Polycyclic aromatic hydrocarbons (PAHs) are known to be activated by the cytochrome P450 (P450) 1 family. However, the precise role of individual P4501 family members in PAH bioactivation remains to be fully elucidated. We therefore investigated the formation of PAH-DNA adducts in the epidermis of Cyp1a2(-/-), Cyp1b1(-/-), and Ahr(-/-) knockout mice. A panel of different PAHs was used, ranging in carcinogenic potency. Mice were treated topically on the dorsal skin with the following tritium-labeled PAHs: dibenzo[a,l]pyre-ne (DB[a,l]P), 7,12-dimethylbenz[a]anthracene (DMBA), benzo[a]pyrene (B[a]P), dibenzo[a,h]anthracene (DB[a,h]A), benzo[g]chrysene (B[g]C), and benzo[c]phenanthrene (B[c]P). At 24 h after treatment, mice (two male and two female mice per group) were sacrificed, and epidermal DNA was isolated and hydrolyzed with DNase I; subsequently, DNA adducts were quantitated by liquid scintillation counting. In the DB[a,l]P-treated mice, levels of DNA adducts were significantly lower in Cyp1a2(-/-) and Cyp1b1(-/-) mice by 57 and 46%, respectively, as compared to wild-type (WT) mice (C57BL/6 background). The levels of DB[a,l]P DNA adducts formed in Ahr(-/-) mice were 26% lower, but this was not statistically significant. The levels of DMBA-DNA adducts in Cyp1a2(-/-) mice were not different than the WT mice but were significantly lower in Cyp1b1(-/-) and Ahr(-/-) mice by 64 and 52%, respectively. DMBA-DNA adduct samples were further analyzed by HPLC following further digestion to deoxyribonucleosides. HPLC analysis of individual DMBA-DNA adducts revealed differences in the ratio of syn-DMBA-diol epoxide- to anti-DMBA-diol epoxide-derived adducts in the Ahr(-/-) and Cyp1b1(-/-) mice. The ratio of syn-/anti-derived adducts in WT mice was 0.49. This ratio was 0.23 in the Cyp1b1(-/-) mice and 0.87 in the Ahr(-/-) mice. In contrast to the results with DB[a,l]P and DMBA, the levels of B[a]P-, DB[a,h]A-, B[g]C-, and B[c]P-DNA adducts were significantly lower in Ahr(-/-) mice by 73, 75, 50, and 81%, respectively, as compared to WT mice but were not significantly lower in the Cyp1a2(-/-) or Cyp1b1(-/-) mice. Collectively, these and other results support a role for both P4501A1 and P4501B1 in the bioactivation of DMBA; P4501A2, P4501B1, and possibly P4501A1 in the bioactivation of DB[a,l]P; and P4501A1 in the bioactivation of B[a]P, DB[a,h]A, B[g]C, and B[c]P in mouse epidermis. Furthermore, in the metabolic activation of DMBA in mouse epidermis, P4501B1 shows a preference for the formation of syn-DMBA-diol epoxide adducts, whereas P4501A1 shows a preference for the formation of anti-DMBA-diol epoxide adducts.