Release of 3H2O from regiospecifically labeled estradiol was measured during 2-hydroxylation of this estrogen by rat liver microsomes. The amount of tritium remaining in the isolated catechol estrogen was also determined. Virtually all the tritium was removed from C-2 during the reaction confirming the absence of an NIH shift. About 20% of the tritium at C-1 was also lost without any such change occurring at C-4 or C-6,7 of the steroid molecule. These findings provide evidence for the formation of an arene oxide or o-semiquinone intermediate during the conversion of estradiol to 2-hydroxyestradiol. No indication of adduct formation at either C-1 or C-4 during this biotransformation was obtained although the 2-hydroxylated product was able to react with a nucleophile such as glutathione. The different regiospecificity of tritium loss in the generation of catechol estrogens and in their subsequent reaction leads to the important conclusion that the reactive intermediates in the two processes must be different. The possible role of catechol estrogens in neoplastic transformation is discussed.