The sensitivity for DNA adduct formation by antitumor alkylating agents (mechlorethamine, cisplatin and adozelesin) of the postlabeling technique and thin-layer chromatography was studied. Three DNAs were used: a double-stranded 20-bp oligonucleotide of defined sequence, calf thymus DNA and murine leukemia L1210 cellular DNA. With high concentrations of mechlorethamine, there was a marked decrease in normal dGp, a lesser decrease in dAp and dCp and no change in dTp. Using 2D mapping PEI-cellulose thin-layer chromatography analyses, it was found that six mechlorethamine: DNA adducts were produced after a short exposure to mechlorethamine. After an extended time at relatively high drug concentrations there was an alteration in the mechlorethamine: DNA adduct pattern that may reflect the conversion of monoadducts to crosslinked adducts. Similar observations were made with cisplatin and adozelesin. When murine leukemia L1210 cells were treated with 50 microM mechlorethamine or 50 microM cisplatin for 1 h, six or more mechlorethamine: DNA adducts and five cisplatin: DNA adducts were detected. After allowing 6 h. for repair of potentially lethal damage, several adducts were no longer detectable and others appeared with diminished intensity. Nuclease P(1) dephosphorylates normal nucleotides at relatively low enzyme concentrations with variation depending upon the nucleotide. In general, considerably lower concentrations of nuclease P1 were required to dephosphorylate the normal nucleotides than to dephosphorylate the antitumor alkylating agent: nucleotide adducts, thus allowing increased sensitivity of the postlabeling assay. The sensitivity of detection of antitumor alkylating agent: DNA adducts in DNA from treated L1210 cells approached one adduct per 10(7)-10(8) nucleotides. These results suggest that the postlabeling technique may be sufficiently sensitive and specific for the study of the clinically effective levels of antitumor alkylating agents.