A chemiluminescence immunoassay (CIA) utilizing antiserum elicited against DNA modified with (+/-)-7beta, 8alpha-dihydroxy-9alpha,10alpha-epoxy-7,8,9,10-tetrahydrobenzo[a]- pyrene (BPDE) has been developed and validated to study the formation of polycyclic aromatic hydrocarbon (PAH)-DNA adducts in human tissues. Advantages include a low limit of detection for 10b-(deoxyguanosin-N(2)-yl)-7beta,8alpha,9alpha-trihydroxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPdG, approximately 1.5 adducts/10(9) nucleotides using 20 micro g DNA) and a high signal-to-noise ratio (> or =100). The CIA BPDE-DNA standard curve gave 50% inhibition at 0.60 +/- 0.08 fmol BPdG (mean +/- SE, n = 30), which was a 10-fold increase in sensitivity compared with the dissociation-enhanced lanthanide fluoroimmunoassay (DELFIA). Calf thymus DNA modified with [1,3-(3)H]BPDE was assayed by radiolabeling, (32)P-postlabeling, DELFIA and CIA, and all assays gave similar values. Liver DNAs from mice exposed to 0.5 and 1.0 mg [7,8-(3)H]benzo[a]pyrene (BP) were assayed by the same four assays and a dose-response was obtained with all assays. The BPDE-DNA CIA was further validated in MCL-5 cells exposed to 4 micro M BP for 24 h, where nuclear and mitochondrial DNA adduct levels were associated with an increase in DNA tail length measured by the Comet assay. Human peripheral blood cell (buffy coat) DNA samples (n = 43) obtained from 25 individuals who were either colorectal adenocarcinoma patients or controls were assayed by BPDE-DNA CIA. Three samples (7%) were non-detectable, and the remaining 40 samples had values between 0.71 and 2.21 PAH-DNA adducts/10(8) nucleotides. The intra-assay coefficient of variation (CV), for four wells on the same microtiter plate, was 1.85%. Sufficient DNA for two assays, on separate plates, was available for 38 of the 43 samples, and the PAH-DNA adduct values obtained were highly correlated (r(2) = 0.95). Coded duplicate DNA samples from 15 individuals were assayed four times gave an inter-assay CV of 13.8%.