The relationship between the formation of O6-methylguanine (O6MG) and the induction of lung, liver, and nasal tumors in the Fisher 344 rat by the tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) was examined in a dose-response study. Animals were treated for 20 wk (3 times/wk) with concentrations of NNK ranging from 0.03 to 50 mg/kg to induce tumors. Steady-state concentrations of O6MG were quantitated, and cytotoxicity was assessed in target cells and tissues after 4 wk of treatment with NNK. No cytotoxicity was detected in the lung during treatment with NNK. The formation of O6MG was greatest in Clara cells compared with macrophages, type II cells, small cells, and whole lung at all doses examined. The difference in adduct concentration between the Clara cell and other pulmonary cell types was most pronounced with low doses of carcinogen. The O6MG:dose ratio, an index of alkylation efficiency, increased 29-fold as the dose of NNK was decreased from 50 to 1 mg/kg of carcinogen. In contrast, only a small increase in alkylation efficiency was observed in type II cells and whole lung. A significant number of tumors were induced in the lung at doses of 0.1 to 50 mg/kg with incidences ranging from 10% at the lowest dose up to 87% in the group of animals which received 50 mg/kg of NNK. A linear relationship was observed when the concentration of O6MG in Clara cells as a function of dose was plotted against the corresponding tumor incidence. This relationship was not observed using DNA adduct concentrations in type II cells or whole lung. The development of pulmonary tumors appeared to involve the formation of alveolar hyperplasias which progressed to adenomas and finally to carcinomas. The majority of adenomas were solid, whereas carcinomas were mainly papillary. Examination of the ultrastructure of the hyperplasias, adenomas, and carcinomas revealed morphological structures (e.g., lamellar bodies, tubular myelin) which are associated with type II cells. Thus, these data suggest that the majority of neoplasms in the lung begin as type II cell proliferations with progression to adenomas and carcinomas within the areas of hyperplasia. The lack of agreement between biochemical and morphological findings makes it difficult to hypothesize a cell of origin for the pulmonary neoplasms. In contrast to the lung, tumors were induced in the liver and nasal passages only after exposure to high doses of NNK. Moreover, both the formation of DNA adducts and cytotoxicity appear obligatory for the generation of tumors in these tissues.(ABSTRACT TRUNCATED AT 400 WORDS)