Structure/Function of the Human Glucocorticoid Receptor: Tyrosine 735 Is Important for Transactivation
D. W. Ray, C.-S. Suen, A. Brass, J. Soden, and A. White Department of Medicine (D.W.R., J.S., A.W.) and School of Biological Sciences (C.-S.S., A.B., A.W.) University of Manchester Manchester, M13 9PT, United Kingdom
Ligand-induced activation of the glucocorticoid receptor (GR) is not well understood. The GR ligandbinding domain was modeled, based on homology with the progesterone receptor. Tyrosine 735 interacts with the D ring of dexamethasone, and substitution of D ring functional groups results in partial agonist steroids with reduced ability to direct transactivation. Loss of the Tyr735 hydroxyl group by substitution to phenylalanine (Tyr735Phe) did not reduce ligand binding affinity [dissociation constant (Kd) 4.3 nM compared with Kd 4.6 nM for wild-type] and did not alter transrepression of an nuclear factor-kB (NF-kB reporter. But, there was a significant 30% reduction in maximal transactivation of a mouse mammary tumor virus (MMTV) reporter, although with an unchanged EC50 (8.6 nM compared with 6 nM). Substitution to a nonaromatic hydrophobic amino acid, valine (Tyr735Val), retained high-affinity ligand binding for dexamethasone (Kd 6 nM compared with 4.6 nM) and did not alter transrepression of NF-kB. However, there was a 36% reduction in MMTV activity with a right shift in EC50 (14.8 nM). The change to serine, a small polar amino acid (Tyr735Ser), caused significantly lower affinity for dexamethasone (10.4 nM). Maximal transrepression of NF-kB was unaltered, but the IC50 for this effect was increased. Tyr735Ser had a major shift in EC50 (118 nM) for transactivation of an MMTV reporter. Maximal transactivation of MMTV induced by the natural ligand cortisol was reduced to 60% by Tyr735Phe and Tyr735Val and was completely absent by Tyr735Ser. These data suggest that tyrosine 735 is important for ligand interpretation and transactivation. (Molecular Endocrinology 13: 1855–1863, 1999)
0888-8809/99/$3.00/0 Molecular Endocrinology Copyright © 1999 by The Endocrine Society
INTRODUCTION The molecular mechanism of glucocorticoid-induced transcription is well understood. After ligand binding the glucocorticoid receptor (GR) migrates to the nucleus and binds to glucocorticoid response elements in the regulatory region of target genes. The activated GR then recruits cofactors, including the GR interacting protein 1 (GRIP-1) and steroid receptor coactivator 1 (SRC1) (1–6). These accessory proteins bind to the receptor and link the GR with the general transcription machinery. The GR has a modular structure, with a central DNA binding motif, which is well conserved among the nuclear receptor superfamily. The N-terminal region contains a transactivation domain, and the C-terminal region contains the ligand binding function as well as a further transactivation domain and nuclear localization signal. Members of the nuclear receptor superfamily whose crystal structures have been solved have a ligand-binding domain (LBD) consisting of 12 a-helices (7–9). These 12 helices are thought to form a pocket with a hydrophobic lining into which the ligand binds. The process whereby ligand binding results in an activated receptor is uncertain. The crystal structure of the estrogen receptor (ER) with agonist and antagonist has been revealing. It appears that the C terminal helix moves across to enclose the ligand and thereby generates a new composite surface on the receptor. The crystal structure of the ER with antagonist bound, however, shows that helix 12 is in a different position, and so the activated surface is not formed. Presumably, this results in a receptor that is unable to efficiently recruit the necessary coactivators to augment transcription (8, 10). Glucocorticoid effects on transcription may be mediated by both the direct binding of activated GR dimers to target DNA, but also by binding of receptor monomers to other transcription factors, including AP-1, nuclear factor-kB (NF-kB) and NUR77 (11–21). It has been shown that these two modes of receptor activity are dissociable, i.e. negative effects on NF-kB 1855
MOL ENDO · 1999 1856
activity can be retained when there is loss of transactivation. Specific mutations in the GR cause selective loss of GR dimerization and so prevent dimer-dependent transactivation. Such dissociated receptors (22) retain the ability to oppose other transcription factor function. It is clear that the GR dimer-independent mechanism is capable of subserving more glucocorticoid actions than previously thought – a total knockout of the GR prevents neonatal survival, but specific abrogation of receptor dimerization results in mice with no gross phenotype (23). A number of synthetic GR ligands have been identified with similar activity to dexamethasone in opposing NF-kB signaling but reduced ability to transactivate a GR dimer-dependent reporter gene (23). These dissociating glucocorticoid ligands provide an interesting insight on receptor function. The GR is capable of binding these ligands with high affinity, and the activated receptor is capable of nuclear translocation. However, the GR is presumably incapable of recruiting the necessary cofactors for transactivation. This suggests that there are signals encoded within the ligand, which, if detected by the ligand-binding pocket, promote the receptor to undergo full conformational change into a transactivating transcription factor. Therefore, the partial agonist ligands would be capable of binding to the receptor, but only of inducing a partial conformational change. Further, true antagonists at the receptor would be predicted to bind receptor but fail to induce the conformational changes needed for either transactivation or interaction with other transcription factors. The partial agonist RU24858 is the most specific yet characterized (23). The molecule differs from dexamethasone or cortisol in the D ring, suggesting that this part of the ligand conveys the activation signal. The A, B, and C rings, which are identical to the full agonists and similar to receptor-binding antagonists, may convey affinity information. The aim of this study was to model the LBD of the GR, by its homology with the progesterone receptor (PR), and further to identify amino acids within the ligand-binding pocket that may differentiate between agonists, partial agonists, and antagonists. In this work we have identified tyrosine 735 as important for ligand binding and ligand-dependent transactivation. The hydroxyl side chain contributes to receptor activation but not to ligand binding as evidenced by mutation to phenylalanine, which results in unchanged ligand binding affinity and transrepressive capacity, but reduced transactivation potential. Mutation to valine results in a minimal reduction in ligand binding affinity, no change in transrepression, but similar changes in transactivation as the phenylalanine mutant. In contrast, mutation to the small polar side chain amino acid serine results in reduced ligand binding affinity, unaltered ability to transrepress NF-kB, but a marked right shift in the transactivation dose response and a
Vol 13 No. 11
further blunting of the dose response. Therefore, the hydrophobic benzene ring of tyrosine contributes to ligand binding, and the tyrosine hydroxyl side chain contributes to transactivation.
RESULTS Human GR LBD We used the crystal structure of the human PR as a template for modeling the human GR because the two molecules have more than 50% identity in the ligand-binding region and strong sequence similarity. The 12 helices observed in the ER, PR, and retinoic acid receptor-g are all preserved in the human GR (Fig. 1). Furthermore, the human GR shares the highly conserved amino acids that make up the nuclear receptor signature (7). In common with the PR, helices 2 and 3 form a contiguous helix as do helices 10 and 11 (9). The tertiary structure GRa LBD containing 251 amino acids (Q527-K777) was predicted using SWISS-MODEL (24). The result included structural alignment against templates and three-dimensional (3D) coordinates. It used the 1.8 Å PR LBD dimer (1A28) as a template, and the identity between these two LBDs is more than 50%. Evaluation of the 3D Structure The quality verification of this structure was performed using the PROCHECK program (25), which generates a Ramachandran plot. About 99% of the f-w angles in this model were placed within the
Fig. 1. Structure of the LBD of the Human GR with Dexamethasone Helices 2 and 3 are contiguous as are helices 10 and 11. The N and C terminals are indicated as are helices 1 and 12.
GR:Tyrosine735
most favored and also within allowed regions of the conformational space. The Protein Structure Analysis (Prosa) program (26) was used for energy analysis of this model. This predicts that each residue interaction energy in the structure was negative. The 3D structure predicted for the LBD of the human GR could be superimposed upon that for the PR with a root mean square deviation (rms) of 0.38 Å for 250 Ca atoms. The LBD of the hGR is outlined by helices 5, 7, 11, and 12, the b-turn, and loops L6–7 and L11–12 (Fig. 1). The ligand-binding pocket is predicted to be lined by 18 amino acids. Of these, 15 are predicted to contribute to the hydrophobic environment of the pocket: Met560, Leu563, Leu566, Gly567, Trp600, Met601, Met 604, Ala605, Leu608, Phe623, Met 646, Leu732, Tyr735, Thr739, Phe749. There are three polar residues, two at one end of the pocket, Gln570 and Arg611, and the other, Cys736, at the opposite end (Fig. 2a). Fitting of Ligand within the Predicted Structure The structure of dexamethasone (27) was obtained from the Cambridge Structural Database (CSD). We predict that the ligand-binding pocket is a longitudinal cleft with two polar residues at one end and a single polar residue at the opposite end. We cannot determine with absolute certainty the orientation of the ligand within the pocket, but several lines of evidence support the orientation of the steroid A ring at the Arg611 end of the pocket (Fig. 2a). The ligands for both the ER and the PRs are orientated in this manner, and both cortisol and aldosterone within the mineralcorticoid receptor are predicted to be in this orientation, based on modeling and mutagenesis studies (28). The ligand interaction with the receptor was detected using the Ligplot program, which was also used to plot the interaction (29). Based on this analysis we predict that the LBD of the human GR has a three-layered, antiparallel, 12ahelical structure, which is highly homologous to that of the PR. The ligand is bound by three hydrogen bonds; Arg611 and Gln570 interact with the ketone group on C3 of the steroid A ring, and Cys736 interacts with the ketone group on C20 of the D ring. An additional 15 amino acids are found to contribute to hydrophobic interactions with the ligand (Fig. 2a). Arg611, Gln570, and Cys736 are conserved residues whose importance for ligand binding has been verified by studying either natural or engineered mutants (9, 28, 30, 31). The aromatic ring of Tyr735 appeared to form a hydrophobic interaction with the ligand. The distance from the aromatic ring to the C22 substituent of the steroid D ring was calculated to be 2.98 A, and to the D ring 3.98 A. However, the hydoxyl group of tyrosine 735 was orientated away from the ligand-binding pocket (Fig. 2b).
1857
Affinity of Mutant Human GR for Dexamethasone As Tyr735 was predicted to have hydrophobic interaction with the D ring of dexamethasone, it was important to identify changes in ligand binding affinity caused by the mutations at position 735. Site-directed mutagenesis to phenylalanine (Tyr735Phe) resulted in no alteration in ligand binding affinity (4.3 nM compared with 4.6 nM for wild-type), in keeping with the hypothesis that the benzene ring is sufficient to generate a hydrophobic surface for ligand interaction. Change to valine (Tyr735Val) resulted in lower affinity binding compared with wild-type, but only to a minor degree (6 nM). The change to serine (Tyr735Ser), however, resulted in a 2-fold reduction in ligand binding (10.4 nM) (Table 1 and Fig. 3). All the mutant receptors were expressed in COS 7 cells, and the mutations did not alter receptor numbers per cell. Transactivation by Mutant GR All three mutant GR molecules were capable of transactivating the mouse mammary tumor virus (MMTV) promoter (Fig. 4 and Table 1). In response to dexamethasone, Tyr735Phe had a similar EC50 to the wildtype (8.6 nM compared with 6 nM) and Tyr735Val a lower EC50 14.8 nM. However, the maximal transactivation potential of both of these two mutant GR molecules was less than the wild type (Fig. 4 and Table 1). The EC50 of Tyr735Ser was increased to 118 nM, and an accurate estimate of the maximal effect was not possible (Fig. 4). It is clear that this effect is disproportionate to the doubling of Kd (Fig. 3), suggesting that disruption of ligand binding is insufficient alone to explain the observed change in receptor transactivation (Fig. 4 and Table 1). Further, the physiological ligand hydrocortisone (100 nM) induced 18-fold induction of MMTV via the wild-type GR, 11-fold with the Tyr735Phe, 10.6-fold with the Tyr735Val, and failed to induce the reporter via the Tyr735Ser. Thus, the mutated receptors have the same rank order of activity with both agonist ligands. Transrepression by Mutant GR Transrepression by activated GR usually has a lower EC50 than transactivation and places less stringent requirements on the receptor. Hence, ligands have been identified that promote transrepression in the absence of transactivation, but not vice versa. We examined the ability of the mutant GR molecules to inhibit NF-kB p65-mediated transactivation through a NF-kB response element linked to luciferase (Fig. 5). The reporter was driven by cotransfection of a p65 expression vector. The wild-type receptor achieved significant suppression at 0.1 nM dexamethasone and maximal suppression at 1 nM, as did Tyr735Phe and Tyr735Val. Tyr735Ser had a minor, but consistent, increase in IC50 for this effect, with
MOL ENDO · 1999 1858
Vol 13 No. 11
Fig. 2. Orientation and Amino Acid Contacts of Dexamethasone in the Ligand-Binding Pocket of the GR a, Coordination of dexamethasone within the ligand-binding pocket of the human GR. Black, Carbon; red, oxygen; green, fluorine; blue, nitrogen; and yellow, sulfur. Plot generated using Ligplot. Hydrogen bonds are indicated as are their lengths. The hydrophobic interaction between Tyr735 and the C22 substituent of the steroid D ring is marked, as is its calculated distance in Angstroms. b, Orientation of the side chain of tyrosine 735 in relation to the structure of dexamethasone. The GR peptide backbone is indicated by lines. The side chain of Tyr735 is depicted in ball and stick form. The structure of dexamethasone is located in the ligand-binding pocket. The distance between the aromatic ring of Tyr735 and the C22 of the steroid is marked. Colors as for panel a.
no suppression at 0.1 nM dexamethasone (Fig. 5). These data are compatible with the observed Kd for binding to dexamethasone and show the dose-response curve for transrepression to be left-shifted in comparison with transactivation (Figs. 4 and 5). In contrast to the transactivation data, Tyr735Phe and Tyr735Val performed similarly to the wild-type GR, showing that the substitution of Tyr735 results in
selective impairment of transactivation in the absence of significant changes in ligand binding affinity and transrepression. Tyr735Ser has a slightly higher IC50 for transrepression compared with the other three GR molecules examined, compatible with the observation that its affinity for Dex is reduced (Fig. 5 and Table 1). However, at 100 nM Dex Tyr735Ser has achieved maximal suppression,
GR:Tyrosine735
1859
Fig. 2. Continued.
Table 1. Affinity of Glucocorticoid Receptors for Dexamethasone (Dex) and Their Activity on Transactivation of an MMTV Promoter, and Transrepression of an NFkB-Responsive Promoter (NRE) in Response to Dex MMTV Kd for Dx (nM)
Wild-type Tyr735Phe Tyr735Val Tyr735Ser
4.6 4.3 6 10.4
NRE
EC50 for Dex (nM)
Vmax for Dex (3106)
IC50 for Dex (nM)
Vmax for Dex (% Suppression)
6 8.6 14.8 118
3.3 2.4 2.1 1.9
0.04 0.07 0.08 0.4
65 52 60 55
which is close to that observed with the wild-type GR, in striking contrast to the results seen on transactivation (Figs. 4 and 5 and Table 1).
DISCUSSION Synthetic glucocorticoids have proved powerful drugs for treatment of human disease. There is a concern that use of such drugs results in an indiscriminate pattern of activity, as most cells express GRs and glucocorticoid action regulates so many cellular processes. Much has been learned about GR function
(1–5), but the mechanism of receptor activation by ligand remains mysterious. This work attempts to gain insight into this mechanism, by modeling the GR structure and using this model as a template for sitedirected mutagenesis studies. The C3 ketone group of the steroid A ring formed two hydrogen bonds with Arg611 (helix 5), bond length 3.00 A, and Gln570 (helix 3), bond length 3.16 A. These two amino acids are found to be conserved in the mineralocorticoid receptor (MR) (Arg817 and Gln776), and the orientation of ligand within both the MR, determined by mutagenesis experiment (28), and the PR, determined by crystal structure (9), is the same as in
MOL ENDO · 1999 1860
Fig. 3. Ligand Binding Affinity of Wild-Type and Tyr735Ser GR Molecules for Dexamethasone Representative Scatchard plot of dexamethasone binding to wild-type and Tyr735Ser GR molecules expressed in COS 7 cells.
Fig. 4. Dexamethasone Dose Response for the GR Molecules Wild-type, Tyr735Phe, Tyr735Ser, and Tyr735Val were all expressed in COS 7 cells with an MMTV-luc reporter gene and were treated with dexamethasone as shown. The response light units/U lacZ is depicted on the y axis. Error bars show SD, n 5 3.
our model. It is relevant that mutation of the Arg817 to Ala in the MR resulted in undetectable binding to cortisol, and Gln776 to Ala resulted in 40-fold lower affinity for cortisol, strongly suggesting their involvement in ligand binding (28). The model predicts that Cys736, within helix 11, makes contact with the ligand by formation of a hydrogen bond with the steroid D ring C20 ketone group. The bond length is calculated to be 3.04 A. Cys736 is conserved in the related MR (Cys942), and the PR. Affinity labeling experiments suggest that Cys736 lies in close proximity to the ligand. Natural mutation of the
Vol 13 No. 11
Fig. 5. Dexamethasone Dose Response for the GR Molecules Wild-type, Tyr735Phe, Tyr735Ser, and Tyr735Val were all expressed in COS 7 cells with an NRE-luc reporter gene driven by expression of the p65 subunit of NF-kB. Cells were divided and treated with dexamethasone at the indicated concentrations. The relative response in light units/U lacZ is depicted relative to the normalized control value for each experiment. Error bars show SD, n 5 3.
homologous residue in mouse, mGR (Cys742Gly), which was found in dexamethasone-resistant lymphoma, results in a receptor molecule with reduced activity (30). In addition, extensive mutagenesis of Cys736 has been performed in yeast and mammalian cells (31). Change to alanine results in no change in receptor function. In contrast, change to Ser results in near-absent transactivation in response to cortisol, and a right shift in dose response to triamcinolone. Interestingly, Cys736Thr results in a molecule with enhanced transactivation response to triamcinolone and reduced response to cortisol (31). These data suggest that different steroids may interact with the pocket-lining amino acids in subtly different ways. All other amino acid substitutions resulted in absent receptor activity. These studies make it clear that the relatively hydrophobic residues, Ala and Cys, function better than Ser or Thr, which have polar hydroxyl side groups, except that Thr functions better with triamcinolone as the ligand. Affinity labeling experiments with the GR have suggested that Met604 (helix 5), Cys638 (loop 6–7) and Cys736 (helix 11) are in close proximity to the ligand binding surface (32). In our model Met604 and Cys736 are within 4.5 A of dexamethasone, but Cys638 is not. Met604 and Cys736 were identified using labeled triamcinolone and the Cys638 by dexamethasone mesylate, and the different molecular structure of this steroid may explain the discrepancy. The adjacent amino acid Tyr735 appears to contribute to the hydrophobic internal surface of the ligandbinding pocket, but the hydroxyl group is too far away to interact with ligand. It is relevant that Tyr735 is tightly conserved through vertebrate evolution, with substitution to Phe in Xenopus and Tilapia, and to
GR:Tyrosine735
isoleucine in guinea pig and rainbow trout. The change to Phe is conservative, in that the benzene ring is preserved. The human PR also has Tyr at the equivalent position. It appears that the steroid D ring carbons C15, C16, and C22 of dexamethasone would make close contact with Tyr735, but the hydroxyl group could not be accommodated in the ligand binding interaction. Dexamethasone is a more potent glucocorticoid agonist than cortisol and has a larger, hydrophobic region in close proximity to the Tyr735, with the C22 attached to C16. Cortisol has a single hydrogen in this position and so would be expected to make weaker contact with the Tyr735 benzene ring. Progesterone, a steroidal GR antagonist, also lacks a C16 substituent, but also has a smaller, less polar C17 attachment with no hydroxyl groups on C21, C17, and C11, which are all orientated on the opposite side of the molecule. Progesterone, therefore, is a less bulky molecule, with a calculated Van der Waal’s volume of 304.8 A3 compared with 318 A3 for cortisol (33), and this may reduce the ligand interaction with Tyr735. To identify amino acid residues that could potentially distinguish agonists from antagonists, we looked at residues predicted to interact with the steroid D ring. The steroid recognizing amino acid should also potentially be capable of transmitting a signal to the rest of the molecule, perhaps by hydrogen bond formation, and Tyr735 fulfilled these criteria. Mutation of such a residue would be predicted to alter transactivation but not ligand binding affinity. Previously identified mutants with this characteristic include mutation of a highly conserved Gln755 within the AF-2 helix, which reduces binding to coactivators (34–36). Clearly Gln755 could not be involved in ligand recognition as it is directed away from the ligand-binding pocket. Mutation of Tyr735 confirms the importance of this residue for ligand binding and transactivation. Substitution by phenylalanine results in no change to ligand binding affinity, as predicted by the model, but does reduce ligand-mediated transactivation. More severe disruption of Tyr735 to valine and to serine indicates the importance of the aromatic ring and hydrophobicity, respectively, for efficient binding to dexamethasone. The change in transactivation ability shown by the mutants is not predictable from their ligand binding affinities, which are, however, concordant with their ability to transrepress. This suggests that the Tyr735 does subserve two functions, not only in mediating ligand binding but also contributing to the conformational change of helix 12 that is predicted to be required for efficient recruitment of transcriptional coactivator molecules. We propose that there are two separate and separable signals encoded within the ligand: one directs high-affinity binding to a specific receptor, and the second induces receptor conformational change. Therefore, steroidal GR antagonists, such as progesterone, contain signal 1, but not signal 2. It is proposed further that partial agonists, such as RU24858, encode high-affinity binding but only allow partial receptor conformational change such that they can interact
1861
with other transcription factors but not efficiently with the coactivators. Full agonists and steroidal antagonists share structural features at the steroid A, B, and C rings, but differ at the D ring. Therefore, it seems likely that the signal to activate the receptor lies within the steroid D ring, and that this promotes receptor conformational change by interaction with a recognition motif within the ligand-binding pocket. In summary, we have identified Tyr735 as an important amino acid for ligand binding, and for mediating ligand-dependent transactivation – an activity conferred by its hydroxyl group. Mutations at this residue have a disproportionate impact on transactivation compared with either ligand binding affinity or transrepression, suggesting that Tyr735 forms part of the mechanism for ligand-dependent conformational change.
MATERIALS AND METHODS Visualization of Ligand-Docked Receptor The figures were produced and drawn using Raster3D (www. bmsc.washington.edu/raster 3d/html/raster 3d.html) (37). Generation of Mutant GR An EcoRI fragment of the human GR cDNA (1630–2377) was subcloned into pGEM 11Zf(1), and codon 735 was mutated using a GeneEditor kit from Promega Corp. (Madison, WI) and synthetic mutagenic oligonucleotides synthesized by Perkin Elmer Corp. (Norwalk, CT). The identity of all mutants was verified by sequencing. Mutated fragments were subcloned back into the hGR cDNA in pcDNA3, and orientation was confirmed by enzyme digestion. Ligand Binding COS 7 cells, obtained from European collection of animal cell cultures (ECACC), were transfected in 10-cm plates with 5 mg of the GR expression vectors using Lipofectamine plus reagent, as suggested by the manufacturer. Cells were split after 24 h into 24-well plates and were serum starved for 16 h before study. Cells were incubated with tritiated dexamethasone (Amersham International, Bucks, UK) at increasing concentrations between 0.1 and 20 nM. Incubations were performed in multiples of 3, and at each concentration three wells were incubated with 100-fold excess of cold dexamethasone to measure nonspecific binding. All experiments were performed on four separate occasions. After 1 h incubation at 37 C, cells were washed three times in serum-free medium and were lysed in Tris-Cl (pH 7.8), 150 mM NaCl, 1% Triton-X 100. The cell lysate was counted in a Packard b-counter (Packard Instruments, Meriden, CT). Specifically bound counts were calculated and bound ligand was determined from the measured specific activity of the tritiated dexamethasone. Ligand binding affinity was calculated by nonlinear regression analysis using the Simfit package (WF Bardsley, University of Manchester, UK), as previously described (38). Transactivation Analysis COS 7 cells were transfected with 2 mg MMTV-luc (39), 1 mg CMV-bGAL, and 1 mg CMV-hGRwt, phe, ser, or val using Lipofectamine plus in 10-cm tissue culture dishes. Cells were
MOL ENDO · 1999 1862
divided and treated in triplicate. Cells were harvested after 24 h, and the luciferase activity was measured as previously described (39). Results were normalized to b-galactosidase as measured using the O-nitrophenyl b-D-galactopyranoside assay, by dividing light units by optical density units, as described previously (28). Experiments were repeated on four occasions with similar results. Transrepression Analysis COS 7 cells were transfected with an NRE-luc reporter (17), an NF-kB p65 expression vector (17), CMV-bgal, and GR expression vectors. Cells were then split and treated with dexamethasone for 48 h before harvest, and luciferase/lacZ assays. Results are expressed as light units per unit lacZ, to control for transfection efficiency (28).
Acknowledgments Received May 21, 1999. Revision received July 27, 1999. Accepted August 4, 1999. Address requests for reprints to: D. W. Ray, Department of Medicine, Endocrine Sciences Research Group, University of Manchester, Stopford Building, Manchester, United Kingdom M13 9PT. E-mail:
[email protected]. David Ray was supported by a Glaxo-Wellcome Research Fellowship.
REFERENCES 1. Evans RM 1988 The steroid and thyroid hormone receptor superfamily. Science 240:889–895 2. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P, Evans RM 1995 The nuclear receptor super-family: the second decade. Cell 83:835–839 3. Tsai MJ, O’Malley BW 1994 Molecular mechanisms of action of steroid/thyroid receptor superfamily members. Annu Rev Biochem 63:451–486 4. Horwitz KB, Jackson TA, Bain DL, Richer JK, Takimoto GS, Tung L 1996 Nuclear receptor coactivators and corepressors. Mol Endocrinol 10:1167–1177 5. Perlmann T, Evans RM 1997 Nuclear receptors in Sicily: all in the famiglia. Cell 90:391–397 6. Hollenberg SM, Evans RM 1988 Multiple and cooperative trans-activation domains of the human glucocorticoid receptor. Cell 55:899–906 7. Wurtz JM, Bourguet W, Renaud JP, Vivat V, Chambon P, Moras D, Gronemeyer H 1996 A canonical structure for the ligand-binding domain of nuclear receptors [published erratum appears in Nat Struct Biol 1996 Feb;3(2): 206]. Nat Struct Biol 3:87–94 8. Brzozowski AM, Pike AC, Dauter Z, Hubbard RE, Bonn T, Engstrom O, Ohman L, Greene GL, Gustafsson JA, Carlquist M 1997 Molecular basis of agonism and antagonism in the oestrogen receptor. Nature 389:753–758 9. Williams SP, Sigler PB 1998 Atomic structure of progesterone complexed with its receptor. Nature 393:392–396 10. Nolte RT, Wisely GB, Westin S, Cobb JE, Lambert MH, Kurokawa R, Rosenfeld MG, Willson TM, Glass CK, Milburn MV 1998 Ligand binding and co-activator assembly of the peroxisome proliferator-activated receptor-g. Nature 395:137–143 11. Jonat C, Rahmsdorf HJ, Park KK, Cato AC, Gebel S, Ponta H, Herrlich P 1990 Antitumor promotion and antiinflammation: down-modulation of AP-1 (Fos/Jun) activity by glucocorticoid hormone. Cell 62:1189–1204 12. Schule R, Evans RM 1991 Functional antagonism between oncoprotein c-Jun and steroid hormone receptors. Cold Spring Harbor Symp Quant Biol 56:119–127
Vol 13 No. 11
13. Miner JN, Yamamoto KR 1991 Regulatory crosstalk at composite response elements. Trends Biochem Sci 16: 423–426 14. Heck S, Kullmann M, Gast A, Ponta H, Rahmsdorf HJ, Herrlich P, Cato AC 1994 A distinct modulating domain in glucocorticoid receptor monomers in the repression of activity of the transcription factor AP-1. EMBO J 13: 4087–4095 15. Konig H, Ponta H, Rahmsdorf HJ, Herrlich P 1992 Interference between pathway-specific transcription factors: glucocorticoids antagonize phorbol ester-induced AP-1 activity without altering AP-1 site occupation in vivo. EMBO J 11:2241–2246 16. Auphan N, Didonato JA, Rosette C, Helmberg A, Karin M 1995 Immunosuppression by glucocorticoids: inhibition of NF-kB activity through induction of I kappa B synthesis. Science 270:286–290 17. Caldenhoven E, Liden J, Wissink S, van-de Stolpe A, Raaijmakers J, Koenderman L, Okret S, Gustafsson JA, van-der Saag PT 1995 Negative cross-talk between RelA and the glucocorticoid receptor: a possible mechanism for the antiinflammatory action of glucocorticoids. Mol Endocrinol 9:401–412 18. Ray A, Prefontaine KE 1994 Physical association and functional antagonism between the p65 subunit of transcription factor NF-kB and the glucocorticoid receptor. Proc Natl Acad Sci USA 91:752–756 19. Scheinman RI, Cogswell PC, Lofquist AK, Baldwin Jr AS 1995 Role of transcriptional activation of I kappa B a in mediation of immunosuppression by glucocorticoids. Science 270:283–286 20. Scheinman RI, Gualberto A, Jewell CM, Cidlowski JA, Baldwin Jr AS 1995 Characterization of mechanisms involved in transrepression of NF-k B by activated glucocorticoid receptors. Mol Cell Biol 15:943–953 21. Philips A, Maira M, Mullick A, Chamberland M, Lesage S, Hugo P, Drouin J 1997 Antagonism between Nur77 and glucocorticoid receptor for control of transcription. Mol Cell Biol 17:5952–5959 22. Reichardt HM, Kaestner KH, Tuckermann J, Kretz O, Wessely O, Bock R, Gass P, Schmid W, Herrlich P, Angel P, Schutz G 1998 DNA binding of the glucocorticoid receptor is not essential for survival. Cell 93:531–541 23. Vayssiere BM, Dupont S, Choquart A, Petit F, Garcia T, Marchandeau C, Gronemeyer H, Resche-Rigon M 1997 Synthetic glucocorticoids that dissociate transactivation and AP-1 transrepression exhibit antiinflammatory activity in vivo. Mol Endocrinol 11:1245–1255 24. Peirsch MC 1996 ProMod and Swiss-Model: Internetbased tools for automated comparative protein modelling. Biochem Soc Trans 24:274–279 25. Laskowski RA, MacArthur MW, Moss DS, Thornton JM 1993 PROCHECK: a programme to check the stereochemical quality of protein structures. J Appl Crystall 26:283–291 26. Sippl MJ 1993 Boltzmann’s principal, knowledge based mean fields and protein folding. An approach to the computational determination of protein structures. J Comput Aided Mol Des 7:473–501 27. Rohrer DC, Duax WL 1977 9 a Fluoro-16a-methyl11 b ,17 a ,21-trihydroxy-1,4-pregnadiene-3,20-dione (Dexamethasone) (C22H29FO5). Crystal Struct Commun 6:123–126 28. Fagart J, Wurtz JM, Souque A, Hellal-Levy C, Moras D, Rafestin-Oblin ME 1998 Antagonism in the human mineralocorticoid receptor. EMBO J 17:3317–3325 29. Wallace AC, Laskowski RA, Thornton JM 1995 LIGPLOT: a programme to generate schematic diagrams of protein-ligand interactions. Protein Eng 8:127–134 30. Byravan S, Milhon J, Rabindran SK, Olinger B, Garabedian MJ, Danielsen M, Stallcup MR 1991 Two point mutations in the hormone-binding domain of the mouse
GR:Tyrosine735
31.
32.
33. 34.
35.
glucocorticoid receptor that dramatically reduce its function. Mol Endocrinol 5:752–758 Lind U, Carlstedt Duke J, Gustafsson JA, Wright AP 1996 Identification of single amino acid substitutions of Cys736 that affect the steroid-binding affinity and specificity of the glucocorticoid receptor using phenotypic screening in yeast. Mol Endocrinol 10:1358–1370 Carlstedt Duke J, Stromstedt PE, Perrson B, Cederlund E, Gustafsson JA, Jornvall H 1988 Identification of hormone-interacting amino acids within the steroid-binding domain of the glucocorticoid receptor in relation to other steroid hormone receptors. J Biol Chem 263:6842–6848 Bogan AA, Cohen FE, Scanlan TS 1998 Natural ligands of nuclear receptors have conserved volumes. Nat Struct Biol 5:679–681 Danielian PS, White R, Lees JA, Parker MG 1999 Identification of a conserved region required for hormone dependent transcriptional activation by steroid hormone receptors. EMBO J 11:1025–1033 Durand B, Saunders M, Gaudon C, Roy B, Losson R, Chambon P 1994 Activation function 2 (AF2) of retinoic
1863
36.
37. 38.
39.
acid receptor and 9-cis retinoic receptor: presence of a conserved autonomous constitutive activating domain and influence of the nature of the response element on AF-2 activity. EMBO J 11:5370–5382 Feng X, Peng ZH, Di W, Li XY, Rochette Egly C, Chambon P, Voorhees JJ, Xiao JH 1997 Suprabasal expression of a dominant-negative RXRa mutant in transgenic mouse epidermis impairs regulation of gene transcription and basal keratinocyte proliferation by RAR-selective retinoids. Genes Dev 11:59–71 Merritt EA, Murphy MEP 1994 Raster3D Version 2.0-A program for photorealistic molecular graphics. Acta Crystallogr 50:869–873 Ray DW, Littlewood AC, Clark AJ, Davis JR, White A 1994 Human small cell lung cancer cell lines expressing the proopiomelanocortin gene have aberrant glucocorticoid receptor function. J Clin Invest 93:1625–1630 Ray DW, Davis JR, White A, Clark AJ 1996 Glucocorticoid receptor structure and function in glucocorticoidresistant small cell lung carcinoma cells. Cancer Res 56:3276–3280
