Detection of DNA alkylphosphotriesters by 32P postlabeling: evidence for the nonrandom manifestation of phosphotriester lesions in vivo.

Many genotoxic carcinogens react with the sugar-phosphate backbone in DNA to form phosphotriester (PTE) adducts. These lesions are relatively abundant and persistent for some alkylating carcinogens and may therefore serve as useful biomarkers with which to assess genotoxic exposure and potential mutagenic risk. In the present study, we have developed a 32p postlabeling method that permits analysis of total methyl and/or ethyl PTE in DNA at the femtomole level. The technique is based on the inability of all known nucleolytic enzymes to cleave the internucleotide PTE bond. Consequently, complete digestion of alkylated DNA with these nucleases in the presence of an alkaline phosphatase yields PTE-dinucleoside phosphates. These species are then converted to the corresponding dinucleoside phosphates (dNpdNs) by treatment with alkali to permit subsequent 32p labeling. The resulting labeled dinucleotides (32pd-NpdN) are then analyzed by PAGE. Validation of this method has been carried out using a polydeoxythymidylic acid oligonucleotide containing a site-specific methyl PTE. The method has been applied to the in vitro analysis of calf thymus (CT) DNA treated with dimethylsulfate (DMS) or diethylsulfate (DES) and to the analysis of liver DNA from mice treated in vivo with nitrosodiethylamine. In each case, autoradiograms of the polyacrylamide gels showed the anticipated five bands representing the sixteen labeled dinucleotides, with proportional increases observed as the concentrations of DMS or DES used in the in vitro treatment of CT DNA were increased. The identity and frequency of the nucleosides located 5' to the PTE lesions were obtained by nuclease P1 digestion of the gel-isolated 32pdNpdN species and by analysis of the released labeled mononucleotides, 32pdN, by high-performance liquid chromatography with radioactivity detection. Results obtained from CT DNA treated with DMS or DES showed that the frequency of the four detected nucleotides reflected the normal nucleoside content of CT DNA, indicating the random formation of methyl and ethyl PTE adducts in the in vitro modified DNA. However, studies using liver DNA from three strains of mice treated in vivo with nitrosodiethylamine indicated that the frequency of the thymidine and the 2'-deoxyguanosine 5' to the ethyl PTE was significantly different from the corresponding normal nucleoside content. These results are indicative of (a) the nonrandom formation of ethyl PTE in vivo and/or (b) base sequence-specific ethyl PTE repair.