Clostridium difficile toxin A in its native form is a high molecular weight (520-540 K) aggregate with five major biological activities. It is lethal, enterotoxic, cytotoxic and cytotonic, and induces hemagglutination of rabbit red blood cells. Possibly these activities are contained in separate components. A major subunit of c. 230-310 K has been defined but lower molecular weight components cannot be excluded. The major component has been cloned, and sequence analysis indicated a complicated pattern of repeating sequences in the C-terminal third of the molecule. This review deals mainly with the effects of toxin A on cultured cells. Most mammalian cells are sensitive to toxin A whose major effect is to stop cell division irreversibly. The toxin binds via its repeat sequences to a trisaccharide receptor expressed on rabbit red cells and on brush border membranes from hamster intestine. This receptor seems to be functional in the hemagglutination reaction and the enterotoxicity. Its role in the cytotoxic effect of the toxin is not clear, but no other receptor structure has as yet been identified. In order to exert its cytotoxic (antiproliferative) effect toxin A must first be internalized by endocytosis. Thus a latency period of at least 30 min after toxin binding to cells is consistently observed, and all cytotoxic effects can be prevented by blocking the endocytosis pathway. The first microscopically visible signs of cytotoxicity consist in retraction and rounding of intoxicated cells. In addition the nucleus becomes polarized to one side of the cell while other cell organelles are not significantly affected. These morphological changes seem to be the consequence of a cytoskeletal rearrangement, mainly involving some components of the microfilament system. Inhibition of macromolecular syntheses as well as permeabilization of the plasma membrane may follow the early cytoskeletal effects and finally lead to cell death. Attempts to identify metabolic pathways of significance in the cytotoxicity suggest that the cytosolic level of Ca2+ is not important, thus excluding certain mechanisms for cell killing. In this respect the cytotoxic mode of action of toxin A clearly differs from that of toxin B. However, the biochemical basis for the antiproliferative effect of toxin A remains unknown.