Two generally applicable [35S]phosphorothioate postlabeling procedures for the HPLC analysis of polycyclic aromatic hydrocarbon (PAH)-DNA adducts have been developed based upon [32P]phosphate postlabeling assays described by Gupta and Randerath et al.In one procedure, benzo[a]pyrene (B[a]P)-modified DNA was digested to nucleoside 3′-phosphates by micrococcal nuclease and spleen phosphodiesterase and the adducted nucleotides were extracted with 1-butanol. The adducted nucleoside-3′-phosphates were 5′-thlophosphorylated by T4 polynuckotide kinase (T4PNK) and adenosine 5′-O-(3-[35S]thiotriphosphate) to yield [35S]B[a]P-nucleoside-5′-phosphorothioate-3′-phosphate adducts. Although thiophosphorylation of B[a]P-DNA adducts was slower than the corresponding phosphorylation reaction, similar recoveries of the postlabed adducts were achieved with longer incubation times and higher concentrations of T4PNK. A major advantage of this procedure over the 32P-postlabeling procedure is that the resistance of phosphorothioates to degradation by phospha-tases allows selective removal of the unlabeled 3′-phosphate from the [35S]B[a]P-nudeoside-5′-phosphorothioate-3′-phosphate adducts by brief treatment with alkaline phosphatase. [35S]B[a]P-nucleoside-5′-phosphorothioate adducts were also prepared using a nuclease P1/prostatic add phosphatase DNA degradation method. For anti-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE)-modified DNA, overall adduct recoveries were substantially higher with the nuclease P1/prostatic add phosphatase method (48–51%) than with the micrococcal nudease/spleen phosphodiesterase/ alkaline phosphatase method (22–29%). There were no significant differences in the HPLC profiles of the [35S]phos-phorothioate-postlabeled adducts obtained from these two procedures. HPLC analysis of B[a]P-DNA adducts formed in B[a]P-treated hamster embryo cell cultures demonstrated the formation of two major adducts, (+)syn-BPDE-deoxy-guanosine-5′-phosphorothioate and (+)anti-BPDE-deoxy-guanosine-5′-phosphorothioate, along with other minor adducts. Based upon an overall adduct recovery of 20% and 0.5 mol as the detection limit of this 35S-postlabeling/HPLC assay, the sensitivity of this assay is 1 adduct/108 nucleotides for a 60 μg DNA sample. This method offers the advantages of using 35S which has a longer half-life and lower radioactive decay ene4rgy than 32P and the ability to prepare PAH-DNA adducts at the monophosphorothioate level which greatly facilitates separation of individual 35S-postlabeled PAH-DNA adducts by HPLC.