Characteristically, enterohemorrhagic Escherichia coli (EHEC) strains produce Shiga-like toxin type I (SLT-I), SLT-II, or both of these immunologically distinct cytotoxins. No antigenic or receptor-binding variants of SLT-I have been identified, but a number of SLT-II-related toxins have been described. Because EHEC O91:H21 strain B2F1, which produces two SLT-II-related toxins, is exquisitely virulent in an orally infected, streptomycin-treated mouse model (oral 50% lethal dose [LD50], < 10 organisms), we asked whether the pathogenicity of strain B2F1 was a consequence of SLT-II-related toxin production. For this purpose, we compared the lethality of orally administered E. coli DH5 alpha (Strr) strains that produced different cytotoxic levels of SLT-II, SLT-IIvha (cloned from B2F1), SLT-IIvhb (also cloned from B2F1), or SLT-IIc (cloned from EHEC O157:H7 strain E32511) on Vero cells. We also calculated the specific activities of purified SLT-IIvhb and SLT-II in intraperitoneally injected mice and on Vero cells. The two purified toxins were equally toxic for mice, but SLT-IIvhb was approximately 100-fold less active than SLT-II on Vero cells and bound to the glycolipid receptor Gb3 with lower affinity than did SLT-II. In addition, characterization of SLT-II-related toxin-binding (B) subunit mutants generated in this study revealed that the reduced in vitro cytotoxic levels of the SLT-II-related toxins were due to Asn-16 in the B subunit. Taken together, these findings do not support the idea that B2F1 is uniquely virulent because of the in vivo toxicity of SLT-II-related toxins but do demonstrate differences in in vitro cytotoxic activity among the SLT-II group produced by human EHEC isolates.