Because of the importance of the inactivation of GABA aminotransferase to the design of anticonvulsant agents, a seemingly wide variety of inactivators has been investigated; all of the compounds, however, are analogues of GABA, beta-alanine, or delta-aminovaleric acid, which are substrates for the enzyme. Relatively minor modifications in the inactivator structures result in major differences in inactivation mechanisms and enzyme adduct structures. Compounds that inactivate GABA aminotransferase by a Michael addition mechanism, leading to modification of an active-site residue are Class I inactivators. Those that proceed by an enamine mechanism and give ternary adducts are Class II inactivators. Class III inactivators modify only the PLP cofactor; if the inactivation involves aromatization of the inactivator, it is a Class IIIA inactivation, and if no aromatization is involved, then it is a Class IIIB inactivation. The last class of inactivators (Class IV) are not classified on the basis of the mechanism, but, rather, that they require the enzyme to be in the PMP form. There appears to be no trend in partition ratio values when comparing Class I with Class II inactivators. Class III inactivations alter only the cofactor, so it may be possible for these adducts to diffuse slowly out of the active site; reactivation of the apoenzyme would require additional PLP. These inactivators also inactivate a variety of other PLP-dependent enzymes. At this point there does not seem to be a therapeutic advantage of one class of inactivators over another, although the only current example of these inactivators to be useful clinically is gamma-vinyl GABA (vigabatrin), a Class I inactivator recently approved for the drug market in France and the U.K. There is a mechanistic significance, however, for one class over another. If labeling of an active-site amino acid residue is desired, then Class I inactivators should be selected; desire for attachment of the inactivator to both the protein and the cofactor or just to the cofactor would determine whether Class II or Class III inactivators would be chosen. The classification presented here should allow us to think about inactivator structures in terms of their mechanistic potential and, as a result of this, should afford us the opportunity to be able to make predictions regarding inactivation mechanisms for hypothetical new structural classes of inactivators. Since the different mechanistic pathways lead to different types of enzyme adducts, inactivator design may be driven by the class of adduct that is desired.(ABSTRACT TRUNCATED AT 400 WORDS)