Virulence characteristics of diarrheal outbreak-associated Escherichia coli O55:NM, O126:NM, and O111:NM were examined. The E. coli O55:NM strains were atypical enteropathogenic E. coli (EPEC), while the E. coli O126:NM and O111:NM strains should be classified as enteroaggregative E. coli (EAggEC). The contributions of EPEC and EAggEC to the human disease burden in Japan might be significantly greater than is currently appreciated.There are six categories of Escherichia coli that cause diarrhea: enterotoxigenic E. coli (ETEC), enteropathogenic E. coli (EPEC), enterohemorrhagic E. coli, enteroaggregative E. coli (EAggEC), enteroinvasive E. coli, and diffusely adherent E. coli (21). EPEC causes characteristic attaching-and-effacing lesions (A/E), which can be observed by intestinal biopsy in both human patient (19) and animal (29) models. A/E is characterized by loss of microvilli, intimate adherence of bacteria between epithelial cell membranes (27,30), and cytoskeletal changes such as actin polymerization directly beneath the adherent bacteria (15). Generally, EPEC causes infantile diarrhea in developing countries and sporadic diarrhea in developed countries (21). EAggEC, on the other hand, is an enteric pathogen defined by its distinctive aggregative or "stackedbrick" pattern of adherence to cultured human epithelial cells (22). EAggEC associates mainly with persistent diarrhea in developing countries (21). Only two reports in Japan have described diarrheal outbreaks caused by EAggEC or EPEC. Itoh et al. (11) reported the isolation of EAggEC from the stools of patients with severe diarrhea in elementary and junior high schools. Makino et al. (18) reported the isolation of EPEC from a mass outbreak. In this paper, we describe three cases of diarrheal outbreaks in Japan caused by E. coli belonging to the traditional EPEC serotype.Chromosomal DNA-embedded agarose plugs for pulsedfield gel electrophoresis (PFGE) analysis were prepared by using the CHEF Bacterial DNA Plug Kit (Bio-Rad, Hercules, Calif.) and were digested with XbaI (Nippon gene; Osaka, Japan) at a concentration of 30 U/plug for 4 h at 37°C. The plugs were applied to a 1% PFC Grade Agarose (Bio-Rad) gel. Electrophoresis was performed in 0.5ϫ Tris-Borate EDTA buffer at 14°C using a CHEF DR-II PFGE apparatus (BioRad) under the following conditions: voltage, 6 V/cm; block 1, 11 h, with initial switching time of 4 s to final switching time of 8 s; block 2, 9 h, with initial switching time of 8 s to final switching time of 50 s. The HEp-2 cell assay was performed following the method described by Craviotto et al. (4), with modifications involving 3 or 6 h of incubation (15). The E. coli isolates were examined for the presence of the following virulence genes by PCR: stx1 (Shiga toxin) and stx2 (16), eaeA (E. coli attaching and effacing) (12), bfpA (bundle-forming pilus) (9), perA (EPEC plasmid-encoded regulatory region) (8), astA (EAggEC heat-stable enterotoxin) (28), aggR (transcriptional activator for EAggEC aggregative adherence fimbria I expression) ...