Protein serine/threonine phosphatases (PP1, PP2A and PP2B) play important roles in intracellular signal transductions. The immunosuppressant drugs FK506 and cyclosporin A (CsA) bind to immunophilins, and these complexes selectively inhibit PP2B (calcineurin), leading to the suppression of T-cell proliferation. Both FK506 and CsA must, however, form complexes with immunophilins to exert their inhibitory action on PP2B. Thus, it is of interest to find a direct and selective inhibitor of PP2B that does not involve the immunophilins as a biological tool for studies of PP2B and also as a candidate therapeutic agent. We selected the simple natural product cantharidin, a known PP2A-selective inhibitor, as a lead compound for this project. Primary SAR indicated that norcantharidin (7-oxabicyclo[2.2.1]heptane-2,3-dicarboxylic anhydride) inhibits not only PP1 and PP2A but also PP2B, and a binding model of norcantharidin carboxylate to the PP2B catalytic site was computationally constructed. Based on this binding model, we designed and synthesized several cantharidin derivatives. Among these compounds, 1,5-dibenzoyloxymethyl-substituted norcantharidin was found to inhibit PP2B without inhibiting PP1 or PP2A. To our knowledge, this is the first highly selective catalytic site-directed inhibitor of PP2B.