The 3' and 5'-monophosphates of O6-methyldeoxyguanosine and N7-methyldeoxyguanosine were chemically synthesized. Using these standards with deoxyguanosine-3'-monophosphate (dGp) as an internal standard, conditions were optimized to quantify O6-methyldeoxyguanosine-3'-monophosphate (O6-MedGp) and N7-methyldeoxyguanosine-3'-monophosphate (N7-MedGp) by 32P-postlabelling. Under optimal conditions, the labelling efficiencies of O6-MedGp and N7-MedGp were respectively approximately 100 and approximately 15%, with detection limits of approximately 1.1 and approximately 6.0 fmol respectively using 10 pmol dGp or 0.8 fmol of O6-MedGp if 2 pmol of dGp was used. The assay developed for O6-MedGp was then applied to the quantitation of [3H]-O6-MedGp and O6-MedGp isolated from DNA digests by immunoaffinity separation. The standard curve generated from the use of [3H]-O6-MedGp, thus isolated, was identical to that generated previously using the chemically synthesized O6-MedGp, indicating that no inhibitory factors co-eluted with the O6-MedGp. After passage through two immunocolumns, recovery of 4 and 40 fmol of O6-MedGp was approximately 30%. Four human stomach samples were analysed by combining this immunoaffinity purification with 32P-post-labelling: levels ranged from 0.21 to 0.86 mumol O6-MedGp/mol dG. Further DNA samples, isolated from the human colon, were fractionated by anion-exchange HPLC and the N7-MedGp and O6-MedGp containing fractions were purified by reverse-phase HPLC and immunoaffinity chromatography respectively. Adduct-containing fractions were dried and 32P-postlabelled. Whereas O6-MedGp was detected at levels between 0.3 and 3.4 mumol O6-MedGp/mol dG, no N7-MedGp was detected in these samples, probably due to depurination of N7-MedGp to N7-methylguanine or reduced assay sensitivity resulting from contaminating nucleotides and/or unidentified radioactivity eluting close to the N7-methyldeoxyguanosine-5'-monophosphate.