Differential interactions of estrogens and antiestrogens at the 17 beta-hydroxyl or counterpart hydroxyl with histidine 524 of the human estrogen receptor alpha.

We investigated the role of H524 of the human estrogen receptor alpha (ERalpha) for the binding of various estrogens [estradiol (E(2)), 3-deoxyestradiol (3-dE(2)), and 17beta-deoxyestradiol (17beta-dE(2))] and antiestrogens [4-hydroxytamoxifen (OHT), RU 39 411 (RU), and raloxifene (Ral)], which possess the 17beta-hydroxyl or counterpart hydroxyl (designated: 17beta/c-OH), with the exception of 17beta-dE(2) and OHT. The work involved a comparison of the binding affinities of these ligands for wild-type and H524 mutant ERs, modified or not with diethyl pyrocarbonate (DEPC), a selective histidine reagent. Alanine substitution of H524 did not significantly change the association affinity constant (relative to OHT) of 17beta-dE(2), whereas those of RU, Ral, E(2), and 3-dE(2) were decreased 3-fold, 14-fold, 24-fold, and 49-fold, respectively. Values of the two ligands available in radiolabeled form (E(2) and OHT) were correlated with the dissociation rate constants, which were increased 250-fold and 2-fold, respectively. The action of DEPC on wild-type ER led to a homogeneous ER population which still bound antiestrogens and 17beta-dE(2) with practically unchanged affinities (less than 4-fold decreases in relative affinity constants), while E(2) and 3-dE(2) displayed markedly decreased affinities (56-fold decrease for E(2)). Conversely, DEPC treatment of H524A mutant ER did not induce marked decreases in the relative affinities of any of the checked compounds (decreases </=3-fold). All of these effects appeared to involve H524 as the H516A mutant behaved as wild-type ER. These combined data relative to mutated or DEPC-modified ER converged to support that the interaction of 17beta/c-OH of ER ligands with H524 is strong for estrogens and weaker for antiestrogens, with quantitative or qualitative differences between the binding modes of the latter, as illustrated by RU and Ral. The abilities of E(2) and OHT to protect the various ER types against inactivation by DEPC were strikingly different: OHT totally prevented the effect of DEPC on wild-type, H516A, and H524A ERs, while E(2) only partially protected wild-type and H516A ERs (H516A ER > wild-type ER) and very weakly protected H524A ER. Molecular modeling was tentatively used to interpret the biochemical results.