A single amino acid that determines the sensitivity of progesterone receptors to RU486.

The progesterone analog RU486, an abortifacient, inhibits the action of progestins in humans but not in chickens or hamsters. Substitution of cysteine at position 575 by glycine in the hormone binding domain (HBD) of the chicken progesterone receptor (cPR) generated a cPR that binds RU486 and whose activity is antagonized by that compound. In fact, all receptors that bind RU486 have a glycine at the corresponding position. The hamster PR, like cPR, has a cysteine. Only glycine--not methionine or leucine--at position 575 allowed binding of RU486 to cPR. Substitution of this glycine by cysteine in the human PR (hPR) abrogated binding of RU486 but not that of an agonist. The corresponding mutation in the human glucocorticoid receptor resulted in a loss of binding of both dexamethasone and RU486. Examination of a series of 11 beta-substituted steroids showed that antagonism is not an intrinsic property of an antihormone, because one hPR antagonist acted as an agonist for a mutated hPR. The positioning of an aromatic 11 beta-substitution in the PR HBD appears to be critical for generating agonistic or antagonistic activity.

Switching agonistic, antagonistic, and mixed transcriptional responses to 11 beta-substituted progestins by mutation of the progesterone receptor.

The study of transcription activation by a series of RU486-related 11 beta-substituted progestins revealed three types of ligands: agonists, antagonists, and a novel type of compounds that exerted a mixed activity. These ligands conferred to the human progesterone receptor (hPR) only weak activation properties despite high affinity binding and, hence, acted as agonists and, at the same time, as partial antagonists of pure agonists. When the same series of ligands was tested with mutant PRs, transcriptional activation/inactivation profiles were different from those seen with the wild-type PR, since several steroids initially classified as antagonists switched to mixed responses. One compound that acted as an antagonist for the hPR was an agonist for a mutated PR in which 15 amino acids of the hormone-binding domain were replaced by the corresponding divergent residues of the chicken homolog. In analyzing a series of steroids with wild-type and mutant PRs, we observed that a phenyl group (or a phenyl derivative) in the 11 beta position of RU486-related steroids generates antagonism with hPR, but has to be bound in a critical position in the hormone-binding domain to exert its antagonistic activity. Apparently, a deviation from this positioning by mutations in the hormone-binding domain can generate mixed or even agonistic activities.

Residues in the ligand binding domain that confer progestin or glucocorticoid specificity and modulate the receptor transactivation capacity.

To localize regions conferring ligand binding specificity of the human glucocorticoid (hGR) and progesterone (hPR) receptors, we constructed chimeras comprising the DNA-binding domain of the yeast transcription factor GAL4, linked to the ligand binding domain of hGR or hPR. Replacement of a sequence of hGR encompassing helices H6 and H7 with the homologous sequence from hPR creates a chimeric protein GP3, which binds the progestin RU 27987 with high affinity, and results in a concomitant loss of glucocorticoid binding [dexamethasone (DEX), RU 43044]. Moreover, GP3 is not able to mediate RU 27987-induced transactivation. A detailed mutational analysis of this sequence and the study of the recently solved hPR crystal structure revealed five residues that confer progestin responsiveness to GR or modulate ligand binding and transcriptional activation. Notably, the simultaneous presence of residues Ser637 and Phe639 on GP3, lining the ligand binding pocket, is specifically involved in RU 27987 binding. The absence of residues Asp641, Gln642, and Leu647 on GP3 is accountable for the lack of glucocorticoids binding on GP3. Unlike residue 642, residues 641 and 647 are not in direct contact with the ligand and most likely act through steric-mediated interactions. The presence of Gln642 and Leu647 are determinant for transcriptional activation in response to DEX and RU 27987, respectively. DEX-dependent transactivation is further enhanced by the presence of Leu647.

Type II antagonists impair the DNA binding of steroid hormone receptors without affecting dimerization.

Two types of steroid antagonists exert their activity by different mechanisms when bound to the cognate receptor. Type I anti-progestins, such as RU486, induce DNA binding of the human progesterone receptor (hPR), while no hPR/DNA complexes were seen in gel shift assays in the presence of the type II anti-progestin ZK98,299 or RU50,331. ZK98,299-liganded hPR exerted significantly less tight nuclear binding than receptor complexes formed with RU486. Also a type II anti-glucocorticoid (RU43,044) was detected which completely abrogated DNA binding of its cognate receptor in vivo. In keeping with the existence of two different classes of anti-progestins, agonist- or RU486-induced hyperphosphorylation of the two hPR isoforms present in the T47D breast cancer cells was not induced by ZK98,299. This lack of hyperphosphorylation was, however, not the cause but most likely the consequence, of the reduced ability of the hPR/ZK98,299 complex to interact with DNA. No "mixed-ligand" heterodimers were formed in vitro between hPR isoform A bound to ZK98,299 and R5020-liganded isoform B, but nuclear co-translocation studies indicated that ZK98,299 efficiently induced hPR dimerization in vivo.