There is a need for methodology to predict clinically significant drug-drug interactions so that clinical studies can be directed toward interactions which are likely to be clinically relevant. To this end, we evaluated selective assays for the seven drug-metabolizing cytochrome P450 (P450) isozymes 1A2 (caffeine N3-demethylation), 2A6 (coumarin 7-hydroxylation), 2C9 (tolbutamide hydroxylation), 2C19 (S-mephenytoin 4-hydroxylation), 2D6 (dextromethorphan O-demethylation), 2E1 (chlorzoxazone 6-hydroxylation), and 3A4/5 (dextromethorphan N-demethylation). Using initial rate conditions, we determined the Km and Vmax values of each reaction in human liver microsomes from three individuals. Because organic solvents (usually methanol) are frequently used as solubilization aids for drugs/inhibitors, we also screened several solvents for inhibitory activity. Methanol was the least inhibitory toward P450s 2A6, 2D6, and 3A4, dimethylformamide was the least inhibitory toward P450s 1A2 and 2C9, and acetonitrile was the least inhibitory toward P450s 2C19 and 2E1. Using substrate concentrations close to the determined Km and an appropriate solvent (where necessary), we used the selective inhibitors furafylline (1A2), 8-methoxypsoralen (2A6), sulfaphenazole (2C9), S-mephenytoin (2C19), quinidine (2D6), diethyldithiocarbamate (2E1), and troleandomycin (3A4) to assess the limitations of each probe assay as an indicator of the P450 isoform in question. Our results were consistent with these inhibitors and probes, being selective tools for studying P450 drug metabolism.