.0%

5.0-100.0% 5.0- 20.0%

with diarrhoea without diarrhoea

3 9

27.6% 33.0%

2.0- 75.0% 5.0- 95.0%

No. of tested E. coli colonies/sample

No. of

=

10-35

SLT production assayed by cytotoxicity for Vera cells and HeLa cells To determine if E. coli colonies positive for SLT I or II genes were capable of producing the toxins, lysates of these colonies were assayed for cytotoxicity in vitro. All SLTEC lysates were toxic for Vero cells (CDso/ml > 10) (Table 3). Alilysates of SLTEC isolates from diarrhoeic calves were cytotoxic for HeLa cells, while two of the SLT II positive strains isolated from healthy calves (72/90-23 and 103/90) were not cytotoxic for HeLa cells. There was no significant difference in CDso/ml between strains isolated from healthy or diarrhoeic calves.

248

L. H. Wieler, R. Bauerfeind, and G. Baljer

Table 3. Properties of shiga-like toxin producing E. coli (SLTEC) strains a Strain

Calves with diarrhoea

Calves without diarrhoea

SLTgenotype

248/89 249/89 250/89 253/89 274/89 303/89 410/89 413/89 459/89 531/89 537/89 555/89 556/89 557/89 558/89 570/89 642/89 666/89 78/90 539/89 246/89 562/89 631/89

II II II II

15/90 72/90-56 83/90 93/90 109/90

I I I I I

12/90 30/90 37/90 68/90 69/90 72/90-23 103/90 110/90

II II II II II II II II

Cytotoxicity EHlyb for Vero- HeLa cells cells

0- and Kantigen type

Api 20profile

+ + + + + + + + + + + + + + + + + + + + + + +

+ + + + + + + + + + + + + + + + + + + + + + +

+ + + +

0111 n. typable n. typable 0111 n. typable 0153:K045:K? 026 n. typable n. typable 084:K? 0128 0123:K0118:K0123:K0111 0111 0118:K084:K?

104457 514456 514456 104457 514456 514457 514457 514456 504456 514456 514454 504454 504454 514456 504454 104457 104457 514456 514455

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

+ +

n. typable 039:K? 022:K? 091:K-

514457 514455 414457 514457

2 2 2 2

1 1 1 1

+ + + + + + + + + + + + +

+ + + + + + + + + +

04:K? 05:K0150:K0123:K0123:K-

504454 514454 514455 504454 504454

2 2 2 2 2

1 1 1 1 1

n. typable 02:K? 091:K032:K? 032:K? 032:K021:K021:K-

514457 504457 514457 514457 514457 514457 514457 514457

2 2 2 2 2 2 2 2

1 1 1 1 1 1 1 1

+

+ + + + + + + + + + +

+ +

+

Biotype

All strains listed were negative for production of heat-labile and heat-stable E. coli enterotoxin, and a-hemolysin. All strains fermented sorbitol b EHly = enterohemolysin

a

Characterization of Shiga-like Toxin Producing Escherichia coli

249

Detection of hemolysis and enterohemolysis, assay for heat-labile (LT) and heatstable enterotoxins (ST) None of the E. coli strains produced a-hemolysin (Table 4). In contrast, 55.6% of SLTEC showed enterohemolysis (20/36) (Table 3, Table 4). The simultaneous production of enterohemolysin and SLT was markedly more frequent in SLT positive strains isolated from diarrhoeic calves (17/23 strains = 73.9%) than in SLT positive strains isolated from healthy calves (23.1 %). Enterohemolysis was significantly more frequent in SLT I positive than in SLT II positive strains; 17/24 SLT I positive E. coli produced enterohemolysin. None of the SLTEC produced LT- or STA- enterotoxin, as determined by EIA and VET-RPLA (Oxoid) or with the Snapr hybridization assay (DuPont) (Table 4). Table 4. Production of other toxins by SLTEC SLTgenotype

II

No. of SLTEC strains producing STAa EHlya HIP

Isolated from calves

No. of SLTEC+ strains

LTa

with diarrhoea without diarrhoea

19 5

0 0

0 0

15 2

0 0

with diarrhoea without diarrhoea

4 8

0 0

0 0

2 1

0 0

a, EHLy = enterohemolysin, Hly stable enterotoxin

= a-hemolysin, LT = heat-labile enterotoxin, STA =

heat-

Determination of biotype and 0- and K-antigens All 36 bovine SLTEC were of biotype 1 (Table 3) and all were capable of fermenting sorbitol. The SLTEC contained 17 different O-groups. Six SLTEC strains carried 0antigens characteristic 6f human enteropathogenic E. coli (EPEC): 0111 (4 strains), 026 (1 strain), and 0128 (1 strain). None of the isolated strains was positive for 0157, which is of interest since E. coli 0157: H7 and 0157: H- are strains known to cause severe clinical syndroms in man, including haemolytic uraemic syndrome. K99antigens, often found in neonatal calf diarrhoea, were not detectable (Table 5). 0Table 5. SLTEC O-groups, associated with classical E.coli diseases O-group

No. of SLTEC

Associated with a

0111 0128 026

4 1 1

human EPEC human EPECIETEC human EPEC and bov. diarrhoea/septicemia

02 045

1 1

bov. diarrhoea/septicemia porcine ETEC

a 0rskov and 0rskov, 1979, 0rskov et aI., 1990

250

L. H. Wieler, R. Bauerfeind, and G. Baljer

groups not yet associated with disease which were most oftenly found were 0123 (4 strains) and 0 32 (3 strains). Some SLTEC (7/36) were not typable (Table 3). Discussion In this study, the production of SLT by E. coli strains isolated from calves with and without diarrhoea was examined. In calves with diarrhoea, the predominant SLT produced was SLT I, while in calves without diarrhoea, SLT II was most often found. The ratio of SLT I : SLT II in diarrhoeic calves was over 4 : 1. In non-diarrhoeic calves, the ratio was 1 : 2, with a total of 12.9% bearing SLTEC. A similar ratio was found by Suthienkul et al. (35) in 145 healthy buffalos and cattle, and Montenegro et al. (26) in 181 healthy cattle. Moreover, Montenegro et al. reported 11.6% of healthy dairy cows in the Federal Republic of Germany to be SLTEC positive, a result agreeing well with our findings. While the total number of animals positive for SLTEC did not vary significantly be~een healthy and diarrhoeic calves (12.9% and 21.9%, respectively), a significantly higher colony-percentage of SLT I producing E. coli was found in diarrhoeic calves, suggesting a pathogenic role in neonatal calf diarrhoea. SLT I producing strains had a mean colony percentage of 47.3% in diarrhoeic calves, while the percentage was only 9.0% in healthy calves. There was no statistical difference in the colony percentage of SLT II producing E. coli obtained from calves with or without diarrhoea. Therefore, SLT I producing E. coli seems to be more important in calf diarrhoea. This is in coherence with data from Mohammed et al. (25), who found a significant correlation between diarrhoea and SLTEC in cattle (27% in cattle with diarrhoea, 6% in cattle without diarrhoea). However, these authors examined just one colony/sample and did not distinguish between SLTEC I and SLTEC II strains. The putative pathogenic role of SLTEC has yet to be proven in infection models. Only one report exists where diarrhoea was induced in calves by infection with a SLTEC strain (14). The occurrence of SLTEC in the samples examined increased from 1985 to 1989/90, where 21.9% of the diarrhoeic calves were SLTEC positive. The detection rates found in 1985-1989, when we could only test one colony/sample, give a somewhat misleading impression. As we and others have shown, the percentage of SLTEC positive colonies in a positive faecal sample varies widely (1, 18, 26). For a reliable assessment of SLTEC presence, more than one colony should routinely be assayed, ideally 10-20. Nevertheless, the absolute number of SLTEC in calf faeces was lower in 1985-1989. Possibly, the frequency of SLTEC harbouring cattle has increased. The observed change in epidemiology may be relevant to human disease and deserves further study since cattle may be a reservoir for human SLTEC infections (7, 8, 17). The examination of other toxins did only reveal the production of enterohemolysin. This enterohemolysin has not been characterized further and nothing is known about a pathogenic role in disease (5). None of the 36 SLT positive E. coli produced STA or LT. Enterotoxigenic E. coli (ETEC) are well known to cause diarrhoea in calves (2, 12). The missing of these pathogens further supports the pathogenic role of SLTEC. In contrast to our results, Seriwatana et al. (33) found 3/34 ETEC (LT II positiv) to bear the SLT II gene, and Mohammad et al. (25) found 6/21 SLT I positive E. coli to produce ST simultaneously. Detection of enterohemolysin may prove to be a useful marker for routine diagnostics of SLTEC, since a strong positive correlation betwen SLT I production and en-

Characterization of Shiga-like Toxin Producing Escherichia coli

251

terohemolysin was found in diarrhoeic calves. An apparent difference exists between characteristics of human and bovine SLTEC. In man, 89% of all SLTEC produced enterohemolysin (5), while in our bovine samples, enterohemolysin was produced within 20 out of 36 SLTEC (55.6%), and this ability was mainly true for SLT I producing E. coli (17/24). None of the SLTEC showed a-hemolytic activity. This is in concern with the properties of bovine ETEC, which show a negative correlation between pathogenicity and a-hemolysis (3). None of the 36'SLTEC produced both kinds of SLT simultaneously. These results agree with those of Marques et al. (22) and Smith et al. ( 34) who also could not find simultaneous SLT I and SLT II production. In contrast, Montenegro et al. (26) found this ability in 26/57 SLTEC from cattle (45.6%).This difference may be due to the different age of animals, for we examined calves, while those authors adult animals. A great diversity of bovine SL TEC O-groups was found. The 36 SL TEC yielded 17 different O-groups (Table 3 and 5). These data agree with results from Mohammed et al. (24), concerning cattle, and Karmali (17), concerning man. Six strains (4xOll1, 1 X 026,1 X 0128) beared O-groups that are grouped as classical human EPEC strains (20, 31, 32), 0128 is grouped as classical ETEC, additionally (13). All six strains produced SLT I. The occurrence of those O-groups in calves strongly supports the theory, that cattle are a SLTEC reservoir for human infections. However, O-group 0157 was not detected. This is in contrast to other reports, where 0157: H7 or 0157: H- were isolated frequently from cattle (6, 7, 21). However, these authors concentrated on 0157: H7, thus the results give a misleading impression in regard to the overall occurrence in cattle. In summary, SLTEC I seem to be involved in calf neonatal diarrhoea. The mechanism of pathogenesis needs further investigations. The occurrence of SLTEC with 0groups of classical human EPEC supports the hypothesis of cattle as a reservoir for human SLTEC infections.

Acknowledgements. The authors gratefully acknowledge the technical assistance of Miss

E. Hirsch. We owe special thanks to Miss E. Grady for her assistance in the preparation of

this manuscript. References

1. Appel, G., C. Ewald, A. Heer, G. v. Mickwitz, S. Aleksic, H. Riissmann, T. Meyer und

H. Karch: Vorkommen und Bedeutung von Verotoxin-(Shiga-like-Toxin-) produzierenden Escherichia coli-Stammen beim Schwein. Tierarztl. Umschau 44 (1989) 410-420 2. Baljer, G. und P. Bachmann: Nachweis enteropathogener Escherichia coli-Stamme und Rotaviren in Kotproben von Kalbern mit Diarrhoe. Zbl. Vet. Med. B 27 (1980) 608-615 3. Baljer, G., W. Eichhorn, G. Gobel, M. Wolf und P. Bachmann: Vorkommen und Verbreitung wichtiger Durchfallerreger bei neugeborenen Kalbern in Siiddeutschland im Zeitraum 1984 bis 1986. Tierarztl. Umschau 42 (1987) 56-62 4. Baljer, G., L. Wieler, R. Bauerfeind, S.-B. Ludwig und A. Mayr: Nachweis von Vero(Shiga-like-)toxinbildenden E. coli-Keimen (VTEC) mittels Zellkulturtest und DNAHybridisierung bei durchfallkranken Kalbern. Tierarztl. Umschau 45 (1990) 71-78 5. Beutin, L., M. Montenegro, I. Orskov, F. Orskov, J. Prada, S. Zimmermann, and R. Stephan: Close association of Verotoxin (Shiga-like Toxin) production with Enterohemolysin production in strains of Escherichia coli. J. Clin. Microbiol. 27 (1989) 2559-2564

1. H. Wieler, R. Bauerfeind, and G. Baljer

252

6. Borczyk, A., M. Karmali, H. Liar, and L. Duncan: Bovine reservoir for verotoxinproducing Escherichia coli 0157: H7. Lancet i (1987) 98 7. Chapman, P., D. Wright, and P. Norman: Verotoxin-producing Escherichia coli infections in Sheffield: cattle as a possible source. Epidem. Infect. 102 (1989) 439-445

8. Clarke, R., S. McEwen, V. Gannon, H. Liar, and C. Gyles: Isolation of verocytotoxinproducing Escherichia coli from milk filters in south-western Ontario. Epidem. Infect. 102 (1989) 253-260

9. Dobrescu, L.: New biological effect of edema disease principle (Escherichia colineurotoxin) and its use as an in-vitro assay for this toxin. Am. J. Vet. Res. 44 (1983) 31-34

J. Dalrymple: .Quantitative microtiter J. CIin. Microbiol. 12 (1980) 361-366

10. Gentry, M. and toxin.

cytotoxicity assay for Shigella

11. Guinee, P., C. Agterberg, and W. Jansen: Escherichia coli 0 antigen typing by means of a mechanised microtechnique. Appl. Microbiol. 24 (1972) 127-131

12. Guinee, P. and W. Jansen: Detection of enterotoxigenicity and attachment factors in Escherichia coli strains from human, porcine and bovine origin; a comparative study. Zbl. Bakt. Hyg., I. Abt. Orig. A 243 (1979) 245-257

13. Guth, B., M. Silva, 1. Scaletsky, M. Toledo, and L. Trabulsi: Enterotoxin production, presence of colonization factor antigen I, and adherence to HeLa cells by Escherichia coli 0128 strains. Infect. Immun. 47 (1985) 338-340 14. Hall, G., D. Reynolds, N. Chanter, J. Morgan, K. Parsons, T. Debney, A. Bland, and J. Baridger: Dysentery caused by Escherichia coli (S102-9) in calves: natural and experimental disease. Vet. Path. 22 (1985) 156-163

15. Johnson, W., R. Pollard, H. Liar, S. Tyler, and K. Rozee: Differentiation of genes coding for Escherichia coli Verotoxin 2 and the Verotoxin associated with porcine

edema disease (VTe) by the polymerase chain reaction. J. CIin. Microbiol. 28 (1990) 2351-2353 16. Karch, H. and T. Meyer: Evaluation of oligonucleotide probes for identification of Shiga-like-toxin-producing Escherichia coli. J. Clin. Microbiol. 27 (1989) 1180-1186 17. Karmali, M.: Infection by verocytotoxin-producing Escherichia coli. Clin. Microbiol. Rev. 2 (1989) 15-38 18. Karmali, M., M. Petrie, C. Lim, R. Cheung, and G. Arbus: Sensitive method for detecting low numbers of Verotoxin-producing Escherichia coli in mixed cultures by the use of colony sweeps and Polymyxin extraction of Verotoxin. J. CIin. Microbiol. 22 (1985) 614-619 19. Konowalchuck, J., J. Speirs, and S. Stavric: Vero response to a cytotoxin from Escherichia coli. Infect. Immun. 18 (1977) 775-779 20. Levine, M., J.-G. XU, J. Kaper, H. Liar, V. Prado, B. Tall,]. Nataro, H. Karch, and K. Wachsmuth: A DNA probe to identify enterohemorrhagic Echerichia coli of 0157 : H7 and other serotypes that cause hemorrhagic colitis and hemolytic uremic syndrome. J. Infect. Dis. 156 (1987) 175-182

21. Martin, M., L. Shipman, J. Wells, M. Potter, K. Hedberg, I. Wachsmuth, R. Tauxe, J. Davis, J. Arnoldi, and]. Tilleli: Isolation of Escherichia coli 0157: H7 from dairy cattle associated with two cases of haemolytic uraemic syndrome. Lancet ii (1986) 1043

22. Marques, L.,J. Peiris, and A. O'Brien: Escherichia coli strains isolated from pigs with

edema disease produce a variant of Shiga-like toxin II. FEMS Microbiol. Lett. 44 (1987) 33-38 23. March, S. and S. Ratnam: Sorbitol-MacConkey medium for detection of Escherichia coli 0157: H7 associated with hemorrhagic colitis. J. CIin. Microbiol. 23 (1986) 869-872 24. Mohammed, iI.., J. Peiris, and E. Wijewnata: Serotypes of verotoxigenic Escherichia coli isolated from cattle and buffalo calf diarrhoea. FEMS Microbiol. Lett. 35 (1986) 261-265

Characterization of Shiga-like Toxin Producing Escherichia coli

253

25. Mohammad, A., J. Peiris, E. Wijewanta, S. Mahalingam, and G. Gunasekara: Role of verotoxigenic Escherichia coli in cattle and buffalo calf diarrhoea. FEMS Microbiol. Lett. 26 (1985) 281-283 26. Montenegro, M.,M. Bulte, T. Trumpf, S. Aleksic, G. Reuter, E. Bulling, and R. Helmuth: Detection and characterization of fecal verotoxin-producing Escherichia coli from healthy cattle. J. Clin. Microbiol. 28 (1990) 1417-1421 27. Morrison, D., D. Tyrrell, and L. Jewell: Colonic biopsy in verotoxin-induced hemorrhagic colitis and thrombotic thrombocytopenic purpura (TIP). Am. J. Clin. Path. 86 (1986) 108-112 28. Moseley, S., I. Huq, A. Alim, M. Samapour-Motalebi, and S. Falkow: Detection of enterotoxigenic Escherichia coli by DNA colony hybridization. J. Infect. Dis. 142 (1980) 892-898 29. Mundell, D., C. Anselmo, and R. Wishnow: Factors influencing heat-labile Escherichia coli enterotoxin activity. Infect. Immun. 14 (1976) 383-388 30. O'Brien, A. and R. Holmes: Shiga and Shiga-like toxins. Microbiol. Rev. 51 (1987) 206-220 31. 0rskov, F. and I. 0rskov: Special Escherichia coli serotypes isolated from enteropathies in domestic animals and man. Fortschr. Vet.-med. 29 (1979) 7-14 32. 0rskov, F., T. Whittam, A. Cravioto, and I. 0rskov: Clonal relationships among classic enteropathogenic Escherichia coli (EPEC) belonging to different 0 groups. J. Infect. Dis. 162 (1990) 76-81 33. Seriwatana, J., P. Echeverria, D. Taylor, L. Rasrinaul, J. Brown, J. Peiris, and C. Clayton: Type II heat-labile enterotoxin-producing Escherichia coli isolated from animals and humans. Inf. Immun. 56 (1988) 1158-1161 34. Smith, H., S. Scotland, G. Willshaw, I. McLaren, T. Cheasty, and B. Rowe:' Vero cytotoxin production and presence of VI genes in Escherichia coli strains of animal origin. J. Gen. Microbiol. 134 (1988) 829-834 35. Suthienkul, 0., J. E. Brown, J. Seriwatana, S. Tienthongdee, S. Sastravaha, and P. Echeverria: Shiga-like toxin-producing Escherichia coli in retail meats and cattle in Thailand. Appl. Environ. Microbiol. 56 (1990) 1135-1139 Dr. Lothar H. Wieler, Institut fur Hygiene und Infektionskrankheiten der Tiere, Frankfurter Str. 89-91, W-6300 GieBen