Advertisement

An Active Nuclear Retention Signal in the Glucocorticoid Receptor Functions as a Strong Inducer of Transcriptional Activation*

  • Author Footnotes
    1 Supported by an Ontario Graduate Scholarship.
    Amanda Carrigan
    Footnotes
    1 Supported by an Ontario Graduate Scholarship.
    Affiliations
    Graduate Program in Biochemistry, University of Ottawa, Ottawa, Ontario K1Y 4K9
    Search for articles by this author
  • Author Footnotes
    2 Supported by an Ontario Graduate Scholarship in Science and Technology.
    Rhian F. Walther
    Footnotes
    2 Supported by an Ontario Graduate Scholarship in Science and Technology.
    Affiliations
    Graduate Program in Biochemistry, University of Ottawa, Ottawa, Ontario K1Y 4K9
    Search for articles by this author
  • Author Footnotes
    3 Supported by a graduate studentship from the Heart and Stroke Foundation of Ontario.
    Houssein Abdou Salem
    Footnotes
    3 Supported by a graduate studentship from the Heart and Stroke Foundation of Ontario.
    Affiliations
    Graduate Program in Biochemistry, University of Ottawa, Ottawa, Ontario K1Y 4K9
    Search for articles by this author
  • Dongmei Wu
    Affiliations
    Ottawa Health Research Institute, Ottawa, Ontario K1Y 4K9
    Search for articles by this author
  • Ella Atlas
    Affiliations
    Ottawa Health Research Institute, Ottawa, Ontario K1Y 4K9
    Search for articles by this author
  • Yvonne A. Lefebvre
    Affiliations
    Ottawa Health Research Institute, Ottawa, Ontario K1Y 4K9

    Providence Health Care, University of British Columbia Vancouver, British Columbia V6Z 1Y6, Canada
    Search for articles by this author
  • Robert J.G. Haché
    Correspondence
    Holds a University of Ottawa Health Research Chair. To whom correspondence should be addressed: Ottawa Health Research Institute, Ottawa Hospital, 725 Parkdale Ave., Ottawa, Ontario K1Y 4E9, Canada. Tel.: 613-762-5142; Fax: 613-761-5036
    Affiliations
    Ottawa Health Research Institute, Ottawa, Ontario K1Y 4K9

    Providence Health Care, University of British Columbia Vancouver, British Columbia V6Z 1Y6, Canada

    Departments of Medicine and Biochemistry, Microbiology, and Immunology, Ottawa, Ontario K1Y 4K9
    Search for articles by this author
  • Author Footnotes
    * This work was supported in part by a grant from the Canadian Institutes of Health Research (to Y. A. L. and R. J. G. H.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
    1 Supported by an Ontario Graduate Scholarship.
    2 Supported by an Ontario Graduate Scholarship in Science and Technology.
    3 Supported by a graduate studentship from the Heart and Stroke Foundation of Ontario.
Open AccessPublished:February 20, 2007DOI:https://doi.org/10.1074/jbc.M602931200
      The glucocorticoid receptor (GR) cycles between a naive chaperone-complexed form in the cytoplasm and a transcriptionally active steroid-bound nuclear form. Nuclear import of GR occurs rapidly and is mediated through the importin α/β karyopherin import pathway. By contrast, nuclear export of GR occurs only slowly under most conditions, despite a dependence on active signaling. In this study we have defined a nuclear retention signal (NRS) in the hinge region of GR that actively opposes the nuclear export of GR as well as the nuclear export mediated through an ectopic CRM1-dependent nuclear export signal (NES). The GR NRS overlaps closely with the basic NL1 nuclear localization signal (NLS) but can be distinguished from NL1 by targeted mutagenesis. Substitution of the classical NLS from SV40 T antigen for the GR NL1 results in a receptor in which nuclear export is accelerated. Remarkably, although the SV40-modified GR remains predominantly nuclear in the presence of steroid and is recruited to transcriptional regulatory regions indistinguishably from wild-type GR, the substitution dramatically weakens the ability of GR to activate transcription of a mouse mammary tumor virus reporter gene. These results suggest that active nuclear retention of GR plays an integral role in glucocorticoid signaling.
      The glucocorticoid receptor (GR)
      The abbreviations used are: GR, glucocorticoid receptor; NRS, nuclear retention signal; NES, nuclear export signal; NLS, nuclear localization signal; DBD, DNA binding domain; LBD, ligand binding domain; MMTV, mouse mammary tumor virus; WT, wild type; ChIP, chromatin immunoprecipitation; GFP, green fluorescent protein; FRAP, fluorescence recovery after photobleaching; GST, glutathione S-transferase; CMV, cytomegalovirus.
      5The abbreviations used are: GR, glucocorticoid receptor; NRS, nuclear retention signal; NES, nuclear export signal; NLS, nuclear localization signal; DBD, DNA binding domain; LBD, ligand binding domain; MMTV, mouse mammary tumor virus; WT, wild type; ChIP, chromatin immunoprecipitation; GFP, green fluorescent protein; FRAP, fluorescence recovery after photobleaching; GST, glutathione S-transferase; CMV, cytomegalovirus.
      is a nuclear hormone receptor of the steroid receptor subfamily whose activity is tightly regulated by glucocorticoids (
      • Beato M.
      • Herrlich P.
      • Schutz G.
      ). GR is a modular protein that features a central DNA binding domain (DBD) and a C-terminal ligand binding domain (LBD). Two activation functions (AF-1 and AF-
      • Gustafsson J.A.
      • Carlstedt-Duke J.
      • Wrange O.
      • Okret S.
      • Wikstrom A.C.
      ) are located in the N terminus and the LBD of the receptor, respectively (
      • Gustafsson J.A.
      • Carlstedt-Duke J.
      • Wrange O.
      • Okret S.
      • Wikstrom A.C.
      ,
      • Giguere V.
      • Hollenberg S.M.
      • Rosenfeld M.G.
      • Evans R.M.
      ). GR is directly and indirectly involved in both the activation and suppression of genes that are involved in cell differentiation, glucose uptake and homeostasis, inflammation, response to stress, neuronal function, lipid metabolism, and cancer (
      • Charmandari E.
      • Kino T.
      • Chrousos G.P.
      ,
      • Chrousos G.P.
      • Charmandari E.
      • Kino T.
      ).
      Transcriptional regulation by GR is tightly controlled. The regulation of function by segregation in subcellular compartment has been proposed to act as an important regulatory checkpoint for a number of transcription factors, including steroid receptors (
      • Nguyen T.
      • Sherratt P.J.
      • Nioi P.
      • Yang C.S.
      • Pickett C.B.
      ,
      • Ghisletti S.
      • Meda C.
      • Maggi A.
      • Vegeto E.
      ). Unlike other steroid receptors, such as estrogen receptor and progesterone receptor, that are nuclear in the absence of steroid, naive GR is largely cytoplasmic (
      • Picard D.
      • Kumar V.
      • Chambon P.
      • Yamamoto K.R.
      ). In the cytoplasm GR is maintained in an inactive hsp90 cochaperone complex that includes hsp70, immunophilins, p23, and other factors (
      • Pratt W.B.
      • Toft D.O.
      ). Upon ligand binding, GR dissociates from the chaperone complex, homodimerizes, and rapidly translocates to the nucleus to regulate target gene expression (
      • Madan A.P.
      • DeFranco D.B.
      ,
      • Sackey F.N.
      • Hache R.J.
      • Reich T.
      • Kwast-Welfeld J.
      • Lefebvre Y.A.
      ,
      • Yang J.
      • Liu J.
      • DeFranco D.B.
      ). Nuclear import of GR is accomplished through two nuclear localization signals as follows: NL1, located in a hinge-like region of GR that separates the DBD from the LBD; and NL2, which is within the LBD (
      • Picard D.
      • Yamamoto K.R.
      ). NL1 is a basic NLS that mediates the nuclear import of GR through interaction with importin α and importin 7 (
      • Jewell C.M.
      • Webster J.C.
      • Burnstein K.L.
      • Sar M.
      • Bodwell J.E.
      • Cidlowski J.A.
      ,
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ,
      • Freedman N.D.
      • Yamamoto K.R.
      ).
      NL2 is strictly steroid-dependent, and both the sequence within the LBD that includes NL2 and the karyopherins that determine NL2-mediated nuclear import remain to be identified. Using a mutant GR that lacks NL1 (GRNL1–), we have previously demonstrated that NL2 is an agonist-specific NLS that mediates the incomplete localization of GR to the nucleus in cells treated with cortisol or the synthetic steroid dexamethasone (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). Furthermore, this reduced nuclear occupancy correlates with a strong decrease in GR transcriptional regulatory potential.
      Upon hormone withdrawal, GR reassociates rapidly into a cochaperone complex, but only slowly redistributes to the cytoplasm over periods that can extend over 12–24 h (
      • Sackey F.N.
      • Hache R.J.
      • Reich T.
      • Kwast-Welfeld J.
      • Lefebvre Y.A.
      ,
      • Hache R.J.
      • Tse R.
      • Reich T.
      • Savory J.G.
      • Lefebvre Y.A.
      ,
      • Qi M.
      • Hamilton B.J.
      • DeFranco D.
      ). Furthermore, fluorescence recovery after photobleaching (FRAP), digitonin cell permeabilization, and other assays have shown that this slow export from the nucleus occurs for both liganded and steroid-withdrawn receptor (
      • Hache R.J.
      • Tse R.
      • Reich T.
      • Savory J.G.
      • Lefebvre Y.A.
      ,
      • Walther R.F.
      • Lamprecht C.
      • Ridsdale A.
      • Groulx I.
      • Lee S.
      • Lefebvre Y.A.
      • Hache R.J.
      ). The mechanism through which this slow transfer of GR from the cytoplasm to the nucleus is accomplished is not understood. For other transcription factors and nuclear proteins, export from the nucleus is accomplished when nuclear export signals (NESs) interact with karyopherins, collectively referred to as exportins (
      • Mosammaparast N.
      • Pemberton L.F.
      ). Most of the known NESs are composed of hydrophobic amino acid; however, they are highly diverse with only a loose consensus (
      • Fried H.
      • Kutay U.
      , ).
      The prototypical exportin for nuclear protein export is exportin1 or CRM1. CRM1 binds to short leucine- or isoleucine-rich sequences. CRM1 recognition of an NES requires the cooperative binding of RanGTP and is inhibited by leptomycin B (
      • Fornerod M.
      • Ohno M.
      • Yoshida M.
      • Mattaj I.W.
      ,
      • Stade K.
      • Ford C.S.
      • Guthrie C.
      • Weis K.
      ). Previous studies from our group and other groups suggest that nuclear export of GR may be accomplished by both CRM1-dependent and -independent mechanisms (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ,
      • Liu J.
      • DeFranco D.B.
      ). Export of steroid-withdrawn GR from the nucleus has been found to be sensitive to leptomycin B treatment and thus is likely CRM-1-dependent, whereas export of liganded GR appears to occur independently of CRM1 (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). Recently we noted that a KKK-NNN substitution of amino acids 513–515 of rat GR within NL1 not only impaired NL1 activity but also led to an accelerated redistribution of GR to the cytoplasm upon the withdrawal of steroid treatment (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). This suggested that the GR NL1 either functioned directly as, or overlapped with, an active nuclear retention signal (NRS) for the receptor.
      Active retention of GR within the nucleus had been proposed previously to be an additional regulatory step for the receptor, and enhanced export of GR has been shown to reduce its transcriptional potential (
      • Kino T.
      • Souvatzoglou E.
      • De Martino M.U.
      • Tsopanomihalu M.
      • Wan Y.
      • Chrousos G.P.
      ). Factors such as hsp90 and the 14-3-3σ protein have been suggested to promote nuclear and cytoplasmic retention of GR, respectively (
      • Kino T.
      • Souvatzoglou E.
      • De Martino M.U.
      • Tsopanomihalu M.
      • Wan Y.
      • Chrousos G.P.
      ,
      • Tago K.
      • Tsukahara F.
      • Naruse M.
      • Yoshioka T.
      • Takano K.
      ). However, specific determinants on GR required for retention have not been defined.
      In this study we have sought to define the NRS activity in GR that overlaps with the receptor NL1 and to determine its relevance to GR function. Our findings indicate that the GR NRS overlaps closely with NL1 and can be partially separated in peptides. In-line replacement of across the hinge region of GR containing NL1 with the NLS region from the SV40 T antigen compromised NRS activity in full-length GR. Notably, this substitution strongly decreased the activation of a mouse mammary tumor virus (MMTV) reporter gene by GR without affecting the recruitment of GR to the MMTV promoter. Therefore, active retention of GR in the nucleus inhibits nuclear export and appears to play an important role in determining the transcriptional regulatory potential of the receptor.

      EXPERIMENTAL PROCEDURES

      Cell Culture—COS-7 cells (ATCC) were maintained in Dulbecco's modified Eagle's medium (Invitrogen), supplemented with 10% fetal calf serum, nonessential amino acids, and sodium pyruvate. Cells were replated the day prior to transient transfection with FuGENE (Roche Applied Science) or Lipofectamine (Invitrogen), with transfections performed according to the manufacturers' recommendations.
      Plasmids—pNES-GST-GFP-NLS (see Fig. 3) was from pcDNA3.1-derived construct, FVHL-GFP, a gift from S. Lee. Rat GR and control inserts were introduced between the GST and GFP using an XhoI site. Point mutants L507A/E508A, R510A/K511A/T512A, and I516A/K517A/G518A were generated by direct cloning of oligonucleotides into pNES-GST-GFP-NLS using the same XhoI site. pGFP-GR and PGFPNL1– were described previously (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). pGFP-GRWT-SV40122–139 was created by removing amino acids 506–523 in full-length rat GR and inserting the SV40 large T antigen sequence amino acids 122–139 into a PstI site (see Fig. 5). pGFP-SV40-GRNL1– was created by fusing SV40 large T antigen amino acids 127–133 into the XhoI site at the GR N terminus. The L507A/E508A substitution in the full-length GR was inserted using Pfu (Stratagene). All constructs were confirmed by sequencing. pQE60-hCRM1 and pQE32-hRanQ69L were a gift from I. W. Mattaj (
      • Fornerod M.
      • Ohno M.
      • Yoshida M.
      • Mattaj I.W.
      ,
      • Askjaer P.
      • Jensen T.H.
      • Nilsson J.
      • Englmeier L.
      • Kjems J.
      ).
      Figure thumbnail gr3
      FIGURE 3Leu507 and Glu508 are important for NRS activity in GR hinge, but their substitution is insufficient to abrogate NRS activity in full-length GR. A, schema of GFPGSTNLSNES fusion proteins employed highlighting the sequence of the GR insert. Shading marks the NL1; underlining indicates sites of mutation; italics show the position of the NL1–substitution. B, representative photomicrographs of the constructs pre- and post-FRAP. C, initial recovery rates post-FRAP (% fluorescence return/min) were calculated using linear regression to determine the slope of the graph representing average percentage of total nuclear fluorescence in the bleached nuclei over the first 5 min. A minimum of three independent experiments was performed, with 2–3 cells of each construct analyzed per experiment. Error bars represent S.E. Asterisk indicates p < 0.05 compared with wild-type hinge control, using Student's t test. D, representative images (left) and quantification (right) of FRAP assays of full-length GR constructs. Cells were treated with 1 μm cortisol and 20 ng/μl cycloheximide 1 h prior to FRAP. At least three independent experiments were performed, with two cells analyzed per construct each time. Error bars represent S.E.
      Figure thumbnail gr5
      FIGURE 5Transcriptional effect of the SV40 hinge replacement. A, transcriptional activity of cells transfected with 6 ng of GFP-tagged full-length GR constructs from an MMTV promoter, 16 h after treatment with 10 nm dexamethasone. Activity of each construct is quantified relative to the fold activation of the wild-type receptor and is normalized to CMV Renilla internal control for each sample. Graphs show average fold activation over three experiments, each containing triplicate samples. Error bars represent S.E. with asterisk indicating p < 0.01 compared with wild-type GR, using Student's t test. B, ChIP of COS-7 cells containing a stably integrated MMTV promoter and transfected with GFP-GRWT or GFP-GRSV122–139. Binding of the various factors was compared in the absence or presence of 60 min of treatment with 10 nm dexamethasone. The results shown are representative of three independent repetitions. C, transcriptional activity of cells transfected with indicated amounts of wild-type and mutant GR, 16 h after treatment with 10 nm dexamethasone quantified as in A. Error bars represent S.E. with asterisk indicating p < 0.03 relative to activation of 18 ng of wild-type GR.
      Confocal Microscopy and FRAP—COS-7 cells were plated on 40-mm round coverslips coated with poly-l-lysine and transfected the following day with 0.5 μg of DNA and 8 μl of Lipofectamine (Invitrogen). Transfection was stopped after 16 h by adding 1 ml of phenol red-free media containing 20% fetal bovine serum. 16 h pre-FRAP, cells were withdrawn from serum. 1–2 h pre-FRAP, the cells were treated with 1 μm cortisol to induce nuclear localization of GR, and 20 μg/ml cycloheximide was added 1 h pre-FRAP to prevent synthesis of new protein over the course of the experiment. During FRAP, coverslips with cells were placed in a Bio-Rad FCS2 chamber system and maintained at 37 °C in cortisol- and cycloheximide-containing media, over the course of the experiment. Cells were imaged using 3% laser power on a Bio-Rad MRC 1024 microscope equipped with an argon-ion laser and LaserSharp software. Photobleaching of GFP was done at 100% laser power for 5–20 passes. Image analysis was performed using Image-J software (National Institutes of Health).
      Transcription Assays—COS-7 cells were transfected in triplicate with FuGENE (Roche Applied Science), as per the manufacturer's instructions, using 6–18 ng of GR expression vector, 250 ng of–237 MMTV-luciferase reporter, 25 ng of CMV-Renilla luciferase as an internal control, and sufficient pEGFP-C1 empty vector to bring the total DNA to 0.5 μg. Cells were treated with 10 nm dexamethasone for 16 h and lysed in Passive Lysis Buffer (Promega), and the luminescence was evaluated using a Lumistar luminometer. Activity was normalized to the CMV internal control and compared with fold induction of WT GR. All experiments were repeated independently a minimum of three times.
      Localization and Hormone Withdrawal Assays—COS-7 cells were plated on poly-l-lysine-coated coverslips in 6-well dishes. 16 h after plating the cells were transfected with 0.3 μg of DNA using Lipofectamine. Cells were withdrawn from serum for 16 h before treatment was initiated as described in individual experiments. Steroid withdrawal was accomplished by rinsing the cells five times for 5 min in phosphate-buffered saline containing 5% bovine serum albumin at 37 °C and then transferring cells to serum- and phenol red-free media. Cells were fixed in 4% paraformaldehyde and mounted on slides with Vectashield 4,6-diamidino-2-phenylindole staining solution (Vector Laboratories). 100–200 cells were scored per slide in each experiment, using a Zeiss Axiovert microscope equipped with a xenon lamp to visualize GFP construct-containing cells. Each experiment was performed a minimum of three times in duplicate.
      In Vitro Binding Assays and Purification of Proteins—GST and NES-GST-GR constructs were produced with the TnT® quick-coupled transcription/translation systems (Promega) in the presence of [35S]methionine for NES-GST-GR proteins as per the manufacturer's instructions. One-third of the in vitro translated mixture was incubated with 10 μl of glutathione-Sepharose beads (Amersham Biosciences) in 500 mm NaCl, 20 mm Tris-HCl, pH 7.5, 5 mm MgCl2, 5 mg/ml bovine serum albumin, 0.1% Nonidet P-40 (binding buffer), rotated for 2 h at 4 °C, followed by three washes in binding buffer. CRM1 and RanQ69L were expressed and purified under native conditions as described previously (
      • Fornerod M.
      • Ohno M.
      • Yoshida M.
      • Mattaj I.W.
      ,
      • Askjaer P.
      • Jensen T.H.
      • Nilsson J.
      • Englmeier L.
      • Kjems J.
      ). For binding assays, 500 ng of CRM1 and 500 ng of GTP-loaded RanQ69L were incubated with NES-GST-GR proteins in binding buffer for 1 h at 4°C. The beads were then washed with binding buffer and resuspended in SDS buffer.
      Western Blotting—Whole cells lysates from transfected COS-7 cells in 60-mm dishes were prepared in lysis buffer (50 mm Tris, 150 mm NaCl, 1 mm dithiothreitol, 1 mm EDTA, 10% glycerol, 0.5% Triton X-100, and Complete protease inhibitor (Roche Applied Science)). Protein yields were determined by the Bradford assay. Samples were separated on 6–10% SDS-acrylamide transferred to polyvinylidene difluoride membranes and probed with JL-8 Living Colors antibody (BD Biosciences). For the analysis of CRM1 binding to the GR-hinge constructs, eluted proteins were separated on 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and probed with anti-Exportin-1/CRM1 (BD Biosciences). The membrane was then dried, and radiolabeled GST fusion proteins were visualized by PhosphorImager.
      Chromatin Immunoprecipitation (ChIP)—ChIPs were performed essentially as described previously (
      • Wiper-Bergeron N.
      • Wu D.
      • Pope L.
      • Schild-Poulter C.
      • Hache R.J.
      ). Cells were treated with 10 nm dexamethasone for 60 min and cross-linked with 1% formaldehyde at room temperature for 10 min. Cells were then rinsed twice with ice-cold phosphate-buffered saline and washed sequentially with buffer I (0.25% Triton X-100, 10 mm EDTA, 0.5 mm EGTA, 10 mm HEPES, pH 6.5) and buffer II (200 mm NaCl, 1 mm EDTA, 0.5 mm EGTA, 10 mm HEPES, pH 6.5). Cell pellets were resuspended and sonicated. Supernatants were collected and diluted, followed by immunoclearing with 4 μg of sheared salmon sperm DNA, 65 μl of protein A-Sepharose slurry for 2 h at 4 °C. Immunoprecipitation was performed with antibodies to Gal4 DBD, GR (M-20) (Santa Cruz Biotechnology), polymerase II (N-20) (Santa Cruz Biotechnology), and anti-acetyl-histone 4 (Upstate) at 4 °C overnight. After immunoprecipitation, 45 μl of protein A-Sepharose and 2 μg of sheared salmon sperm DNA were added, and the incubation was continued for another 1 h. Precipitates were washed and extracted three times in 100 μl of 1% SDS, 0.1 m NaHCO3. Eluates were kept at 65 °C overnight to reverse cross-links. DNA was purified and amplified by PCR using the following primers for the MMTV promoter: –270 and –247, +61 and +84.

      RESULTS

      The Addition of a Strong NLS to the GRNL1 Receptor Fails to Reconstitute GR Nuclear Retention—We have previously shown that the redistribution of GR to the cytoplasm from the nucleus in COS7 cells following the withdrawal of steroid treatment occurs only very slowly, taking over 24 h for complete redistribution (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). By contrast, the return of a GR containing a K513N/K514N/K515N substitution that inactivates NL1 (GRNL1–) was much more rapid, reaching completion within 2–4 h, albeit from a starting point in which the receptor was only 50–60% nuclear (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ).
      One simple potential explanation for this result is that the GR NL1 opposed nuclear export, although this is a property that has not been previously described for other NLSs. To test whether the addition to GRNL1– of a strong NLS alone would be sufficient to promote GR nuclear retention, we introduced the classical SV40 T antigen NLS sequence into a GFP GRNL1– construct between the N-terminal GFP and the GR (Fig. 1A, GFPSV40 GRNL1–).
      Figure thumbnail gr1
      FIGURE 1Substitution of lysines 513–515 in rat GR reveal a nuclear retention activity coincident with NL1. A, schema of GFP-GR constructs employed. B, bar graph summarizing the localization of the constructs expressed in COS7 cells prior to (–) and 2 h subsequent to treatment (+) with 1 μm cortisol (Cort). Cells were fixed and counted as described under “Experimental Procedures.” C, graphical presentation of the decrease in GFP construct nuclear localization upon hormone withdrawal after 4 h treatment with 1 μm cortisol. The percentage of mostly nuclear cells at each time point is compared with localization before withdrawal (w/d) for each construct. D, schema of FRAP assay. Shading in nuclei represents fluorescent GFP fusion proteins. E, FRAP assay of GR constructs. Top, representative photomicrographs of transfected cells pre- and post-FRAP. Bottom, quantification of the recovery of fluorescence in N2. Cells were treated with 1 μm cortisol and 20 ng/μl cycloheximide 1 h prior to FRAP. The assay was performed on at least two cells of each construct in experiment over the course of at least three independent experiments. Error bars represent S.E.
      GFPSV40GRNL1– was cytoplasmic in its naive state prior to steroid treatment, and like WT GFPGR, it translocated rapidly and efficiently to the nucleus following treatment with the natural steroid cortisol (Fig. 1B, Cort) or the synthetic steroid dexamethasone (data not shown). By contrast, upon withdrawal of steroid treatment, which leads to rapid loss of bound ligand and reassociation of GR with the heat shock protein cochaperone complex (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ), redistribution of GFPSV40GRNL1– to the cytoplasm remained markedly accelerated compared with GFPGR, approaching the rate of redistribution seen for GFPGRNL1– (Fig. 1C).
      To directly determine whether this increased rate of redistribution of GFPSV40GRNL1– reflected an enhanced rate of nuclear export, we directly assessed the nuclear export rates of our constructs in live cells (Fig. 1, D and E). Our assay, which we have previously validated in detail (
      • Walther R.F.
      • Lamprecht C.
      • Ridsdale A.
      • Groulx I.
      • Lee S.
      • Lefebvre Y.A.
      • Hache R.J.
      ), employs FRAP to study the movement of GFP-labeled proteins between nuclei in naturally multinucleated cells (for example, up to 10% of COS7 cells may contain more than one nucleus). As illustrated schematically in Fig. 1D, a fluorescent nuclear protein with a strong NLS that is undergoing rapid nuclear export will be seen to transfer from its original nucleus to its photobleached twin, with the rate of transfer reflecting 50% of the rate of export plus reimport (once in the cytoplasm, the protein could be reimported into either nucleus).
      As shown previously (
      • Walther R.F.
      • Lamprecht C.
      • Ridsdale A.
      • Groulx I.
      • Lee S.
      • Lefebvre Y.A.
      • Hache R.J.
      ), liganded GFPGR redistributed only very slowly from the untreated nucleus to the photo-bleached nucleus in the presence of cortisol, such that little transfer was detected 1 h following photobleaching (Fig. 1E). By contrast, the liganded GFPSV40GRNL1–, which localized as strongly to the nucleus as GFPGR, displayed an accelerated rate of nucleo-cytoplasmic exchange comparable with that of GFPGRNL1–. Moreover, we observed the same result for the constructs remaining in the nucleus when FRAP was performed 30 min subsequent to cortisol withdrawal (data not shown).
      Together, these results indicated that the GR NL1 NLS overlaps with a nuclear retention activity that is distinct from a strong NLS, but which inhibits the export of GR from the nucleus. Furthermore, they suggested that the NRS activity inhibited both the CRM1-dependent and CRM1-independent nuclear export of GR.
      Amino Acids 500–525 of GR Contain an NRS That Overlaps Closely with NL1—The GR NL1 motif is localized to the hinge region of the receptor that separates the DBD and LBD of the receptor (
      • Picard D.
      • Yamamoto K.R.
      ). To determine whether this region of GR was sufficient for NRS activity and to determine whether this activity could actively oppose nuclear export, we expressed variations of amino acids 500–525 of the GR hinge containing NL1 in the context of a GST-GFP fusion construct in which nuclear import was driven by the SV40 NLS and nuclear export mediated by the human immunodeficiency virus reverse CRM1-dependent nuclear export signal (Fig. 2A).
      Figure thumbnail gr2
      FIGURE 2NRS activity within the GR hinge region. A, schemas of the GFP fusion proteins employed, highlighting the positions of the GFP, glutathione S-transferase (GST), the reverse NES, SV40 NLS, FLAG tag, and GR peptide insert (amino acids 500–525). GST was included in the constructs to bring the total molecular weight of the expressed proteins (55 kDa) above the limit for passive diffusion through the nuclear pore complex (40–50 kDa). GR = wild-type hinge sequence, amino acids 500–525; RG = inversion of the GR 500–525 DNA sequence; GRNL1 = GR amino acids 500–525 with the K513–515N substitution. B, representative photomicrographs (top) and graphical presentation of movement in FRAP (bottom) of the constructs described in A. ▪, RG; ▵, GRNL1–; ♦, GRWT. Cells were treated with 20 ng/μl cycloheximide for 1 h prior to FRAP. The assay was performed on at least three cells of each construct per experiment, with three independent experiments. Error bars represent S.E. C, [35S]methionine-labeled GST-GFP constructs containing variants of the hinge region of GR, as indicated, were bound to recombinant CRM1 in the presence of GTP bound RanQ69L as described under “Experimental Procedures.” Binding of CRM1 levels was detected by Western blot (top) and total GST-GFP fusion proteins used for binding was assessed by PhosphorImager (bottom).
      In the first instance, a GSTGFP fusion protein containing the reversal of GR cDNA sequence that encodes hinge amino acids 500–525 (NESGSTGR525–500GFPNLS) was observed to move rapidly between nuclei in FRAP (Fig. 2B). Indeed, the rate of movement of NESGSTGR525–500GFPNLS was indistinguishable from that of a construct lacking a cDNA insert (data not shown and see Ref.
      • Walther R.F.
      • Lamprecht C.
      • Ridsdale A.
      • Groulx I.
      • Lee S.
      • Lefebvre Y.A.
      • Hache R.J.
      ). A construct containing the K513N/K514N/K515N substitution in NL1 (NESGSTGR500–525NL1–GFPNLS) moved between nuclei with similar ease. By contrast, movement of the fusion protein containing the WT GR hinge was markedly retarded. Inhibition of nuclear export did not simply result from the presence of two strong NLSs on this construct as an NESGSTGRGFP construct lacking the SV40 NLS was also completely nuclear but similarly impaired in nuclear export (data not shown). The reduced nuclear export of the WT GR 500–525 construct appeared unlikely to reflect a substantial alteration of the interaction between the fusion proteins and CRM1, as all three constructs when translated in vitro showed the same ability to interact with GSTCRM1 in a pulldown assay (Fig. 2C).
      Next, in an attempt to distinguish between NL1 and the NRS within the GR sequence, we performed alanine-scanning mutagenesis within the GR hinge and measured the effect of these substitutions on nuclear export within the context of NES-GST-GR-GFP-NLS (Fig. 3). Two of the AAA substitutions made (510–512 and 507–508) appeared to have at least a modest effect on NRS activity at 30 min post-FRAP (Fig. 3B). However, only the 507–508 AA substitution (509 is already Ala in WT GR), which showed the most complete transfer at 30 min visually (Fig. 3B), also showed a significantly enhanced initial rate of transfer following FRAP compared with the WT GR hinge control (Fig. 3C).
      These results suggested that L507A/E508A substitution might substantially cripple NRS activity in full-length GR, which would then allow us to investigate the importance of NRS activity for GR function. GFPGR507,508AA was cytoplasmic in the presence of steroid and transferred rapidly to the nucleus upon cortisol or dexamethasone treatment in a manner that was indistinguishable from WT GR, indicating that this substitution did not impact on the GR NL1 activity (data not shown). However, the rate of nuclear export in FRAP for cortisol-treated GFPGR507,508AA was indistinguishable from that of GFPGR (Fig. 3D). The return of GFPGR to the cytoplasm following withdrawal of cortisol treatment was also unaffected by the LE/AA substitution (data not shown).
      Replacement of GR NL1 with the SV40 NLS Impairs NRS Activity and Cripples Transcriptional Activation without Affecting Promoter Recruitment—The GR hinge region is an unstructured region of the GR that allows flexible juxtaposition of the receptor DNA and ligand binding domains (
      • Kumar R.
      • Thompson E.B.
      ). Our results with GFPGR507,508AA suggested that NRS activity would be difficult to separate from NL1 by point mutagenesis within full-length GR. Nonetheless, we reasoned that a more extensive substitution replacing the entire GR NL1 with another strong NLS lacking NRS activity offered the possibility of regenerating a GR with nuclear localization activity equivalent to WT receptor and intact spacing and functionality of the DBD and LBDs.
      As the SV40 NLS was verified in our experiments to lack NRS activity, we performed an in-line replacement of the 18 amino acids of GR beginning at 506 (506–523) with the sequence from SV40 in a manner that preserved the positioning of the basic amino acid clusters of the NLSs (Fig. 4A, GRSV122–139). The hinge region of GR between amino acid 506 and 540 exhibits a disordered structure that does not contact DNA or influence GR DNA binding specificity (
      • Luisi B.F.
      • Xu W.X.
      • Otwinowski Z.
      • Freedman L.P.
      • Yamamoto K.R.
      • Sigler P.B.
      ), and previous studies have shown that truncation of the C terminus of GR to 505 does not significantly affect its DNA binding (
      • Hollenberg S.M.
      • Giguere V.
      • Segui P.
      • Evans R.M.
      ). GFPGRSV122–139 expressed at the same level as GFPGR and GFPGRNL1– (Fig. 4B) and transferred to the nucleus as efficiently as GFPGR in response to cortisol treatment. Moreover, GFPGRSV122–139 transferred between nuclei in FRAP at a rate significantly faster than GFPGR, albeit still somewhat slower than GFPGRNL1– (Fig. 4C). Similar results also were observed following treatment with dexamethasone (data not shown). Thus our extended SV40 substitution substantially compromised the NRS activity revealed by the NL1–substitution, yet still did not eliminate it completely.
      Figure thumbnail gr4
      FIGURE 4Effect of SV40 NLS and surrounding sequence on GR export rate. A, GR residues 506–523 were replaced by SV40 large T antigen 122–139. This placed the SV40 NLS in a similar position to that of WT GR within the receptor. B, Western blot of whole cell extract WCE from COS-7 cells transiently transfected with GR full-length constructs. GFP antibody (JL-8, BD Biosciences) was used to detect GFP fusion proteins, with Ku70 levels used as a loading control. C, representative pre- and post-FRAP images (top) and graph (bottom) of FRAP assays. Cells were treated with 1 μm cortisol and 20 ng/μl cycloheximide 1 h prior to FRAP. A minimum of three independent experiments was performed, with 2–3 cells of each construct analyzed per experiment. Error bars represent S.E.
      To determine whether this decrease in NRS activity affected the ability of the receptor to activate transcription, we compared the ability of WT GFPGR, GRNL1–, and GRSV122–139 to activate transcription from the mouse mammary tumor virus promoter when expressed at the same levels in COS 7 cells (Fig. 5A). Steroid-mediated transcriptional activation from the MMTV promoter has been shown previously to be completely dependent on a complex series of glucocorticoid-response elements that occur between –70 and –180. Previously, we had shown that the GRNL1– substitution strongly decreased the GR transcriptional activation potential without affecting the DNA binding affinity of GR (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). This correlated with the slower and incomplete nuclear transfer of GRNL1– to the nucleus in response to steroid. Remarkably, despite the complete nuclear localization of GRSV122–139 in response to steroid, this substitution reduced the transcriptional activation of the MMTV promoter in response to 10–8 m dexamethasone to about 20% of the activity of WT GR at 16 h. This level of transcriptional activity was similar to that observed with GRNL1–, despite the reduced nuclear localization of GRNL1–.
      Notably, the decrease in the level of transcriptional activity observed was not because of a decrease in the recruitment of GR to the MMTV promoter. Chromatin immunoprecipitation analysis of the MMTV promoter integrated into the COS7 cell genome showed similar recruitment of GR and GRSV122–139 to the promoter 60 min following steroid treatment (Fig. 5B). However, consistent with the decrease in transcriptional activation mediated by GRSV122–139, the recruitment of RNA polymerase II to the MMTV promoter was dramatically reduced in the presence of this mutant compared with WT GR. Similarly, only WT GR induced significant acetylation of histone H4 at the promoter.
      This indicates that the substitution in the hinge region did not discernibly affect the potential for GR to interact with its DNA-responsive elements in the nucleus, but it suggests that inhibition of the nuclear export of GR through NRS activity made an important contribution to modulation of the transactivation potential of the receptor.
      As GR forms heterodimers in solution and binds to palindromic DNA-response elements, we next assessed the effect of coexpression of WT GR or GRSV122–139 on the steroid-mediated activation of MMTV transcription (Fig. 5C). In the first instance, increasing the amount of WT GR transfected from 6 to 18 ng provided an expected proportional increase in GR-dependent transcription from 15- to ∼33-fold. Similarly, increasing the amount of GRSV122–139 transfected from 6 to 12 ng increased the fold activation of transcription in response to steroid from 3- to 5-fold (data not shown). Interestingly, cotransfection of 12 ng of GRSV122–139 with 6 ng of WT GR, a ratio that would result in a predominance of GR/GRSV122–139 heterodimers and GRSV122–139/GRSV122–139 homodimers, yielded a 12-fold induction of transcription, which is less than the 20-fold that would have been expected if the effects of each construct were simply additive. A similar result was observed upon expression of a 1:1 ratio of the two constructs (data not shown).
      These results suggest that WT GR and GRSV122–139 largely retain their individual transcriptional activation potentials when activating transcription as heterodimers, with GRSV122–139 showing mild if any dominant negative activity over WT GR.

      DISCUSSION

      GR, which exchanges between an active nuclear form and an inactive cytoplasmic form, has long been appreciated to have the potential to be regulated through control of its movement between nucleus and cytoplasm. Indeed, GR has previously been shown to be actively retained in the nucleus, and addition of an NES that increases the rate of GR nuclear export has been reported to decrease GR transcriptional activation potential (
      • Kino T.
      • Souvatzoglou E.
      • De Martino M.U.
      • Tsopanomihalu M.
      • Wan Y.
      • Chrousos G.P.
      ,
      • Tago K.
      • Tsukahara F.
      • Naruse M.
      • Yoshioka T.
      • Takano K.
      ). However, the signals that promote the active nuclear retention of proteins such as GR remain to be defined. In this study we present compelling evidence that the hinge region of GR surrounding the NL1 NLS includes an active nuclear retention signal or NRS, which promotes the retention of GR in the nucleus and represses nuclear protein export through the CRM1-dependent nuclear export pathway. Remarkably, inhibition of the GR NRS activity through mutagenesis dramatically compromised the transcriptional activation potential of GR at the MMTV promoter.
      Previous results suggest that the GR NL1 overlaps with an NRS activity (
      • Savory J.G.
      • Hsu B.
      • Laquian I.R.
      • Giffin W.
      • Reich T.
      • Hache R.J.
      • Lefebvre Y.A.
      ). Here we have determined that amino acids 500–525 of the GR hinge region that encompass the NLS are both necessary and sufficient to impede CRM1-mediated nuclear export of GFP reporter proteins and to inhibit the nuclear export of steroid-stimulated nuclear GR. Although the peptide appeared to be sufficient for NRS activity when taken out of context, mutations that abrogated NRS activity in synthetic constructs were insufficient to significantly affect nuclear retention of full-length GR. Given that the GR DBD and LBD effectively surround and position the hinge region, it is possible that additional determinants beyond the 25 amino acids delimited as the minimal NRS motif contribute to NRS activity within the context of the full-length receptor. However, more simply, it would also seem possible that the LE/AA substitution has a less dramatic effect on presentation of the NRS when integrated into the native receptor where the surrounding amino acids in the hinge are under greater constraint.
      Nonetheless, our observation that the L507A/E508A substitution within the hinge abrogated NRS activity in synthetic constructs without affecting NL1 activity suggests that NRS and NL1 activities are distinct properties of GR. Furthermore, it is quite clear that at least most basic NLSs do not cohabit with NRSs. What then might the NRS receptor or binding site be? GR is highly mobile in the nucleus, both when liganded and following ligand withdrawal (
      • Walther R.F.
      • Lamprecht C.
      • Ridsdale A.
      • Groulx I.
      • Lee S.
      • Lefebvre Y.A.
      • Hache R.J.
      ,
      • Nagaich A.K.
      • Rayasam G.V.
      • Martinez E.D.
      • Becker M.
      • Qiu Y.
      • Johnson T.A.
      • Elbi C.
      • Fletcher T.M.
      • John S.
      • Hager G.L.
      ,
      • Stavreva D.A.
      • Muller W.G.
      • Hager G.L.
      • Smith C.L.
      • McNally J.G.
      ). This was also true of the GRSV122–139 receptor mutant (data not shown). Therefore, it seems unlikely that the NRS would dictate the static interaction of GR with a nuclear structure such as the nuclear matrix for example, even though such an interaction has been demonstrated for GR previously in permeabilized cells (
      • Tang Y.
      • Getzenberg R.H.
      • Vietmeier B.N.
      • Stallcup M.R.
      • Eggert M.
      • Renkawitz R.
      • DeFranco D.B.
      ). However, the rapid movement of the receptor within the nucleus would not necessarily preclude a network of transient interactions with structures such as the nuclear matrix that would form a virtual cage that could functionally segregate GR from the nuclear export machinery. Alternatively, it is possible that the NRS exerts its function through interaction with a soluble and mobile nuclear factor that interferes with the nuclear export machinery or somehow inhibits the assembly of CRM1-RanGDP-cargo nuclear export complex. It would seem unlikely however, even given the close overlap between NL1 and NRS, that NRS activity reflects a modulation of the means through which GR interacts with the importins that mediate its transfer to the nucleus. The availability of an NRS peptide and a mutation that abrogates NRS activity within the peptide provide the means to seek out the NRS receptor through biochemical or genetic screening approaches.
      Substitution of an 18-amino acid peptide centered over the SV40 NLS into GR inhibited NRS activity, but it did not result in an export rate that fully matched that of GRNL1–. It may be premature to consider that this substitution did not fully compromise the NRS activity, given that GRNL1– only ever becomes partially nuclear and thus may not be provide an accurate reference point for comparison. However, it is also possible that additional determinants to the NRS surround the substituted region or that GR contains additional independent NRS activities in the same manner in which it contains two NLSs.
      The strong decrease in the transcriptional activation potential for GRSV122–139 exceeded what might have been predicted, given that the overall localization of GR remained strongly nuclear despite the increase in nuclear export rate. Furthermore, the modest reduction of the total expected level of transcriptional activation upon cotransfection of WT GR and GRSV122–139, suggested that abrogation of NRS activity in one molecule of GR modestly reduced the transcriptional activation potential of its heterodimerization partner.
      Although it is tempting to speculate that continuous rapid export of GR to the cytoplasm, even in the context of receptor that remains predominantly nuclear, results in a biochemical change in the receptor that dramatically impedes its transcriptional regulatory potential, there are many additional potential contributing factors to this effect. For example, if the NRS represents an additional nuclear target for GR, it is possible that this targeting is directly important for receptor activity (
      • Luisi B.F.
      • Xu W.X.
      • Otwinowski Z.
      • Freedman L.P.
      • Yamamoto K.R.
      • Sigler P.B.
      ,
      • Archer T.K.
      • Hager G.L.
      • Omichinski J.G.
      ). Given that GR and GRSV122–139 were recruited to the MMTV promoter to the same extent, it would seem that additional targeting mediated through the NRS would have a substantial effect on the ability of GR to promote chromatin remodeling and nucleate transcription complexes on the promoter, as was reflected by the dramatic decrease in RNA polymerase II recruitment and histone H4 acetylation at the MMTV promoter in the presence of GRSV122–139 compared with WT GR.
      Recently it has been demonstrated that rapid cycling of GR and coregulatory factors and the modulation of HDAC1 both occur at the MMTV promoter during activation (
      • Qiu Y.
      • Zhao Y.
      • Becker M.
      • John S.
      • Parekh B.S.
      • Huang S.
      • Hendarwanto A.
      • Martinez E.D.
      • Chen Y.
      • Lu H.
      • Adkins N.L.
      • Stavreva D.A.
      • Wiench M.
      • Georgel P.T.
      • Schiltz R.L.
      • Hager G.L.
      ). Although additional experiments suggested that both WT and mutant GR were recruited to the promoter as early as 30 min following hormone stimulation (data not shown), it is possible that the differences in receptor transactivation potential reflect differences in the temporal profile of receptor-promoter interactions or in the recruitment of coactivators and corepressors that cycle through the promoter. For example an increased recruitment of enzymatically active HDAC1 by GRSV122–139 might explain the decreased histone H4 acetylation and the reduced RNA polymerase II recruitment.
      It will also be important to determine in future studies whether this change in GR transcriptional activation potential is promoter-specific and whether changes in DNA-independent means of transcriptional regulation by GR, through NF-κB or AP-1 for example, are also impacted by the changes to the NRS.

      Acknowledgments

      We thank S. Lee for plasmids and helpful advice.

      References

        • Beato M.
        • Herrlich P.
        • Schutz G.
        Cell. 1995; 83: 851-857
        • Gustafsson J.A.
        • Carlstedt-Duke J.
        • Wrange O.
        • Okret S.
        • Wikstrom A.C.
        J. Steroid Biochem. 1986; 24: 63-68
        • Giguere V.
        • Hollenberg S.M.
        • Rosenfeld M.G.
        • Evans R.M.
        Cell. 1986; 46: 645-652
        • Charmandari E.
        • Kino T.
        • Chrousos G.P.
        Ann. N. Y. Acad. Sci. 2004; 1024: 1-8
        • Chrousos G.P.
        • Charmandari E.
        • Kino T.
        J. Clin. Endocrinol. Metab. 2004; 89: 563-564
        • Nguyen T.
        • Sherratt P.J.
        • Nioi P.
        • Yang C.S.
        • Pickett C.B.
        J. Biol. Chem. 2005; 280: 32485-32492
        • Ghisletti S.
        • Meda C.
        • Maggi A.
        • Vegeto E.
        Mol. Cell. Biol. 2005; 25: 2957-2968
        • Picard D.
        • Kumar V.
        • Chambon P.
        • Yamamoto K.R.
        Cell Regul. 1990; 1: 291-299
        • Pratt W.B.
        • Toft D.O.
        Endocr. Rev. 1997; 18: 306-360
        • Madan A.P.
        • DeFranco D.B.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 3588-3592
        • Sackey F.N.
        • Hache R.J.
        • Reich T.
        • Kwast-Welfeld J.
        • Lefebvre Y.A.
        Mol. Endocrinol. 1996; 10: 1191-1205
        • Yang J.
        • Liu J.
        • DeFranco D.B.
        J. Cell Biol. 1997; 137: 523-538
        • Picard D.
        • Yamamoto K.R.
        EMBO J. 1987; 6: 3333-3340
        • Jewell C.M.
        • Webster J.C.
        • Burnstein K.L.
        • Sar M.
        • Bodwell J.E.
        • Cidlowski J.A.
        J. Steroid Biochem. Mol. Biol. 1995; 55: 135-146
        • Savory J.G.
        • Hsu B.
        • Laquian I.R.
        • Giffin W.
        • Reich T.
        • Hache R.J.
        • Lefebvre Y.A.
        Mol. Cell. Biol. 1999; 19: 1025-1037
        • Freedman N.D.
        • Yamamoto K.R.
        Mol. Biol. Cell. 2004; 15: 2276-2286
        • Hache R.J.
        • Tse R.
        • Reich T.
        • Savory J.G.
        • Lefebvre Y.A.
        J. Biol. Chem. 1999; 274: 1432-1439
        • Qi M.
        • Hamilton B.J.
        • DeFranco D.
        Mol. Endocrinol. 1989; 3: 1279-1288
        • Walther R.F.
        • Lamprecht C.
        • Ridsdale A.
        • Groulx I.
        • Lee S.
        • Lefebvre Y.A.
        • Hache R.J.
        J. Biol. Chem. 2003; 278: 37858-37864
        • Mosammaparast N.
        • Pemberton L.F.
        Trends Cell Biol. 2004; 14: 547-556
        • Fried H.
        • Kutay U.
        Cell. Mol. Life Sci. 2003; 60: 1659-1688
        • Weis K.
        Cell. 2003; 112: 441-451
        • Fornerod M.
        • Ohno M.
        • Yoshida M.
        • Mattaj I.W.
        Cell. 1997; 90: 1051-1060
        • Stade K.
        • Ford C.S.
        • Guthrie C.
        • Weis K.
        Cell. 1997; 90: 1041-1050
        • Liu J.
        • DeFranco D.B.
        Mol. Endocrinol. 2000; 14: 40-51
        • Kino T.
        • Souvatzoglou E.
        • De Martino M.U.
        • Tsopanomihalu M.
        • Wan Y.
        • Chrousos G.P.
        J. Biol. Chem. 2003; 278: 25651-25656
        • Tago K.
        • Tsukahara F.
        • Naruse M.
        • Yoshioka T.
        • Takano K.
        Mol. Cell. Endocrinol. 2004; 213: 131-138
        • Askjaer P.
        • Jensen T.H.
        • Nilsson J.
        • Englmeier L.
        • Kjems J.
        J. Biol. Chem. 1998; 273: 33414-33422
        • Wiper-Bergeron N.
        • Wu D.
        • Pope L.
        • Schild-Poulter C.
        • Hache R.J.
        EMBO J. 2003; 22: 2135-2145
        • Kumar R.
        • Thompson E.B.
        J. Steroid Biochem. Mol. Biol. 2005; 94: 383-394
        • Luisi B.F.
        • Xu W.X.
        • Otwinowski Z.
        • Freedman L.P.
        • Yamamoto K.R.
        • Sigler P.B.
        Nature. 1991; 352: 497-505
        • Hollenberg S.M.
        • Giguere V.
        • Segui P.
        • Evans R.M.
        Cell. 1987; 49: 39-46
        • Nagaich A.K.
        • Rayasam G.V.
        • Martinez E.D.
        • Becker M.
        • Qiu Y.
        • Johnson T.A.
        • Elbi C.
        • Fletcher T.M.
        • John S.
        • Hager G.L.
        Ann. N. Y. Acad. Sci. 2004; 1024: 213-220
        • Stavreva D.A.
        • Muller W.G.
        • Hager G.L.
        • Smith C.L.
        • McNally J.G.
        Mol. Cell. Biol. 2004; 24: 2682-2697
        • Tang Y.
        • Getzenberg R.H.
        • Vietmeier B.N.
        • Stallcup M.R.
        • Eggert M.
        • Renkawitz R.
        • DeFranco D.B.
        Mol. Endocrinol. 1998; 12: 1420-1431
        • Archer T.K.
        • Hager G.L.
        • Omichinski J.G.
        Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 7560-7564
        • Qiu Y.
        • Zhao Y.
        • Becker M.
        • John S.
        • Parekh B.S.
        • Huang S.
        • Hendarwanto A.
        • Martinez E.D.
        • Chen Y.
        • Lu H.
        • Adkins N.L.
        • Stavreva D.A.
        • Wiench M.
        • Georgel P.T.
        • Schiltz R.L.
        • Hager G.L.
        Mol. Cell. 2006; 22: 669-679