The protein processing enzyme, methionine aminopeptidase-2 (MetAP-2), has been identified as a molecular target of fumagillin and its derivative, TNP-470, compounds known to inhibit endothelial cell proliferation and angiogenesis. A high-throughput screening program was undertaken to identify selective, reversible inhibitors of MetAP-2 in an attempt to discover structurally novel anti-angiogenic agents for potential therapeutic use in oncology. Approximately 90 small-molecule, reversible, selective inhibitors of rhMetAP-2 were identified. The most potent of these compounds contained a singly-substituted triazole moiety which exhibited an IC50 of 8 nM (95% confidence limits 5 to 13 nM) and was highly selective for MetAP-2 over MetAP-1 (approximately 60-fold difference in IC50 values). Unlike fumagillin, these MetAP-2 inhibitors failed to significantly inhibit growth factor-stimulated endothelial cell (HUVEC) proliferation or to suppress angiogenesis in the in vitro aortic ring explant model of microvessel outgrowth. The MetAP-2-inhibitory activity of these compounds was dependent on the divalent cation used as the metalloenzyme activating cofactor for MetAP-2. These inhibitors were identified using cobalt(II)-activated recombinant human MetAP-2 for screening compound libraries. When manganese (Mn2+) was substituted for cobalt following EDTA treatment and extensive dialysis of the MetAP-2 protein, these inhibitors were significantly less potent (40-fold increase in IC50) as inhibitors of MetAP-2. These results support the recent hypothesis that cobalt may not be the relevant divalent metal ion cofactor for MetAP-2 in cells and may explain the observed absence of cell-based activity using potent triazole inhibitors of cobalt-activated MetAP-2.