Conversion of thyroxine (T(4)) to 3,5,3'-triiodothyronine (T(3)) in rat brain has recently been shown in in vivo studies. This process contributes a substantial fraction of endogenous nuclear T(3) in the rat cerebral cortex and cerebellum. Production of T(4) metabolites besides T(3) in the brain has also been suggested. To determine the nature of these reactions, we studied metabolism of 0.2-1.0 nM [(125)I]T(4) and 0.1-0.3 nM [(131)I]T(3) in whole homogenates and subcellular fractions of rat cerebral cortex and cerebellum. Dithiothreitol (DTT) was required for detectable metabolic reactions: 100 mM DTT was routinely used. Ethanol extracts of incubation mixtures were analyzed by paper chromatography in t-amyl alcohol:hexane:ammonia and in 1-butanol:acetic acid. Rates of production of iodothyronines from T(4) and T(3) were greater at pH 7.5 than at 6.4 or 8.6 and greater at 37 degrees C than at 22 degrees or 4 degrees C. Lowering the pH, reducing the protein or DTT concentrations, and preheating homogenates to 100 degrees C all increased excess I(-) production but reduced iodothyronine production. In cerebral cortical homogenates from normal rats, products of T(4) degradation were as follows (percent added T(4)+/-SEM in nine experiments): T(3), 1.9+/-0.5%; 3,3',5'-triiodothyronine (rT(3)), 34.0+/-2.4%; 3,3'-diiodothyronine (3,3'-T(2)), 5.8+/-1.6%; 3'-iodothyronine (3'-T(1)), </=2.5%; and excess I(-), 4.7+/-1.2%. In the same experiments, products of T(3) degradation were 3,3'-T(2), 63.3+/-5.5%, and 3'-T(1), 12.6+/-1.4%. Cerebral cortical homogenates from hyperthyroid rats and normals were similar in regard to T(4) to T(3) deiodination. In contrast, in cerebral cortical homogenates from hypothyroid rats, phenolic ring deiodination rates were increased and tyrosyl ring deiodination rates were decreased compared with normals.T(4) to T(3) conversion rates in cerebellar homogenates were greater than rates in cerebral cortical homogenates from the same normal rats and less than rates in cerebellar homogenates from hypothyroid rats. T(4) and T(3) tyrosyl ring deiodination rates were greatly diminished in cerebellar homogenates compared with cerebral cortical homogenates in normal and hypothyroid rats. High-speed (1,000-160,000 g) pellets from cerebral cortical homogenates were enriched in phenolic and tyrosyl ring deiodinating activities relative to cytosol. Fractional conversion of T(4) to T(3) was inhibited by T(4), iopanoic acid, and rT(3), but not by T(3). Tyrosyl ring deiodination reactions were inhibited by T(3), T(4), and iopanoic acid, but not by rT(3). These studies demonstrate separate phenolic and tyrosyl ring iodothyronine deiodinase enzymes in rat brain. The brain phenolic ring deiodinase serves in vivo as a T(4) 5'-deiodinase and closely resembles anterior pituitary T(4) 5'-deiodinase in physiological and biochemical characteristics. The physiological significance of the tyrosyl ring iodothyronine deiodinase enzyme is unclear; it shares several properties with rat hepatic T(4) 5-deiodinase.