This study addressed the hypothesis that the vestibular or lethal toxicities of allylnitrile depend on CYP2E1-mediated bioactivation. Wild-type (129S1) and CYP2E1-null male mice were exposed to allylnitrile at doses of 0, 0.5, 0.75, or 1.0 mmol/kg (po), following exposure to drinking water with 0 or 1% acetone, which induces CYP2E1 expression. Induction of CYP2E1 activity by acetone in 129S1 mice and lack of activity in null mice was confirmed in liver microsomes. Vestibular toxicity was assessed using a behavioral test battery and illustrated by scanning electron microscopy observation of the sensory epithelia. In parallel groups, concentrations of allylnitrile and cyanide were assessed in blood after exposure to 0.75 mmol/kg of allylnitrile. Following allylnitrile exposure, mortality was lower in CYP2E1-null than in 129S1 mice, and increased after acetone pretreatment only in 129S1 mice. This increase was associated with higher blood concentrations of cyanide. In contrast, no consistent differences were recorded in vestibular toxicity between 129S1 and CYP2E1-null mice, and between animals pretreated with acetone or not. Additional experiments evaluated the effect on the toxicity of 1.0 mmol/kg allylnitrile of the nonselective P450 inhibitor, 1-aminobenzotriazole, the CYP2E1-inhibitor, diallylsulfide, and the CYP2A5 inhibitor, methoxsalen. In 129S1 mice, aminobenzotriazole decreased both mortality and vestibular toxicity, whereas diallylsulfide decreased mortality only. In CYP2E1-null mice, aminobenzotriazole and methoxsalen, but not diallylsulfide, blocked allylnitrile-induced vestibular toxicity. We conclude that CYP2E1-mediated metabolism of allylnitrile leads to cyanide release and acute mortality, probably through alpha-carbon hydroxylation, and hypothesize that epoxidation of the beta-gamma double bond by CYP2A5 mediates vestibular toxicity.