HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 24, 22053-22062, June 14, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/24/22053    most recent M202311200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Chen, J. Articles by Simons, S. S. Search for Related Content PubMed PubMed Citation Articles by Chen, J. Articles by Simons, S. S., Jr. Social Bookmarking                      What's this?

Structure/Activity Elements of the Multifunctional Protein, GMEB-1

CHARACTERIZATION OF DOMAINS RELEVANT FOR THE MODULATION OF GLUCOCORTICOID RECEPTOR TRANSACTIVATION PROPERTIES^*

Jun Chen, Sunil Kaul, and S. Stoney Simons Jr.

From the Steroid Hormones Section, NIDDK/Laboratory of Molecular and Cellular Biology, National Institutes of Health, Bethesda, Maryland 20892

Received for publication, March 8, 2002, and in revised form, March 25, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

GMEB-1 was initially described as a component of a 550-kDa heteromeric DNA binding complex that is involved in the modulation of two properties of glucocorticoid receptor (GR) transactivation, the dose-response curve of agonists and the partial agonist activity of antagonists. Subsequently, GMEB-1 was also found to bind to hsp27, to associate with the coactivator TIF2 in yeast cells, and to participate in Parvovirus replication. To understand these multiple activities of GMEB-1 at a molecular level, we have now determined which regions are associated with the various activities associated with the modulation of GR transactivation properties. These activities include, homooligomerization, heterooligomerization, DNA binding, binding to GR and the transcriptional cofactor CBP, and GR modulation. Complex activities such as DNA binding and GR modulation, are found to require the physical combination of those domains that would be predicted from the involved biochemical processes. We have previously documented that GMEB-1 possesses both GR modulatory and intrinsic transactivation activity. However, the domains for these two activities of GMEB-1 are found not to overlap. This separation of activities provides a structural basis for our prior biological observations that the modulation of the dose-response curve and partial agonist activity of GR complexes is independent of the total levels of gene activation by the same GR complexes.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A major function of steroid receptors is to translate the intracellular concentration of cognate ligands into graded amounts of specific biological responses. These responses most commonly involve either the induction or repression of gene expression due to changes in transcription. The particular response of a given cell depends upon on a variety of tissue and developmental controls that dictate which genes are capable of regulation by receptor-steroid complexes (1-3). Among those genes that are programmed to respond, many of the details of steroid hormone action have only recently emerged and include binding to specific DNA sequences called hormone response elements (HREs),¹ reorganization of chromatin, recruitment of a burgeoning assortment of transcription cofactors, and interaction with components of the transcription complex (4-7).

Two major classes of ligands bind to steroid receptors, agonists and antagonists. Agonists unleash the full activity of the receptor to cause either induction or repression of expression of the responsive gene. In both cases, the dose-response curve defines the relationship between ligand concentration and biological response. Unless exogenous steroids are administered, the concentration of circulating steroid in the bloodstream is rarely high enough to saturate the binding capacity of the intracellular steroid receptors. Therefore, the expression levels of most genes are sub-maximal and reflect the changing percentages of the total amount of a given steroid receptor in target cells that is bound by the fluctuating levels of endogenous steroids. A theoretically useful concentration is the EC₅₀, which is the concentration of steroid required for half the maximal response and also corresponds in theory to the equilibrium dissociation constant for steroid binding to the receptor. The levels of endogenous steroid are usually similar to the EC₅₀. The value of the EC₅₀ is determined experimentally by subtracting the basal level expression of a given gene from the activities induced by a variety of steroid treatments and then calculating what percentage of full induction is produced by each steroid concentration. Thus, the EC₅₀ is independent of both the basal and fully induced levels. This independence has been documented upon numerous occasions (8-13).

Antagonists, or antisteroids, block the action of agonist steroids but often display some residual agonist activity. The amount of partial agonist activity of an antisteroid is a major consideration for antihormone therapies of such conditions as inflammation (14), conception (15), and hormone-dependent cancers (16, 17). To the extent that the antisteroid blocks only the target gene and retains agonist activity for all other responsive genes, one can minimize the side effects that result from an antisteroid indiscriminately preventing the expression of all responsive genes. However, it is not as much the absolute level of partial agonist activity that is important as the percentage of the maximal response elicited by an agonist steroid. In this respect, as for the EC₅₀ of the dose-response, the partial agonist activity of an antisteroid is independent of the basal and fully induced levels of the responsive gene (8-13).

The determinants of the dose-response curve and the EC₅₀ are still not known despite playing a central role in steroid hormone action. The EC₅₀was initially thought to be determined by the affinity of steroid binding to cognate receptor (18-20). However, reports of different dose-response curves for various genes within a cell for the same receptor-steroid complex indicate the involvement of additional parameters (21-27). Furthermore, the EC₅₀ can even change for the same gene under different cellular conditions (28, 29). Although initially perplexing as to how such changes in EC₅₀ might occur, this variability offers tremendous benefits to a cell or organism. Because the levels of endogenous steroid are similar to the value of the EC₅₀, changes in the EC₅₀ afford a simple mechanism for the differential control of gene expression during development, differentiation, homeostasis, and the endocrine treatment of disease states. Similarly, the partial agonist activity of antisteroids is known to depend upon numerous factors such as promoter, cell type, tissue, and growth conditions (26, 28, 30-34). Thus, a better understanding of these phenomena might allow one to harness these variations for clinical purposes.

We have recently succeeded in identifying several factors that can modulate both the EC₅₀ and dose-response curve of receptor-agonist complexes and the partial agonist activity of receptor-antagonist complexes. The first is a cis-acting element of the rat tyrosine aminotransferase gene, which we call a glucocorticoid modulatory element, or GME (8, 29, 35, 36). Changing concentrations of the homologous receptor, coactivators such as TIF2/GRIP1 (37, 38), and the corepressors NCoR (39) and SMRT (40) modulate the above induction properties of both glucocorticoid receptors (GRs) (9-11, 13) and progesterone receptors (13, 41). Interestingly, the precise effects of some of these modulators are quite different for GR and progesterone receptor, even within the same cell (13). More recently, additional modulators acting via different pathways have been discovered (42).²

The GME was initially isolated from the rat tyrosine aminotransferase gene (29) and binds two novel proteins called GMEB-1 and -2 (43). We have cloned GMEB-1 (44), which is the larger of these two proteins. GMEB-1 is located on human chromosome 1 (45) and is a member of a new family of transcription factors called KDWK proteins (44, 46, 47), or SAND domain proteins (48). GMEB-1 exists as a large, heterooligomeric complex (molecular mass of about 550 kDa) with GMEB-2 in intact cells (43). GMEB-1 also homooligomerizes, possesses intrinsic transactivation activity, and modifies the induction properties of GR-agonist and -antagonist complexes in a reversible manner (44). GMEB-1 binds specifically to the GME sequence and interacts with GR and with CBP (44). Thus, GMEB-1 is a multi-functional protein.

Interestingly, GMEB-1 has been identified in other contexts that are completely unrelated to the GME and modulation of GR transactivation activities. GMEB-1 has been cloned on the basis of its binding to hsp27 (49), which is an antiapoptotic protein that acts in part by delaying the release of cytoplasmic cytochrome c (50) and that increases the tumorigenic potential of rat colon carcinoma cells (51). GMEB-1 was also isolated by its binding to the second activation domain (AD2) of the coactivator TIF2 (37) in a yeast two-hybrid assay, although this binding may be mediated by an adapter molecule in yeast (52). Finally, GMEB-1 is involved in Parvovirus replication and has been called PIF (Parvovirus initiation factor (53, 54) and activates the viral nickase NS1 (55). GMEB-1 is present in highest abundance in fetal and reproductive tissues (45), which suggests that it could be a developmentally important protein for many actions. GMEB-1 is found in puffer fish (GenBank^TM accession number AL300721) and mammals but not prokaryotes and, thus, appears to be an evolutionarily recent protein.

In view of the many different processes involving GMEB-1, each of which may require specialized activities, we sought to determine which domains of GMEB-1 are relevant for the modulation of GR transactivation properties. Thus, we wanted to identify the sequences required for homo- and heterooligomerization, DNA binding, intrinsic transactivation, GR modulation, interacting with CBP, and binding to GR. We find that several of these activities require multiple domains, consistent with the presence of the GMEB-1 as a heterooligomer in intact cells (43). This represents a first step in understanding at a molecular level how GMEB-1 functions as a component of GME action.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Unless otherwise indicated, all operations were performed at 0 °C.

Chemicals-- The following chemicals were purchased from the indicated sources: [³⁵S]Met, Amersham Biosciences; prestained molecular weight markers, LipofectAMINE plus, herring sperm DNA, and oligonucleotides, Invitrogen; acrylamide and bisacrylamide, National Diagnostics (Atlanta, GA); TNT-coupled reticulocyte lysate system, Promega (Madison, WI); cross-linking reagent (ethylene glycol bissuccinimidylsuccinate), Pierce; restriction enzymes and DNA polymerase, New England Biolabs (Beverly, MA), Invitrogen, and Promega.

Plasmids-- For HisGEMB1, cDNA was inserted into pcDNA3.1HisA. For HisGEMB1(152C), BspMII-(fill-in) and NotI of HisGMEB1 were inserted into EcoRV and NotI of pcDNA3.1HisC. For HisGEMB1(177C), BclI and NotI of HisGMEB1 were inserted into BamHI and NotI of pcDNA3.1HisA. For HisGEMB1(325C), XmnI and NotI of HisGMEB1 were inserted into EcoRV and NotI of pcDNA3.1HisA. For HisGEMB1(412C), SspI and NotI of HisGMEB1 were inserted into EcoRV and NotI of pcDNA3.1HisB. For HisGMEB1(Y45S), ATTTATG was changed to AtaTcTG using a site-specific mutagenesis kit (GeneEditor from Promega). For HisGEMB1(46C), EcoRV and XbaI of HisGMEB1(Y45S) were inserted into EcoRV and XbaI of pcDNA3.1HisA. For HisGEMB1(E91I), GGAGAGC was changed to GatatcC using a site-specific mutagenesis kit (GeneEditor from Promega). For HisGEMB1(92C), EcoRV and XbaI of HisGMEB1(E91I) were inserted into EcoRV and XbaI of pcDNA3.1HisA. For HisGEMB1(N115), BclI-(fill-in) and HindIII of HisGMEB1 were inserted into EcoRV and HindIII of pcDNA3.1HisA. For HisGEMB1(N153), BspMII-(fill-in) and NotI of HisGMEB1 were inserted into EcoRV and NotI of pcDNA3.1HisA. For HisGEMB1(N171), BspMI-(fill-in) and HindIII of HisGMEB1 were inserted into EcoRV and HindIII of pcDNA3.1HisA. For HisGEMB1(N229), MscI and BstXI of HisGMEB1 were inserted into NotI-(fill-in) and BstXI of HisGMEB1. For HisGEMB1(N306), SapI and PflFI-f and PvuI of HisGEMB1 (N412) were inserted into EcoRV and PvuI of pcDNA3.1HisA. For HisGEMB1(N324), XmnI and KpnI of HisGMEB1 were inserted into EcoRV and KpnI of pcDNA3.1HisA. For HisGEMB1(N412), SspI and HindIII of HisGMEB1 were inserted into EcoRV and HindIII of pcDNA3.1HisA. For HisGEMB1(116-152), BclI and BspMII of HisGMEB1 were deleted, filled-in, and religated.

For pmB1, EcoRI and XbaI of HisGMEB1 were inserted into EcoRI and XbaI of pm. For pmB1(154C), EcoRI and XbaI of HisGMEB1(154C) were inserted into EcoRI and XbaI of pm. For pmB1(177C), BclI and XbaI of HisGMEB1 were inserted into BamHI and XbaI of pm. For pmB1 (232-306), MscI and SapI of pmB1(N306) were inserted into XmaI and SapI of pm (and CCACGG was deleted). For pmB1(307C), PflFI and XmaI of pmB1(177C) was deleted, filled-in, and religated. For pmB1(412C), SspI and XbaI of HisGMEB1(412C) were inserted into BamHI and XbaI of pm. For pmB1(N115), EcoRV and XbaI of HisGMEB1(N115) were inserted into EcoRI and XbaI of pm. For pmB1(N153), EcoRV and XbaI of HisGMEB1(N153) were inserted into EcoRI and XbaI of pm. For pmB1(N171), EcoRV and XbaI of HisGMEB1(N171) were inserted into EcoRI and XbaI of pm. For pmB1(N229), MscI and NotI of pmB1(N412) were deleted and religated. For pmB1(N230), MscI and XbaI-f of pmB1(N412) were deleted, filled-in, and religated. For pmB1(N306), PflFI and NotI of pmB1(N412) were deleted and religated. For pmB1(N412), EcoRV and XbaI of HisGMEB1(N412) were inserted into EcoRI and XbaI of pm. For pmB1(N519), AvaII (filled-in) and EcoRI of HisGMEB1 were inserted into EcoRI and SmaI of pm. For pmB1(116-152), EcoRI and XbaI of HisGMEB1(116-152) were inserted into EcoRI and XbaI of pm.

For pVP16 B1, EcoRI and XbaI of pmB1 were inserted into EcoRI and XbaI of VP16. For pVP16-B1(N171), EcoRI and XbaI of pmB1(N171) were inserted into EcoRI and XbaI of VP16. For pVP16/B1(N230), EcoRI and SapI of pmB1(N230) were inserted into EcoRI and SapI of pVP16. For pVP16-B1(N306), EcoRI and SapI of pmB1(N306) were inserted into EcoRI and SapI of pVP16. For pVP16-B1(N412), EcoRI and XbaI of pmB1(N412) were inserted into EcoRI and XbaI of pVP16. For pVP16/B1(46C), BamHI and XbaI of HisGMEB1(46C) were inserted into BamHI and XbaI of pVP16. For pVP16-B1(177C), EcoRI and XbaI of pmB1(177C) were inserted into EcoRI and XbaI of VP16. For pVP16-B1(307C), EcoRI and XbaI of pmB1(307C) were inserted into EcoRI and XbaI of pVP16. For pVP16/B1(116-152), EcoRI and XbaI of HisGMEB1(116-152) were inserted into EcoRI and XbaI of pVP16.

For GEX-4T-B1, BamHI and NotI of HisGMEB1 were inserted into BamHI and NotI of GEX-4T-2. For GEX-4T-B1(177C), EcoRI and NotI of pmB1(177C) were inserted into EcoRI and NotI of GEX-4T-B1. For GEX-4T-B1(307C), EcoRI and NotI of pmB1(307C) were inserted into EcoRI and NotI of GEX-4T-B1. For GEX-4T-B1(46C), EcoRI and NotI of HisGMEB1(46C) were inserted into EcoRI and NotI of GEX-4T-2. For GEX-4T-B1(N171), BamHI and NotI of HisGMEB1(N171) were inserted into BamHI and NotI of GEX-4T-2. For GEX-4T-B1(N230), EcoRI and HincII of pmB1(N230) were inserted into EcoRI and SspI of GEX-4T-B1(N171). For GEX-4T-B1(N306), EcoRI and Bsp120I of HisGMEB1(N306) were inserted into EcoRI and NotI of GEX-4T-B1(N171). For GEX-4T-B1(N412), BamHI and NotI of HisGMEB1(N412) were inserted into BamHI and NotI of GEX-4T-2.

In Vitro Expression of Proteins-- All cDNAs of the protein to be expressed were cloned so that they were under the control of either the T7 or SP6 promoter. For each reaction, 1 µg of plasmid DNA was mixed with 2 µl of 1 mM methionine and 40 µl of TNT T7 (or SP6) master mix (Promega) and brought up to a total volume of 50 µl with H₂O. The reaction was conducted at 30 °C for 2 h. For radiolabeling the protein, 2 µl of 1 mM methionine was replaced with 2 µl of [³⁵S]methionine (Amersham Biosciences). When in vitro translating GR, 1 µl of 50 µM Dex was added in the reaction mix to bind and thereby stabilize the newly translated GR.

Bacterial Expression of Proteins-- GST-GMEB1 constructs were transformed into BL21 bacteria (Amersham Biosciences) according to the manufacturer's procedure. A single colony was selected, inoculated into 20 ml of LB broth with 100 mM ampicillin, and cultivated overnight at 37 °C in a shaking incubator. A portion (15 ml) of the overnight culture was inoculated into 150 ml of LB broth (with 100 µg/ml ampicillin), shaken at 37 °C for 3-4 h, and then induced with 2-3 mM isopropyl-1-thio--D-galactopyranoside for 3-4 h. Cells were washed once with PBS, resuspended in 10 ml of PBS, sonicated for 30 cycles (1 cycle = 10 s on and 10 s off). Cell lysates were obtained by spinning the sonicated cells at 20,800 × g for 15 min and collecting the supernatant.

Pull-down Assay-- Glutathione-Sepharose 4B beads (30 µl; Amersham Biosciences) was added to each tube and washed twice with 500 µl of PBS (spin at 6000 × g for 2 min). Bacterial lysates (500-1500 µl) containing GST with and without fused heterologous protein were mixed gently with the beads by rotation (12-15 rpm) for 1-2 h at 4 °C. The pellets were isolated by centrifugation (2 min at 6000 × g) with 2 washes of 500 µl of PBS containing 10 mM mercaptoethanol. Dex-bound ³⁵S-labeled GR (7-10 µl) was added and incubated at 4 °C overnight with rotation. The beads were washed twice with 500 µl of PBS (+10 mM mercaptoethanol) and then 5 times with PBS (+100 mM NaCl). Proteins were removed from the beads by heating at 90 °C for 10-15 min in 40 µl of 2× sample buffer (Quality Biological, Inc.). Aliquots (10-15 µl) of the supernatant were loaded on to 10% SDS-PAGE mini-gels (200 V for 45 min). One gel was dried for 25 min and then used to expose film or phosphorimaging screen. The other gel was analyzed by Western blotting with anti-GST antibody to detect the GST fusion proteins.

Western Blotting-- SDS-PAGE gels were equilibrated in transfer buffer for 15 min at room temperature before electrophoretic transfer of proteins to nitrocellulose membranes in a Bio-Rad small (150-200-mA overnight) or large (350-mA overnight) Transblot apparatus. The nitrocellulose was stained in Ponceau S (0.02% Ponceau S and 0.04% glacial acetic acid in water) to localize the molecular weight markers, incubated with 5% Carnation nonfat dry milk in TBS for 30-45 min, and washed three times with TBS containing 0.2% Tween (0.1 TTBS) for 5 min. Primary antibody was diluted in 0.2 TTBS (1:10,000 for anti-VP16, 1:10,000 for anti-Gal, 1:10,000 for Xpress, and 1:5,000 for anti-GST) and added to the nitrocellulose for 1-2 h at room temperature. Biotinylated anti-rabbit secondary antibody and ABC reagents (each was diluted 1:5,000, except anti-goat, which was 1:50,000; Vector Laboratories, Burlingame, CA) were each added for sequential 30-min incubations at room temperature. After the incubation periods with primary antibody, secondary antibody, and ABC reagents, the nitrocellulose was washed 4 times for 5 min each with 0.1 TTBS. The signals were detected by enhanced chemiluminescence using the recommended protocol of the supplier (Amersham Biosciences).

Cross-linking Assay-- In vitro translated protein (5 µl) in 40 µl of water or PBS at 0 °C was treated with 5 µl of Me₂SO containing 0, 10, or 100 mM cross-linking reagent (ethylene glycol bissuccinimidylsuccinate; Pierce). After brief vortexing (2-5 s), the mixture was incubated for 15-30 min before adding 10 µl of 1 M Tris-HCl to stop the reaction. Samples (10 µl) were then assayed on 10% SDS-PAGE gels as described above.

Transient Transfection Assays-- COS-7 and CV-1 cells were grown on 60-mm dishes in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum. Cells were seeded 1 day before the transfection at a density of 2 × 10⁵ for CV-1 cells and 2 × 10⁶ for COS-7 cells. Cells were transfected using 5 µl of LipofectAMINE Plus reagent and 8 µl of LipofectAMINE with 1 µg of reporter plasmid and other plasmids as indicated and adjusted to a total of 3 µg/plate with herring sperm DNA. After incubating the cells at 37 °C for 5 h, the transfection mixture was replaced with normal medium. The cells were incubated at 37 °C overnight before being induced with the appropriate steroid for 24 h. The cells were lysed and assayed for reporter gene activity using the luciferase assay reagent according to the manufacturer's instructions (Promega). Luciferase activity was measured in an EG&G Berthhold luminometer (Microlumat LB 96 P).

Gel Shift Assay-- The oligonucleotides 5'-CTTCTGTATGAGCGCCAGTAT-3' and 3'-GAAGACATACTCGCGGTCATA-5' were annealed and ³²P-end-labeled by Lofstrand Laboratories (Gaithersburg, MD). Gel shift experiments were performed as described (47) with minor modifications. Briefly, the in vitro transcription/translation product (1 µl) was incubated with 20,000 cpm of the ³²P -end-labeled GME (0.5 fmol) in a total volume of 20 µl for 15 min with sheared, non-denatured herring sperm DNA (0.3 µg) as a nonspecific competitor. After electrophoresis in a 5% non-denaturing polyacrylamide gel at 150 V in 0.4× Tris/Borate/EDTA electrophoresis buffer, the dried gels were autoradiographed for 12-24 h at room temperature with Kodak X-Omat XAR-5 film. Alternatively, the gels were exposed to the phosphorimaging screen for the Molecular Dynamics ImageQuant system for 16-48 h at room temperature The amount of each specific band was calculated as the intensity of that band (calculated by the Molecular Dynamics software) minus the constant background value of the same area from an unrelated region of the gel.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Transactivation Domain of GMEB-1-- We have previously used a mammalian one-hybrid assay in COS-7 cells to establish that GMEB-1 contains an intrinsic transactivation domain (44). This same assay (Fig. 1A) was used to determine which region encodes the transactivation activity of GMEB-1. A variety of chimeras were prepared that contain the GAL4 DNA binding domain (GAL4-DBD = pm) fused to portions of the GMEB-1 (Fig. 1B). The full-length GMEB-1 has good activity compared with the pm control without any GMEB-1 sequence (lanes 7 versus 8 of Fig. 1C). GMEB-1 constructs with C-terminal truncations all display less activity than the full-length GMEB-1 (lanes 4-6 versus 7) even though the whole cell expression levels of each of the chimeras is about the same (see the inset of Fig. 1C). Conversely, the C-terminal fragments show high activity (lanes 1-3 versus 7 of Fig. 1C). Thus, the amino acids 412-563 are sufficient for expression of the intrinsic transactivation activity of GMEB-1.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Transactivation domain of GMEB1. A, cartoon depicting the one-hybrid assay system used to evaluate the intrinsic transactivation activity of GMEB-1 sequences (pm = DNA binding domain of GAL4, or GAL4-DBD). B, structure of the GAL/GMEB-1 chimeras used in the one-hybrid assay. C, biological activities of GAL/GMEB-1 chimeras in the one-hybrid assay. Duplicate samples of COS-7 cells were transiently transfected with 1 µg of the indicated GAL/GMEB-1 chimeras and 1 µg of reporter (GAL4-E1B-LUC), and the induced luciferase values were determined 20 h after transfection as described under "Materials and Methods." The relative luciferase activities were normalized for total lysate protein. The plot gives the average values ± S.D. of three independent experiments. The Western blot (inset, with positions of molecular mass markers (kDa) indicated on the left) with anti-GAL antibody shows the expression level of the GAL/GMEB-1 (pmGMEB-1) chimeras after transient transfection in COS-7 cells.

Heterooligomerization Domain of GMEB-1-- A distinguishing feature of GMEB-1 is that it forms a large heterooligomer with GMEB-2 in intact cells with a size of 550-600 kDa (43). This formation of heterooligomers has been confirmed by mammalian two-hybrid assays (44). This same two-hybrid assay (Fig. 2A) was used to identify the region of GMEB-1 that is required for heteroligomerizing with GMEB-2. Specifically, GAL4-DBD chimeras with GMEB-1 (Fig. 2B) were examined for their ability to induce the luciferase reporter as a result of interacting with another chimera composed of the full-length GMEB-2 fused to the VP16 activation domain. To compensate for the changing amounts of intrinsic activity with each of the pmGMEB-1 constructs, the data are expressed as the fold increase in luciferase activity when the pmGMEB-1 constructs interact with the VP16/GMEB-2 chimera as opposed to just VP16 ( = (activity of VP16/GMEB-2 + pmGMEB-1)/(activity of VP16 + pmGMEB-1)). Using this method of data normalization, it can be seen that neither the N- nor C-terminal fragments of GMEB-1 associate with GMEB-2 (lanes 1 and 4 versus 5 in Fig. 2C). This inactivity is not due to poor levels of expression of these chimeras, as shown by the Western blot in the inset of Fig. 2C. In contrast, the middle sequences of GMEB-1 afford a very strong interaction with GMEB-2 (lanes 2 and 3 versus 5). The much larger signal produced by these chimeras compared with the wild type GMEB-1 is due partially to the fact that the basal level activity of the truncated GMEB-1s of lanes 2 and 3 is so much lower (data not shown) due to the absence of the intrinsic transactivation domain that resides in the C-terminal region of GMEB-1 (see Fig. 1C). Therefore, we conclude that amino acids 230-306 of GMEB-1 are required for its heterooligomerization with GMEB-2.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Heterooligomerization domain of GMEB1. A, cartoon depicting the two-hybrid assay system used to evaluate the interaction between regions of GMEB-1 and full-length GMEB-2 (pm = DNA binding domain of GAL4 or GAL4-DBD; VP16 = activation domain of VP16). B, structure of the GAL/GMEB-1 chimeras used in the two-hybrid assay. C, biological activities of GAL/GMEB-1 chimeras in two-hybrid assay with VP16/GMEB-2. Duplicate samples of COS-7 cells were transiently transfected as described under "Materials and Methods" with 1 µg each of the indicated GAL/GMEB-1 chimeras and reporter (GAL4-E1B-LUC), 1 µg of either VP16 or VP16/GMEB-2 chimera plasmid, and 20 ng of Renilla plasmid as an internal control. The relative luciferase activities were normalized for Renilla expression. The fold increase in luciferase activity caused by the presence of GMEB-2 was then determined by dividing the normalized activity of each GAL/GMEB-1 construct with VP16/GMEB-2 by the normalized activity of the same GAL/GMEB-1 construct with VP16. Similar results were obtained in two other experiments. The inset shows the expression levels of the indicated GAL/GMEB-1 (pmGMEB-1) chimeras after transient transfection in COS-7 cells, as determined by Western blotting with anti-GAL antibody (the positions of the molecular mass markers (kDa) are indicated on the left).

Homooligomerization Domain of GMEB-1-- The existence of GMEB-1 homooligomers was suggested by the ability of purified protein to bind to the GME as a multimer and by mammalian two-hybrid assays involving pmGMEB-1 and VP16/GMEB-1 chimeras (43, 44). To understand this homooligomerization in greater detail and to compare this process to the heterooligomerization of Fig. 2, we examined the ability of wild type and truncated His-tagged GMEB-1 molecules to be chemically cross-linked with ethylene glycol bissuccinimidylsuccinate (linker length = 16.1 Å) after being synthesized by in vitro translation. We know that in vitro translated GMEB-1 is functionally active because it binds to the GME oligonucleotide in gel shift assays in a manner that is indistinguishable from the native protein (43, 44). We elected to use cross-linking as opposed to mammalian two-hybrid assays due to the relatively weak signal that is generated by GMEB-1/GMEB-1 interactions in the latter assay (44) (data not shown). Also, chemical cross-linking permits us to distinguish between dimeric and higher order oligomeric complexes, something that is not possible with the two-hybrid assay using a reporter with five tandem repeats of UAS, which is the GAL4-DBD binding site. As shown in Fig. 3, the sequence of amino acids 177-324 is sufficient for homooligomerization. It should be noted that the size of the major cross-linked species is usually about three times the size of the monomeric species. For example, the most abundant cross-linked species of HisGMEB-1(N324) is 158 ± 7 versus 58 kDa (n = 2) for the monomeric species (Fig. 3B, lanes 7-9). This result suggests that the cross-linked species are larger than a dimer and are probably a trimer.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Homooligomerization domain of GMEB1. A, structure of the His/GMEB-1 chimeras used in the cross-linking assay. B, determination of homooligomerization activity of GMEB-1 chimeras as indicated by their ability to be cross-linked. The indicated His/Xpress-tagged GMEB-1 proteins were prepared by in vitro translation. The proteins were then cross-linked with 10 or 100 mM ethylene glycol bissuccinimidylsuccinate (EGS), separated by electrophoresis on 10% SDS-PAGE gels, and detected by Western blotting with anti-Xpress antibody as described under "Materials and Methods." The positions of molecular mass markers (kDa) are indicated. Similar results were obtained for each GMEB-1 chimera in at least one additional experiment.

DNA Binding Domain of GMEB-1-- We next asked if the homooligomerization domain was sufficient for the binding of GMEB-1 oligomers to DNA or whether additional sequences are needed. To answer this question, we determined the ability of in vitro translated, His-tagged GMEB-1 constructs (Fig. 4A) to bind to ³²P-labeled GME oligonucleotides in our gel shift assays (29, 43, 44). Most truncations eliminated the ability of GMEB-1 to bind to the GME oligonucleotide (lanes 3-6 versus 1 of Fig. 4B). The minimum region for DNA binding corresponds to amino acids 46-324 (lanes 2 and 7). The 39 amino acids at positions 115-153 are particularly important as their deletion destroys the DNA binding activity of GMEB-1 (lane 4 of Fig. 4B). In all cases, the lack of DNA binding activity of a particular construct was not due to abnormally low levels of protein expression (Fig. 4C).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   DNA binding domain of GMEB1. A, structure of the His/GMEB-1 chimeras used in the gel shift assay. B, DNA binding activity of GMEB-1 chimeras as assessed by their ability to bind to GME in a gel shift assay. The indicated His/Xpress-tagged GMEB-1 proteins were prepared by in vitro translation. Aliquots (1 µl) of the programmed lysate were incubated with 32P-end-labeled GME oligonucleotide, and the resulting complexes were detected by autoradiography as described under "Materials and Methods." Similar results were obtained for each GMEB-1 chimera in at least one additional experiment. C, the expression level of the His/GMEB-1 chimeras after in vitro translation was determined by Western blotting with anti-Xpress antibody (positions of molecular mass markers (kDa) are indicated on the left).

GR Binding Domain-- The mechanism by which the GME modulates the dose-response curve of GR-agonist complexes and the partial agonist activity of GR-antagonist complexes is thought to involve GMEB binding both to the GME and to GR. In support of this hypothesis, GMEB-1 is known to bind to GR under cell-free and whole cell conditions (44). To further define this biologically relevant interaction, we used pull-down assays to locate the domain of GMEB-1 that is required to bind GR. This assay affords a stronger signal than that of the two hybrid assay, which has higher backgrounds due to the intrinsic transactivation activity of pmGMEB-1 (44). GST-fused constructs of GMEB-1 (Fig. 5A) along with a GST control were overexpressed in Escherichia coli and then immobilized on anti-GST antibody beads. Constructs lacking N-terminal sequences beyond position 46 of GMEB-1 do not immobilize the full-length GR (Fig. 5B) even though these GMEB-1 constructs are being expressed at approximately equivalent levels (Fig. 5C). Conversely, the first 171 amino acids are sufficient for binding GR (Fig. 5B). It should be noted that the increased mobility of the [³⁵S]methionine-labeled GR bound by GST-B1(N412) in Fig. 5B appears to be an artifact caused by the large amount of GST-B1(N412) that is co-extracted from the matrix (see Fig. 5C) and displaces the GR from its normal migration position in the SDS gel. Also, overexpression of several of the GST-B1 chimeras often results in C-terminal truncated species for unknown reasons (Fig. 5C). However, these fragments do not affect our final assignment of the C-terminal boundary of the GR binding domain of GMEB-1 because the smallest of the C-terminal truncated chimeras (GST-B1(N171)) is not contaminated by many smaller fragments (Fig. 5C). Therefore, we conclude that amino acids 46-171 of GMEB-1 are sufficient for the binding of GR.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   GR interaction domain of GMEB1. A, structure of the GST/GMEB-1 chimeras used in the pull-down assay. B, immobilization of full-length GR by GMEB-1 fragments in a pull-down assay. Bacterially expressed GST/GMEB-1 chimeras that had been immobilized on glutathione-Sepharose beads were incubated with in vitro translated [35S]methionine-labeled, activated, Dex-bound complexes of glucocorticoid receptor for 20 h. After extensive washes, bound GR was eluted with 2× sample buffer, analyzed on 10% SDS-PAGE gels, and detected by autoradiography as described under "Materials and Methods." Similar results were obtained for each GMEB-1 chimera in at least one additional experiment. C, the amount of each GST/GMEB-1 chimera that was immobilized on the anti-GST matrix was directly determined by Western blotting with anti-GST antibody of the material eluted from the matrix with 2× sample buffer after separation on 10% SDS-PAGE gels (positions of molecular mass markers (kDa) are indicated on the left).

CBP Binding Domain-- In addition to GMEB binding to GR, our current model of GME action involves GMEB interactions with other transcription factors and/or components of the transcriptional machinery (12, 56). One attractive target molecule is the comodulator CBP (57), which is known to participate in GR transactivation (58, 59) and to interact with GMEB-1 (44). Current models of steroid hormone action emphasize the role of histone acetylation in causing chromatin reorganization and altered levels of gene expression (5-7). GMEB-1 is devoid of histone acetylation activity but binds to CBP (44), which does possess histone acetylation activity (60). Thus, the binding of GMEB-1 to CBP may be intimately related to the modulatory activity of the GME. We, therefore, prepared a variety of VP16/GMEB-1 chimeras (Fig. 6A) to examine their whole cell interaction with a GAL-DBD/CBP (pmCBP; amino acids 1678-2441 of CBP) chimera in a mammalian two-hybrid assay similar to that of Fig. 2A. Most deletions are detrimental for GMEB-1 interacting with CBP (lanes 3, 4, and 8-10 of Fig. 6B). The inactivity of these truncations was not due to a lack of expressed protein (Fig. 6C). Productive interactions of GMEB-1 with CBP require the extended sequence of amino acids 46-306 (lanes 1 and 5-7 of Fig. 6B).

View larger version (23K): [in this window] [in a new window]   Fig. 6.   CBP interaction domain of GMEB1. A, structure of the VP16/GMEB-1 chimeras used in the two-hybrid assay. B, biological activities of VP16/GMEB-1 chimeras in two-hybrid assay with GAL/CBP. Duplicate samples of COS-7 cells were transiently transfected as described under "Materials and Methods" with 1 µg each of the indicated VP16/GMEB-1 chimeras and reporter (GAL4-E1B-LUC), 1 µg of either GAL or GAL/CBP chimera plasmid, and 20 ng of Renilla plasmid as an internal control. The relative luciferase activities were normalized for Renilla activity. The fold increase in luciferase activity caused by the presence of CBP was then determined by dividing the normalized activity of each VP16/GMEB-1 construct with GAL/CBP by the normalized activity of the same VP16/GMEB-1 construct with GAL. The average values from 2-4 experiments (±S.E.) are plotted. C, the expression levels of the VP16/GMEB-1 chimeras after transient transfection in COS-7 cells was determined by Western blotting with anti-VP16 antibody (positions of molecular mass markers (kDa) are indicated on the left).

Domain Required for Modulation of GR Transactivation Properties-- We have previously documented that the addition of the GME upstream of the GREs of a glucocorticoid-responsive reporter gene both lowers the EC₅₀ of the dose-response curve for GR-agonist complexes and increases the partial agonist activity of GR-antagonist complexes (12, 27, 29, 35, 36). Consistent with a role of GMEB-1 in the action of the GME, GMEB-1 can modify the EC₅₀ and the partial agonist activity of GR complexes. Overexpressed GMEB-1 increases the EC₅₀ of agonists and decreases the partial agonist activity of antagonists, presumably due to the squelching of other limiting factors (44). The sequences that are responsible for the modulatory effects of GMEB-1 were identified by determining the ability of transiently transfected His-tagged GMEB-1s to modify the EC₅₀ and partial agonist activity in an abbreviated assay. In this assay, we quantitated the effects of each construct on the activity of a single subsaturating concentration of Dex, expressed as percent of maximal induction by saturating concentrations of Dex, instead of performing complete dose-response curves. We have previously established that a decrease in the activity of a subsaturating concentration of Dex (such as 4 nM Dex), when expressed as percent of maximal response, corresponds to a right shift in the dose-response curve (8-11, 29). Therefore, this abbreviated method is just as revealing as the full dose-response curve assay. Using this abbreviated assay with the His-tagged GMEB-1 constructs of Fig. 7A, we found that the majority of GMEB-1 is required to cause a decrease in the activity of 4 nM Dex and, thus, a right shift in the dose-response curve to higher values for the EC₅₀ (lanes 2 versus 6 and 7 in Fig. 7B, p values <0.05-0.005). These same sequences are required to reduce the partial agonist activity of the antisteroid Dex-mesylate (Fig. 7C, p values <0.05). Western blots show that the reduced activity of the constructs used in lanes 3-5 and 8 of Fig. 7, B and C, is not due to lower levels of protein expression than that for His/GMEB-1 in lane 2 (Fig. 7D). From these experiments, we conclude that the region of amino acids 46-412 encodes most of the sequences of GMEB-1 that are required to modulate GR transactivation properties.

View larger version (32K): [in this window] [in a new window]   Fig. 7.   GR modulatory domain of GMEB-1. A, structure of the His/GMEB-1 chimeras used in the biological activity assays. B and C, modulatory activity of His/GMEB-1 chimeras in the abbreviated dose-response curve assay (B) and the partial agonist activity assay (C). Triplicate samples of COS-7 cells were transiently transfected as described under "Materials and Methods" with 1 µg of the GMEGREtkLUC reporter, 40 ng of GR plasmid (pSVLGR), 100 ng of His/GMEB-1 chimera (or the molar equivalent of chimera with mutant GMEB-1), and 20 ng of Renilla plasmid as an internal control. The relative luciferase activities of each chimera, after being normalized for Renilla expression, were expressed as the percent of activity seen with the His control plasmid, which lacks any GMEB-1 sequences (lane 1), for 4 nM Dex (B) or 1 µM Dex-mesylate (C). The average values from 4 to 8 experiments (±S.E.) are plotted. The asterisks indicate those values that are significantly different from the control value at the level of p < 0.05 (*), < 0.005 (**), and < 0.0005 (***). C, the expression level of the His/GMEB-1 chimeras after transient transfection in COS-7 cells was determined by Western blotting with anti-Xpress antibody (positions of molecular mass markers (kDa) are indicated on the right).

Domains Responsible for Increased Apparent Molecular Mass-- When the GMEB-1 was first cloned, the difference between the calculated molecular mass and that observed for the in vitro translated material on SDS gels was quite large (27 kDa) (44). This kind of difference is not uncommon. For example, the rat GR migrates on SDS gels as a protein that is about 10 kDa larger than its predicted 89-kDa size (61). However, the 27-kDa difference for GMEB-1 was so much larger a percentage of its predicted size of 61 kDa that we initially questioned the identity of the clone (44). We were, therefore, interested to see if such a large discrepancy between the observed and calculated sizes could be localized to a particular sequence of GMEB-1. In an effort to localize the molecular weight differences to small regions, we determined the molecular weights of two His-tagged proteins that differed by the region of interest as opposed to trying to visualize the small region directly (e.g. M_r of the amino acid sequence of 1-46 is determined by subtracting the M_r of His/46C from the M_r of His/GMEB-1). Any contribution of the His tag would disappear as its effect is subtracted along with the other common regions of the two proteins. Using this method, a significant portion of the abnormal M_r of GMEB-1 is seen to be due to amino acids 1-46 (Fig. 8). No other region of similar size causes increased M_r values. When larger sequences are examined such as 177-307 and 325-563, more aberrant sizes are observed. This suggests that these larger segments of GMEB-1 may form unusual and/or highly asymmetric structures.

View larger version (35K): [in this window] [in a new window]   Fig. 8.   Sequences responsible for abnormally high apparent molecular weight of GMEB-1. The theoretical molecular weight of the indicated GMEB-1 fragments was calculated using the program DNA Strider (version 1.2) with its amino acid composition. The observed molecular weight of the indicated sequences was determined from SDS-PAGE gels either for the actual protein (for 1-563 = GMEB-1) or by subtraction among pairs of His/GMEB-1 constructs (e.g. Mr of the sequence of 1-46 = Mr of His/GMEB-1 minus Mr of His/46C). In all cases, molecular weights were determined by interpolation of a curve of Mr versus SDS-PAGE gel mobility that was generated by CricketGraph III (Computer Associates International, Inc., version, 1.5.1) from standard molecular mass markers ranging from 200 to 20 kDa.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

GMEB-1 was originally identified as one of two proteins in an ~550-kDa heterooligomeric complex that binds to a tyrosine aminotransferase gene sequence called a GME, or glucocorticoid modulatory element (43). This binding of GMEB-1, as part of the larger protein complex, to the GME is closely associated with the ability of the GME to modulate selected transcriptional properties of GR-agonist and -antagonist complexes (29). In an effort to understand the modulatory activity of GMEB-1 at a molecular level, we have determined the amino acid sequences that are required for many of the properties that are thought to be involved: homooligomerization, heterooligomerization, DNA binding, intrinsic transactivation, and binding to GR and CBP in addition to GR modulation itself (Fig. 9). Our current results indicate that almost all of GMEB-1 is associated with some specialized activity.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Summary of GMEB1 domains. The acidic and basic regions of GMEB1 were identified by DNA Strider. Gln (Q)- and Ser/Thr (S/T)-rich boxes are defined as sequences of 10 amino acids that are 30% glutamine or serine/threonine respectively. The boundaries for the domains (depicted by rectangles) of the various indicated activities are those that have been determined by the above experiments. The domain for homology with GMEB-2 is taken from Kaul et al. (44). The darkened region in several domains shows the position of the deletion mutant (amino acids 115-162) that eliminates the DNA binding activity of GMEB-1. Two of the domains for molecular weight have dashed borders to indicate that their contributions are less concentrated than that for the N-terminal sequence of amino acids 1-46. The potential -helical secondary structure of GMEB-1 sequences was calculated using three programs (GCG, PHD (www.embl-heidelberg.de/predictprotein/predictprotein.html), and NNPREDICT (www.cmpharm.ucsf.edu/~nomi/nnpredict.html)). The displayed result was predicted by at least two of the three programs.

Although future studies will undoubtedly refine the boundaries of some of these domains, the large size of many domains is not unexpected because several activities require the cumulative activity of more selective properties. The mere presence of GMEB-1 in a 550-kDa heterooligomer suggests that both homooligomerization and heterooligomerization domains are required. The precise stoichiometry of this complex is not yet known. However, the observation that GMEB-1 readily affords cross-linked species that are larger than dimers and are probably trimers (Fig. 3) is consistent with GMEB-1 and -2 forming structures that are larger than heterodimers. The fact that the homooligomerization and heterooligomerization domains overlap and correspond to a region of high -helical content (Fig. 9) suggests that coiled-coil interactions are largely responsible for oligomer formation. It should be noted that amino acids 88-181 of GMEB-1, which are 80% identical to a region in its heterooligomerizing partner, GMEB-2, are outside of both oligomerization domains of GMEB-1 (Fig. 9). Thus, neither homo- nor heterooligomerization are likely to involve self-complimentary interactions. GMEB-1 binding to DNA appears to occur after the formation of oligomeric complexes (43, 44). As expected, the DNA binding domain includes the regions required for homo- and heterooligomerization (Fig. 9). This overlap of oligomerization and DNA binding domains is similar to that seen with several other DNA binding proteins such as the steroid receptors (62) and the Rel/NFkB family (63). In addition, the DNA binding of GMEB-1 requires the KDWK sequence that was predicted to be involved in the DNA binding of proteins containing this conserved sequence (44, 46, 48). The observation that deletion of the 39-amino acid sequence of 115-153, which includes the KDWK sequence, does eliminate the DNA binding capacity of GMEB-1 (Fig. 4) is consistent with the observation of Bottomley et al. (48) that mutations of the KDWK motif eliminate the DNA binding of related proteins (48). The DNA binding of these KDWK, or SAND domain, proteins has been found by NMR spectroscopy to involve a novel /-fold (48). Finally, the domain specifying the GR modulatory activity, i.e. the ability of GMEB-1 to modulate the transcriptional properties of GR-mediated transactivation, might be predicted to involve the above domains plus the ability to interact with GR and/or necessary transcription factors such as CBP. Thus, it is not surprising that the GR modulatory domain includes all of these sub-domains and may encompass as much as 65% of the protein sequence of GMEB-1 (Fig. 9). This also suggests that some of the surfaces of GMEB-1 that are involved in interactions with other proteins are formed only after the formation of higher order structures.

The anti-GMEB-1 antibody is unable to supershift or immunoprecipitate DNA-bound complexes containing GMEB-1 (47). This antibody was raised against amino acids 125-147. In view of the intimate involvement of this sequence in the DNA binding of GMEB-1 (Fig. 4) (48), it seems likely that the antigenic surface is occluded in the DNA-bound complex.

Acidic domains are often associated with transcriptional activity (64, 65). The transactivation domain of GMEB-1 does contain a clustering of acidic amino acids (Fig. 9). However, the significance of this is not yet clear because other acidic domains exist that are transcriptionally inactive such as amino acids 1-177 (Fig. 1).

Because of the different spatial localizations of the intrinsic transactivation and the GR modulatory domains in GMEB-1 (Fig. 9), it is unlikely that the activities encoded by these two domains participate in the same biochemical processes. We have consistently noted that the total level of transactivation by GR is independent of changes in both the position of the dose-response curve (or EC₅₀) of GR-agonist complexes and partial agonist activity of GR-antagonist complexes (8-13, 42). These combined observations indicate that the effects of GMEB-1 on the total level of GR transactivation product and the EC₅₀/partial agonist activity of GR complexes are caused by GMEB-1 contacting different molecular partners, which are components of distinctly different molecular pathways.

The observed molecular mass of GMEB-1 on SDS-PAGE gels (88 kDa) is 44% greater than the calculated size of 61 kDa and was of major concern when it was first cloned (44). The reason for this very large increase in apparent molecular mass is not known and does not appear to be associated with any obvious chemical, biological, or structural feature. However, the molecular causes for this phenomenon are not restricted to just one sequence but are distributed throughout GMEB-1 (Fig. 9).

GMEB-1 has also been identified as being involved in hsp27 binding (49), Parvovirus replication (53-55), and interaction with the C-terminal activation domain of TIF2 (52). Because the various domains encoding the various activities of GMEB-1 cover most of GMEB-1, it is probable that many of the activities of GMEB-1 with hsp27, Parvovirus replication, and TIF2 colocalize with those in Fig. 9. The oligomerization and DNA binding domains would be expected to be common. What other domains are shared remains to be determined.

	ACKNOWLEDGEMENTS

We thank members of the Steroid Hormones Section for helpful discussions and Paul Yen (NIDDK, National Institutes of Health) for critical review of the manuscript.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Bldg. 8, Rm. B2A-07, NIDDK/LMCB, NIH, Bethesda, MD 20892. Tel.: 301-496-6796; Fax: 301-402-3572; E-mail: steroids@helix.nih.gov.

Published, JBC Papers in Press, April 4, 2002, DOI 10.1074/jbc.M202311200

² S. Chen and S. S. Simons Jr., submitted for publication.

	ABBREVIATIONS

The abbreviations used are: HRE, hormone response element; GME, glucocorticoid modulatory element; GR, glucocorticoid receptor; CBP, cAMP-response element-binding protein (CREB)-binding protein; DBD, DNA binding domain; Dex, dexamethasome; GST, glutathione S-transferase; PBS, phosphate-buffered saline; TBS, Tris-buffered saline; TTBS, TBS containing 0.2% Tween; GAL, galactosidase.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES