Abstract

The herbicide 2,6-dichlorobenzonitrile (DCBN) is known to cause tissue-specific toxicity at very low doses in the olfactory mucosa of rodents. The toxicity of DCBN is reportedly cytochrome P450 (P450) dependent, but the isoforms involved have not been identified, and the effects of this agent on humans are not known. In the present study, DCBN metabolism was examined with microsomes and with purified P450s in a reconstituted system. Rat and rabbit olfactory microsomes act on DCBN to form DCBN-protein adducts as well as two metabolite peaks, designated M1 and M2, identified through high performance liquid chromatography with radiometric detection. The activity of rat olfactory microsomes in DCBN metabolism is much higher than that of liver or lung microsomes. Of seven purified rabbit P450s known to be expressed in the olfactory mucosa, including 1A2, 2A10/11, 2B4, 2E1, 2G1, and 3A6, the 2A10/11 preparation is the most active, producing M2 as well as DCBN-protein adducts; P450 2E1 is the only other active isoform. The addition of purified epoxide hydrolase (EC 4.2.1.63) to the reconstituted enzyme system leads to the formation of M1 and decreased formation of M2. It seems that M1 and M2 are derived from an epoxide intermediate that also forms covalent protein adducts. Gas chromatography- and liquid chromatography-mass spectrometry analyses of nasal microsomal DCBN metabolites and DCBN-glutathione conjugates indicated that the major reactive intermediate may be 2,3-oxo-DCBN and that M1 may be 2,3-dihydroxy-6-chlorobenzonitrile, whereas M2 may correspond to a monohydroxy-DCBN. Interestingly, heterologously expressed human P450s 2A6 and 2E1, but not 1A2, are active in the metabolism of DCBN, forming protein adducts as well as M2. Thus, the preferential expression of P450s of the 2A subfamily in olfactory tissue suggests a molecular basis for the tissue-specific toxicity of the herbicide and may have important implications for risk assessment in humans.