To determine the feasibility of using human sperm cells for DNA 32P-postlabeling analyses, and to evaluate the baseline level and the possible presence of smoking-related DNA adducts in these cells, sperm DNA was isolated from specimens obtained from 12 heavy smokers, 12 light smokers, and 12 nonsmokers. Background levels of radioactivity were minimized by using magnet transfer of 32P-labeled mononucleotides to new polyethyleneimine cellulose plates. Compared with placental tissues, few adducts were observed. Diffuse radioactivity observed in some of the autoradiograms was minimally above background but the level of radioactivity expressed as putative adducts/nucleotide was not related to smoking status. It was not clear, in some cases, whether this radioactivity was associated with chemically bound adducts or was from nonspecifically bound chemicals, radiolabeled enzymes, or other proteins. One major discrete DNA adduct of unknown chemical structure was detected in three of the 36 samples analyzed (one nonsmoker and two smokers). Based on the level of radioactivity associated with various dilutions of a benzo(a)pyrene-derived adduct, our limit of sensitivity was at least 1.2 adducts/10(9) nucleotides. Our study emphasizes the need to more clearly define the significance of background radioactivity associated with DNA adduct maps where the measured adduct levels approximate detection limits defined by visual observance of adduct spots. This point is particularly relevant given that the 32P-postlabeling procedures rely, in part, on visual verification of the presence of DNA adducts.