Analysis of oestrogen receptor dimerisation using chimeric proteins

Differential requirements of Hsp90 and DNA for the formation of estrogen receptor homodimers and heterodimers

The two estrogen receptor (ER) subforms, ERalpha and ERbeta, are capable of forming DNA-binding homodimers and heterodimers. Although binding to DNA is thought to stabilize ER dimers, how ERalpha/alpha, ERbeta/beta, and ERalpha/beta dimerization is regulated by DNA and the chaperone protein Hsp90 is poorly understood. Using our highly optimized bioluminescence resonance energy transfer assays in conjunction with assays for transcriptional activation of ERs, we determined that DNA binding appears to play a minor role in the stabilization of ER dimers, especially in the case of ERbeta/beta homodimers. These findings suggest that ER dimers form before they associate with chromatin and that DNA binding plays a minor role in stabilizing ER dimers. Additionally, although Hsp90 is essential for the proper dimerization of ERalpha/alpha and ERalpha/beta, it is not required for the proper dimerization of ERbeta/beta. Despite this, Hsp90 is critical for the estrogen-dependent transcriptional activity of the ERbeta/beta homodimer. Thus, Hsp90 is implicated as an important regulator of distinct aspects of ERalpha and ERbeta action.

Current status of estrogen receptors

Increasing knowledge on structure and function of estrogen receptors is providing information on the mechanism of action of estrogen agonists, as well as antagonists, and in understanding their tissue-selective action. However, there are still many factors associated with estrogen response which are poorly understood. Therefore, the task of designing a tissue-selective estrogen for use as a pharmaceutical in estrogen-dependent disorders remains an uncertain game. This review provides information on the current status of estrogen receptors for a better understanding.

In vitro and in vivo interactions of bisphenol A and its metabolite, bisphenol A glucuronide, with estrogen receptors alpha and beta

The estrogenic activities of bisphenol A (BPA) and its major metabolite BPA glucuronide (BPA-G) were assessed in a number of in vitro and in vivo assays. BPA competed with [3H]-17beta-estradiol (E2) for binding to mouse uterine cytosol ER, a glutathione S-transferase (GST)-human ER D, E, and F domain fusion protein (GST-hERalphadef) and full-length recombinant hERbeta. The IC(50) values for E2 were similar for all three receptor preparations, whereas BPA competed more effectively for binding to hERbeta (0.96 microM) than to either mouse uterine cytosol ER (26 microM) or GST-hERalphadef (36 microM). In contrast, BPA-G did not competitively displace [3H]E2 from any of the ER preparations. In MCF-7 cells transiently transfected with Gal4-hERalphadef or Gal4-hERbetadef, BPA induced reporter gene activity with comparable EC(50) values (71 and 39 microM, respectively). No significant induction of reporter gene activity was seen for BPA-G. Cotreatment studies showed that concentrations of (10 microM) BPA and BPA-G did not antagonize E2-induced luciferase mediated through either Gal4-hERalphadef or Gal4-hERbetadef. In vivo, the uterotropic effect of gavage or subcutaneous (sc) administration of 0.002-800 mg of BPA/kg of body weight/day for three consecutive days was examined in immature rats. Dose-related estrogenic effects on the rat uterus were observed at oral doses of 200 and 800 mg/kg and at sc doses of 10, 100, and 800 mg/kg. These results demonstrate that BPA competes more effectively for binding to ERbeta, but induces ERalpha- and ERbeta-mediated gene expression with comparable efficacy. In contrast, BPA-G did not exhibit any in vitro estrogenic activity. In addition, there was a clear route dependency on the ability of BPA to induce estrogenic responses in vivo.

Identification and characterization of estrogen receptor variants in prostate cancer cell lines

A sensitive semi-nested reverse transcriptase-polymerase chain reaction (RT-PCR) amplification was performed to evaluate estrogen receptor-alpha (ER-alpha) mRNA expression in prostate cancer cell lines. We demonstrated the presence of wild-type ER-alpha (wt ER-alpha) and five ER-alpha variants, designated ER-alphaA, B, C, D, and E. Unlike ER-alphaA and D, ER-alphaB, C, and E were not previously reported in normal or cancerous mammalian cells. DNA sequencing analysis of these ER-alpha variants revealed the genetic changes to be either in-frame or out-of-frame deletions. The expression of each ER-alpha variant differs significantly depending on the androgen responsiveness, tumorigenic and metastatic potentials of each prostate cancer cell line. The potential functional significance of ER-alpha variants was assessed in yeast two-hybrid and ERE promoter-reporter mammalian transcription assay systems. The results of these studies indicated that none of the ER-alpha variants can form homo- or heterodimers either with wt ER-alpha or among themselves in vivo, and that these ER-alpha variants have no demonstrable transcriptional or dominant-negative activity, as assessed in vitro.

Dimerizing the estrogen receptor DNA binding domain enhances binding to estrogen response elements

In this work, we provide a rationale for the finding that the estrogen receptor (ER) binds to its DNA response element as a homodimer in vivo. Binding of the monomer estrogen receptor DNA binding domain (ER DBD) to a palindromic, consensus estrogen response element (ERE) is increased 5-6-fold when the ER DBD is dimerized either by a monoclonal antibody that recognizes an attached epitope tag or by expressing the ER DBD as a single molecule in which the two monomers are joined by a peptide linker. Most of the increase in binding is due to stabilization of the ER DBD.ERE complex. We observed only an approximately 2.5-fold reduction in binding when a consensus ERE was replaced with widely spaced ERE half-sites, suggesting that the interaction between ER DBDs on the ERE is relatively weak, and that in full-length ER the DBDs can move independently of each other. To test binding to an imperfect palindrome, typical of the imperfect EREs found in almost all natural estrogen receptor responsive genes, we used the pS2 ERE. Even at high concentrations of ER DBD, specific binding of the ER DBD to the imperfect pS2 ERE was undetectable. Both of the dimerized ER DBDs exhibited efficient binding to the imperfect pS2 ERE, with an affinity at least 25-fold greater than monomer ER DBD. These data support the view that steroid receptor dimerization provides an important mechanism facilitating the recognition of naturally occurring, imperfect hormone response elements.

Regulation of Cre recombinase activity by mutated estrogen receptor ligand-binding domains

Ligand-dependent chimeric Cre recombinases are powerful tools to induce specific DNA rearrangements in cultured cells and in mice. We report here the construction and characterization of a series of chimeric recombinases, each consisting of Cre fused to a mutated human oestrogen receptor (ER) ligand-binding domain (LBD). Two new ligand-dependent recombinases which contain either the G400V/M543A/L544A or the G400V/L539A/L540A triple mutation of the human ER LBD are efficiently induced by the synthetic ER antagonists 4-hydroxytamoxifen (OHT) and ICI 182,780 (ICI), respectively, but are insensitive to 17 beta-oestradiol (E2). Both chimeric recombinases should be useful for efficient spatio-temporally controlled site-directed somatic mutagenesis.

Site-directed estrogen receptor antibodies stabilize 4-hydroxytamoxifen ligand, but not estradiol, and indicate ligand-specific differences in the recognition of estrogen response element DNA in vitro

Conformational differences between type I antiestrogen-liganded estrogen receptor and estradiol (E2)-liganded estrogen receptor (ER) are thought to be responsible for differentiating agonist versus antagonist ER activity at individual genes. To examine the impact of ER ligand on estrogen-response element (ERE) binding kinetics and receptor conformation, we quantitated the effect of site-directed, ER-specific antibodies raised against synthetic peptides corresponding to the DNA-binding domain of human ER on ER-ERE binding in vitro. Although 4-hydroxytamoxifen-liganded-ER (4-OHT-ER) and E2-ER bind a consensus ERE with equal high affinity, the stoichiometry of 4-OHT-ER-ERE binding at saturation is approximately 50% lower than that of E2-ER binding to all ERE sequences tested. In contrast, the ERE binding stoichiometry of tamoxifen aziridine-liganded ER (TAz-ER) is identical to that of E2-ER: one receptor dimer bound per ERE. The difference in binding stoichiometry is caused by dissociation of one molecule of 4-OHT from the ER as the dimeric receptor binds DNA. Addition of low concentrations of ER-specific polyclonal antibodies AT3A or AT3B prevented 4-OHT ligand dissociation, yielding an increase in specific 4-OHT-ER-ERE binding to a level equal to that of E2-ER- or TAz-ER-ERE binding. However, higher amounts of AT3A or AT3B inhibited specific ERE binding of both 4-OHT- and E2-ER. We conclude that differences in ER conformation when liganded with 4-OHT versus E2 are revealed by these antibodies and that such differences in receptor conformation may influence subsequent interaction of the receptor with other proteins necessary for transactivation.

Functional analysis of six androgen receptor mutations identified in patients with partial androgen insensitivity syndrome

Partial androgen insensitivity syndrome (PAIS) is caused by defects in the androgen receptor gene and presents with a wide range of undervirilization phenotypes. We studied the consequences of six androgen receptor ligand-binding domain mutations on receptor function in transfected cells. The mutations, Met742Ile, Met780Ile, Gln798Glu, Arg840Cys, Arg855His and Ile869Met, were identified in PAIS patients with phenotypes representing the full spectrum seen in this condition. In all cases the androgen receptor was found to be defective, suggesting that the mutation is the cause of the clinical phenotype. The Gln798Glu mutation is exceptional in that it did not cause an androgen-binding defect in our system, although the mutant receptor was defective in transactivation assays. This mutation may affect an aspect of binding not tested, or may be part of a functional subdomain of the ligand-binding domain involved in transactivation. Overall we found milder mutations to be associated with milder clinical phenotypes. There is also clear evidence that phenotype is not solely dependent on androgen receptor function. Some of the mutant receptors were able to respond to high doses of androgen in vitro, suggesting that patients carrying these mutations may be the best candidates for androgen therapy. One such mutation is Ile869Met. A patient carrying this mutation has virilized spontaneously at puberty, so in vivo evidence agrees with the experimental result. Thus a more complete understanding of the functional consequences of androgen receptor mutations may provide a more rational basis for gender assignment in PAIS.

Yeast two-hybrid system demonstrates that estrogen receptor dimerization is ligand-dependent in vivo

Previous studies using in vitro procedures have not clearly established whether the estrogen receptor (ER) acts as a monomer or dimer in the cell. We have used the yeast two-hybrid system as an in vivo approach to investigate the dimerization of the estrogen receptor in the absence and presence of estrogen and anti-estrogens. This system is independent of ER binding to the estrogen response element. Two vectors, expressing GAL4 DNA binding domain-human ER and GAL4 transactivation domain-human ER, were constructed. Control experiments showed that each fusion protein had a high affinity binding site for estradiol-17 beta and could transactivate an ERE-LacZ reporter gene in yeast similar to the wild type ER. The two fusion proteins, GAL4 DB-hER and GAL 4 TA-hER, were expressed in the yeast strain, PCY2, which carries a GAL1 promoter-lacZ reporter. ER dimerization was measured via reconstitution of GAL4 through interaction of the fusion proteins, which transactivates LacZ through the GAL1 promoter. When both ER fusion proteins were expressed, beta-galactosidase activity was estradiol-17 beta-inducible. Furthermore, we showed that both tamoxifen and ICI 182,780 also induced beta-galactosidase activity, albeit lower than that induced by estradiol-17 beta. These results strongly argue that ER dimerization is ligand-dependent and the dimer can be induced by estradiol-17 beta, tamoxifen, or ICI 182,780. We also treated the yeast containing the two fusion proteins with estradiol-17 beta and tamoxifen or ICI 182,780 simultaneously to determine the effects on ER dimerization. beta-Galactosidase activity was lower when the yeast was treated with a higher ratio of tamoxifen or ICI 182,780 to estrogen than estradiol-17 beta alone. Taken together, we conclude that ER dimerization is ligand (estradiol-17 beta, tamoxifen, or ICI 182, 780)-dependent, and we suggest that estradiol-17 beta-induced dimers are destabilized when estradiol-17 beta is used with tamoxifen or ICI 182,780 simultaneously.

The side chains responsible for the dimerization of the estradiol receptor by ionic bonds are lost in a 17 kDa fragment extending downstream from aa 303

Fragments of 32, 26 and 17 kDa of the porcine estradiol receptor were prepared, all of which contain the ligand-binding site. While dimers of the 32 and 26 kDa fragments like those of intact receptor can be dissociated by protonation, the dimer of the 17 kDa fragment obtained by trypsination of the 26 kDa fragment is resistant to lowering the pH from 7.0 to 6.5 and below. Its dissociation can be achieved by 0.5 M MgCl2 at pH 7.0. All fragments are recognized by the MAB 13H2 in Western blots. The antibody also reacts with native receptor and the three fragments, both in their monomer and dimer states. The combining ratios of antibody with receptor, or its fragments, in the monomer and dimer states and the weakening of the estradiol-receptor bond by antibody attachment support the back to back and head to toe model of receptor dimers.

The human estrogen receptor hormone binding domain dimerizes independently of ligand activation

High level expression of biochemically active human estrogen receptor hormone binding domain (hER-HBD) was achieved using a Saccharomyces cerevisiae expression system. Using dissociation kinetic analysis, density gradient centrifugation and cross-linking studies, a spontaneous dimerization activity of hER-HBD independent of the presence of the DNA binding domain, ligand, and of elevated temperature is demonstrated.

Autoinduction of nuclear receptor genes and its significance

Although downregulation of receptors by their respective hormonal ligands is a well-studied phenomenon, relatively less is known about autoupregulation of receptors. However, an increasing number of observations of the latter process are now being reported. Here, we discuss the phenomenon of autoinduction of nuclear receptors of the steroid/thyroid hormone gene family, and its significance in the context of the developmental and gene regulatory function of the ligands. Much of this review is illustrated by recent work from our laboratory on the autoregulation of Xenopus estrogen (ER) and thyroid hormone (TR) receptors and their transcripts, accompanying or anticipating vitellogenesis and metamorphosis, respectively. The activation by estrogen (E2) of the silent vitellogenin genes and the induction of FOSP-1 genes in primary cultures of hepatocytes from male Xenopus and oviduct cells, respectively, are tightly coupled to a substantial upregulation of ER protein and its transcript. The developmental competence to activate vitellogenin in response to E2 was found to be acquired during late metamorphosis. Since the latter process is obligatorily controlled by thyroid hormones (TH), we extended our studies to the developmental and hormonal regulation of Xenopus TR genes. Although very low levels of TR alpha and beta mRNAs are detectable in embryos and early larvae, there is a large increase in the accumulation of both transcripts before the onset of metamorphosis (stage 54 tadpoles), by which time the larval thyroid gland has first begun to secrete TH. Filter and in situ hybridization revealed that the two transcripts were differentially regulated and were not equally distributed in all regions or tissues of the tadpole. Their concentration peaks at metamorphic climax and drops to low levels in froglets and adult Xenopus. Exogenous TH given to pre-metamorphic tadpoles is known to induce metamorphosis precociously. Administration of triiodothyronine (T3) to early tadpoles (stages 50/52) caused a rapid upregulation of TR alpha and beta mRNAs which was particularly marked for the beta transcript (20- to 50-fold increase in steady-state levels). This autoinduction, which is the earliest response to T3, is mimicked to variable degrees in some Xenopus cell lines. In XTC-2 cells, in which the in vivo process is fully reproduced, it was possible to show with cycloheximide that the increase in TR mRNA requires protein synthesis. It was also possible to show by transfection of XTC-2 cells with a reporter-promoter construct of Xenopus albumin gene, which is a target for T3, that the extra TR mRNA increases functional receptor in the cell. Although the role of TH is well-known in metamorphosis, we found that TR is also autoinduced in primary culture of adult male Xenopus hepatocytes. The significance of this finding lies in the fact that T3 potentiates the autoinduction of ER and the de novo activation of vitellogenin genes by E2. Prolactin (PRL) is known to exert a "juvenilizing" action by preventing the induction of amphibian metamorphosis by TH. It is therefore highly significant that PRL prevented both the autoinduction of TR alpha and beta mRNAs in whole tadpoles and organ cultures and the activation of TR target genes, such as those encoding albumin and 63 kDa adult-type keratin. Although how PRL exerts its antimetamorphic effect is not known, these findings lead us to propose a dual threshold model for the autoinduction of TR, whereby the autoinduction of TR genes requires a very low level of TR and TH to rapidly augment the amount of functional TR. This higher amount of receptor would be required to achieve a higher threshold to activate "downstream" or target genes which specify the adult phenotype at the end of metamorphosis. Finally, a survey of recent literature shows that the phenomenon of autoinduction is not restricted to Xenopus ER and TR but is more widespread among members of the nuclear receptor family.