Glutathione-S-transferase (GST) isozyme-selective inhibitors were designed by an empirically guided strategy. In the first phase, literature data were used to select C-terminal modifications which generated maximum variation in the catalytic efficiency (Vmax/Km) for glutathione (GSH) analogs used as substrates with different rat GSTs. Also, on the basis of literature data, the sulfhydryl group was functionalized with a selection of alkyl and aryl groups to maximize potential isozyme specificity. Affinity chromatography sorbents were prepared from these which showed isozyme selectivity for both rat tissue and recombinant human GST isozymes. Some of these compounds also showed selective inhibition of GST activity in catalysis of the reaction of 1-chloro-2,4-dinitrobenzene with GSH. In the second phase, electronic effects were explored through synthesis of an isostructural series of S-benzyl GSH ligands with different substituents on the aromatic ring. GST isozyme specificity for these ligands, measured by binding to derivatized sorbents, varied substantially, with hydrophobic substituents favoring the human GST M1a isozyme and electronegative moieties favoring GST P1. In the third phase, information obtained from testing both series of compounds was combined and used to prepare GSH analogs with chemical features responsible for isozyme specificity at both the C-terminus and the sulfur. This approach gave two new compounds which showed improved potency while still maintaining selectivity in the inhibition of GSTs. A detailed discussion of the logic used in the selection of functional groups for maximum potency and selectivity is included.