Our previous studies have shown that 2,3-epoxy-4-hydroxynonanal, a reactive epoxy aldehyde capable of forming etheno adducts with DNA bases, is mutagenic and tumorigenic (Carcinogenesis, 14, 2073). The epoxy aldehyde can be generated from trans-4-hydroxy-2-nonenal, a lipid peroxidation product of omega-6 polyunsaturated fatty acids, by autoxidation or by incubation with fatty acid hydroperoxides or hydrogen peroxides (Chem. Res. Toxicol., 9, 306). These are plausible in vivo pathways for the formation of 2,3-epoxy-4-hydroxynonanal. The possibility that 2,3-epoxy-4-hydroxynonanal is a tumorigen of endogenous origin is suggested by recent observations that etheno bases are detected as background DNA lesions in untreated rodents and humans. A metabolic pathway critical for detoxification of 2,3-epoxy-4-hydroxynonanal involves the ring-opening by epoxide hydrolase, which abolishes its ability to form cyclic etheno DNA adducts. In this study, we examined whether 2,3-epoxy-4-hydroxynonanal is a substrate of cDNA expressed human epoxide hydrolase. Human epoxide hydrolase was expressed in TK- 143 cells (thymidine kinase-deficient human embryoblast) infected with recombinant vaccinia virus encoding human epoxide hydrolase cDNA. Controls consisted of the cells infected with vaccinia virus in the absence of human epoxide hydrolase cDNA. No hydrolysis occurred when [2,3-(3)H]2,3-epoxy-4-hydroxynonanal was incubated at 37 degrees C for 30 min at pH 7.4 with cells expressing human epoxide hydrolase, as indicated by the presence of a pair of radioactive peaks in reversed-phase HPLC chromatography, which comigrated with the UV standards of the two diastereomers of the epoxy aldehyde. The identity of these compounds as the intact epoxy aldehyde was further supported by derivatization to the 2,4-dinitrophenylhydrazones followed by reversed phase HPLC analysis. Similar results were observed with the control cells or with the heat deactivated human epoxide hydrolase. The epoxide hydrolase activity in the expressed cells was demonstrated by their ability to convert benzo[a]pyrene-4,5-dihydroepoxide to benzo[a]pyrene-trans-4,5-dihydrodiol under the same conditions. These results clearly indicate that 2,3-epoxy-4-hydroxynonanal is not a substrate of human epoxide hydrolase, and, thus strengthen its possible endogenous role in the formation of promutagenic exocyclic etheno adducts in vivo.