With the increased availability of human liver tissue, recombinant (cDNA-expressed) cytochrome P450 proteins (rCYPs), and knowledge of the human CYP pool (e.g. immunoquantitated levels of each CYP form in native liver microsomes), it is now possible to carry out in vitro "CYP reaction phenotyping" in an integrated manner. Reaction phenotyping allows one to identify which CYP form(s) is (are) involved in the metabolism of a given drug, using a combination of data obtained with native human liver microsomes and rCYP proteins. The following describes how one can attempt to integrate such data. A total of ten drugs are included in the analysis, represented by twelve reactions (six hydroxylations, two O-demethylations, one N-demethylation, one O-deethylation, and two sulfoxidations) that are largely catalyzed (> or =20%) by various combinations of CYPs (CYP3A4, CYP2C9, CYP1A2, and CYP2D6), and characterized by a wide range of apparent Km values (12-820 microM). Briefly, reaction rates measured with individual rCYPs are normalized with respect to the nominal specific content of the corresponding CYP in native human liver microsomes. In turn, the normalized rates for each rCYP are summed, yielding a "total normalized rate" (TNR), and the normalized rate for each rCYP is expressed as a percent of the TNR (% TNR). Finally, % TNR is related to inhibition (percent inhibition in the presence of CYP form selective chemical inhibitors; % I) and univariate regression analysis (r > or = 0.63; P < or = 0.05; N > or = 10 different livers) data obtained with native human liver microsomes. Therefore, the reaction phenotype of a drug is assigned by integrating all three data sets (r, % TNR, and % I).