The (32)P-postlabeling assay is an extremely sensitive technique for detecting carcinogen-DNA adducts. However, for the assignment of DNA adduct structures and the accurate determination of DNA adduct levels by (32)P-postlabeling, authentic adduct standards are needed. For most (32)P-postlabeling applications, such verified synthetic standard compounds are required in the form of their deoxynucleoside 3'-phosphates because they represent substrates for the polynucleotide kinase for transfer of [(32)P]phosphate from [gamma-(32)P]ATP. Three N-(deoxyguanosin)-4-aminobiphenyl 3'-phosphate adducts were prepared and fully characterized by (1)H NMR and mass spectroscopy to serve as standards for the (32)P-postlabeling assay. Apart from the C8- and the N(2)-deoxyguanosine 3'-phosphate adducts of the mutagenic human bladder carcinogen 4-aminobiphenyl (dG3'p-C8-4-ABP and dG3'p-N(2)-4-ABP), the C8-deoxyguanosine 3'-phosphate adduct of the nonmutagenic 4'-tert-butyl-4-aminobiphenyl (dG3'p-C8-4'tBu-4-ABP) was included in the study. Both C8-deoxyguanosine 3'-phosphate adducts were prepared by the in situ formation of deoxyguanosine 3'-phosphate and its subsequent reaction with the appropriate electrophilic amination agent (N-acetoxy compound). The N(2)-deoxyguanosine 3'-phosphate adduct was obtained by a modification of a previously described procedure for the synthesis of N(2)-deoxyguanosine adducts of aromatic amines. The three adduct standards were added at different concentrations to calf thymus DNA, and adduct recoveries were determined by the (32)P-postlabeling assay under conditions routinely used in the standard methodology, enhancement by nuclease P1 digestion and enhancement by butanol extraction. The dG3'p-C8-4-ABP adduct was recovered irrespective of the concentration with approximately 30% in both the standard and the butanol extraction version of the assay. Both C8-deoxyguanosine 3'-phosphate adducts were sensitive to nuclease P1 digestion resulting in recoveries of only 1-3%. In contrast, the dG3'p-N(2)-4-ABP adduct was resistant to nuclease P1 digestion; however, recovery in all three versions was poor (1-2%) resulting in a detection limit of one adduct/10(6) nucleotides. These results demonstrate that the (32)P-postlabeling assay underestimates the level of DNA adducts formed by 4-ABP and indicates that there is a need to determine the recovery for each adduct to be analyzed by the (32)P-postlabeling technique.