All of the selective COX-2 inhibitors described to date inhibit the isoform by binding tightly but noncovalently at the substrate binding site. Recently, we reported the first account of selective covalent modification of COX-2 by a novel inactivator, 2-acetoxyphenyl hept-2-ynyl sulfide (70) (Science 1998, 280, 1268-1270). Compound 70 selectively inactivates COX-2 by acetylating the same serine residue that aspirin acetylates. This paper describes the extensive structure-activity relationship (SAR) studies on the initial lead compound 2-acetoxyphenyl methyl sulfide (36) that led to the discovery of 70. Extension of the S-alkyl chain in 36 with higher alkyl homologues led to significant increases in inhibitory potency. The heptyl chain in 2-acetoxyphenyl heptyl sulfide (46) was optimum for COX-2 inhibitory potency, and introduction of a triple bond in the heptyl chain (compound 70) led to further increments in potency and selectivity. The alkynyl analogues were more potent and selective COX-2 inhibitors than the corresponding alkyl homologues. Sulfides were more potent and selective COX-2 inhibitors than the corresponding sulfoxides or sulfones or other heteroatom-containing compounds. In addition to inhibiting purified COX-2, 36, 46, and 70 also inhibited COX-2 activity in murine macrophages. Analogue 36 which displayed moderate potency and selectivity against purified human COX-2 was a potent inhibitor of COX-2 activity in the mouse macrophages. Tryptic digestion and peptide mapping of COX-2 reacted with [1-14C-acetyl]-36 indicated that selective COX-2 inhibition by 36 also resulted in the acetylation of Ser516. That COX-2 inhibition by aspirin resulted from the acetylation of Ser516 was confirmed by tryptic digestion and peptide mapping of COX-2 labeled with [1-14C-acetyl]salicyclic acid. The efficacy of the sulfides in inhibiting COX-2 activity in inflammatory cells, our recent results on the selectivity of 70 in attenuating growth of COX-2-expressing colon cancer cells, and its selectivity for inhibition of COX-2 over COX-1 in vivo indicate that this novel class of covalent modifiers may serve as potential therapeutic agents in inflammatory and proliferative disorders.