Possible reasons for the observed differences in metabolic behavior and drug interaction liability between the structurally similar oral anticoagulants warfarin and phenprocoumon were explored. Incubating (S)-phenprocoumon with human liver microsomes and cDNA-expressed CYP2C9 and determining its metabolism both in the absence and presence of the CYP2C9 inhibitor, sulfaphenazole, confirmed that phenprocoumon is a substrate for CYP2C9. Comparing the metabolic behavior of (S)- and (R)-warfarin, (S)- and (R)-phenprocoumon, and fixed structural mimics of the various tautomeric forms [(S)- and (R)-4-methoxyphenprocoumon, (S)- and (R)-2-methoxyphenprocoumon, (S)- and (R)-4-methoxywarfarin, (S)- and (R)-2-methoxywarfarin, and 9(S)- and 9(R)-cyclocoumarol] available to these two drugs with expressed CYP2C9 provides compelling evidence indicating that the ring closed form of (S)-warfarin and the ring opened anionic form of (S)-phenprocoumon are the major and specific structural forms of the two drugs that interact with the active site of CYP2C9. The conclusion that (S)-warfarin and (S)-phenprocoumon interact with CYP2C9 in very different structural states provides a clear basis for the significant differences observed in their metabolic profiles. Moreover, in accord with a previously established CoMFA model these results are consistent with the hypothesis that the active site of CYP2C9 possesses at least two major substrate binding sites, a pi-stacking site for aromatic rings and an ionic binding site for organic anions. An additional electrostatic binding site also appears to contribute to the orientation of coumarin analogs in the CYP2C9 active site by interacting with the C2-carbonyl group of the coumarin nucleus.
Copyright 1999 Academic Press.