Many toxic effects are not caused by the administered compound itself, but are due to metabolites. All cell types express some xenobiotic-metabolizing enzymes, but levels and patterns are very variable. Critical metabolic steps may occur within the target cell and/or at other sites. This complex situation is difficult to mimic in vitro. The further problem is that cells that are taken into culture tend to rapidly cease the expression of important xenobiotic-metabolizing enzymes. Part of the problem may be solved by the addition of exogenous metabolizing systems, for example, in the form of freshly isolated hepatocytes, crude subcellular preparations, or purified enzymes. In these systems, the plasma membrane of the target cell may act as a barrier for the active metabolite and thereby lead to false negative results. The alternative is the use of metabolically active target cells. We therefore screened 18 cell lines for monooxygenase, cytochrome P-450 reductase, epoxide hydrolase, glutathione transferase, and UDP-glucuronosyl transferase activities. In further studies, IEC-17, IEC-18, and HuFoe-15 cells showed their capabilities of activating a broad spectrum of structurally heterogenous promutagens, as indicated by the induction of micronuclei. These cells, however, were not suited for the study of a more relevant genetic end point, the induction of hereditary functional changes (gene mutations), implying that a compromise had to be made on the level of the toxicodynamics. In the second approach, cDNAs encoding the rat cytochromes P-450IA1 and P-450IIB1, set under the control of a constitutive promoter, were transfected into V79 Chinese hamster cells, which do not express cytochromes P-450 but are ideal target cells for gene mutation assays. The resulting substrains (XEM1, XEM2, XEM3; SD1) stably expressed cytochromes P-450IA1 and P-450IIB1, respectively, and showed the corresponding monooxygenase activities. Aflatoxin B1, cyclophosphamide, dibutylnitrosamine, and benzo[a]pyrene mutated SD1 and/or XEM1 and XEM2 cells, but were inactive in parental V79 cells. The mutagenicity of benzo[a]pyrene 7,8-trans-dihydrodiol was about 1000 times more potent in XEM1 and XEM2 cells than in SD1 and V79 cells. Other promutagens were inactive in V79 as well as in the genetically engineered daughter lines. This system therefore is not yet optimal in general screening for the detection of new mutagens, but appears ideal in the identification of critical xenobiotic-metabolizing enzymes for a given mutagen.