Sensitive methods for quantifying DNA adducts from (i) benzo[a]pyrene (BP), (ii) alkylation exposure, and (iii) etheno(epsilon)-DNA adduct-forming chemicals were developed and applied to humans and animal models. The aims were to identify hitherto unknown sources and mechanisms of exogenous and endogenous DNA damage, to examine the effect of drug polymorphism on BP adduct levels, and to develop QSAR between tumorigenic potency, heritable genetic damage and structural elements of alkylating carcinogens (Vogel and Nivard (1994) Mutation Res., 395, 13-32). (i) BP-DNA adducts: An HPLC/fluorimetry assay suitable for measuring (+)-anti-BP-diol-epoxide (BPDE) adducts in human tissues and white blood cells (WBC) was developed (Alexandrov et al. (1992) Cancer Res., 52, 6248-6253). In smokers, a positive correlation was found between pulmonary CYP1A1-related catalytic activity (AHH) and the level of lung BPDE-DNA adducts. In coke oven workers, an enhancing effect of smoking on BPDE-adduct levels in WBC was demonstrated (Rojas et al. (1995) Carcinogenesis, 16, 1373-1376). (ii) 3-Alkyladenines (3-alkAde): Alkylating carcinogens form 3-alkAde adducts in DNA which depurinate to yield 3-alkAde in urine, for which a detection method was developed (Friesen et al. (1991) Chem. Res. Toxicol., 4, 102-106; Prevost et al. (1990) Carcinogenesis, 11, 1747-1751), using immunoaffinity purification and GC-MS analysis. The usefulness of 3-alkAde analysis for the determination of the whole-body dose of alkylating agents derived from exogenous and endogenous sources was demonstrated. (iii) Etheno-DNA adduct-forming agents: Etheno(epsilon)-DNA base adducts (epsilon A, epsilon dC, epsilon dG) are promutagenic DNA lesions that are formed by occupational (vinyl halides) and environmental (urethane) carcinogens. An ultrasensitive detection method was developed (Nair et al. (1995) Carcinogenesis, 16, 613-617), based on immunoaffinity purification and 32P-postlabelling of epsilon-nucleoside 3'-monophosphates. Liver DNA from unexposed rats, mice and from human samples contained background levels of epsilon dA and epsilon dC (Bartsch et al. (1994) Drug. Metab. Rev., 26, 349-371). As formation of epsilon dA and epsilon dC adducts by lipid peroxidation products was demonstrated (El Ghissassi et al. (1995) Chem. Res. Toxicol., 8, 278-283), they may serve as markers for oxidative stress. Results from testing this hypothesis are presented.