Cytotoxins in non-enterotoxigenic strains of Escherichia coli isolated from feces of diarrheic calves

Veterinary Microbiology, 15 (1987) 137-150 Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands

137

C y t o t o x i n s in N o n - E n t e r o t o x i g e n i c S t r a i n s of E s c h e r i c h i a coli Isolated from Feces of D i a r r h e i c Calves J. DE RYCKE 1, J.F. GUILLOT2 and R. BOIVIN 1

lInstitut National de la Recherche Agronomique, Station de Pathologie de la Reproduction, 37380 Nouzilly (France) 2Institut National de la Recherche Agronomique, Station de Pathologie Aviaire-Parasitologie, 37380 NouziUy (France) ( Accepted for publication 17 March 1987 )

ABSTRACT De Rycke, J., Guillot, J.F. and Boivin, R., 1987. Cytotoxins in non-enterotoxigenic strains of Escherichia coli isolated from feces of diarrheic calves. Vet. Microbiol., 15: 137-150. We have examined the cytotoxic responses produced in HeLa and Vero cell cultures by sonicates from 15 non-enterotoxigenic ( S T a - , L T - ) strains of E. coli, highly lethal for mice parenterally LD,~o< 3 X 107 CFU ), which had been isolated from feces of diarrheic calves. Three types of cytotoxic responses were observed. Type 1 ( five strains) consisted of enlargment, rounding and polynucleation of HeLa cells, an effect previously reported with cytotoxic necrotizing factor ( CNF ) in E. coli from infant and piglet enteritis. Type 2 toxicity (three strains and the control Vir strain $5) was also characterized by enlargement and polynucleation of HeLa cells, but in contrast to Type 2 effect, cells were elongated. Sonicates from the latter strains were lethal for chickens, producing the lesions previously described with Vir strains. Type 3 toxicity (two strains and the control VT strain H19), produced an extensive destruction of both Vero and HeLa cell cultures. Cytotoxic effects were completely abolished upon heating for 1 h at 60 ° C for Type 1 and 2 extracts and at 80 ° C for Type 3 extracts. Seroneutralization assays showed that cytotoxins of the same type were closely related antigenically. In addition, a slight cross-neutralization was observed between Type 1 (CNF) and Type 2 (Vir) toxins.

INTRODUCTION

Two main groups ofEscherichia coli have been shown to be virulent for newborn calves: enterotoxigenic strains ( E T E C ) and bacteremic strains ( W r a y and Morris, 1985). E T E C can be specifically characterized by their attachment factors ( K99, F41 ) and the production of a heat-stable enterotoxin (STa). Bacteremic strains belong to a limited number of serotypes and some of them possess either a plasmid Col V, which specifies an iron-sequestering mecha-

~78-1145/87/$03.50

© 1987 Elsevier Science Publishers B.V.

138

nism, or a plasmid Vir, which codes for the production of a lethal toxin and a surface antigen (Smith, 1974; Lopez-Alvarez and Gyles, 1980). In addition to enterotoxigenic and bacteremic strains, strains producing a verocytotoxin (VT), also named Shiga-like toxin, have been detected in feces of diarrheic calves (Kashiwazaki et al., 1981; Mohammad et al., 1985; Sherwood et al., 1985; Marques et al., 1986). When administered orally to calves, atypical VT strains of E. coli were shown to elicit mucoid enteritis in two separate experiments (Chanter et al., 1986; Moxley and Francis, 1986). During two previous field investigations on calf enteritis, we have reported the significant occurrence, in the feces of diarrheic calves, of non-enterotoxigenic strains of E. coli highly lethal for mice intraperitoneally ( LD~o < 3 X 107 CFU ) ( de Rycke et al., 1982, 1986 ). Lethality for mice is known to be exhibited by E. coli strains with bacteremic or toxemic potentialities, such as septicemic strains (Jacks and Glantz, 1967), Vir strains (Smith, 1974), VT strains responsible for edema disease in swine (Schimmelpfennig and Weber, 1979) and hemolytic strains from different sources (Em5dy et al., 1980). The present investigation was carried out to detect the possible toxic properties of these lethal strains of E. coli, by means of cytotoxicity assays in HeLa and Vero cell cultures. In addition to VT toxins, we detected two other types of toxins, namely Vir toxins, whose cytotoxic effect has not been described previously and CNF toxins, first observed in E. coli from infant enteritis by Caprioli et al. (1983) and recently reported in E. coli from piglet enteritis ( Gonz~iles and Blanco, 1985). MATERIALS AND METHODS

Origin of E. coli strains E. coil isolates were collected during two field investigations on calf diarrhea (De Rycke et al., 1982, 1986). Both studies included unrelated cases of diarrhea in calves < 21 days of age. A total of 58 diarrheic and 45 healthy calves were sampled. In the first investigation, 5-10 E. coli isolates were collected per fecal sample and typed according to carbohydrate fermentation profile and resistance to antimicrobial agents. All the dominant types within each fecal sample were subsequently tested for lethality in mice. In the second investigation, 13 isolates of E. coli were collected per fecal sample and typed by means of multiple replication on solid agar plates, according to the method of Hinton et al. (1982). All the different types in each fecal sample were subsequently tested for lethality in mice.

139

Screening for strains highly lethal for mice parenterally For each strain of E. coli, 0.3 ml of a 1:50 dilution of an 18-h broth culture in trypticase soy broth (TSB; Biom~rieux, Charbonni~res-les-Bains, France) was inoculated intraperitoneally into three Swiss female mice OF1 ( Iffa-Credo, St. Germain l'Arbresle, France), 5-6 weeks old. This inoculum contained 6 X 108-107 CFU. Any strain that killed at least one mouse after a period of observation of 7 days was labeled lethal. The lethal dose 50 (LDso) was determined in the above model of infection for a random selection of lethal strains. Four serial dilutions of the broth culture separated by a dilution factor of 0.6 were used, with six mice per dilution. LD~o values were calculated according to the method of Reed and Muench (1938). Three control reference strains, lethal for mice according to the above definition, were also examined. Strain V2019 (O78:K80), isolated from a septicemic calf, was provided by L. Renault (Sanders, Athis-Mons, France) ; Vir strain $5 (O15:K?:H21) from a septicemic lamb and VT strain H19 (O26:K?:H?) from infant diarrhea, were obtained from H.W. Smith (Houghton Poultry Station, Huntingdon, U.K.).

Preliminary characterization of the lethal strains of E. coli Hemolysin production was detected after growth on trypticase soy agar (TSA; Biom~rieux) containing 5% of washed sheep erythrocytes. STa toxin was detected in the infant mouse assay (Dean et al., 1972) and LT and VT toxins in the Vero cell culture assay (Konowalchuk et al., 1977). Serotyping was carried out by I. and F. Orskov (International Escherichia and Klebsiella Reference Center, Copenhagen, Denmark).

Toxicity of bacterial extracts for HeLa and Vero cell cultures Production of bacterial extracts Cultures were grown at 37 °C for 24 h in 250-ml Erlenmeyer flasks containing 20 ml of TSB, on a giratory apparatus (160 RPM ). Cells were concentrated by centrifugation (10 000xg, 30 min), resuspended in 5 ml phosphate-buffered saline (PBS; pH 7.2) and disrupted by sonication at 4°C (MSE ultrasonic disintegrator, Crawley, U.K. ). The lysate was centrifuged during 2 h at 10 000 g ( Beckman ultracentrifuge) and the clarified supernatant was filtered through a 0.2-~m membrane (Sartorius, Gottingen, G.F.R.) then stored at 70 ° C. The total protein content of the extracts was determined by the method of Bradford (1976). -

140

Cell culture assays H e L a cells, strain CCL2 (Gibco, Courbevoie, France) and Vero cells (BioM~rieux) were cultured in Eagle's M i n i m u m Essential Medium (Eurobio, Paris) plus 2 m M L-glutamine, 10% fetal bovine serum, 100 units of penicillin G m1-1 and 1000 zg of streptomycine m l - L Cytoxicity assays were performed into 96-well-microtiter plates (Falcon 3072, Becton Dickinson, Oxnard, U.S.A.). After trypsinisation, a volume of 180 zl of the appropriate cell suspension was first distributed into microtiter wells, immediately followed by 20 zl of bacterial extracts or their dilutions in PBS. Plates were incubated for 72 h at 38°C in an atmosphere containing 5% CO2. Two types of toxic effects were detected and quantitated: cell viability and morphological alterations. Cell viability testing was performed with 5 X 104 V e r o cells and 2 X 105 H e L a cells, respectively, per milliliter of growth medium, according to the method of Gentry and Dalrymple (1980). In this method, viable cells are stained with crystal violet and stain uptake is measured spectrophotometrically after elution. Serial 2-fold dilutions of the extracts were first tested, then 10-fold dilutions with extracts producing an extensive killing of the cell cultures. Morphological alterations were detected after staining with Giemsa stain (Biolyon, Dardilly, France). Preliminary assays showed that alterations were more pronounced in H e L a than in vero cell cultures. Therefore, all the assays were performed in H e L a cell cultures, with an initial concentration of 4 X 104 cells m l - 1 of growth medium. Titration was performed with serial 2-fold dilutions. For each specific effect observed, the definition of cytotoxic titer is given in the Results section. To determine the thermolability of the toxic products, aliquots of bacterial extracts were heated for 1 h in a water bath at 40, 50, 60, 70 and 80 ° C. Seroneutralization assays Seroneutralization assays were carried out with antisera prepared in rabbits immunized with bacterial extracts representative of each type of toxicity. For the first immunization, 0.5 ml of the bacterial extracts (average protein concentration: 2.5 mg m1-1) were emulsified with an equal volume of Freund's complete adjuvant (Difco) and injected intradermally in the back of rabbits at 20 different sites. Subsequent immunizations were performed every 3 weeks in the same manner with incomplete Freund's adjuvant (Difco). At least two sera were produced per toxic extract. Neutralization tests were performed in microtiter plates by mixing 50/~l of the dilutions of serum in P B S with 50 pl of the appropriate concentrations of toxic extract. The plates were then incubated at 37 ° C for 1 h and placed at 4 ° C overnight. The toxicity of the mixture was then tested in cell cultures as described above. The neutralization titer was defined as the highest dilution of serum that completely inhibited the cytotoxic effect.

141

TABLE I Frequencyof E. colistrains lethal for mice parenterally in the fecesof diarrheic and healthy calves Calves

No. of calves No. of strains tested No. of strains lethal for mice No. of calveswith lethal strains

Diarrheic

Healthy

58 154 14 12

45 113 1 1

Lethality of bacterial extracts for chickens Toxicity for chickens was used for the detection of Vir strains of E. coli as recommended by Smith (1974) and by Lopez-Alvarez and Gyles (1980). Bacterial extracts, prepared as described above (average protein concentration: 2.5 mg m l - 1), were administered IV in a volume of 0.2 ml, to specific pathogenfree (SPF) Leghorn chickens, line PAl2 ( INRA, Station de Pathologie Aviaire et Parasitologie), 18-20 days old, raised in microisolators. Lethality was recorded over a period of 7 days and the lesions of dead birds were examined. RESULTS

Origin and preliminary characterization of E. coli strains lethal for mice parenterally A total of 15 strains of E. coli lethal for mice were detected during the two field investigations, of which 14 originated from 12 diarrheic calves and one from a healthy calf ( Table I). The estimated prevalence rates of these lethal strains were thus 21% in diarrheic calves and 2% in healthy calves. The main characteristics of these 15 lethal strains are given in Table II. Five strains were hemolytic. No strain produced the enterotoxins STa or LT and two (3AM1 and VA) produced a VT toxin, as detected in TSB culture supernatants. The LD~o for mice inoculated intraperitoneally was determined for some lethal strains. It ranged between 3 X 106 and 3 X 107 CFU.

Toxicity of bacterial extracts for HeLa and Vero cell cultures The toxicity of the 15 selected lethal isolates, of the three reference lethal strains and of a random selection of 20 non-lethal strains was assessed by the type of morphological alteration produced and the effect on cell viability. A

142 TABLE II Characteristics of E. coli strains lethal for mice parenterally Strains a

Clinicalb origin

Serotype

Hemolysin

VT toxin

log (LDso)

BM2-1 BM3-1 BM4-1 9AM-1 DM2-2

D D D D D

02:K53,93:H1 O8:K87:H08:K87:H O8:K87:HOll:H-

+ + + + +

-

6.5 ND ¢ ND 7.3 6.9

3AM-1 VA

D

O5:K?:H-

-

+

D

0127:H-

-

+

6.9 7.0

LS18-2 DM2-10 13AM-1 BM2-10 14AM-1

H D D D D

02:K?:H7 Oll:K15:H15 045:H19 O88:H25 O117:K98:H4

-

-

ND ND 7.4 ND 7.4

LMll-2 B9S2 LM1-6

D D D

O117:K98:H4 O168.'K?:H8 0?:H25

-

-

ND 6.5 ND

No strain produced STa or LT enterotoxin. aStrains BM2-1 and BM2-10 were isolated from the same calf, as well as strains DM2-2 and DM210. Strain 3AM1, referenced under the no. 125.80 in the Pasteur Institute collection, is anaerogenic and urease positive. bD: diarrheic calf; H: healthy calf. eND = not done. m a j o r t o x i c effect w a s d e t e c t e d i n 10 s e l e c t e d s t r a i n s a n d i n t w o r e f e r e n c e s t r a i n s (Table III).

Morphological alterations M o r p h o l o g i c a l a l t e r a t i o n s a r e s h o w n i n Fig. 1. A s i d e f r o m t h e t y p i c a l e f f e c t o f V T t o x i n s ( T y p e 3 ), w h i c h m a i n l y a f f e c t t h e v i a b i l i t y o f Vero a n d H e L a cells, t w o o t h e r d i s t i n c t t y p e s o f m o r p h o l o g i c a l a l t e r a t i o n s w e r e i d e n t i f i e d i n H e L a cell c u l t u r e s , n a m e l y T y p e 1 a n d T y p e 2. T y p e 1 a l t e r a t i o n w a s c h a r a c t e r i z e d b y t h e f o r m a t i o n o f e n l a r g e d ( ~ 25% i n l e n g t h a n d 4 0 0 % i n w i d t h ) , r o u n d e d , p o l y n u c l e a t e d cells. T h e o u t l i n e o f t h e cells a p p e a r e d f a i n t c o m p a r e d t o c o n t r o l cells. T h e h i g h e s t c o n c e n t r a t i o n s of t o x i c e x t r a c t s p r o d u c e d a l m o s t 100% p o l y n u c l e a t e d cells ( F i g . 2) a n d w i t h 2 CDso ~ 70% o f m u l t i n u c l e a t e d c e l l s c o n t a i n e d m o r e t h a n t w o n u c l e i ( F i g . 3 ) . T h i s t y p e o f m o r p h o l o g i c a l a l t e r a t i o n w a s d i s p l a y e d b y t h e five h e m o l y t i c strains. E x t r a c t from s t r a i n B M 2 - 1 was repeatedly the m o s t toxic, produci n g ~ 35 × 10 e CDso m g - 1 p r o t e i n . T y p e 1 c y t o t o x i c e f f e c t w a s c o m p l e t e l y a b o l -

143 T A B L E III Titration of toxic responses produced in HeLa and Vero cell cultures by cell-free sonicates of E.

coli Strains

Morphological alterations in HeLa cell cultures

Loss of viability: titers mg 1 protein"

Type a

Number of CD.~ob mg 1 protein

HeLa cells

BM2-1 BM3-1 BM4-1 DM2-2 9AM-1

1 1 1 1 1

35 × 11 × 7× 9× 6×

102 102 102 10 ~ 102

30 <20 20 < 30 < 20

30 <20 40 < 30 < 20

S5 (Vir) d B9S2 LM1-6 BM2-10

2 2 2 2

6× 5× 3× 3×

102 102 102 102

20 20 20 < 20

80 80 320 < 20

H19 (VT) d 3AM-1 VA

3 3 3

ND" ND ND

2 X 10 4 2 × 10 4 1 × 10 ~

2 × 10 ~ 2 X 10 ~ 1 × 10 ~'

Vero cells

~Type of toxicity refers to Fig. 1. bCD~o: lower dose of sonicate producing at least 50% of multinucleated cells (2-fold dilutions). CCell viability titers: for Type 1 and Type 2, lower dose of sonicate producing a significant loss of viability as compared to non*treated cells (2-fold dilutions); for Type 3, lower dose of sonicate producing at least 50% of loss of viability (10-fold dilutions). dControl strains. eND: not determined.

ished upon heating for 1 h at 60 ° C. It was also detectable in undiluted T S B culture supernatants, but only 20-50% of the cells were multinucleated. Type 2 alteration consisted of the formation of enlarged ( ~ 50% in length and 100% in width) and elongated cells, with clearly defined outlines. Polynucleation of H e L a cells was also a characteristic feature of this effect, but it differed from Type 1 polynucleation in three ways: (1) it affected at the most 70-80% of the cells (Fig. 2), (2) only ~ 30% of the multinucleated cells had more than two nuclei when treated with 2 CD~o (Fig. 3 ) and (3) a significant proportion of giant cells had a single nucleus. Three lethal strains and the reference strain $5 Vir produced this effect. Extracts contained 300-600 CDso m g - 1 protein and the toxic effect was completely abolished upon heating for 1 h at 60 ° C. We did not detect this toxic activity in T S B culture supernatants of the same strains.

144

Fig. 1. Morphological aspects of HeLa cell cultures after 72 h of incubation with extracts from strains V2019 (control, non-toxic), BM2-1 (Type 1 toxicity), $5 Vir (Type 2) and H19 VT (Type 3 ). Giemsa staining. Magnification X 10 (left) and X 20 (right). Bar = 50 pm.

145 I00

0 BM2-1

2 ¢9

/

50 4.J E

~9

0

~'/~!/] -3

-2

-I

I

I

I

I

0

1

2

4

Log 2 (CD50) Fig. 2. Percentage of multinucleated cells (at least two nuclei) in relation to number of CD.~o in H e L a cell cultures treated with extracts from strain B M 2 - 1 (Type I) and from strain $5 Vir (Type 2 ). Each count was performed on at least I00 multinucleated cells;9 5 % confidence intervals are shown.

Cell viability The two VT strains and the reference VT strain H19 produced an extensive degeneration of cell cultures, up to 2 X l0 s CD~o in Vero cell cultures and 2 × 104 CD~o in H e L a cell cultures, respectively, per milligram of protein. This effect was entirely destroyed upon heating at 80 °C for 1 h. A much lower degenerative effect was observed in two strains of Type 1 and in three strains of Type 100

-

[]

BM2-1

C3 0 50,.....,

OJ

o _ 2

a

4

>s

Number of nuclei per cell Fig. 3. Distribution of mean number of nuclei per cell in multinucleated HeLa cells treated with extracts from strain BM2-1 (Type 1 ) and from strain $5 Vir ÷ (Type 2). Counts were performed on 289 and 242 multinucleated cells, respectively; 95% confidence intervals are shown.

146 T A B L E IV Titers of seroneutralization in cytotoxicity assays Toxic extracts

Antisera against extracts BM2-1 ( Type 1 )

$5 Vir ( Type 2 )

H19 V T (Type 3 )

Type 1 strains a BM2-1 BM3-1 BM4-1 DM2-2 9AM-1

512 256 256 128 128

16 16 32 16 16

<2 <2 <2 <2 <2

Type 2 strains ~ S5 B9S2 LM1-6 BM2-10

4 4 8 8

256 256 1024 1024

<2 <2 <2 <2

Type 3 strains b H 19 3AM- 1 VA

<2 <2 <2

<2 <2 <2

2048 2048 2048

Titer: highest dilution of a n t i s e r u m t h a t completely i n h i b i t e d the toxic response. aNeutralization of the specific alterations produced in H e L a cell cultures by 4 CDso. bNeutralization of the loss of viability produced in Vero cell cultures by 10 CD.~o.

2. A similar reduced effect was also detectable in one lethal strain which displayed no other major cytotoxicity and in five of the 20 non-lethal strains (not shown). This last effect was not destroyed at 80 ° C.

Neutralization assays The results of neutralization assays are shown in Table IV. The results obtained after the fifth immunization with the most potent serum of each group are shown. Pre-immune sera did not neutralize cytotoxic effects. Within each group of toxicity, the specific antiserum neutralized all the extracts at similar titers, within the limit of two 2-fold dilutions. A partial cross-neutralization was also observed between Type 1 and 2 toxins. For both Type 1 and 2 extracts, the cross-neutralization was ~ 16-32 times less effective t h a n neutralization with the homologous antiserum.

Lethality of toxic extracts for chickens The lethality of toxic extracts for chickens is shown in Table V. Extract from strain $5 (Vir reference strain) was lethal and the toxic effect

147 TABLE V Lethality of E. coli extracts for chickens inoculated intravenously E. coli extracts

Number of chickens Inoculated

Dead

Type 2 strains $5 Vir $5 heated (60°C, I h) B9S2 BM2-10 LM1-6

5 5 5 5 5

5 0 5 4 5

Other strains BM2-1 (Type 1) BM3-1 (Type 1) H19 VT (Type 3) V2019 (non-cytotoxic)

5 4 5 5

3 4 0 0

was destroyed upon heating for 1 h at 60 ° C. The three Type 2 extracts were also lethal, together with the two extracts of Type 1 that we tested. With both types of extracts, death occurred within 12 h after inoculation and large amounts of serous fluid were observed in the body cavities, associated with an intense congestion of the liver. Conversely, no lethality or sign of toxicity were observed with extracts from strains V2019 and H19 VT.

DISCUSSION

From the data presented here, we can estimate at 14% the prevalence of diarrheic calves < 21 days of age harbouring strains of E. coli with cytotoxic properties. We were able to characterize three distinct types of cytotoxins in 10 of the 15 strains highly lethal for mice parenterally that we have examined. Type 1 cytopathic effect appears similar to the response elicited by the cytotoxic necrotizing factor ( C N F ) , a thermolabile toxic product originally described in E. coli strains from infant enteritis (Caprioli et al., 1983) and subsequently reported in E. coli from pig enteritis ( Gonz~ilesand Blanco, 1985). To our knowledge, this type of E. coli has not previously been reported in calves. In the present study, C N F strains were detected in five diarrheic calves out of 58 (9%) and in no healthy calves, which suggests that they may have an epidemiologic importance. Seroneutralization studies showed that they were closely related antigenically. All the strains producing the C N F toxin were hemolytic, both in the present study and in the studies mentioned above, but

148

Caprioli et al. (1983) demonstrated, by partially purifying the CNF toxin, that two different proteins accounted for the hemolytic and the toxic activity. Although Type 2 cytopathic effect shared common features with Type 1 (thermolability, polynucleation of cells), it was clearly distinct from Type 1, both morphologically and on the basis of seroneutralization studies. The production of Type 2 cytoxicity by the reference Vir strain $5 suggested that the Vir toxin itsef, or a closely associated toxin, accounted for the cytopathic effect. This hypothesis was strengthened by our observation that sonicates from all Type 2 strains were lethal for the chicken inoculated intravenously, a test originally proposed for the detection of E. coli strains producing the Vir toxin ( Smith, 1974; Lopez-Alvarez and Gyles, 1980 ). Yet, Type 1 sonicates were also lethal for chicken, which indicates that the chicken assay is not specific for Vir strains. In this regard, the cytotoxicity assay can be proposed as a more specific detection test. Since the preliminary publication of Smith (1974), there has been, to our knowledge, only one published report on the occurrence of Vir strains in calves (Lopez-Alvarez and Gyles, 1980). The above authors, by using the chicken test, detected two Vir strains out of 18 strains from calf Septicemia. In the present study, the overall isolation rate of Vir strains in feces was ~ 5% (3 out of 58) in diarrheic calves versus 0% in healthy calves. The partial cross-neutralization that was observed between Type 1 and 2 toxins may be attributed to some common antigenic determinants. This hypothesis is consistent with the fact that both toxins produced multinucleation in cell cultures and were toxic for chickens. A partial homology between these two types of toxins should be further investigated with purified toxins or monoclonal antibodies. Such methods would also rule out the possible nonspecific reactions due to the presence, in a strain producing predominantly one toxin, of trace levels of the other toxin. Type 3 toxicity is typical of VT strains of E. coll. Such strains have previously been isolated from calf diarrhea (Kashiwazaki et al., 1981; M o h a m m a d et al., 1985; Sherwood et al., 1985; Marques et al., 1986). In the present study, VT toxins from bovine strains were of the same antigenic type as the VT toxin from the h u m a n reference strain H19. Such an antigenic similarity was also observed by Sherwood et al. (1985). The moderate degenerative cytotoxic effect observed in some other strains cannot be attributed to low levels of VT toxin, since it was resistant to heating at 80 ° C. This minor toxicity, which might be due to residual endotoxin, deserves further investigation. The remaining lethal strains, in particular the reference bacteremic strain V2019, had no major toxic activity for cell cultures. Two of these lethal nontoxic strains belonged to the serogroup Ol17 and one to the serogroup 02. These serogroups have been reported in strains from calf septicemia ((~rskov and (~rskov, 1979). The virulence of these strains for mice is possibly related to their ability to multiply in vivo.

149

In conclusion, the significant association of CNF, Vir, and VT strains of E. coli with feces of diarrheic calves suggests that these strains may be involved in diarrhea. Experimental data supporting an enteropathogenic role are available for VT strains only. Human VT strains from neonatal diarrhea, such as strain H19 which was used in the present study, were able to cause fluid accumulation in rabbit ligated intestinal loops (Smith et al., 1983). In addition, calf atypical VT strains were shown to be enteropathogenic in experimentally infected gnotobiotic (Moxley and Francis, 1986) and conventional calves (Chanter et al., 1986). These strains attached to the epithelium of the large intestine and caused a degeneration of enterocytes, leading to dysentery. Our strain 3AM1 (CIP125-80) has the same characteristics as the strains used by the above authors, (serogroup 05, anaerogenic, urease and VT producer) and, therefore, most likely belongs to the same enteropathogenic group. As for Vir and CNF strains, there is not yet published experimental evidence for their direct involvment in enteritis. The lethal properties exerted by these two types of toxic strains in the mice model suggest that they may cause toxemia, as do toxigenic strains of E. coli responsible for edema disease o f swine (Schimmelpfennig and Weber, 1979). ACKNOWLEDGMENTS

We are grateful to I. and F. (~rskov for the serotyping of E. coli strains. We also wish to thank M. Duchet-Suchaux, T.T. Kramer and P. Pardon for their critical reading of the manuscript and S. Benoist for her technical help.

REFERENCES Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem., 72: 248-254. Caprioli, A., Falbo, V., Roda, L.G., Ruggeri, F.M. and Zona, C., 1983. Partial purification and characterization of an Escherichia coli toxic factor that induces morphological cell alterations. Infect. Immun., 39: 1300-1306. Chanter, N., Hall, G.A., Bland, A.P., Hayle, A.J. and Parsons, K.R., 1986. Dysentery in calves by an atypical strain of Escherichia coli (S102-9). Vet. Microbiol., 12: 241-253. Dean, A.G., Ching, V.C., Williams, R.G. and Harden, L.B., 1972. Test for Escherichia coli enterotoxin using infant mice: application in a study of diarrhea in children in Honolulu. J. Infect. Dis., 125: 407=411. De Rycke, J., Boivin, R., Le Roux, Ph., 1982. Mise en ~vidence de souches d'Escherichia coli caract~re septic~mique dans les feces de veaux atteints d'ent~rite muco'fde. Ann. Rech. V~t., 13: 385-397. De Rycke, J., Bernard, S., Laporte, J., Naciri, M., Popoff, M.R. and Rodolakis, A., 1986. Prevalence of various enteropathogens in the feces of diarrheic and healthy calves. Ann. Rech. V~t., 17: 159-168. EmSdy, L., Palt, T., Safonova, N.V., Kuch, B. and Golutva, N.K., 1980. Alpha-haemolysin: an additive virulence factor in Escherichia coll. Acta Microbiol. Acad. Sci. Hung., 27: 333-342.

150 Gentry, M.K. and Dalrymple, J.M., 1980. Quantitative microtiter cytotoxicity assay for ShigeUa toxin. J. Clin. Microbiol., 12: 361-366. Gonz~iles, E.A. and Blanco, J., 1985. Production of cytotoxin VT in enteropathogenic and nonenteropathogenic Escherichia coil strains of porcine origin. FEMS Microbiol. Lett, 26: 127-130. Hinton, M., Allen, V. and Linton, A.H., 1982. The biotyping of Escherichia coli isolated from healthy farm animals. J. Hyg., 88: 543-555. Jacks, T.M. and Glantz, P.J., 1967. Virulence of Escherichia coli serotypes for mice. J. Bacteriol., 93: 991-995. Kashiwazaki, M., Ogawa, T., Nakamura, K., Isayama, Y., Tamura, K. and Sakazaki, R., 1981. Verocytotoxin produced by Escherichia coli strains of animal origin. Natl. Inst. Anim. Health Q., 21: 68-72. Konowalchuk, J., Speirs, J.I. and Stavric, S., 1977. Vero response to a cytotoxin of Escherichia coll. Infect. Immun., 18: 775-779. Lopez-Alvarez, J. and Gyles, C.L., 1980. Occurrence of the Vir plasmid among animal and human strains of invasive Escherichia coll. Am. J. Vet. Res., 41: 769-774. Marques, L.R.M., Moore, M.A., Wells, J.G., Wachsmuth, I.K. and O'Brien, A. D., 1986. Production of Shiga-like toxin by Escherichia coll. J. Infect. Dis., 154: 338-341. Mohammad, A., Peiris, J.S.M., Wijewanta, E.A. Mahalingam, S. and Gunasekara, G., 1985. Role of verocytotoxinogenic Escherichia coli in cattle and Buffalo calf diarrhea. FEMS Microbiol. Lett., 26: 281-283. Moxley, R.A. and Francis, D.H., 1986. Natural and experimental infection with an attaching strain of Escherichia coli in calves. Infect. Immun., 53: 339-346. (~rskov, F. and (~rskov, I., 1979. Special Escherichia coli serotypes from enteropathies in domestic animals and man. In: H. Willinger and A. Weber (Editors), E. coli Infections in Domestic Animals. Verlag Paul Parey, Berlin, pp. 7-14. Reed, L.J. and Muench, H., 1938. A simple method of estimating fifty per cent endpoints. Am. J. Hyg., 27: 493-497. Schimmelpfennig, H. and Weber, R., 1979. Studies on the oedema disease producing toxin of Escherichia coli. In: H. Willinger and A. Weber (Editors), E. coli Infections in Domestic Animals. Verlag Paul Parey, Berlin, pp. 25-32. Sherwood, D., Snodgrass, D.R. and O'Brien, A.D., 1985. Shiga-like toxin production from Escherichia coli associated with calf diarrhea. Vet. Rec., 116: 217. Smith, H.W., 1974. A search for transmissible pathogenic characters in invasive strains of Escherichia coli: the discovery of a plasmid-mediated toxin and a plasmid-controlled lethal character closely associated, or identical, with colicine V.J. Gen. Microbiol., 83: 95-111. Smith, H.W., Green, P. and Parsell, Z., 1983. Vero cell toxins in Escherichia coli and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chickens and pigs. J. Gen. Microbiol., 129: 3121-3137. Wray, C. and Morris, J.A., 1985. Aspects of colibacillosis in farm animals. J. Hyg., 95: 577-593.