A method, using HPLC combined with electrospray tandem mass spectrometry (ES-MS/MS), was developed and validated to detect and quantify the major DNA adduct resulting from exposure to the ultimate tumorigenic benzo[a]pyrene (BP) metabolite, trans-7,8-dihydroxy-anti-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE). Calf thymus DNA was reacted with BPDE, digested enzymatically to nucleosides, and the major DNA adduct, 10-(deoxyguanosin-N(2)-yl)-7,8,9-trihydroxy-7,8,9,10-tetrahydrobenzo[a]pyrene (dG-BPDE), was purified by HPLC. Similar procedures were applied to prepare dG-BPDE-d(8) from [1,2,3,4,5,6,11,12-(2)H(8)]BPDE for use as an internal standard. The HPLC-ES-MS/MS method was validated using a mixture of hydrolyzed salmon testis DNA (82 mug) and 10 pg dG-BPDE (analogous to 6.9 adducts/10(8) nucleotides). The results indicated an inter- and intraday accuracy of 99-100% and precision of 1.6-1.7% (relative standard deviation). When applied to a calf thymus DNA sample modified in vitro with [1,3-(3)H]BPDE, the method gave a value very similar to those obtained by radiolabeling, (32)P-postlabeling, and immunoassay. HPLC-ES-MS/MS analysis of hepatic DNA from mice treated intraperitoneally with 0.5 and 1.0 mg of [7,8-(3)H]BP gave values comparable to those determined by (32)P-postlabeling and immunoassay. Lung DNA from mice fed a 0.3% coal tar diet (containing approximately 2 mg BP/g coal tar) for one month had 0.6 +/- 0.04 dG-BPDE adducts/10(8) nucleotides. This value is much lower than the 102 +/- 14 total DNA adducts/10(8) nucleotides determined by (32)P-postlabeling, which suggests that dG-BPDE makes only a minor contribution to the DNA adducts formed in lung tissue of mice administered coal tar. The HPLC-ES-MS/MS method was used to assess human lung DNA samples for the presence of dG-BPDE. Based upon a limit of detection of 0.3 dG-BPDE adducts/10(8) nucleotides, when using 100 mug of DNA, dG-BPDE was detected in only 1 out of 26 samples. These observations indicate that HPLC-ES-MS/MS is suitable to assess the contribution of BP to DNA damage caused by exposures to polycyclic aromatic hydrocarbon (PAH) mixtures. The results further suggest that dG-BPDE may contribute only a small fraction of the total DNA adducts detected by other DNA adduct methodologies in individuals exposed to PAHs.