Data obtained from human studies in vivo show that the dispositions of the tricyclic antidepressant drugs desmethylimipramine (DMI) and nortriptyline are related to the debrisoquine hydroxylation phenotype. To obtain insight into the enzymic mechanisms behind this, the metabolism of debrisoquine and antidepressant drugs by human liver preparations have been studied. The 2-hydroxylation of DMI in vitro correlates with the 4-hydroxylation of debrisoquine among various livers (rs = 0.90). Debrisoquine inhibits DMI hydroxylation competitively, and DMI inhibits debrisoquine hydroxylation, suggesting that DMI hydroxylation is catalysed by the debrisoquine hydroxylase in human liver. By monitoring the hydroxylation of DMI in various fractions during separation and purification of cytochrome P-450 from human liver microsomes we have purified a cytochrome P-450 which efficiently hydroxylates this drug. The apparently electrophoretically homogeneous enzyme had a molecular weight of 51,500 and hydroxylated DMI and debrisoquine at rates of up to 0.95 and 0.45 nmol/min . nmol P-450, respectively. This is probably the major debrisoquine hydroxylating cytochrome P-450 in man. Nortriptyline 10-hydroxylation correlates strongly (r = 0.96) with debrisoquine hydroxylation in human liver microsomes. Nortriptyline inhibits DMI-hydroxylation competitively, and the drug also inhibits the 4-hydroxylation of debrisoquine. Thus it is probable that nortriptyline is hydroxylated by debrisoquine hydroxylase. Imipramine N-demethylation did not correlate significantly (P greater than 0.1) with debrisoquine hydroxylation among microsomes from nine livers. However, if a liver from a subject, which was a poor metabolizer of debrisoquine in vivo, was included, a correlation was obtained (r = 0.79, P less than 0.01, N = 10). Imipramine demethylation also correlated with DMI-hydroxylation only if the 'poor metabolizer' liver was included (r = 0.75, P less than 0.05, N = 10). Debrisoquine inhibited imipramine demethylation competitively. The data indicate that imipramine can interact with debrisoquine- and DMI-hydroxylase, but it is uncertain if this enzyme plays an important quantitative role in its demethylation. Ethoxyresorufin O-deethylation correlated with DMI hydroxylation (r = 0.80) in human liver preparations, and DMI inhibited the former reaction in what is probably a mixed competitive-non-competitive inhibition. Liver preparations from a subject who was a poor oxidizer of debrisoquine both in vivo and in vitro had unusually low capacity to metabolize ethoxyresorufin. Thus ethoxyresorufin, at least partly, seems to interact with an enzyme that can metabolize DMI in human liver.(ABSTRACT TRUNCATED AT 400 WORDS)