Evidence supporting a membrane site of action for progesterone includes the rapidity of its effects when directly infused into tissue containing mainly nerve terminals, the absence of functional intracellular progesterone receptors in vitro and the fact that progesterone conjugated to bovine serum albumin (BSA) in the C-3 position (P-3-BSA) activates the release of hypothalamic luteinizing hormone releasing hormone (LHRH) or modulates amphetamine-evoked striatal dopamine release. In addition, P2 membrane fractions from different areas of the CNS but not P1 fractions or P2 membranes from peripheral progesterone targets have specific binding sites for P-11-125I-BSA. Among several BSA-conjugated steroids tested for competition displacement P-3-BSA had the highest affinity with an estimated inhibition constant of 28.5 +/- 2.1 nM. This binding depends on the presence of cations such as Ca2+ and Mg2+ and after chemical depolarization of the P2 membranes the binding curve of P-3-BSA shifts to the right. While progesterone is effective in releasing LHRH from the hypothalamus, 5 beta-pregnan-3 beta-ol-20-one (a 5 beta reduced metabolite) is at least 1000-fold more potent than the parent compound when tested in vitro and in vivo. This action is indirect because tetrodotoxin at 10(-6) M blocks the LHRH releasing action, although 5 beta-prenan-3 beta-ol-20-one is still capable of releasing noradrenaline. Although 5 beta-pregnan-3 beta-ol-20-one can replace progesterone in activating the LHRH neural apparatus this is not true for the nigro-striatal dopamine system where only progesterone or P-3-BSA is effective, an action which is also indirect since tetrodotoxin blocks the effect of either compound. These results indicate that progesterone acts at membrane sites to modulate specific functions of the CNS and that site-specific mechanisms exist within the CNS which may differentially control its conversion to more active compounds.