Glutathione (GSH) plays a major role in cytoprotection, acting as a nucleophile trap for reactive species derived from xenobiotics. This has led to the development of an assay for the detection of reactive species generated by liver microsomal metabolism of xenobiotics. This assay has been used extensively to study reactive metabolites which initiate toxicity through a direct (non-immunological) mechanism, but there are few data on its ability to detect reactive metabolites that initiate toxicity through neo-antigen formation, or to detect xenobiotics that cause GSH loss by oxidation mediated by a redox cycling process. Accordingly, the ability of rat and human liver microsomes to metabolize xenobiotics to GSH-depleting metabolites has been investigated further. Of the five neo-antigen-forming xenobiotics tested, four (amodiaquine, phenobarbitone, procainamide, and sulphanilamide) displayed GSH reactivity that was either dependent or independent (amodiaquine) on metabolism. The other neo-antigen-forming xenobiotic (carbamazepine) was inactive in all microsomal samples tested. Four quinones believed to exert toxcity through arylation (1,4-benzoquinone) and/or redox cycling (duroquinone, menadione, mitomycin c) displayed GSH reactivity, as did nitrofurantoin and diquat, two other redox cycling xenobiotics. Induction of the mixed function oxidase system with Aroclor afforded little advantage when using rat liver microsomes, whilst there was considerable inter-individual variation in the ability of human liver microsomes to mediate metabolism-dependent GSH depletion. It is concluded that the liver microsome GSH depletion assay may be of general utility as a screen for a number of xenobiotic-derived reactive species.