Prevalence of diarrheagenic Escherichia coli strains detected by PCR in patients with travelers' diarrhea

ORIGINAL ARTICLE

Prevalence of diarrheagenic Escherichia coli strains detected by PCR in patients with travelers' diarrhea M . Ihrgas', J. Gasc6n2, F Gallardol, 111.T j i m e n e x de A n t a ' andJ. Vila' Departments of 'Microbiology and 'Infectious Diseases (Tropical Medicine Unit), Institut d'Investigacions BionGdiques Agusti Pi i Sunyer, Hospital Clinic, University of Barcelona. School of Medicine, Barcelona, Spain

Objective: To determine the prevalence of the different categories of diarrheagenic Escherichia coli, enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), verotoxin-producing E. coli (VTEC), enteroaggregative E. coli (EAggEC), diffusely adherent E. coli (DAEC), and enteropathogenic E. co/i (EPEC), associated with travelers' diarrhea. Methods: Stool specimens from 350 patients with travelers' diarrhea were collected between 1994 and 1996. The virulence factors of the diarrheagenic E. coli isolated were detected by PCR technique, in subcultures of single colonies of all morphotypes of E. coli observed in culture on MacConkey agar. Results: ETEC (15.7%),EAggEC (13.4%) and DAEC (9.14%) are significantly more prevalent than EIEC (3.4%),EPEC (2.86%) and VTEC (0.86%) (pc0.05;z-test). The prevalence of ETEC and EAggEC was similar in all geographic areas visited.

Conclusions: PCR is a rapid and specific technique to use i n the identification of the different categories of diarrheagenic E. coliand greatly increases the yield of potential enteropathogens from cases of travelers' diarrhea. Not only ETEC but also EAggEC and DAEC strains play a major role i n the etiology of travelers' diarrhea, whereas EIEC, EPEC, and VTEC strains play a minor role, leading t o the question of whether it is necessary to search routinely for these three types of E. coli in diagnostic laboratories.

Key words: ETEC, EPEC, EAggEC, DAEC, VTEC, EIEC, travelers' diarrhea, Escherichia coli

I NTRODUCTlON Travelers' diarrhea (TLI) is a coiiiiiion illness among travelers to tropical and subtropical areas, with a prevalence of 20-6094 [ 1,2]. The microbiological causc of x u t e diarrhea in travelers is diagnosed in only 00-70%1 of the cases. Rotavirus, Sh
Corresponding author and reprint requests: J. Vila, Laboratori de Microbiologia, Hospital Clinic, Facultat de Medicina, Universitat de Barcelona, Villarroel 170, 08036 Barcelona, Spain.

Tel: +34 3 2275522

Fax: 4 4 3 2275454

E-mail: [email protected] Accepted 26 May 1998

682

been among the most frequently detectecj etiologic agents causing TI) [ 3] . Several categories of diarrheagenic E. roli have been described. Enterotoxigeriic E. roli (ETEC) strains, which produce A heat-labile toxin (LT) (similar to that of Vihrio clzolerne) [4] and/or a heat-stable toxins (ST) [S],are the most frequent cause of TI) [6-91. Other types of E . roli related to diarrhea are the enteropathogenic E.coli (EPEC), enteroinvasive E. roli (EIEC), eriteroadherent E . ccdi (EAEC) [I 01, and verotoxinproducing E . coli (VTEC) [ 1 11. EPEC strains are widely distributed throughout the world and are considered to be among the main etiologic agents causing infantile diarrhea outbreaks. The mechanism of pathogenicity of this category of E. cmli is not totally understood, although several virulence factors have been associated with this type. One is an enteropathogen adherence factor (EAF), coded for by a plasmid of 50-70 MDa

Vargas e t al: Diarrheagenic E s c h e r i c h i a coli in p a t i e n t s w i t h travelers’ d i a r r h e a

[12]. With the use of a DNA probe specific for EAF [13], EPEC strains have been divided into two classes: class I (EAF-positive) and class I1 (EAF-negative). Bundle-forming pili (bfp) [ 141, which are coded for by the b f p A gene, have been related to the initial adherence of the bacteria to the enterocyte. Another virulence factor is a membrane protein (94 kDa) outer-membrane protein (OMP), coded for by a chromosomal eaeA gene, which has been implicated in the expression of the attaching and effacing phenotype, characterized by an effacement of microvilli. The product of the eaeA gene mediates intimate attachment, resulting in accumulation of filamentous actin and other cytoskeletal proteins [15]. EIEC strains provoke a bacilliary dysentery similar to that caused by Slzigella. Their capacity to penetrate into the intestinal cells is mediated in part by a long plasmid of 120-140 MDa [16]. VTEC strains produce two cytotoxins, which are very similar, biologically and structurally, to the toxins synthesized by Shigella dysenteriae. These toxins have been called verotoxins (VT1 and VT2) or Shiga-like toxins (SLTI and SLTII) [17]. EAEC strains display binding properties in vitro to HEp-2 or HeLa cells [18,19]. In the presence of Dmannose, three characteristic adherence patterns can be observed: localized adherence, diffuse adherence and aggregative adherence [20]. The localized adherence E . coli (LAEC) strains are associated with EPEC. Moreover, gene clusters of the gfa operon have been found to encode an adhesin involved in diffuse adherence [21,22]. O n the other hand, the enteroaggregative E. coli (EAggEC) strains possess plasmid genes encoding adhesins responsible for the aggregative adherence [20]. In spite of the fact that ETEC strains have been described in many reports as comprising an important cause of TD, the association of the remaining categories of diarrheagenic E. coli with diarrhea has not been investigated in depth. The main aim of this study was to determine the prevalence of the different categories of diarrheagenic E. coli (ETEC, EIEC, VTEC, EPEC, EAggEC and DAEC) associated with TD.

MATERIAL AND METHODS Patients During the period 1994-96, stool specimens from 350 patients with T D were analyzed. T D was defined as the occurrence between 12 h after arrival in, and 5 days after departure from, the country visited, of three or more episodes of watery diarrhea within a 24-h period with or without other symptoms, or as the occurrence of unformed stools accompanied by one of the following: abdominal cramps, tenesmus, vomiting, nausea,

683

fever, chills or prostration. All the sample were collected from travelers with diarrhea after their return to Spain and were processed within 2 h of collection Microbiological tests The stool specimens were cultured for E. coli, and other bacterial enteropathogens, including Salmonella, Shigella, Campylobacter, Aeromonas, Pleriomonas, Vibrio and Yersinia spp., using conventional methods [23]. The samples were also examined for ova and parasites. Nevertheless, the different categories of diarrheagenic E. coli formed the focus of this study. Single-colony subcultures of all different lactosefermenting colonial morphotypes growing on MacConkey agar were identified following conventional biochemical criteria. The E. coli isolates were tested by P C R to detect the different virulence factors, and with polyvalent antisera in order to establish the serogroup. If they presented a positive reaction with these polyvalent antisera, colonies were tested with monovalent antisera against the following groups: 0 2 6 , 0 5 5 , 0 8 6 , 0 1 1 1 , 0 1 1 9 , 0 1 2 5 , 0 1 2 6 , 0 1 2 7 and 0128. PCR assay The virulence factors of the isolated E. coli strains were detected by a PCR technique. In the same reaction, two different genes were amplified, the size of the amplicon being taken into account in interpretation. The sets of primers were used in six combinations: LTIb/STI, VTl/VT2, SLTI/SLTII, eae/bfp, EAF/ EAggEC, and EIEC/DAEC. Two different sets of primers were used to detect VT1- and/or VT2producing E. coli strains in order to increase the specificity. We have called them VTl/VT2 and SLTI/SLTII for differentiation purposes only. The bacteria were grown on MacConkey agar overnight. One colony of each isolate was suspended in 25 pL of reaction mixture containing 20 m M Tris-HC1 (pH 8.8), 100 mh4 KCl, 3.0 mh4 MgClz, gelatin 0.1%, 400 p M dNTPs, and 1 p M primer was added, together with 2.5 U Taq polymerase (Gibco, Life Technologies, Inc., Gaithersburg, MD, USA). The reaction mixture was overlaid with oil and subjected to the following program: 30 cycles at 95°C for 50 s, 55°C for 1.5 min, and 72°C for 2 min. Twelve sets of primers (Boehringer Mannheim, Mannheim, Germany), for the different virulence factors were used, as indicated in Table 1. AU primer sequence were obtained from the references in Table 1, except for the primers used to amplify the eae and bfp genes, which were designed by us according to the gene sequences mentioned in references 15 and 26. P C R assays were performed in a D N A Thermal Cycler 480 (Perkin-Elmer Cetus, Emeryvde, CA, USA). A

C l i n i c a l M i c r o b i o l o g y a n d I n f e c t i o n , V o l u m e 4 N u m b e r 1 2 , D e c e m b e r 1998

684

Table 1 Primers used for the identification of the different virulence factors of E. roli E . ioii

Virulence factor

ETEC'

LTlb

LTIb-1 LTIb-2

ETEC

STI

EPEC

Oligonucleotide sequence (5' to 3')

with gene

Amplified product (bp)

Keference

TCTCTATGTGCATACGGAGC CCATACTAGTTGCCGCAAT

46-65 349-367

322

21

STI-1 STI-2

TTAATAGCACCCGGTACAAGCAGC, CTTGACTCTTCAAAAGAGAAAATTA[~

186-218 72-97

147

24

EAF

EAF-I EAF-25

CAGGGTWGAAAGATGATAA TATGGGGACCATGTATTATCA

546-567 922-942

3'17

23

EPEC

btp

bfp-upper bfp-lower

ACAAAGATACAACAAACAAAAA TTCAGCAGGAGTAAAAGCAGTC

260

20

EPEC/ VTEC

Cat

ear- 1 Cde-2

CCCGGACCCGGCACAAGCATAAGC TTGAAATAGTCTCGCCAGTATTCG

383-404 621-642 22-47 880-002

88 1

15

VTEC

SLTI

SLTI-1 SLTI-3

CAGTTAATGTGGTGGCGAAG CTGCTAATAGTTCTGCGCATC

21 5-234 1089- I 109

8'15

2-1

VTEC

VTl

VTla VTlb

GAAGAGTCGTGGGATTACG AGCGATGCAGCTATTAATAA

1191-1210 1301-1320

130

27

VTEC

SLTII

SLTI I-1 SLTII-2

CTTCGGTATCCTATTCCCGG GGATGCATCTCTGGTCATTG

288-307

479

24

VTEC

VT2

VT2a VT2b

TTAACCACACCCACXXCAGT GCTCTGGATGCATCTCTGGT

42(1-44.i

752-771

346

2728

EAggEC

AA

AM-I AGG-2

CTGGCGAAAGACTGTATCAT CAATGTATAGAAATCCGCTGTT

64-83 h72-(1')3

(730

20

EIEC

ipaH

Shig-l Shg-2

TGGAAAAACTCAGTGCCTCT CCAGTCCGTAAATTCATTCT

123

.io

DAEC

IIA

aL-1 3f3-2

GCTGGGCAGCAAACTGATAACTCT CATCAAGCTGTTT(;TTCGTCCG~:~~;

731

77 __

For

Primers

MI explanation of E. mli types, virulence

Location

747-766

11163%I083

1466- I4b.i 11'1-142

814-868

factor5 and primer n a n m , we text

reagent blank, which contained all components of the reaction mixture with the exception of the bacteria, was included in every P C R procedure. Amplification products were submitted to gel electrophoresis in agarose 2% and detected by staining with ethidium bromide. T h e E.coli strains used as positive controls for the PCR reaction were as follows. Strain E-2348, EAF probe+control, strain 17-2 EAggEC probe + control and strain 51-1 DA probe+control were kindly provided by Dr. J. I! Nataro (Center for Vaccine Development, School of Medicine, University of Maryland, USA), strain 11 1/87 (serotype 0 1 11) EAF+ was kindy provides by D r H. Karch from the Institut fur Hygiene und Mikrobiologie, Universitat Wiirzburg (Germany), and strains A T C C 35401 ST+LT+, A T C C 43890 VT-l+ and A T C C 43889 VT-2+ were puchased from the American Type Culture Collection (Rockville, MI)).

RESULTS T h e amplified products of the different PCR reactions are shown in Figure 1. In the amplification reaction of the genes encoding the LTb and S T toxins, two bands, of 322 bp (LT) and 147 bp (ST), respectively, were

M

1

2

3

4

5

6

7

8

Figure 1 Agarore gel electrophoresis showing PCR amplification products: lane M. DNA niolecular weight marker 100-bp DNA ladder from Gibco-BKL: lane 1. LT (322 bp) and ST (147 bp) genes; lane 2. detection of EAggEC (630 bp); lane 3, detection of EAF (397 bp): lane 1,e m gene (881 bp) and &J gene (260 bp); lane 5, +uH gene (423 bp); lane 6, 4f;l gene (750bp); lane 7, detection of VTI- and VT2-producing E. roli strains using SLTI/SLTII primers (895 bp (VT1) and 179 bp (VT2)); lane 8, detection of VT1 and VTZproducing E. coli strains using VTI/VT2 primers (130 bp (VT1) and 346 bp (VT2)).

Vargas e t al: Diarrheagenic Escherichia c o l i in patients w i t h travelers‘ d i a r r h e a

685

Table 2 Distribution of diarrheagenic E. coli among 350 patients with travelers’ diarrhea ~

Enterotoxigenic E. coli (ETEC) Heat-stable toxin (ST) Heat-labile toxin (LT) ST + LT Enteroaggregative E. coli (EAggEC) Diffusely adherent E. coli (DAEC) Enteroinvasive E. coli (EIEC) Enteropathogenic E. coli (EPEC) EPEC attaching-effacing (eae) Bundle-forming factor (bfp) EPEC adherence factor (EAF) Verotoxin-producing E. coli (VTEC) VTl toxin VT2 toxin V T l + VT2 Total no. of patients with at least one diarrheagenic E. coli

1994-95 (n = 154)

1996 ( a = 196)

25 17 6 2 24 14 7 5 5 0 0

30

Total

n (“3) 55 (15.71) 38 (10.86) 11 (3.11)

21

5 4

6 (1.71) 47 (13.43) 32 (9.14)” 12 (3.43)a.b 10 (2.86)”’ 9 (2.57) I (0.29) 0 (0.00) 3 (0.86)”,’ 2 (0.57) 0 (0.00) 1 (0.29) 146 (41.71)

23 18 5 5 4 1 0 2 1 0 1

1 1 0 0

”2-test, pi0.05 versus ETEC and EAggEC hz-test, p<0.05 versus DAEC.

Table 3 Total number of enteropathogens found in 350 patients with travelers’ diarrhea Enteropathogens

E . coli alone (one type) E. coli associated with other enteropathogens Bacteria’ Parasites” Association of enteropathogens other than E. coli Other enteropathogens‘ Total enteropathogens

1994-95 (no. patients)

1996 (no. patients)

50 21 18 3 14 36 121

57 18

Total (no. patients)

(“A)

107 (30.57) 39 (11.14) 33 (9.43) 6 (1.71) 35 (10.0) 79 (22.5) 260 (74.3)

15 3 21 43 137

’Bacteria (including different categories of diarrheagenic E. coli), and parasites found together with diarrheagenic E . mli. bacteria or parasites found alone, not associated with E. coli.

observed. EAF and EAggEC were detected by amplifying two bands of 397 bp and 630 bp, respectively. The P C R products for eae and bfp genes, were of 881 bp and 260 bp, respectively. The amplified products for the ipuH gene (EIEC) and the afd gene (DAEC) were of 423 bp and 750 bp, respectively. Amplification of VT1 and/or VT2 genes using SLTUSLTII primers showed two PCR products of 895 bp and 479 bp, respectively, whereas amplification of VT1 and VT2 genes using VTlIVT2 primers showed two P C R products of 130 bp and 346 bp, respectively. Strains belonging to one of the different categories of diarrheagenic E. coli were found in 146 (41.71%) of 350 travelers presenting a diarrheic episode during and after their trip, whereas in 13 patients two different diarrheagenic E. coli strains were found (Table 2). No significant differences were found between the prevalence of diarrheagenic E. coli in 1994-95 and

Table 4 Geographic distribution of ETEC and EAggEC among 196 patients with travelers’ diarrhea

Mica Asia Central Anirrica South America

ETEC n (“A)

EAggEC n (Yo)

12 (40.0) 10 (33.3) 6 (20.0) 2 (6.7)

6 (26.1) 10 (43.5) 5 (21.7) 2 (8.7)

1996. With respect to the distribution of the different categories of diarrheagenic E. coli, ETEC, 55 strains (15.71%), and EAggEC, 47 strains (13.43%), were the etiologic agents most frequently isolated with no significant differences being observed between them. DAEC, 32 strain (9.14%) were the third most frequently detected etiologic agents, showing no significant differences from EAggEC (p=O.O45; z-test).

686

C l i n i c a l M i c r o b i o l o g y a n d I n f e c t i o n , V o l u m e 4 N u m b e r 12, D e c e m b e r 1 9 9 8

From the 35 ETEC strains isolated, 38 strains (69.1%) produced only the ST toxin, 11 (20%) only the LT toxin, and six (11(%) both toxins. The prevalences ofElEC (12 strains, 3.4%), EPEC (1 0 strains, 2.9%) and VTEC three strains, O.S%i) were significantly lower (p<0.05; z-test). Nine of 10 EPEC strains were P C R positive for eue gene, whereas only one strain was positive for the bfp gene and none showed amplification of EAF. Among the three VTEC strains, two possessed only the VT1 toxin and the other strain produced both VT1 and VT2 toxins. In 39 patients (11.14%), the different types of enterovirulent E . coli were associated with other bacteria or parasites (Table 3). The identification of the different categories of diarrheagenic E. coli produced an important decrease (30.57%) in the cases of diarrhea without etiologic diagnosis. None of the positive strains agglutinated with the classic EPEC antisera. The geographic distribution of ETEC and EAggEC during 1996 is shown in Table 4. ETEC and EAggEC are the etiologic agents most frequently isolated from travelers coming from Africa, Asia and Central America, and there are no significant differences among these areas.

DISCUSSION The classical methods for the detection of the different categories of diarrheagenic E. roli are based on detection of phenotypic characteristics. Current methods such as biological assay or cell-culture assays are cumbersome, and are not useful for the rapid identification of diarrheagenic E. c01i. The P C K technique provide5 specific and rapid diagnosis. However, the appropriate choice and design of the priniers is important to increase the specificity of the technique. For instance, in the detection of EIEC, the amplification of the ipaH gene, which is present in niultiple copies on both the plasmid and the chromosome, is more specific than the aniplification of the invasive plasmid antigen (idgene), which is located on a plasmid which may be lost during storage [30].In our study, the P C R conditions used to amplifj. two different genes in the same reaction were set up using specific controls of diarheagenic E . roli strains. An annealing temperature of 55°C was chosen, since it does not differ very much from the optimal Tafor each primer, and the controls showed a pure P C R product without unspecific amplification. Several authors have found ETEC to be one of the most frequent etiologic agents causing TD, and they have been isolated from patients from all the geographic areas studied [6-91. Black [7]and Taylor and Echevarria [9] showed that the prevalence of ETEC is lower in

the Asian continent. However, in our study a high prevalence of ETEC was observed in Asia, in agreement with the results published by Vila et a1 1311. Nevertheless, a lower overall prevalence of ETEC (15.71%) was observed than xvas reported by Black [6,7] and Vila et a1 [31]. One possible explanation for this result is the increase in tourism to developing countries in recent years. Therefore, some people may have obtained protection due to previous diarrhea episodes caused by ETEC. A similar suggestion was made by Echevarria et a1 1321 in studies of US troops in Thailand. Many reports [33,34] have demonstrated the association of EAggEC with children's diarrhea, although few studies have been carried out in adults. In our study, a high prevalence of EAggEC was observed, in agreement with the results published by Cohen et a1 [35], who found EAggEC in 12-1396 of the patients who had traveled to Central or South America, the Caribbean, or Mexico. ILecently, in a case-control study designed to elucidate the role of EAggEC as a cause of TD, Gasc6n et a1 [36] found that EAggEC strains were isolated significantly more often in cases than in controls. The importance of VTEC. mainly 0 1 57:H7 E. coli, in public health has increased in the last few years. The incidence of diarrhea caused by VTEC in Alberta (Canada) and Scotland (UK) between 1987 and 1991 w a s very high (64-82'%,) [37]. Likewise, in Austria, VTEC has been found to be the third most frequently isolated bacterial cause of diarrhea in children [38]. Meanwhile, in the cohort study of diarrheagenic E.coli in 340 children in Chile, VTEC detected by a DNA probe for the 60-MDa virulence plasniid was the least frequently found category of diarrheagenic E. coli 1391. Similarly, in a study describing the epidemiology of diarrheagenic E . roli in New Caledonian children no VT-producing strains were detected in a total of 1542 E. coli isolates studied [K].I n our study, using two different sets of primers to aniplifiy the VT1 and VT2 genes in order to increase the sensitivity of the method, we found a very low (0.8%) prevalence of this category of E . roli among patients with diarrhea who visited tropical or subtropical countries. These studies performed in developing countries are in agreement with our results, suggesting a minor role for this category of diarrheagenic E. coli in developing nations. There are few reports concerning the role of EIEC as a cause of TD. The low prevalence of EIEC, is in agreement with the study by Mattila et a1 1411. The roles of DAEC and EPEC as causes of diarrhea in children have also been studied, with controversial results being found for the former 121,421. However, their roles as causes of T D have been poorly

V a r g a s e t a l : D i a r r h e a g e n i c E s c h e r i c h i a coli i n p a t i e n t s w i t h t r a v e l e r s ‘ d i a r r h e a

investigated. DAEC was not found to cause diarrhea when administered to healthy volunteers [43]. Mathewson et al. [44] administered DAEC isolated from travelers to Mexico to eight volunteers and found that only one developed diarrhea. Nevertheless, more recent data suggest a potential pathogenic role for DAEC, particularly among older children. Recently, DAEC has been significantly associated with diarrhea in children over 18 months of age in Western Australia [45]. Similarly, in a cohort of 340 children in Chile, DAEC was identified in 16.6% of the cases and in 11.9%of controls (p=0.002) [40]. In a previous study, Mathewson et a1 [46] found a high prevalence of EAEC in travelers with TD. These results are similar to those obtained in the present study. A gene probe for the bundle-forming pilus gene (bfp), of EPEC, developed by Giron et a1 [47], demonstrated a slightly greater sensitivity than the EAF probe among EPEC strains with localized adherence. In our study, only one strain was positive for the P C R directed to amplift. the bfp gene and no strains were positive for EAE Another gene which has been associated with the pathogenesis of EPEC is the eaeA gene, which encodes a protein mediating intimate attachment of the bacteria to the epithelial cell. This gene has also been found in enterohemorrhagic E. coli. Nine out of the 350 E. coli strains possessed this gene and none were enterohemorrhagic, suggesting that P C R to detect the eae gene may be more useful than P C R to detect bfp or EAF as a method for EPEC detection. A possible methodological limitation in our study is that we only subcultured one colony of each different colonial morphotype growing on MacConkey agar and identified as E. coli, to perform the P C R reaction. Merson et al [48] found that toxin testing of one colony yielded a diagnosis in 92% of the cases, whereas testing two colonies yielded a diagnosis in 95% of the cases. Even assuming this error, our data showed a high prevalence of ETEC, EAggEC and DAEC in patients with TD, and it is probable that the relative prevalence of the various categories of E. coli is reliable. In summary, for rapid and specific detection of the different categories of diarrheagenic E. coli, it is important to use a specific DNA test, in order to initiate adequate treatment of the patient. The P C R technique is currently the most rapid and specific method. The present study shows that ETEC, EAggEC and DAEC are significantly more prevalent than EIEC, VTEC and EPEC in stool samples from patients with TD, leading us to question whether it is necessary to search routinely for these three latter categories of E. coli in diagnostic laboratories.

687

Acknowledgments This work wa supported in part by grant FIS 94/0980 from Fondo de Investigaciones Sanitarias de la Seguridad Social, Spain and Mutual Mkdica, Barcelona, Spain. References 1. Gorbach SL, Edelman R.Traveller’s diarrhea: National Institute of Health Consensus Development Conference. Introduction. Rev Infect Dis 1986; 8(suppl 2): 109-10. 2. MacDonald KL, Cohen ML. Epidemiology of travelers diarrhoea: current perspectives. Rev Infect Dis 1986; 8(suppl2): S 117-2 1. 3. Gasc6n J, Vila J, Valls ME, et al. Aetiology of traveller’s diarrhea in Spanish travellers to developing countries. Eur J Epidemiol 1993; 9: 217-23. 4. Domenighini M, Pizza M, Jobling MG, Holmes RK, Rappuoli R. Identification of errors among database sequence entries and comparison of correct amino acid sequence for the heat-labile enterotoxins of Escherickia coli and Vibrio ckolerae. Mol Microbiol 1995; 15: 1165-7. 5. Yoshimura S, Ikemura H , Watanabe H , et al. Essential structure for full enterotoxigenic activity of heat-stable enterotoxin produced by enterotoxigenic Escherickia coli. FEBS Lett 1985; 181: 1 3 8 4 1 . 6. Black RE. Pathogens that cause travelers’ diarrhea in Latin America and Africa. Rev Infect Dis 1986; B(supp1. 2): S131-5. 7. Black RE. Epidemiology of travelers’ diarrhea and relative importance ofvarious pathogens. Rev Infect Dis 1990; 12(suppl. 1): s73-9. 8. Taylor DN, Blaser MJ, Blacklow N, et al. Polymicrobial etiology of travellers’ diarrhoea. Lancet 1985; 1: 381-3. 9. Taylor DN, Echeverria P. Etiology and epidemiology of travelers’ diarrhea in Asia. Rev Infect Dis 1986; 8(suppl 2): S136-41. 10. Levine MM. Escherickia coli that cause diarrhea: enterotoxigenic, enteropathogenic. enteroinvasive and enteroadherent. J Infect Dis 1987; 155: 377-89. 11. Karmali MA. Infection by verocytotoxin producing Esckerichia coli. Clin Microbiol Rev 1989; 2: 15-38. 12. Levine MM, Nataro JP, Karch H , et al. The diarrhoeal response of humans to some classic serotypes of enteropathogenlc Escherickia coli is dependent on a plasmid encoding an enteroadhesiveness factor. J Infect Dis 1985; 152: 550-9. 13. 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 Infect Dis 1985; 152: 56C-5. 14. Giron JA, Ho AS, Schoolnik GK. An inducible bundle-forming pilus of enteropathogenic Escherickia coli. Science 1991; 254: 710-13. 15. Jerse AE, Yu J, Tall BD, Kaper JB. A genetic locus of enteropathogenic Esckerichia coli necessary for the production of attaching and effacing lesions on tissue culture cells. Proc Natl Acad Sci USA 1990; 87: 7 8 3 9 4 3 . 16. Hale L, Formal SB. Pathogenesis of Shigella infections. Pathol Immunopathol Res 1987; 6: 117-27. 17. Marquez LRM, Moore MA, Wells JG, Wachsmuth IK, O’Brien AD. Production of Shiga-like toxln by Esckerichia coli. J Infect Dis

1986; 154: 338-41. 18. Nataro JP, Yikang D, Giron JA, Savarino SA, Kothary MH, Hall R. Patterns of adherence of diarrheogenic Esrherickia coli to HEp-2. Pediatr Infect Dis 1987; 6: 829-31. 19. Scaletsky I, Silvia ML, Trabulsi, LR. Distinctive patterns of adherence of enteropathogenic Esckerichia coli to HeLa cell. Infect Immun 1984; 45: 534-6.

688

C l i n i c a l M i c r o b i o l o g y a n d I n f e c t i o n , V o l u m e 4 N u m b e r 12, D e c e m b e r 1 9 9 8

20. Vial PA. Robins-Browne K, Lior H . et al. Characterization of enteroadherent-aggregative Escherichia coli, a putative agent of diarrheal disease. J Infect Dis 1988; 158: 70-9. 21. German Y, Begaud E, Ihval P, Le Bouguknec C. Prevalence of enteropathogenic. enteroaggregative and diffusely adherent Eseherichia coli among isolates from children with diarrhea in New Caledonia. J Infect Dis 1996; 174: 1124-6. 22. Le BouguGriec C , Archaiiibaud M, Labigne A. Rapid and specific detection of pap,
34. Forestier C, Meyer M, Favre-Bonte S, et al. Enteroadherent Escherichia coli and diarrhea in children: a perspective case-control study. J Clin Microbiol 1996; 34: 2897-903. 35. Cohen MB, Hawkins JA, Weckbach LS, Staneck JL. Levine MM, J. Heck E. Colonization by enteroaggegative Escheridiia coli in travelers and without diarrhea. J Clin Microbiol 1993; 31: 351-3. 36. Gascbn J, Vargas M, Quint6 L1 et al. Enteroaggregative Esrlierichia coli strain as a cause of traveler’s diarrhea. A prospective case-control study. J Infect Ilis 1998; 177: 1409-12. 37. Waters J R , Sharp JCM, and Dev VJ. Infection caused b>Esclierichia coli 0157:H7 in Alberta, Canada, and in Scotland: a five-year review, 1987-1991. Clm Infect Dis 1994; 19:834-43. 38. Allerberger F, Rossboth ll, Llierich MP, Aleksic S, Schmidt H, Karch H. Prevalence aiid clinical manifestation of Shiga toxinproducing Esclierichia coli infections in Austrian children. Eur J Clin Microbiol Infect Ilis 1996; 15: 535-50. 39. Levinc MM. Ferrercio C, Prado V, et al. Epidemiologic studies of Escherichia coli diarrheal infections in a low socioeconomic level peri-urban community in Santiago, Chile. Am J Epidemiol 1993; 138: 8 4 9 4 9 40. Begaud E, Jourand l’, Morillon M, Mondet D, Gcrmani Y. Iletectioii of diarrheagenic Escherichia coli in children lcrs than ten years old with and without diarrhea iii New Caledonia using seveii acetylaminofluorene-labeled DNA probes. Am J Trop Med Hyg 1993; 48: 26-34. 41. Mattila L, Siitonen A, Kyroriseppd H , et al. Seasonal vai-idtion i n aetiology of travelers’ diarrhea. J Infect Llis 1992; 165: 385-8. 42. Cravioto A. Tello A. Navarro A. Association of Eschcrichia c d i HEp-2 adherence patterns with type aiid duration of diarrhoea. Lancet 1991; 337: 262-4. 43. Tacket C O , Moseley ST, Kay U. Losonsky G , Lcvi~ie MM. Challenge studies i n volunteers using Exherichin ciik strains with diffuse adherence to HEp-2 cells. J Infect I l i s 1990: 162: 550-2. 44. Mathewson JJ, Johnson PC. Uupoiit HL. Satterwhite TK, Widsor DK. Pathogenicity of enteroadhereiit Esclicrichia ruli in adult volunteers. J Infect l l r 1986: 154. 524-7. 45. Gunzburg ST. Chang BJ, Elliot SJ, Burke V, Gracey M . Diffuse and enteroaggregative patterns of adherence of eiiteric Ercliericliia roli isolated from aboriginal chddren from the Kimberley region of Western Australia. J Infect Dis 1993: 167: 755-8. 46. Mathewson JJ, Johnson PC:, Dupont HL. A newly recognized cause of travelers’ diarrhea: enteroadhereiit Eschericiiin ruli. J Infect 111s 1985; 151. 471-5. 17. Giron JA, Donnenberg MS, Martin WC. J a rm KG, Kaper JU. Distribution of the bundle-forming pilus structure gene ([!&A) among enteropathogenic E d e r i c h i a roli. J Infect I l i s 1993; 168: 1037-41. 48. Merson MH, Sack lU3, Kibriva AKMG, et al. Use of colony pools for diagnosis of enterotoxigenic Esckcririira roli diarrhca. J Chi1 Microbiol 1979: 9: 403-7.