Diquat produces hepatic necrosis in the Fischer-344 rat, and although reactive oxygen species generated by redox cycling are thought to mediate the damage, the possibility that covalent binding of diquat or diquat metabolites to tissue macromolecules contributes to the observed hepatotoxicity of diquat needed to be examined experimentally. Intraperitoneal administration of [ethylene-14C]diquat (0.1 mmol/kg) results in distribution of radioactivity to all organs examined. Measurable radioactivity remains associated with hepatic and renal protein even after extensive solvent extraction, but the amount (12-16 pmol/mg protein) is 100-fold less than the extent of covalent binding observed with comparably hepatotoxic doses of other hepatotoxins such as acetaminophen and bromobenzene. Similarly, although small amounts of radioactivity remain in Folch lipid extracts of liver and kidney (56-179 pmol/mg lipid), this is virtually completely removed by transesterification of the lipid (less than 5 pmol/mg lipid), indicating that the radioactivity does not represent an alkylation of electroneutral alkyl residues of the lipid. The diquat-induced increase in biliary excretion of glutathione disulfide temporally parallels the biliary excretion of radioactivity. Although the present results do not prove the absence of a contribution by alkylation mechanisms to diquat hepatotoxicity, the extremely low upper limits placed on possible covalent interactions reinforce the confidence with which the diquat-treated Fischer-344 rat can be employed as an animal model for mechanisms of cell death mediated by reactive oxygen species.