Verotoxins (VT) are a family of Escherichia coli-derived toxins which have been associated with hemolytic uremic syndrome, the leading cause of acute pediatric renal failure, and hemorrhagic colitis. Verotoxins (VT1 and VT2c) both show terminal gal alpha 1-4gal-dependent binding to globotriaosylceramide (Gal alpha 1-4Gal beta 1-4Glc-Cer; Gb3), yet VT2c shows a thousandfold lower specific cytotoxic activity in vitro. Our previous studies have shown this discrepancy is a function of the receptor binding B subunit and that VT1/Gb3 binding in a lipid matrix is affected by heterogeneity in the ceramide fatty acid chain length. The influence of the fatty acid composition of Gb3 on the binding of VT1 and VT2c has now been compared using 14 homogeneous semisynthetic Gb3 molecular species of differing fatty acid chain length and degree of saturation from C12 to C24. The binding of verotoxin was quantitated by Scatchard analysis using a solid-phase binding assay in the presence of auxiliary lipids, which may in some respects approximate to receptor function within the plasma membrane of sensitive cells. Differential binding was observed for several of these species in the lipid matrix, indicating that the fatty acid moiety of Gb3 is important for VT binding under such conditions. The short chain fatty acid containing Gb3 (C12 and C14) showed minimal binding. Middle and long chain fatty acid Gb3 homologues (C16, C18, C20, C22, and C24) were effectively recognized by VTs. The presence of an unsaturated fatty acid in Gb3 significantly increased VT binding in all cases. C20:0 and C22:1 containing Gb3 had the greatest capacity to bind VT1. In contrast, C18:0 and C18:1 homologues showed the greatest capacity for VT2c binding (higher than VT1). These results were, in general, reflected in cell cytotoxicity in that receptor-deficient cells reconstituted with C22:1Gb3 were maximally sensitive to VT1 in vitro whereas cells reconstituted with C18:1Gb3 were maximally sensitive to VT2c. VT2c was an ineffective inhibitor of 125I-VT1 binding to C22:1 Gb3 but in contrast, more effective than VT1 to compete binding to C18:1 Gb3. Similarly, VT1 was less effective than VT2c to compete binding of 125I-VT2c to C18:1 but more effective than VT2c to compete for C22:1 Gb3 binding. These results suggest that VT1 and VT2c bind selectively to different but overlapping carbohydrate epitopes on the Gb3 molecule which are differentially available in these Gb3 fatty acid homologues in a lipid environment.(ABSTRACT TRUNCATED AT 400 WORDS)