J. Med. Microbiol. - Vol. 47 (1998), 781-790 ((, 1998 The Pathological Society of Great Britain and Ireland

BACTERIAL PATHOGENICITY AND CHARACTERISATION

En t ero pat h og en i ci ty m a r kers i n Escrt erichia coli isolated from infants with acute diarrhoea and healthy controls in Rio de Janeiro, Brazil

A. C. P. ROSA, A. T. MARIANO, A. M. S. PEREIRA, A. TIBANA*, T. A. T. GOMEST and J. R . C. ANDRADE

Departamento de Microbiologia e Imunologia, Universidade do Estado do Rio de Janeiro, 20551-030, Rio de Janeiro, * lnstituto de Microbiologia, UniversidadeFederal do Rio de Janeiro, 2 194 1-590, Rio de Janeiro and t Departamen to de Microbiologia, lmunologia e Parasitologia, Universidade Federal de S8o Paulo, 04023-062, Sa’o Paulo, Brazil

Faeces from urban children < 2 years old with acute diarrhoeal illness and from non- diarrhoeal infants (controls) were examined for Escherichia coli and other enteropatho- gens. A total of 990 E. coli isolates from 100 patients and 50 controls was tested for enteropathogenic E. coli (EPEC) serotype (O:H), adherence to HEp-2 cells after incubation for 3 and 6 h, fluorescent actin staining (FAS), DNA hybridisation with EAF, eaeA, STh, STp and EAggEC probes and production of heat-labile enterotoxin (LT) and verocytotoxin (VT) with Y1 and Vero cells. EPEC were the most prevalent enteropathogens in patients (32.7%; and 14% in controls). Enteroinvasive E. coli (EIEC) and Vero cytotoxin-producing E. coli (VTEC) were not detected. The rate of isolation of enterotoxigenic E. coli (ETEC) was identical in both groups. Among the EPEC isolates the prevalent serotypes were Oll l :H2, 055:NM and 01 19:H6. Localised adherence (LA) was found significantly more frequently in isolates from patients (19.6%) than controls (2.1%). All LA-positive EPEC isolates were FASf and eaeAf, but only 75.2% of them hybridised with the EAF probe. Diffusely adhering E. coli (DAEC) and enteroaggregative E. coli (EAggEC) were found with equal frequency in patients and controls. Twenty-seven E. coli isolates were negative for EAF but positive for eaeA and FAS and produced LA in 6-h adherence tests. These EAF-/eaeA+ strains were the only putative enteropathogen identified in seven patients and were not found in controls. The ability of these strains to elicit ultrastructural cell alterations and cell-signalling events was evaluated in Caco-2 cells (human colon carcinoma cell line) by the gentamicin invasion assay and by transmission electron microscopy. The numbers of intracellular bacteria in cell invasion tests varied from 0.4% to 1.6% of the cell- associated bacteria after a 6-h incubation period. Tyrosine phosphorylation of host cell proteins was assessed in HEp-2 cells by immunofluorescence microscopy and all strains gave positive results. EAF-/eaeA+ E. coli strains express most of the virulence properties found among true EPEC strains and can be a relevant cause of infant diarrhoea in developing countries.

Introduction

Acute diarrhoea is a leading cause of mortality in developing countries. Among the diarrhoea-producing Escherichia coli, enteropathogenic E. coli (EPEC) are considered the main cause of epidemic outbreaks and

Received 23 Sept. 1997; revised version accepted 10 Jan. 1998. Corresponding author: Dr A. C. P. Rosa.

endemic cases of infantile diarrhoea [ 1-31. Illness caused by EPEC is often a clinically acute severe diarrhoea. On some occasions, persistent diarrhoea may develop [4]. Mannose-resistant adhesiveness to epithe- lial cells is considered to be a strong indication of enteropathogenicity for E. coli. Three patterns of adherence are recognised: localised, diffuse and ag- gregative [5 , 61. The localised adherence (LA) pattern is often exhibited by strains belonging to certain serotypes of EPEC frequently associated with acute diarrhoea in infants [3, 7-91. Diffbsely adhering E. coli

782 A. C. P. ROSA E T A L .

(DAEC) and enteroaggregative E. coli (EAggEC) strains are usually not related to EPEC serotypes and their role in acute diarrhoeal disease is controversial [6, 8-15]. In most EPEC strains, LA correlates with the presence of an EPEC adherence factor (EAF) plasmid [ 161. A specific DNA probe has been used for the identification of localised adhering E. coli strains [17, 181. Associated with the LA phenotype is the induction of the attaching and effacing (A/E) lesions characterised by intimate attachment of bacteria to the apical enterocyte membrane and localised destruction of brush border microvilli. Accumulation of poly- merised actin beneath bacteria results in the formation of cup-like pedestal structures [ 191. The fluorescent actin staining (FAS) test is a diagnostic assay that detects polymerised actin associated with the A/E lesions [20]. Actin re-arrangement is also involved in the internalisation of a subpopulation of EPEC by epithelial cells [2 1 -241. Signal transduction, including induction of tyrosine phosphorylation of a 90-kDa host cell protein (Hp 90) by the attached EPEC, is the initial event that leads to formation of A/E lesions [25]. The genes implicated in the formation of A/E lesions have been localised in a large (35-kb) region of the EPEC chromosome termed the locus of enterocyte effacement (LEE) [26]. Several loci within the LEE, including eaeA, espA, espB (formerly eaeB) and sep, have been characterised. A DNA probe derived from the chromo- somal gene eaeA encoding intimin can also be used to detect EPEC strains [27]. In the present study, virulence markers such as EPEC serotyping (O:H), adherence to HEp-2 cells, FAS test, DNA hybridisation with EAF, eaeA, Sth, Stp and EAggEC probes and production of heat-labile enterotoxin (LT) and verocy- totoxin (VT) were analysed in E. coli isolates from children < 2 years old with acute diarrhoea and healthy controls in Rio de Janeiro, Brazil.

Materials and methods

Bucteriul struins

E. coli isolates from diarrhoeal stool specimens from 150 urban infants (< 2 years old) attending the outpatient unit or admitted to the paediatric wards of Hospital Universitario Pedro Ernest0 in Rio de Janeiro from July 1990 to April 1993 were examined. Diarrhoea was defined as the occurrence of one or more liquid or watery stools in a 24-h period. E. coli isolates from 50 non-hospitalised healthy infants belonging to the same age group and urban areas as the patients were also examined. Patients and controls in whom E. coli was not detected in the stools were excluded from the study. E. coli strains were isolated on Eosin Methylene-blue Agar (Difco Laboratories). Five randomly selected lactose-fermenting colonies were examined by biochemical tests [28] and E. coli isolates were stored on nutrient agar slants at 15°C until further testing. Shigella spp., Salmonella spp., Aeromonas spp., Yersinia enterocolitica, Campylobac-

ter spp. and Vibrio cholerae were isolated and identified according to Lennette et al. [29]. Rotavirus was identified by a monoclonal latex agglutination test (Slidex Rota-Kit 2, bioMkrieux, Lyon, France). E. coli strains E2348/69 (0127:H6; EAF+, eaeA+), E40705 (0157:H7;VTl+), H1/1 (O?:H?; DAEC), 239 ( 0 1 11 :H?; EAggEC), 40 T (O?:H?; LTh+ - obtained from Dr A. F. P. Castro, Universidade Estadual de Campinas, UNICAMP, Brazil), 4871-1 (078:H12, STh+) and 4511-2 (027:H-, STp+) were used as positive control strains in cell adhesion and invasion tests, cytotoxicity assays and colony hybridisation assays. E. coli DH5a (K12) was used as the negative control in all tests.

E. coli serotyping

For the detection of EPEC serogroups, all E. coli colonies were tested for slide and tube agglutination with polyvalent and monovalent antisera (Difco; Probac, Siio Paulo, Brazil). The serogroups investigated were as follows: 018, 020, 026, 028, 044, 055, 086, 0111, 0112, 0114, 0119, 0125, 0126, 0127, 0129, 0142 and 0158. The actively motile cultures were tested by tube agglutination to determine H antigen with HI, H2, H4-H9, H11, H12, H15, H17, H18, H20, H21, H23, H25, H27, H29, H31-H35, H37, H40 and H41 antisera from the Centers for Disease Control, Atlanta, GA, USA. O:H serotyping of a few selected strains was done at the Laboratory of Enteric Patho- gens, Central Public Health Laboratory, London. Sorbitol-negative E. coli isolates were screened with 0157 and H7 antisera, raised in rabbits in our laboratory. Non-motile, lysine-negative E. coli isolates were tested with EIEC polyvalent antisera (Probac). The EIEC serogroups sought were as follows: 028ac, 029, 0112ac, 0124, 0136, 0143, 0144, 0152, 0164 and 0167. Agglutinating isolates were subjected to the Serbny test [30].

Cytotoxicity assays

Cytotoxicity assays were done essentially as described by Smith and Scotland [31]. Vero and Y-1 adrenal cells were grown in 96-well tissue culture plates (Costar, Cambridge, MA, USA) respectively in Eagle’s Mini- mum Essential Medium (MEM; Sigma) supplemented with fetal calf serum (FCS) 5% v/v and Ham F-10 medium (Sigma) supplemented with L-glutamine 0.5% w/v and FCS 10% v/v. E. coli isolates were inoculated into 10 ml of Tryptic Soy Broth (TSB; Difco) in a 250- ml Erlenmeyer flask and incubated with shaking at 200 rpm for 18 h. The cultures were centrifuged at 9000 g for 10 min. The supernates were filtered through 0.45-,urn membrane filters (Millipore, Bedford, MA,USA) and stored at 4°C. Sterile culture supernates were then tested in the Y-1 cell assay for LT [32] and in the Vero cell assay for VT [31].

VIRULENCE MARKERS IN ENTEROPATHOGENIC E. COLZ 783

HEp-2 adherence assay

All E. coli isolates from patients and controls were tested individually for adherence to HEp-2 cells by the method of Cravioto et al. [33]. The HEp-2 human epithelial cell line CCL 23 (American Type Culture Collection, MD, USA) was cultivated in MEM supplemented with FCS 5% v/v, gentamicin 50 pg/ml and amphotericin B 2.5 pg/ml. Subconfluent cell monolayers were obtained on 13-mm diameter glass coverslips placed in 24-well tissue culture plates (Costar), washed twice with Dulbecco’s phosphate- buffered saline (PBS-D) pH 7.2, and covered with 1 ml of fresh MEM without antibiotics and containing D- mannose 1%. Samples (35 pl) of each bacterial culture grown overnight in TSB (c . lo7 bacteria) were incubated with cell monolayers for 3 or 6 h at 37°C in C02 5%. In 6-h assays, cells were washed with PBS- D and fresh medium was added after 3 h. After two washes with PBS-D to remove non-adherent bacteria, cells were fixed with methanol and stained with Giemsa 5% stain for 30 min. The stained coverslips were removed from wells, washed in water, dried, mounted on glass slides and examined by oil immer- sion microscopy by two independent examiners.

FAS test

All E. coli isolates were examined by the fluorescent actin staining (FAS) test by the method described by Knutton et al. [20]. After 3-h and 6-h HEp-2 adherence assays the cells were washed three times with PBS-D and fixed for 15 min in formalin 3% v/v. Fixed and washed cells were permeabilised by treating coverslips with Triton X-100 0.1% in PBS for 4 min. After washing three times in PBS-D the cells were treated with a solution of fluorescein isothiocyanate-phalloidin (Sigma) 5 pg/ml in PBS-D for 30 min to specifically stain filamentous actin. Coverslips were washed three times and mounted in glycerol-PBS-D. Specimens were examined by alternating fluorescence and phase-con- trast microscopy with a Labophot microscope (Nikon, Tokyo, Japan) equipped with epifluorescence. The FAS test was positive when the foci of intense fluorescence corresponded to areas of bacterial adhesion observed under phase-contrast.

DNA probes and hybridisation analyses

All strains were spotted on to MacConkey Agar (Difco) and incubated at 37°C overnight. Colonies were then transferred to Whatman 54 1 paper (Whatman, Clifton, NJ, USA) and the filters were processed and hybridised as described by Maas [34]. The probe fragments were radiolabelled with [ G ~ - ~ ~ P ] ~ A T P by the random primer method. Results were revealed by autoradiography. The filters were hybridised with radiolabelled probes that detect the EAF (l-kb Barn Hl-SaZ 1 fragment of pJPN 16) [18], eaeA (1-kb Sal 1-Kpn 1 fragment of pCVD 434) [27], human (216-bp EcoR 1 fragment of pCVD 427) and porcine (157-bp Pst 1 fragment of pCVD

426) heat-stable enterotoxin (ST) of E. coli [35], and EAggEC (l-kb EcoRI-PstI fragment of pCVD 432) ~361.

Bacterial invasion

Caco-2 cells from a human colon carcinoma [37] were kindly provided by A. Zweibaum (Institut National de la Sante et de la Recherche Medicale, Unite 178, Villejuif, France). The cells were grown to post- confluence in 24-well plates containing Dulbecco’s Modified Eagle’s Medium (DMEM; Sigma) supple- mented with antibiotics, FCS 10% v/v, essential amino acids mixture (Sigma) 1% and L-glutamine 0.5% at 37°C in C02 5%. The cell monolayers were washed twice in PBS-D and covered with 1 ml of fresh medium containing D-mannose 1%. The cells were incubated at 37°C for 6 h with medium change at 3 h to minimise extracellular bacterial growth. For each bacterial isolate, half of the Caco-2 monolayers were washed twice with PBS-D and incubated for 1 h in the same medium containing gentamicin 250 pglml. Control wells were incubated with medium without antibiotics. After removal of the anti biotic-supplemented medium the cells were washed twice and lysed with trypsin (Trypsin 1:250, Difco) 0.1% and Tween 20 (Sigma) 1% for 30 min at 37°C. Samples of the cell lysates were diluted and plated on Tryptic Soy Agar (TSA; Difco) to quantify viable bacteria. The percentage invasion was calculated by dividing the number of surviving (intra- cellular) bacteria by the number of bacteria in the controls (cell-associated bacteria) [3 81. Experiments were run in duplicate and repeated independently at least three times with positive (EPEC E2348/69) and negative (DH5a K12) controls. All the strains were susceptible to killing by gentamicin as previously established by broth dilution techniques.

Detection of tyrosine phosphorylated proteins

EAF-/eaeA+ E. coli isolates were tested for tyrosine phosphorylation of eukaryotic proteins by immuno- fluorescent microscopy [39]. After 6-h HEp-2 adher- ence assays the coverslips were washed three times with PBS-D to remove non-adherent bacteria. The cells were then fixed in paraformaldehyde 2% for 15 min and washed. The cells were permeabilised by treatment with Triton X-100 0.1% in PBS-D for 4 min and then washed. Accumulation of tyrosine-phosphorylated pro- teins was detected by incubation of cells with a 1 in 2000 dilution of murine monoclonal antiphosphotyr- osine IgG (Sigma) as the primary antibody for 60 min. Following three washes in PBS-D, the cells were incubated with a 1 in 100 dilution of fluorescein isothiocyanate-conjugated anti-mouse rabbit IgG (Sig- ma) for 60 min. Following further washes and mounting in glycerol-PB S-D the coverslips were examined by alternating phase-contrast and fluores- cence microscopy in a Labophot microscope equipped with epifluorescence.

784 A. C. P. ROSA E T A L .

Transmission electron microscopy

Caco-2 cells were grown to post-confluence on the surface of a thin layer of polymerised type 1 collagen spread on coverslips placed in 24-well tissue culture plates. After incubation for 6 h, infected Caco-2 cells were washed with PBS-D and fixed in glutaraldehyde 2.5% in cacodylate-sucrose buffer (0.1 M cacodylate, 0 . 0 9 ~ sucrose, 0.01 M MgC12, 0.01 M CaC12, pH 7.2) for 2 h at 4°C. Cells were post-fixed in OsO4 1%, dehydrated through a series of graded ethanol solutions and embedded in Epon. Small blocks were cut and oriented before resin polymerisation at 60°C for 72 h. Ultra-thin sections were cut with a Reichert Ultracut S microtome, stained in aqueous solutions of uranyl acetate 5% and lead citrate 2.5% and examined in an EM 906 Zeiss transmission electron microscope at 80 Kv.

Sta tis tica I analysis

The x2 test was used to determine the statistical significance of the data. A p value of <0.05 was considered significant.

Results

A total of 990 E. coli isolates from stool cultures, comprising 740 from 150 patients and 250 from 50 controls, was assayed for adherence to HEp-2 cells, FAS test, DNA hybridisation with EAF, eaeA, STh, STp and EAggEC probes and production of verocyto- toxin (VT) and heat-labile enterotoxin (LT). The occurrence of enteropathogens in 1 S O infants with diarrhoea and 50 controls is shown in Table 1. EPEC were far more common in patients than controls (32.7% versus 14.0%; pCO.025). There were no significant differences in the rates of isolation of EAggEC, DAEC and ETEC between patients and controls. EIEC, VTEC and enterohaemorrhagic E. coli (EHEC) strains of the serotype 0157:H7 were not found in either group.

Table 1. lsolation of enteropathogens from the stools of 150 children < 2 years old with acute diarrhoea and 50 heal thy controls

Number (YO) of isolates for

Pathogen Patients Controls P*

EPEC 49 (32.7) 7 (14.0) < 0.025 EAggEC 18 (12.0) 12 (24.0) NS DAEC 15 (10.0) 2 (4.0) NS Rotavirus 14 (9.3) ND . . . ETEC 5 (3.3) 2 (4.0) NS Salmonella spp. 2 (1.3) ND . . . Salmonella typhimurium 1 (0.7) ND . . . Shigella sonnei 1 (0.7) ND . . .

EPEC, enteropathogenic E. coli; EAggEC, enteroaggregative E. coli; DAEC, diffusely adhering E. coli; ETEC, enterotoxigenic E. coli; ND, not determined; NS, not significant. "Determined by x 2 test; p < 0.05 was considered significant.

Patients

Among the 740 E. coli isolates from diarrhoea1 specimens, 158 gave positive agglutination with EPEC antisera. Table 2 presents the serotypes and the virulence markers of these isolates. Among the EPEC serotypes found, the most common serotypes were 0 1 11:H2 (five children), 055:NM (four) and 0 1 19:H6 (four). Of the EPEC isolates, 80 (50.6%) showed localised adherence. All isolates with LA gave a positive FAS assay and hybridised with the eaeA probe. Of these 80 LA-positive EPEC strains, 60 (75%) hybridised with the EAF probe. In this study, the serotypes that produced LA were 055:NM, 055:H40, 055:H?, 086:H34, 0 1 11:H2, 0 1 1 l:H?, 0 1 19:NM, 0 1 19:H2, 0 1 19:H6, 0 1 19:H? and 0242:H6. Twenty EAF-/eaeA+ E. coli isolates that belonged to serotypes 055:NM (three), 0 1 19:H2 (nine), 0 1 19:H? (two), 0 1 19:H6 (one) and 0142:H6 (five) were isolated from six patients (Table 2). Diffuse adherence was displayed by eight EPEC ( 5 .O%) isolates. Thirty-one EAggEC isolates were from 18 (12.00/0) patients and most of them belonged to EPEC serotypes (026:H27, 086:NM, 086:H?, 086:H11, 086:H18, 0 1 1 l:H?, 0 1 19:H17, 0125:Hll , 0125:H31, 0127:H? and 0128:H?) lacking the EPEC virulence factors. All EAggEC probe-positive isolates that belonged to EPEC serotypes showed the aggregative pattern of adherence in HEp-2 cells. However, EAggEC isolates of serotype 0125:Hll produced a typical aggregative phenotype but were negative with the EAggEC probe. On three occasions, EAggEC isolates were associated with rotavirus (one), EPEC (one) or Shigella sonnei (one). Among the 582 E. coli isolates that did not agglutinate with EPEC antisera, 23 (3.9%) showed LA, gave a positive FAS test and hybridised with the eaeA probe (Table 3). Of these, only 16 (69.5%) isolates hybridised with the EAF probe. All EAF-/eaeA+ EPEC and non- EPEC isolates were LA+ only after 6 h, although a positive FAS test had already been observed after incubation for 3 h. EAF-/eaeA+ isolates were the only putative enteropathogens identified in seven patients (Tables 2 and 3). In one patient, EAF- and EAF+ isolates of the same EPEC serotype were co- isolated. Diffuse adherence was observed in 48 (8.2%) non-EPEC isolates. Among the eight (1.3%) EAggEC probe-positive isolates, two showed the aggregative pattern of adherence in HEp-2 cells. However, four probe-positive strains attached to cells without any recognisable pattern of adherence and two strains caused complete detachment of cell monolayer even during 3-h adherence tests.

Controls

Of the 250 E. coli isolates from controls, 29 (11.6%) belonged to EPEC serotypes (Table 4). Five EPEC isolates (055:NM) showed LA and hybridised with the EAF and eaeA probes. EAF-/eaeA+ E. coli isolates were not found in the control group. Diffuse adherence

VIRULENCE MARKERS IN ENTEROPATHOGENIC E. COLZ 785

Table 2. Virulence markers in isolates of EPEC serogroups from 150 patients < 2 years old in Rio de Janeiro, Brazil

Number (%) of strains

HEp-2 adherence Assays

Isolates FAS EAF eaeA EAggEC Seroty pe (patients) LA DA AA A+ NA test probe probe probe

026:H27 0 5 5 : N M 055:H21 055:H40 055:H? 086:NM 086:H 1 1 O86:H 18 086:H34 086:H? 0 1 11:H2 Ol l l :H12 Ol l l :H21 01 1 l:H? 0 1 19:NM 0 1 19:H2 0 1 19:H6 01 19:H17 01 19:H'? 0124:H3 1 0125:NM 0125:H4 0125:H6 0125:Hll 0125:H31 0 125 :H? 0126:H8 0 1 26:H27 01 27:H? 01 28:H7 0 128:H40 0 128:H? 0 142:H6 0142:Hl I 0 142:H?

Total 80 (50.6) 8 (5.0) 23 (14.5) 21 (13.2) 27 (17.0) 80 (50.6) 60 (38.0) 80 (50.6) 18 (11.4)

LA, localised adherence; DA, diffuse adherence; AA, aggregative adherence; A+, adherence without a recognisable pattern; NA, non-adherent. *Total number of EPEC serogroup occurrence in 49 patients. The following associations were found in 11 patients: 055:H? and 055:NM (l), 055:H? and 055:H21 (l), O l l l : H ? and O l l l : H 2 (3), 0119:H? and 0119:H6 (2), 0119:H2 and 0125:H31 (l), 0119:H6 and 0125:H? (l), 0125:H? and 0123:Hll ( l ) , 0125:NM and 0125:H4 ( I ) .

Table 3. Virulence markers in non-EPEC isolates from 150 patients < 2 years old in Rio de Janeiro, Brazil

Number (%) of strains

HEp-2 adherence Assays Non- EPEC FAS EAF eaeA EAggEC isolates Patients LA DA AA A+ NA test probe probe probe

Total 101 23 (3.9) 48 (8.2) 2 (0.3) 110 (18.9) 399 (68.5) 23 (3.9) 16 (2.7) 23 (3.9) 8 (1.3)

LA, localised adherence; DA, diffuse adherence; AA, aggregative adherence; A+, adherence without a recognisable pattern; NA, non-adherent. *Four EAggEC probe-positive isolates from three patients exhibited the detachment of cell monolayer.

was detected in three (1.2%) EPEC isolates and one (0.4%) non-EPEC isolate. Twenty EAggEC probe- positive isolates were detected in nine (18%) control subjects. All of them produced the aggregative

4+ adherence phenotype and two isolates from two patients caused

phenotype in HEp-2 cells, but none reacted with the EPEC antisera. On the other hand, five EAggEC probe- negative isolates that produced a typical aggregative phenotype were isolated from three controls.

786 A. C. P. ROSA E T AL.

Table 4. Virulence markers in EPEC and non-EPEC isolates from 50 controls < 2 years old in Rio de Janeiro, Brazil Number (%) of strains

HEp-2 adherence Assays Patient isolates FAS EAF eaeA EAggEC

5 5 5 5

Serotype (controls) LA DA AA A+ NA test probe probe probe

- - - - -

- - - - - - - - 5

1 5 1 5 4

055:NM 5 (1) 086:H? 5 (1) 0 125:NM 3 (1) 0125:H18 1 (1) 0125:H? 5 (1) 0126:H20 1 (1) 0128:H15 5 (1) 0 142:H2 4 (1)

- - - - - - 3 - -

- - - - - - - -

- - - - - - - -

- - - - - - - - - - - - - - - -

- - - - - - - -

- - - - 1 25 24 171 20" Non-EPEC 221 (43)

Total 250 (51)+ 5 (2) 4 (1.6) 25 (10) 24 (9.6) 192 (76.8) 5 (2) 5 (2) 5 (2) 20 (8)

LA, localised adherence; DA, diffuse adherence; AA, aggregative adherence; A+, adherence without a recognisable pattern; NA, non-adherent. *All EAggEC probe-positive strains exhibited the AA adherence phenotype. tAssociation of the serotypes 0125:NM and 0125:H18 was found in one control.

Comparison of patients and controls

Of 150 infants with diarrhoea, 138 were selected from whom no recognised enteropathogens, except classic enteropathogenic E. coli, were identified. For compari- son, 48 controls were also examined (Table 5). E. coli isolates that agglutinated with EPEC antisera were far more common in patients than controls (31.9% versus 14.6%; p <0.025) and isolates exhibiting LA were more often isolated from patients than controls (19.6% versus 2.1 %; p < 0.005). Diffusely adhering and EAggEC isolates were found with equal frequency in patients and controls, thereby showing no association with diarrhoea. E. coli isolates exhibiting adherence without a recognisable pattern (A+ strains) were slightly more frequent in patients than controls (10.8% versus 4.2%; not significant). On the other hand, non-adherent E. coli isolates were significantly associated with children without diarrhoea (Table 5). LA was strongly correlated with a positive FAS assay and eaeA probe, whereas hybridisation with the EAF probe showed partial correlation with the same virulence markers. In seven (5.1%) children with

Table 5. Virulence markers in E. coli isolates from 138 children with diarrhoea without other recognised enter- opathogens and from 48 controls in Rio de Janeiro, Brazil

Number (%) from

Virulence markers Patients Controls P*

EPEC serology Adherence patterns

LA DA AA A+ NA

Assays FAS EAF probe eaeA probe EAggEC probe

44 (31.9)

27 (19.6) 11 (7.8) 15 (10.8) 15 (10.8) 73 (52.9)

27 (19.6) 20 (14.5) 27 (19.6) 14 (10.1)

7 (14.6)

1 (2.1) 2 (4.2)

10 (20.8) 2 (4.2)

42 (87.5)

1 (2.1) 1 (2.1) 1 (2.1) 9 (18.7)

< 0.025

< 0.005 NS NS NS

< 0.005

< 0.005 < 0.01 < 0.005

NS

*Determined by x2 test; p < 0.05 was considered significant.

diarrhoea, E. coli isolates positive for LA, FAS and eaeA but negative for EAF were detected. E. coli exhibiting this combination of markers were not detected among controls.

Toxin production

LT+ ETEC were isolated from five (3.6%) diarrhoea cases and one (2.1 %) control, whereas LT+/STh+ strains were isolated from one (2.1%) control. STp+ ETEC and VTEC were not found in either group. However, in 10 (6.7%) patients, 13 EPEC and nine non-EPEC isolates were found whose culture filtrates, even after heating ( 100°C, 30 min) produced cytocidal effects in Vero cells characterised by complete detach- ment of the cell monolayer even during 3-h adherence tests. Among these isolates, 10 EPEC and one non- EPEC showed all the EPEC virulence markers (LA+, FAS+, EAF+ and eaeA+).

Pathogenicity of EAF- IeaeA -+ isolates

To evaluate the pathogenic potential of the EAF-/eaeA+ E. coli isolates, the invasive ability of these isolates for Caco-2 cells was investigated by the gentamicin invasion assay and by transmission electron microscopy. After incubation for 6 h, all EAF-/eaeA+ isolates were found in the intracellular compartment. The percentages of intracellular bacteria varied from 0.4% to 1.6% of the cell-associated bacteria (Table 6). Transmission electron microscopy was performed on Caco-2 monolayers after a 6-h invasion assay. As shown in Fig. 1, an intimate adherence was demon- strated by the close association of the bacteria to the cell membrane. Intracellular EAF-/eaeA + E. coli were seen in membrane-bound vesicles. Marked cytoskeletal changes were seen directly beneath the adherent bacterium, including the accumulation of polymerised actin, and the bacteria sometimes sat upon a pedestal-like structure (Fig. 2). Accumulation of tyrosine-phosphorylated proteins was assayed by im-

VIRULENCE MARKERS IN ENTEROPATHOGENIC E. COLI 787

Table 6. Invasion of Caco-2 cells by E A F - l e a d + E. coli isolates from children with diarrhoea in Rio de Janeiro, Brazil

E. coli isolate Percent invasion

Serotype (SD)*

I 3 1 / 5 H32/5 H98/3 H100/1 H 104/2 H111/2 H 126/S H152/2 E2348/69 (positive control) DHSa(K12) (negative control)

0119:H? 0 1 19:H? 0 1 19:H2 0 1 19:H2 045:NM 055:NM O?: H3 3 0 142:H6 0127:H6

1.0 (0.2)+ 0.4 (0.1) 1.6 (0.1) 1.4 (0.2) 0.6 (0.2) 1.3 (0.3)

0.6 (0.1) 1.5 (0.2)

0.04 (0.03)

0.5 (0.1)

*Percentage of intracellular bacteria relative to the cell-associated (intracellular plus extracellular) bacteria after 6-h infection. +Results are the mean (SD) of at least three independent tests, performed in duplicate. 'Untypable with 0 1 -0 173 antisera.

munofluorescence in HEp-2 cells infected with EAF-/eaeA+ E. coli isolates. Results from micro- scopy showed foci of increased fluorescence in HEp-2 cells after infection for 6 h. These foci of fluorescence corresponded with areas of bacterial adhesion observed under phase-contrast microscopy.

Discussion

EPEC were isolated from 49 (32.7%) patients com- pared with seven (14.0%) controls, a statistically

significant difference (p < 0.025). No pathogen other than EPEC was isolated from 44 of these 49 cases. These findings as well as others obtained in Brazil [3,40,42] implicate EPEC as the main cause of infantile diarrhoea in the large Brazilian urban centres and emphasise the role of this micro-organism as an important cause of acute diarrhoea in developing countries [2, 13,431. The incidence of ETEC was very low. The results of the present study differ from those in many developing countries, where ETEC are often the most frequent pathogen found [43-451. VTEC and 0157:H7 (EHEC) strains were not detected. In Siio Paulo, Brazil, Giraldi et al. [46] found only one (0.2%) non-0157 VTEC in 466 isolates from 95 children < 1 year of age with acute diarrhoea. This result suggests that VT production is an uncommon feature among E. coli isolates from endemic acute infantile diarrhoea in Brazil. In fact, both VTEC and EHEC are usually associated with food-borne outbreaks involving patients of different age groups. EHEC strains probably do not have a universal distribution and are presently restricted to some geographical areas. EIEC was not found. In Brazil EIEC has been isolated most frequently in children > 2 years of age [47].

Among the EPEC serotypes, the most commonly isolated were 0 1 11:H2, 055:NM and 0 1 19:H6. These O:H types were also prevalent in several studies conducted in Brazil and other developing nations [3,40,41,48].

Fig. 1. Electron micrograph of a confluent monolayer of Caco-2 cells 6 h after infection with E. coli EAF-/ eaeA+ H 104/2 (045:NM). Bacterial microcolonies are transversely sectioned. Membrane folding and production of digitiform projections suggests intense mobilisation of cell membrane at the sites of bacterial colonisation. Bacteria closely associated with cell membrane and bacteria undergoing internalisation can be seen (X9 150).

788 A. C. P. ROSA E T A L .

Fig. 2. Electron micrograph of Caco-2 cell 6 h after infection with E. coli EAF-/eaeA+ HI 52/2 (0142: H6). Intimate adherence and bacterial internalisation can be seen. Membrane cupping and formation of discrete pedestals can be seen at the sites of bacterial attachment (X25 200).

With regard to the adhesive properties of isolates, LA was found both in EPEC and non-EPEC isolates from patients. Among 103 LA+ isolates, 77% belong to a restricted number of EPEC O:H serotypes. A signifi- cant association between some EPEC serotypes and LA has been shown by several authors [7,8, 41,49,50]. Serotypes 055:NM, 086:H34, 0 1 1 I:H2, 0 1 19:H6 and 0142:H6 were strongly associated with LA in Brazil [7,40-421 and other countries [51, 521. In the control group only one child (2.0%) harboured EPEC LA+ isolates belonging to the 055:NM serotype. Non-EPEC isolates producing LA were detected in patients only at low frequency (3.9%). All these isolates were FAS+ and hybridised with eaeA probe, but 30% of them were negative for the EAF sequence. Other authors have also reported the isolation of these non-EPEC strains in association with diarrhoea [ 12, 131.

In the present work, complete correlation was found between LA, FAS test and hybridisation with the eaeA probe, as also verified by several authors in many parts of the world [27,48,52]. However, a weaker correlation (89%) between EAF and the other virulence markers was found, especially for EPEC strains of serotypes 055:NM, 0 1 19:H2, 0 1 19:H6, 0 1 19:H?, 0142:H6 and for non-EPEC strains. EPEC strains cured of the EAF plasmid retain the ability to form A/E lesions in tissue culture and cause actin aggregation, although much less efficiently than the EAF- expressing parent [20, 521. EPEC serotypes that

do not possess the EAF plasmid have also been incriminated by epidemiological studies and studies in volunteers as causes of diarrhoea, even though their pathogenesis at the molecular level is not as precisely known as that of the EAF-positive EPEC [I]. In the UK, Scotland et al. [12] proposed that EAF probe- negative EPEC may be of greater importance than EAF probe-positive strains.

DAEC isolates showed no association with diarrhoea. All isolates were negative in the FAS test and did not hybridise with the eaeA, EAF and EAggEC probes. Several auth,ors did not find any association of DAEC with acute infantile diarrhoea [9, 13,42,49]. However, Giron et al. [14] and Jallat et al. [53] reported its isolation with significant frequency from cases of acute diarrhoea.

The small number of epidemiological studies per- formed to date give conflicting results with respect to differences in the rate of isolation of EAggEC from patients versus controls and their association with acute and persistent diarrhoea. In this study and in previous work performed in Brazil [9], EAggEC were isolated with similar frequency from cases of acute diarrhoea and controls. Children with persistent diar- rhoea were not included in these studies. Among the strains that produced the aggregative phenotype, the pattern was identified only in 6-h adhesion tests. In 3-h tests, bacteria could occasionally be seen scattered on the periphery of HEp-2 cells, but without the

VIRULENCE MARKERS IN ENTEROPATHOGENIC E. COLI 789

distinctive stacked-brick pattern. The EAggEC probe hybridised with 80.0% of EAggEC isolates from patients and controls in the present study. Previous studies have shown that some EAggEC strains do not hybridise with the EAggEC probe and that this group is heterogeneous [8,36]. In France, among the strains exhibiting an aggregative pattern of adhesion in HEp- 2 cells, only 82.4% hybridised with the EAggEC probe [53]. EAggEC isolates from 11 (61.1%) of 18 patients belonged to EPEC serogroups, and 65.2% of these isolates exhibited the flagellar antigens H1 1, H17, H18, H27 and H31. Some of the EPEC serogroups and serotypes detected, such as 086:H 1 1, 086:H18, 0111, 0125, 0127 and 0128 were also found elsewhere in association with EAggEC [ 10, 121. Geographic variation in predominance of serotypes of EAggEC is suggested by the finding of certain types of H antigen associated with strains isolated from different parts of the world, e.g., HI1 and H18 in India [ l l ] , H33, H2 and H10 in Chile [lo] and H18, H21 and H27 in the UK [54]. All A+ strains detected in this study were clearly distinct from the recognised diarrhoea-producing E. coli categories, as they have different adherence phenotypes and were uniformly negative for the virulence markers investigated. Despite their Hep-2 adherence ability, their isolation rates were similar in patients and controls, indicating that this adhesion is not a characteristic of pathogenic E. coli.

As defined in the 2nd International Symposium on EPEC recently held in Brazil, most true EPEC strains fall into well-recognised O:H types, carry EAF, bfpA and eaeA DNA sequences and do not produce VT [55] . In this study, EPEC and non-EPEC isolates from eight patients were EAF-lead f. These isolates were the sole pathogen isolated from seven patients and were not found in the control group. Previous studies have shown E. coli EAF- /eaeA+ strains significantly associated with diarrhoea [ 12, 561. However, case- control studies in Chile [8], Brazil [3,41], Thailand [48] and Bangladesh [43] found that only EAF positive strains were associated with diarrhoea. At the above-mentioned International Symposium, these EAF - /eaeA + strains were referred to provisionally as 'atypical EPEC'. The EAF-/eaeA+ isolates in the present study showed LA in 6-h adhesion tests, but produced a positive FAS test in 3-h tests. The interaction of these strains with Caco-2 cells was studied further by TEM and by the gentamicin invasion assay - widely used to assess bacterial internalisation by cultured epithelial cells. All isolates showed a significant invasive capability. However, the uptake of most EAF-/eaeA+ E. coli strains is far more discrete than with true EPEC strains. TEM observation of infected Caco-2 cells showed evidence of A/E lesions, such as intimate adherence of bacteria, cytoskeletal re-arrangement and cell mem- brane alterations, with the formation of discrete cup- like pedestals. Internalised bacteria were seen inside

membrane-bound cell vacuoles. In EPEC, the A/E lesion results from the co-ordinated expression of genes encoding a type I11 protein secretion system and esp genes involved in signal transduction events, such as inositol phosphate release and induction of tyrosine phosphorylation of host cell proteins. The atypical EPEC isolates studied showed ability to induce A/E lesions and to trigger cell signalling events character- istic of fully virulent EPEC. The results of this study confirm that EPEC remain the leading pathogen associated with acute diarrhoea in Brazilian children, and also raise questions as to the role of EAF-/eaeA+ E. coli strains as an important agent of infantile diarrhoea.

This study was supported by grants from the Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq), Coordenaqzo de AperfeiCoamento de Pessoal de Nivel Superior (CAPES), Programa de Apoio a Nucleos de Excelincia (PRONEX) and Fundaqio de Amparo a Pesquisa do Estado de SZo Paulo (FAPESP), Brazil.

References

1 .

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

Levine MM. Escherichia coli that cause diarrhea: enterotoxi- genic, enteropathogenic, enteroinvasive, enterohemorrhagic, and enteroadherent. J Infect Dis 1987; 155: 377-389. Katouly M, Jaafari A, Farhoudi-Moghaddam AA, Ketabi GR. Aetiological studies of diarrhoea1 diseases in infants and young children in Iran. J Trop Med Hyg 1990; 93: 22-27. Gomes TAT, Rassi MacDonald KL et al. Enteropathogens associated with acute diarrheal disease in urban infants in Sio Paulo, Brazil. J Infect Dis 1991; 164: 131-137. Clausen CR, Christie DL. Chronic diarrhea in infants caused by adherent enteropathogenic Escherichia coli. J Pediatr 1982;

Scaletsky ICA, Silva MLM, Trabulsi LR. Distinctive patterns of adherence of enteropathogenic Escherichia coli to HeLa cells. Infect Immun 1984; 45: 534-536. Nataro JP, Kaper JB, Robins-Browne R, Prado V, Vial P, Levine MM. Patterns of adherence of diarrheagenic Escherichia coli to HEp-2 cells. Paediatr Infect Dis J 1987; 6: 829-831. Scaletsky ICA, Silva MLM, Toledo M E , Davis BR, Blake PA, Trabulsi LR. Correlation between adherence to HeLa cells and serogroups, serotypes, and bioserotypes of Escherichia coli. Infect Immun 1985; 49: 528-532. Levine MM, Prado V, Robins-Browne R et al. Use of DNA probes and HEp-2 cell adherence assay to detect diarrheagenic Escherichia coli. J Infect Dis 1988; 158: 224-228. Gomes TAT, Blake PA, Trabulsi LR. Prevalence of Escherichia coli strains with localized, diffuse, and aggregative adherence to HeLa cells in infants with diarrhea and matched controls. J Clin Microbiol 1989; 27: 266-269. Vial PA, Robins-Browne RM, Lior H et al. Characterization of enteroadherent-aggregative Escherichia coli, a putative agent of diarrheal disease. J Infect Dis 1988; 158: 70-79. Bhan MK, Raj P, Levine MM et al. Enteroaggregative Escherichia coli associated with persistent diarrhea in a cohort of rural children in India. J Infect Dis 1989; 159: 1061-1064. Scotland SM, Smith HR, Said B, Willshaw GA, Cheasty T, Rowe B. Identification of enteropathogenic Escherichia coli isolated in Britain as enteroaggregative or as members of a subclass of attaching-and-effacing E. coli not hybridising with the EPEC adherence-factor probe. J Med Microbiol 1991; 35:

Cravioto A, Tell0 A, Navarro A et al. Association of Escherichia coli HEp-2 adherence patterns with type and duration of diarrhoea. Lancet 1991 ; 337: 262-264. Giron JA, Jones T, Millan-Velasco F et al. Diffuse-adhering Escherichia coli (DAEC) as a putative cause of diarrhea in Mayan children in Mexico. J Infect Dis 1991; 163: 507-513. Echeverria P, Serichantalerg 0, Changechawalit S et al. Tissue culture-adherent Escherichia coli in infantile diarrhea. J Infect Dis 1992; 165: 141-143.

100: 358-361.

278-283.

790 A. C. P. ROSA E T A L .

16. Baldini MM, Kaper JB, Levine MM, Candy DCA, Moon HW. Plasmid-mediated adhesion in enteropathogenic Escherichia coli. J Pediatr Gastroenterol Nutr 1983; 2: 534-538.

17. Baldini MM, Nataro JP, Kaper JB. Localization of a determinant for HEp-2 adherence by enteropathogenic Escher- ichia coli. Infect Immun 1986; 52: 334-336.

18. Nataro JP, Baldini MM, Kaper JB, Black RE, Bravo N, Levine MM. Detection of an adherence factor of enteropathogenic Escherichia coli with a DNA probe. J Injkct Dis 1985; 152:

19. Moon HW, Whipp SC, Argenzio RA, Levine MM, Giannella RA. Attaching and effacing activities of rabbit and human enteropathogenic Escherichia coli in pig and rabbit intestines. Infect Immun 1983; 41: 1340-1351.

20. Knutton S, Baldwin T, Williams PH, McNeish AS. Actin accumulation at sites of bacterial adhesion to tissue culture cells: basis of a new diagnostic test for enteropathogenic and enterohemorrhagic Escherichia coli. Infict Immun 1989; 57:

21. Andrade JRC, de Santa Rosa MR. Attachment and intracellular penetration of classic enteropathogenic Escherichia coli into HEp-2 cells. Rev Microbiol (Siio Paulo) 1986; 17: 53-57.

22. Andrade, JRC, da Veiga VF, de Santa Rosa MR, Suassuna I. An endocytic process in HEp-2 cells induced by enteropatho- genic Escherichia coli. J Med Microbiol 1989; 28: 49-51.

23. Donnenberg MS, Donohue-Rolfe A, Keusch GT. Epithelial cell invasion: an overlooked property of enteropathogenic Escher- ichia coli (EPEC) associated with the EPEC adherence factor. J Infect Dis 1989; 160: 452-459.

24. Donnenberg MS, Donohue-Rolfe A, Keusch GT. A comparison of HEp-2 cell invasion by enteropathogenic and enteroinvasive Escherichia coli. FEMS Microbiol Lett 1990; 69: 83-86.

25. Rosenshine I, Donnenberg MS, Kaper JB, Finlay BB. Signal transduction between enteropathogenic Escherichia coli (EPEC) and epithelial cells: EPEC induces tyrosine phosphorylation of host cell proteins to initiate cytoeskeletal rearrangement and bacterial uptake. EMBO J 1992; 11: 3551-3560.

26. McDaniel TK, Jarvis KG, Donnenberg MS, Kaper JB. A genetic locus of enterocyte effacement conserved among diverse enterobacterial pathogens. Proc Natl Acad Sci USA

27. Jerse AE, Yu J, Tall BD, Kaper JB. A genetic locus of enteropathogenic Escherichia coli necessary for the production of attaching and effacing lesions on tissue culture cells. Proc Natl Acad Sci USA 1990; 87: 7839-7843.

28. Edwards PR, Ewing WH (eds) Identification of Enterobacter- iaceae, 3rd edn. Minneapolis, Burgess Publishing. 1972.

29. Hausler WJ, Lennette EH, Balows A, Shadomy HJ (eds) Manual of clinical microbiology, 5th edn. Washington, DC, ASM Press. 1991.

30. Sireny B. Experimental Shigella keratoconjunctivitis. Acta Microbiol Acad Sci Hung 1955; 2: 293-296.

31. Smith HR, Scotland SM. Vero cytotoxin-producing strains of Escherichia coli. J Med Microbiol 1988; 26: 77-85.

32. Sack DA, Sack RB. Test for enterotoxigenic Escherichia coli using Y-1 adrenal cells in miniculture. Infect Zmmun 1975; 11:

33. Cravioto A, Gross RJ, Scotland SM, Rowe B. An adhesive factor found in strains of Escherichia coli belonging to the traditional infantile enteropathogenic serotypes. Curr Microbiol

34. Maas R. An improved colony hybridization method with significant sensitivity for detection of single genes. Plasmid

35. Harford N, De Wilde M, Cabezon T. Cloning of two distinct but related ST enterotoxin genes from porcine and human strains of E. coli. In: Levy SB, Clowes RC, Koening EL (eds) Molecular biology, pathogenicity, and ecology of bacterial plasmids. New York, Plenum Press. 1981: 61 1.

36. Baudry B, Savarino SJ, Vial P, Kaper JB, Levine MM. A sensitive and specific DNA probe to identify enteroaggregative Escherichia coli, a recently discovered diarrheal pathogen. JInfect Dis 1990; 161: 1249-1251.

37. Rousset M. The human colon carcinoma cell lines HT-29 and

560-565.

1290-1298.

1995; 92: 1664-1668.

334-336.

1979; 3: 98-99.

1983; 10: 296-298.

Caco-2: two in vitro models for the study of intestinal differentiation. Biochimie 1986; 68: 1035- 1040.

38. Robins-Browne RM, Bennett-Wood V. Quantitative assessment of the ability of Escherichia coli to invade cultured animal cells. Microb Pathog 1992; 12: 159-164.

39. Rosenshine I, Duronio V, Finlay BB. Tyrosine protein kinase inhibitors block invasin-promoted bacterial uptake by epithelial cells. Infect Zmmun 1992; 60: 22 1 1-22 17.

40. Toledo MRF, Alvariza MdoCB, Murahovschi J, Ramos SRTS, Trabulsi LR. Enteropathogenic Escherichia coli serotypes and endemic diarrhea in infants. Infect Immun 1983; 39: 586-589.

41. Gomes TAT, Vieira MAM, Wachsmuth IK, Blake PA, Trabulsi LR. Serotype-specific prevalence of Escherichia coli strains with EPEC adherence factor genes in infants with and without diarrhea in SBo Paulo, Brazil. J Infect Dis 1989; 160:

42. MagalhBes M, Amorim RJM, Takeda Y, Tsukamoto T, Antas MG, Tateno S. Localized, diffuse, and aggregative-adhering Escherichia coli from infants with acute diarrhea and matched- controls. Mem Inst Oswuldo Cruz 1992; 87: 93-97.

43. Albert MJ, Faruque SM, Faruque ASG et al. Controlled study of Escherichia coli diarrheal infections in Bangladeshi children. J Clin Microbiol 1995; 33: 973-977.

44. Wolk M, Ohad E, Shafran R et al. Epidemiological aspects of enterotoxigenic Escherichia coli diarrhoea in infants in the Jerusalem area. Public Health Rev 1995; 23: 25-33.

45. Black RE, Brown KHN, Becker S, Abdul ARMA, Huq I. Longitudinal studies of infectious diseases and physical growth of children in rural Bangladesh. 11. Incidence of diarrhea and association with known pathogens. Am J Epidemiol 1982; 115:

46. Giraldi R, Guth BEC, Trabulsi LR. Production of shiga-like toxin among Escherichia coli strains and other bacteria from diarrhea in SBo Paulo, Brazil. J Clin Microbiol 1990; 28:

47. Toledo MRF, Trabulsi LR. Frequency of enteroinvasive Escherichia coli in children with diarrhea and healthy controls in SBo Paulo, SP, Brazil. Rev Microbiol (SZo Paulo) 1990; 21:

48. Echeverria P, Orskov F, Orskov 1 et al. Attaching and effacing enteropathogenic Escherichia coli as a cause of infantile diarrhea in Bangkok. J Infect Dis 1991; 164: 550-554.

49. Nataro JP, Scaletsky ICA, Kaper JB, Levine MM, Trabulsi LR. Plasmid-mediated factors confering diffuse and localized adherence of enteropathogenic Escherichiu coli. Infect Immun

50. Scotland SM, Willshaw GA, Smith HR, Gross RJ, Rowe B. Adhesion to cells in culture and plasmid profiles of enteropathogenic Escherichia coli isolated from outbreaks and sporadic cases of infant diarrhoea. J Infect 1989; 19: 237-249.

51. Moyenuddin M, Wachsmuth IK, Moseley SL, Bopp CA, Blake PA. Serotype, antimicrobial resistance, and adherence properties of Escherichia coli strains associated with outbreaks of diarrheal illness in children in the United States. J Clin Microbiol 1989; 27: 2234-2239.

52. Knutton S, Phillips AD, Smith HR et al. Screening for enteropathogenic Escherichia coli in infants with diarrhea by the fluorescent-actin staining test. Infect Immun 1991; 59:

53. Jallat C, Livrelli V, Darfeuille-Michaud A, Rich C, Joly B. Escherichia coli strains involved in diarrhea in France: high prevalence and heterogenicity of diffusely adhering strains. J Clin Microbiol 1993; 31: 2031-2037.

54. Scotland SM, Smith HR, Cheasty T et al. Use of gene probes and adhesion tests to characterise Escherichia coli belonging to enteropathogenic serogroups isolated in the United Kingdom. J Med Microbiol 1996; 44: 438-443.

55. Kaper JB. Defining EPEC. Rev Microbiol (YZo Paulo) 1996; 27(Suppl 1): 130-133.

56. Scotland SM, Willshaw GA, Smith HR, Said B, Stokes N, Rowe B. Virulence properties of Escherichia coli strains belonging to serogroups 026, 055, 0111 and 0128 isolated in the United Kingdom in 1991 from patients with diarrhoea. Epidemiol Infect 1993; 111: 429-438.

131- 135.

3 15-324.

1460- 1462.

1-4.

1985; 48: 378-383.

365-37 1.