The endogenous DNA adduct, M 1 dG, has been shown to arise in vitro in reactions of dG with malondialdehyde (MDA), a product of both lipid peroxidation and 4-oxidation of deoxyribose in DNA, and with base propenals also derived from deoxyribose 4-oxidation. We now report the results of cellular studies consistent with base propenals, and not MDA, as the major source of M 1 dG under biological conditions. As a foundation for cellular studies, M 1 dG, base propenals, and MDA were quantified in purified DNA treated with oxidizing agents known to produce deoxyribose 4-oxidation. The results revealed a consistent pattern; Fe 2؉ -EDTA and ␥-radiation generated MDA but not base propenals or M 1 dG, whereas bleomycin and peroxynitrite (ONOO ؊ ) both produced M 1 dG as well as base propenals with no detectable MDA. These observations were then assessed in Escherichia coli with controlled membrane levels of polyunsaturated fatty acids (PUFA). ONOO ؊ treatment (2 mM) of cells containing no PUFA (defined medium with 18:0/stearic acid) produced 6.5 M 1 dG/10 7 deoxynucleotides and no detectable lipid peroxidation products, including MDA, as compared with 3.8 M 1 dG/ 10 7 deoxynucleotides and 0.07 g/ml lipid peroxidation products with control cells grown in a mixture of fatty acids (0.5% PUFA) mimicking Luria-Bertani medium. In cells grown with linoleic acid (18:2), the level of PUFA rose to 54% and the level of MDA rose to 0.14 g/ml, whereas M 1 dG fell to 1.4/10 7 deoxynucleotides. Parallel studies with ␥-radiation revealed levels of MDA similar to those produced by ONOO ؊ but no detectable M 1 dG. These results are consistent with base propenals as the major source of M 1 dG in this model cell system.