Clostridium difficile and its toxins are implicated as the cause of pseudomembranous colitis in patients undergoing antibiotic therapy. Very little information is known about how these toxins bind to cells and cause disease. In an attempt to understand how these toxins work this dissertation research was undertaken to determine whether a receptor for C. difficile enterotoxin (toxin A) exists in the brush border membranes (BBMs) of the hamster, an animal known to be extremely sensitive to the action of the toxin. Toxin A was the only antigen adsorbed by the BBMs from the culture filtrate of C. difficile. Erythrocytes from various animal species were also examined for binding activity. Only rabbit erythrocytes could bind the toxin, and the cells agglutinated. A binding assay based on an enzyme-linked immunosorbent assay method for quantifying C. difficile toxin A was used to compare binding of the toxin to hamster BBMS, rabbit erythrocytes, and to BBMs from rats, which are less susceptible to the action of C. difficile toxin A than hamsters. Results of this comparison indicated the following order of toxin binding frequency: rabbit erythrocytes > hamster BBMs > rat BBMs. Binding of toxin A to BBMS from hamsters at 37°C was comparable to what has been observed with cholera toxin, but binding was enhanced at 4°C. A similar binding phenomenon was observed with rabbit erythrocytes. Examination of the cell surfaces of hamster BBMs and rabbit erythrocytes with lectins and specific glycosidases revealed a high concentration of terminal α-linked galactose. Treatment of both membrane types with α-galactosidase destroyed the binding activity. The glycoprotein, calf thyroglobulin, also bound the toxin and inhibited toxin binding to cells. An efficient, single-step method for isolation of highly purified toxin A from C. difficile culture filtrate has been developed based on toxin association with calf thyroglobulin. Toxin A did not bind to human erythrocytes from blood group A, B, or O donors. However, after fucosidase treatment of human erythrocytes, only blood group B erythrocytes could bind the toxin. This indicated that toxin A was likely binding to Ga1α1-3Ga1ß1-4GlcNAc, a carbohydrate sequence also found on calf thyroglobulin and rabbit erythrocytes. All of the results indicate that hamster BBMs contain a carbohydrate binding site for toxin A that has at least a Ga1α1—3Ga1ßl-4GlcNAc nonreducing terminal sequence. / Ph. D. / incomplete_metadata