Diol epoxides formed by the sequential action of cytochrome P-450 and the microsomal epoxide hydrolase (mEH) in the endoplasmic reticulum (ER) represent an important class of ultimate carcinogenic metabolites of polycyclic aromatic hydrocarbons. The role of the membrane orientation of cytochrome P-450 and mEH relative to each other in this catalytic cascade is not known. Cytochrome P-450 is known to have a type I topology. According to the algorithm of Hartman, Rapoport and Lodish [(1989) Proc. Natl. Acad. Sci. U.S.A. 86, 5786-5790], which allows the prediction of the membrane topology of proteins, mEH should adopt a type II membrane topology. Experimentally, mEH membrane topology has been disputed. Here we demonstrate that, in contrast with the theoretical prediction, the rat mEH has exclusively a type I membrane topology. Moreover we show that this topology can be inverted without affecting the catalytic activity of mEH. Our conclusions are supported by the observation that two mEH constructs (mEHg1 and mEHg2), containing engineered potential glycosylation sites at two separate locations after the C-terminal site of the membrane anchor, were not glycosylated in fibroblasts. However, changing the net charge at the N-terminus of these engineered mEH proteins by +3 resulted in proteins (++mEHg1 and ++mEHg2) that became glycosylated and consequently had a type II topology. The sensitivity of these glycosylated proteins to endoglycosidase H indicated that, like the native mEH, they are still retained in the ER. The engineered mEH proteins were integrated into membranes as they were resistant to alkaline extraction. Interestingly, an insect mEH with a charge distribution in its N-terminus similar to ++mEHg1 has recently been isolated. This enzyme might well display a type II topology instead of the type I topology of the rat mEH. Importantly, mEHg1, having the natural cytosolic orientation, as well as ++mEHg1, having an artificial huminal orientation, displayed rather similar substrate turnovers for the mutagenic metabolite benzo[a]pyrene 4,5-oxide. To our knowledge this is the first report demonstrating that topological inversion of a protein within the membrane of the ER has only a moderate effect on its enzymic activity, despite differences in folding pathways and redox environments on each side of the membrane. This observation represents an important step in the evaluation of the influence of mEH membrane orientation in the cascade of events leading to the formation of ultimate carcinogenic metabolites, and for studying the general importance of metabolic channelling on the surface of membranes.