Naphthalene produces selective necrosis of Clara cells in the mouse but not in the rat. The pulmonary toxicity depends on cytochrome P450-mediated metabolism; however, the selective pulmonary toxicity of naphthalene in the mouse does not correspond to tissue-selective covalent binding of reactive naphthalene metabolites in vivo. These studies compare reactive metabolite binding in target and nontarget cells and in various subcompartments of mouse lung and characterize, by sodium dodecyl sulfate polyacrylamide gel electrophoresis, the proteins to which arylating metabolites are bound. Reactive metabolite binding was substantially higher in incubations of [3H]-naphthalene with distal bronchioles and isolated Clara cells than with explants of trachea or bronchus from the mouse. Likewise, binding was substantially higher in incubations of murine Clara cells than in identical incubations with mouse hepatocytes (nontarget cells) or rat trachea cells (nonsusceptible species). These data show a good correlation between cellular susceptibility to toxicity and the amount of reactive metabolite bound in vitro. Concentrations of adduct were highest in the medium and the nuclear/cell debris fraction (1000 x g pellet) of isolated Clara cells incubated with naphthalene; very small amounts of adduct were noted in pellets isolated at 20,000 or at 100,000 x g (mitochondrial and microsomal fractions) or in cytosol. These observations were consistent with the finding that adduct concentrations in bronchoalveolar lavage were substantially higher than in the lung at low doses of naphthalene and suggest that monitoring adducts in lavage may serve as a useful biomarker of exposure and effect.(ABSTRACT TRUNCATED AT 250 WORDS)