NADPH-sustained N-oxygenation of N,N-dimethylaniline (DMA) was investigated with the aid of a reconstituted membranous cytochrome P-4502B4 system. The N-oxidative process did not appear to be supported by hydroxyl radicals or products arising from lipid peroxidation. However, superoxide dismutase was a very potent scavenger of N-oxide formation, while catalase was ineffective. Superoxide by itself did not bring about N-oxygenation of DMA. Therefore, O2- was presumed to serve as a source of the actual proximate oxidant. The reconstituted hemoprotein system catalyzed N-oxygenation of DMA when excess H2O2 substituted for NADPH/O2. This 'peroxygenase' process was entirely dependent on the presence of native enzyme and was not inhibited by CO or metyrapone. By contrast, cyanide severely blocked metabolic transformation. Among some other hemeproteins tested, only horseradish peroxidase was efficient in producing appreciable amounts of N-oxide in the presence of H2O2. Peroxidatic DMA N-oxygenation in intact liver microsomes fortified with cumene hydroperoxide was 2-fold stimulated by pretreatment of rabbits with phenobarbital, whereas administration of 3-methylcholanthrene or ethanol decreased turnover. Studies with uninduced hepatic microsomes, in which the activity of the flavin-containing monooxygenase had been partially suppressed by thermal treatment, revealed pronounced susceptibility of the NADPH-dependent N-oxide formation to the inhibitory action of both superoxide dismutase and antibody to NADPH-cytochrome P-450 reductase. These findings were interpreted to mean that at least 23% of the total amount of N-oxide produced in these preparations resulted from superoxide-dependent conversion of DMA by the P-450 system.