The aim of the study was to identify the cytochrome P450s (CYPs) that catalyze the biotransformation of clomipramine in vitro. A high-performance liquid chromatography method was developed to assay N-desmethylclomipramine, 8-hydroxyclomipramine, 2-hydroxyclomipramine, 8-hydroxydesmethhylclomipramine, didesmethylclomipramine and 2-hydroxydesmethylclomipramine formed by microsomes prepared from human liver and yeast expressing human CYP1A1, 1A2, 2C8, 2C9, 2C18, 2C19, 2D6 and 3A4. There was a statistically significant correlation between the formation rate of desmethylclomipramine and the immunoquantified concentration of CYP3A4 in 12 human liver microsome preparations (rs = 0.664, P = .028). Ketoconazole was a very potent inhibitor of desmethylclomipramine formation (Ki = 0.054 microM) and microsomes from yeast expressing CYP3A4 were also active in forming the metabolite (formation rate: 25.6 nmol/nmol of CYP per hr). Thus, the results are consistent with the assumption that the N-demethylation of clomipramine is catalyzed by CYP3A4. As expected from in vivo panel studies, CYP2C19 in yeast was also very active in the N-demethylation (formation rate, 145 nmol/nmol of CYP per hr). Fluvoxamine was a potent inhibitor of desmethylclomipramine formation (Ki, 0.15 microM), suggesting that CYP1A2 is a third CYP involved in the N-demethylation. CYP2D6 in yeast microsomes catalyzed the 8-hydroxylation of clomipramine and desmethylclomipramine (formation rates, 65 and 75 nmol/nmol of CYP per hr) and quinidine was a very potent inhibitor (Ki, 0.10 and 0.16 microM). Both results confirm that CYP2D6 catalyzes the 8-hydroxylation in agreement with the results obtained in previous in vivo studies. Besides quinidine, paroxetine, fluoxetine and norfluoxetine, all were potent inhibitors of the 8-hydroxylations (Ki, 0.24-1.5 microM) and sertraline was a less potent inhibitor (Ki, 16 and 27 microM, respectively).