Probing conformational changes in the estrogen receptor: evidence for a partially unfolded intermediate facilitating ligand binding and release.

Because the ligand bound to the ligand-binding domain (LBD) of nuclear hormone receptors is completely enveloped by protein, it is thought that the process of ligand binding or unbinding must involve a significant conformational change of this domain. We have used the intrinsic tryptophan fluorescence of the estrogen receptor-alpha (ERalpha) or estrogen receptor-beta (ERbeta) LBD, as well as bis-anilinonaphthalenesulfonate (bis-ANS), a probe for accessible interior regions of protein, to follow the guanidine-hydrochloride (Gua-HCl)-induced unfolding of this domain. In both cases, we find that the ER-LBD unfolding follows a two-phase process. At low Gua-HCl, the ER-LBD undergoes partial unfolding, whereas at high Gua-HCl, this domain undergoes a global unfolding, with bis-ANS binding preferentially to the partially unfolded state. The partially unfolded state of the ERalpha-LBD induced by denaturant does not bind ligand stably, but it may resemble an intermediate that this domain accesses transiently under native conditions that allow ligands to enter or exit the ligand-binding pocket.

Coactivator peptides have a differential stabilizing effect on the binding of estrogens and antiestrogens with the estrogen receptor.

The effectiveness of estrogens in stimulating gene transcription mediated by the estrogen receptor (ER) appears to depend on ER interactions with coactivator proteins. These coactivators bind to ER when it is liganded with an estrogen agonist, but not when it is liganded with an estrogen antagonist. Because estrogen agonists are known to induce a conformation in ER that stabilizes coactivator binding, we asked whether coactivator binding to ER causes a reciprocal stabilization of agonist ligand binding. We used a fluorescent ligand for ER, tetrahydrochrysene-ketone, to monitor the rates of ligand dissociation from ERalpha and ERbeta, and to see how this process is affected by the p160-class coactivator, steroid receptor coactivator-1 (SRC-1). We used a 15-amino acid peptide corresponding to the second nuclear receptor box LXXLL motif in SRC-1 (NR-2 peptide), which is known to interact with the ER ligand-binding domain, a mutant peptide with an LXXAL sequence (NR-2A peptide), and a 203-amino acid fragment of SRC-1, termed the nuclear receptor domain (SRC1-NRD), embodying all three of the internal NR boxes of this protein. Both the NR-2 peptide and the SRC1-NRD fragment markedly slow the rate of dissociation of the agonist ligands tetrahydrochrysene-ketone, estradiol, and diethylstilbestrol, increasing the half-life of the ER-agonist complex by up to 50- to 60-fold. The SRC1-NRD has much higher potency in retarding ligand dissociation than does the NR-2 peptide; it is maximally effective at 30 nM, and it appears to bind with the stoichiometry of one SRC1-NRD per ER dimer. The peptides had little effect on the dissociation rate of antagonist ligands. Consistent with these results, we find that increasing the concentration of SRC-1 in cells by transfection of an expression plasmid encoding SRC-1 causes a 17-fold increase in the potency of estradiol in an estrogen-responsive reporter gene transcription assay. Thus, there is multifactorial control over receptor-coactivator interaction, its strength being determined by the agonist vs. antagonist nature of the ligand and the particular structure of the agonist ligand, and by the receptor subtype and the NR box sequence. The stabilizing effect of coactivator on ER-agonist ligand complexes may be important in determining the potency of estrogen agonists in a cell and may also underlie the tissue-selective pharmacology of certain synthetic estrogens.