I-compounds are covalent DNA modifications that can be detected and measured by 32P-postlabeling assay because of their DNA-adduct like properties. They accumulate in an age-dependent, highly reproducible manner in tissue DNA of untreated animals in the absence of exogenous carcinogens and, therefore, appear to arise via the interaction of DNA with endogenous reactants formed in the course of normal metabolism. Chromatographically, they exhibit a wide range of polarities, indicative of structural diversity. In addition to age-dependent increases, I-compound profiles exhibit prominent species-, sex-, tissue- and diet-dependent qualitative and quantitative differences. Natural-ingredient (chow) diets produce qualitative differences as well as substantially higher I-compound levels in rat liver and kidney, when compared with purified diets. Modified purified diets containing high carbohydrate, protein, or fat concentrations further modulate I-compound profiles. During liver regeneration, I-compounds behave like DNA adducts rather than m5C in that their levels are not quickly restored. Treatment of rats with the hepatocarcinogens 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), CCl4, and peroxisome proliferators as well as with a choline-devoid hepatocarcinogenic diet depressed the age-related increases of I-compound levels in liver, the target organ. Additional 32P-labeled derivatives were observed only with the peroxisome proliferators and presumably represent DNA adducts of exogenous origin. No I-compounds were detected in a series of Morris hepatomas with different degrees of differentiation. Thus, loss of I-compounds may be associated with altered gene expression/dedifferentiation. On the other hand, the age-dependent accumulation of I-compounds and their adduct-like character suggest potential relations to aging-associated dysdifferentiation and initiation of cancer. Structural complexity indicates different biological roles of I-compounds. © 1990.