Based on our previous observations (H. S. Ramsdell and D. L. Eaton, 1990, Cancer Res. 50, 615-620) that the proportion of aflatoxin B1 (AFB1) converted to the highly reactive AFB1-8,9-epoxide in microsomal incubations varies with substrate concentration, we have examined the hypothesis of T. Shimada and F. P. Guengerich (1989, Proc. Natl. Acad. Sci. USA 86, 462-465) that cytochrome P450 IIIA4 is principally responsible for the activation (epoxidation) of AFB1 by human liver microsomes. The initial rates of formation of AFB1-8,9-epoxide and hydroxylated AFB1 metabolites were determined in microsomes prepared from livers of organ donors (n = 14) at AFB1 concentrations of 124 and 16 microM. Microsomal oxidation of nifedipine, catalyzed primarily by P450 IIIA enzymes, was also determined by HPLC. Rates of formation of AFB1 metabolites and nifedipine oxidation were poorly correlated at either AFB1 concentration (r2 = 0.13-0.41). A somewhat better correlation between AFB1 epoxidation and nifedipine oxidation was observed at 124 microM AFB1 (r2 = 0.41) than at 16 microM AFB1 (r2 = 0.26). Treatment of pooled microsomes with troleandomycin, an apparently specific inhibitor of P450 IIIA enzymes, resulted in 35% inhibition of AFB1-8,9-epoxide formation at the high AFB1 level but had little effect at 16 microM AFB1. An antibody against rat cytochrome P450 IIIA1 significantly inhibited AFB1 epoxidation at high, but not low, AFB1 concentrations, whereas AFQ1 formation was strongly inhibited at all substrate levels examined. These results are consistent with the hypothesis that cytochrome P450 IIIA enzyme(s) can form AFB1-8,9-epoxide, but are effective at only relatively high substrate concentrations. Another P450 enzyme(s) appears to be principally responsible for AFB1-8,9-epoxide formation at the low AFB1 levels that would be typical for dietary exposures.