DNA adducts were quantified in hamster tracheas exposed to benzo(a)pyrene (BP) in organ culture, in basal as well as in non-basal cells, by in situ detection with an adduct-specific rabbit antiserum (W2/01) and with a mouse monoclonal antibody against human cytokeratins 5 and 8 (RCK102) to identify hamster trachea basal cells. Recognition by W2/01 of the adduct of (+)-anti-7,8-dihydroxy-9,10-epoxide of BP (BP-diolepoxide; BPDE) to deoxyguanosine (dG) was checked on human white blood cells (WBCs) exposed to BP together with 3-methylcholanthrene (3MC)-induced rat-liver microsomes. By comparison with the adduct levels determined by 32P post-labeling, a lower detection limit of about 1 adduct per 10(6) nucleotides could be deduced. Next, tracheal rings were exposed to BP (40 microM) in organ culture for 2 days, then washed and cultured without BP for another 3 days. At different time points epithelial cells were isolated and cytospin preparations made. Staining of BP DNA adducts combined with that of cytokeratin (both visualized with fluorescence) allowed detection of adducts in both basal and non-basal cells in the same preparation. BP DNA adduct formation in basal and non-basal cells after 2 days of exposure to BP was not different. However, on removal of BP the adducts disappeared significantly faster from basal cells than from non-basal cells. The combination of the two antibodies mentioned above thus allows selective determination of BP DNA adduct levels in different cell types. This could be of importance with regard to the involvement of specific cell types in the process of tumor initiation.