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

J. Biol. Chem., Vol. 277, Issue 2, 1538-1543, January 11, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/2/1538    most recent M105438200v1 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 Pandit, S. Articles by Sitlani, A. Search for Related Content PubMed PubMed Citation Articles by Pandit, S. Articles by Sitlani, A. Social Bookmarking                      What's this?

Allosteric Effects of Dexamethasone and RU486 on Glucocorticoid Receptor-DNA Interactions^*

Shilpa Pandit, Wayne Geissler, Georgianna Harris, and Ayesha Sitlani

From the Department of Metabolic Disorders, Merck and Company, Inc., Rahway, New Jersey 07065

Received for publication, June 12, 2001, and in revised form, October 9, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The glucocorticoid receptor (GR) is a DNA-binding protein that can regulate the transcription of a large number of genes in a ligand-dependent fashion. Although much progress has been made on the mechanism of transcriptional regulation by GR, a potential allosteric effect of GR-binding ligands on specific GR-DNA interactions is controversial. In this study, gel-shift methods are used to measure the effects of a classical agonist dexamethasone and a prototypical antagonist RU486 on the in vitro interactions of GR with DNA substrates, which contain glucocorticoid response elements (GREs) from promoters of GR-regulated genes. These studies show that cell extracts containing human GR bind specifically and with high affinity to GREs in the absence of ligand. An agonist dexamethasone and antagonist RU486 do not affect the affinity of GR for DNA but subtly alter the electrophoretic mobility of the GR-DNA complex. Importantly, the dissociation rate of GR from DNA increases as a function of the concentration of GRE-containing DNA. At a fixed DNA concentration, dexamethasone-bound GR dissociates from DNA significantly faster than does ligand-free GR or RU486-bound GR. These results are consistent with a model for transcriptional activation in which a dynamic complex is formed between agonist-bound GR and DNA.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The glucocorticoid receptor (GR)¹ is a member of the steroid hormone receptor family of transcription factors that regulates a large number of genes in a hormone- or ligand-dependent manner. GR is predominantly cytosolic and requires association with heat shock proteins for ligand binding, a feature that has complicated in vitro structural studies of GR (1-6). The molecular mechanism by which hormones and small molecule ligands regulate GR-mediated gene expression is not clearly understood. A favored model is that hormone binding induces dissociation of heat shock proteins from the ligand-bound GR complex, which then translocates to the nucleus for sequence-specific DNA binding to GREs (7-10). However, the underlying mechanism for the expression of agonist versus antagonist activity for the steroid-bound GR-DNA complex is not clear.

Attempts to study the allosteric effects of hormones on GR-DNA interactions in vitro have led to contradictory observations. In general, studies indicate that hormones are not required for stable GR binding to DNA in vitro (11-16). Specifically, complexes of agonist triamcinolone acetonide and antagonist RU486 bound to GR have a comparable affinity for glucocorticoid response elements from different promoters (14-16). In contrast to these studies done on asymmetric GR binding sites, gel-shift experiments reported with GR and an artificial palindromic GRE suggest that ligands directly affect GR-DNA binding (17). The agonist dexamethasone enhances GR-DNA binding severalfold, whereas antagonists including RU486 have no effect on GR-DNA binding. However, this observation is contested by a report that also implements gel-shift methods with a palindromic GRE substrate but uses in vitro transcribed GR (13). This work shows that dexamethasone and RU486 predominantly affect the mobility of GR-DNA complex in a gel and only modestly affect the affinity of GR for DNA (13). Therefore, although ligands clearly regulate GR-mediated transcriptional activity, a direct effect of these ligands on specific GR-DNA interactions is controversial.

In vivo, several lines of evidence suggest that GR-binding ligands have an indirect effect on chromatin accessibility and structure by the recruitment of specific co-activators and DNA remodeling factors to transcriptional complexes (18, 19). Several co-activator proteins including CBP, SRC1, and GRIP bind GR in the presence of hormone. These proteins have histone acetyl transferase activity that weakens histone-chromatin interactions and makes chromatin more accessible and amenable to transcription (20-23). In comparison, remodeling proteins (including the Swi/Snf factors that interact with GR) can directly alter the structure and flexibility of chromatin to promote transcription (24-26). Furthermore, chemical footprinting studies suggest that hormone-bound GR transiently binds to its response elements and then leaves its site accessible to other factors including NF1 (27) and HNF5 (28). A model that describes a dynamic transcriptional complex is also supported by recent photobleaching experiments carried out on GR-bound mouse mammary tumor virus (MMTV) chromatin in living cells (29). These studies show that the GR-DNA complex in the presence of cortisol has a short half-life, on the order of seconds, with cortisol-bound GR dissociating and associating rapidly with chromatin.

Footprinting and photobleaching experiments contradict the traditional view that receptors remain bound to DNA sites in the presence of ligands. All of these studies allude to a "hit and run" mechanism of transactivation whereby the hormone-bound receptor is in fast exchange with its chromatin regulatory elements (29-31). The studies that allude to this mechanism all are carried out with agonist-bound GR; however, the results of a comparison of the effects of antagonist versus agonist or no ligand on the kinetics of GR-DNA interactions are not known. In vivo, GR-DNA interactions in the absence of ligand are not readily measured, because cellular GR is predominantly cytosolic in the absence of hormone. However in vitro dissociation kinetic studies, using gel-electrophoretic methods on the DNA binding domain of GR with its response elements, show that dissociation is dependent on DNA concentration (32). These studies predict that at concentrations of DNA present in the nucleus, GR is in fast dissociation with DNA; however, the question of how ligands affect GR-DNA dissociation kinetics remains unanswered.

This work explores the effects of functional ligands on GR-DNA interactions with respect to binding affinity and dissociation kinetics in vitro by gel-shift methods. We investigate in detail the effects of dexamethasone, a potent and classical GR agonist, and RU486, a prototypical GR antagonist, on the affinity and dissociation kinetics of the GR-DNA complex. We observe that these ligands do not alter the highly cooperative binding affinity of GR for asymmetric and palindromic DNA response elements. Moreover, the dissociation of full-length receptor from DNA is directly proportional to the square of the concentration of GRE-containing DNA. Although an agonist like dexamethasone enhances GR-DNA dissociation by 10-fold compared with steroid-free receptor, an antagonist like RU486 has a modest depreciation on GR-DNA kinetics. These observations are discussed with respect to the implications on transcriptional activation and ligand-mediated receptor-DNA search mechanisms.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials and Plasmid Preparation-- All tissue culture flasks were obtained from Falcon, and all chemicals were obtained from Sigma unless otherwise specified. Radiolabeled [³H]dexamethasone (35 Ci/mmol) and [-³³P]ATP were obtained from PerkinElmer Life Sciences. COS-7 cells were obtained from ATCC (CRL-1651). Culture media and sera were products of Life Technologies, Inc. unless otherwise stated. The fetal bovine serum was from Gemini Bio-Products (Calabasas, CA).

The following oligonucleotides containing a single GRE, purified by PAGE, were obtained from Invitrogen: GRE1-TOP, 5'-GCAGGTCTTTATGGTTACAAACTGTTCTTAAAACAAGG-3'; GRE1-BOT, 5'-ATCCTTGTTTTAAGAACAGTTTGTAACCATAAAGACCT-3'; GRE2-TOP, 5'-GCAGGTCCTCTGCTGTACAGGATGTTCTAGCTACAAGG-3'; GRE2-BOT, 5'-ATCCTTGTAGCTAGAACATCCTGTACAGCAGAGGACCT-3'; GRE3-TOP, 5'-GCAGGTCCTCTGCAGAACACAGTGTTCTAGCTACAAGG-3'; GRE3-BOT, 5'-ATCCTTGTAGCTAGAACACTGTGTTCTGCAGAGGACCT-3'; GRE4-TOP, 5'-AGCTTAGAACACAGTGTTCTCTAGAG-3'; GRE4-BOT, 5'-GATCCTCTAGAGAACACTGTGTTCTA-3'.

The top and bottom strands of each pair of dimeric GRE-containing oligonucleotides were labeled with a TOP and BOT suffix. The consensus nonpalindromic dimeric GRE1 and GRE2 substrates were from the MMTV and the tyrosine aminotransferase (TAT) promoters, respectively. The center of oligonucleotides GRE3-TOP and GRE3-BOT contained a palindromic GRE sequence reported to bind GR with high affinity (12). Oligonucleotides GRE4-TOP and GRE4-BOT contained a palindromic GRE designed at the end of the sequence identical to that used in a reported GR-DNA binding gel-shift study (17).

All transfection reagents were purchased from Invitrogen unless specified. A plasmid containing the coding sequence of the human GR gene (flanked by BamHI and XhoI restriction sites) was obtained from ATCC (pRShGR, 67200). To generate efficient expression plasmids, the 5'-untranslated region of the human GR cDNA in pRShGR was removed. The entire GR coding fragment was reassembled in a baculovirus transfer vector, pFastBac (Invitrogen), by ligating the BamHI/EcoRI fragment of the PCR product and the EcoRI/XhoI fragment of pRShGR into BamHI/XhoI-digested pFastBac. The complete GR coding fragment was also subcloned into the mammalian cell transfection vector pcDNA3.1(+) to generate the hGRpcDNA3.1 vector, which was used for transient transfection into COS-7 cells.

Transient Transfection and Lysate Preparation-- COS-7 cells were cultured in Dulbecco's modified Eagle's medium, 10% fetal bovine serum, and 1% penicillin/streptomycin in a 37 °C, 5% CO₂ incubator and split every 3-4 days. Cells were plated in 75-cm² tissue culture flasks or in poly(D-lysine)-coated 6-well plates until they were ~80% confluent for transfection. Cells were transiently transfected with the hGRpcDNA3.1 vector using LipofectAMINE-2000 according to the manufacturer's protocol. After 24 h of transfection, a ligand (dexamethasone or RU486) was added to each well at a final concentration of 100 nM and further incubated for 24 h followed by the addition of enzyme-free dissociation solution, Hanks' based, (Specialty Media, Lavellette, NJ) to detach the cells. The cells were transferred to a tube, centrifuged at 2000 rpm, and washed with cold phosphate-buffered saline. The cell pellets were lysed in the presence of low salt buffer A (20 mM Tris-HCl (pH 7.5), 10 mM NaMoO₄, 5 mM dithiothreitol, 1 mM EDTA, and a 1:100 dilution of protease inhibitor mixture (Sigma)) by three cycles of freezing and thawing. Cell lysates prepared with a high salt buffer B (Buffer A and 500 mM NaCl) or low salt buffer A yielded similar concentrations of active GR. However, high salt extracts contained other contaminating nuclear components as well, which bind nonspecifically to GRE-containing DNA (as shown in gel-shift experiments); therefore, we chose to use low salt buffer A to prepare GR extracts for these studies. To prevent dissociation of ligand from the receptor, ligand (at a final concentration of 100 nM) was included in the lysis buffer for those cells that were pretreated with ligand. The lysate was spun down at 14,000 rpm for 30 min at 4 °C, and the supernatant collected as the whole cell extract.

Electrophoretic Mobility Shift Assay (EMSA)-- For EMSA experiments, 0.1 nmol of oligonucleotides GRE1-TOP, GRE2-TOP, GRE3-TOP, and GRE4-TOP were end-labeled with 2 µl of [-³³P]ATP using T4-polynucleotide kinase (Invitrogen) in a 20-µl reaction. The labeled oligonucleotides were purified to remove unincorporated labels using a Qiaquick nucleotide removal kit (Qiagen, Valencia, CA). To form duplexes, 0.07 nmol of each labeled oligonucleotide (top strand) was mixed with 0.07 nmol of the complementary oligonucleotide (bottom strand) in 100 µl of TE buffer (10 mM Tris-HCl and 1 mM EDTA). The hybridization mixture was heated in a water bath to boiling temperature and was then slow cooled to room temperature. The efficiency of duplex formation was checked by PAGE on a 15% gel.

For binding reactions, whole cell extract (containing ~1 nM active GR, as measured by [³H]dexamethasone binding, Ref. 33) was incubated with 5 nM labeled GRE duplex in a buffer containing 40 mM Tris-Cl (pH 7.5), 0.2 mM EDTA, 8 mM dithiothreitol, 80 mM NaCl, 0.1% bovine serum albumin, 20% glycerol, 0.25% Nonidet P-40, and 12-25 µg of poly(dI-dC). For the comparison of DNA binding by GR, GR-Dex, and GR-RU486 extracts, the amount of extract and protein used in each lane was normalized by measuring the total protein concentration in each extract by a Coomassie Plus protein assay reagent kit (Pierce, Rockford, IL). Each binding reaction mixture (containing increasing concentrations of GR extract) was incubated at 25 °C for 1 h, 45 min until equilibrium was attained. A small aliquot (2-4 µl) of the reaction mixture was loaded onto a 5% gel (75:1 acrylamide/bis-acrylamide) and run at 100 V in 50 mM Tris borate, 1 mM EDTA buffer. The samples were added while the gel was running to avoid precipitation of the protein-DNA complexes in the gel wells.

For the dissociation reactions, excess unlabeled duplex GRE (prepared by hybridizing unlabeled top and bottom strands according to the protocol described above) was added to the binding reactions that had reached equilibrium. Following the addition of an excess duplex competitor, reaction aliquots (3.0 µl) were loaded at different time points (1, 2.5, 5, 7.5, 15, 20, 30, 45, and 60 min) onto a gel run at 100 V. The gels were run for 1 h and dried on a Slab gel-dryer (Forma Scientific).

Quantitation and Analyses-- Dried gels were quantitated by a STORM PhosphorImager (Molecular Dynamics, Sunnyvale, CA), and data were plotted and analyzed using PRISM software (GraphPad, CA). The percentage of bound GR-DNA was plotted as a function of log[GR], and the data fit to a sigmoidal dose-response curve to extract the concentration at half-maximal binding (EC₅₀). Dissociation data were plotted as a function of time (t, s), and the data fit to the first order kinetic equation,

(Eq. 1)

in which the specific [GR-DNA complex] counts were normalized for the total counts in each lane to account for variations in gel loading, and A and B were constants. The fit was used to extract the dissociation constant (k_d) and half-life (t_1/2 = 0.69/k_d) at each condition. Each dissociation curve was fit to data obtained from at least three independent experiments, and the error was represented as standard deviations. Kinetic experiments were carried out at varying duplex substrate concentrations (625-3750 nM), and the k_d values were obtained for each concentration of duplex used.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

GR-DNA Binding-- Gel mobility shift methods were used to measure and compare the thermodynamic parameters of the GR-DNA, Dex-GR-DNA, and RU486-GR-DNA complexes prepared in vitro. Gel-shift experiments are conducted with unpurified cell extracts of GR prepared by transfecting COS-7 cells with a GR expression vector, as described under "Experimental Procedures." Cell extracts of GR prepared in the absence of ligand bind with high affinity and specificity to a duplex (GRE1) containing a dimeric GRE site from the MMTV promoter (Fig. 1). Importantly, the specific GR-DNA band shift (Fig. 1) can be competed with unlabeled duplex GRE1 substrate but not with sequence-scrambled duplex DNA, confirming a specific interaction with the GRE site. In addition, extracts prepared from cells transfected with a vector not containing the GR coding sequence do not show a specific GR-DNA band (data not shown), indicating that the specific gel-shifts identified in Figs. 1 and 3 are GR-dependent. Interestingly, extracts of ligand-free GR bind with similar affinity to nonpalindromic GREs from the TAT and MMTV (GRE1 and GRE2) promoters as well as with a palindromic sequence contained in the center of the duplex GRE3 (Fig. 1A). However, no stable GR-DNA band shift is observed with the GRE4 substrate that contains a palindromic GRE sequence at the end of the duplex (Fig. 1A). This observation points to the importance of nonspecific flanking sequences in stabilizing GR-DNA binding.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   The effect of GRE sequence and ligands dexamethasone and RU486 on GR-DNA binding. A, whole cell extracts of GR bind to GREs in vitro, as measured by standard gel-shift methods. The substrates GRE1, GRE2, GRE3, and GRE4 are defined under "Experimental Procedures." Extracts containing GR bind with similar affinity to asymmetric GRE binding sites GRE1 and GRE2 and to the palindromic GRE contained in GRE3, but the extracts do not bind specifically to substrate GRE4, which contains a palindromic GRE at the end of the duplex. B, the effect of ligands dexamethasone and RU486 on GR binding to MMTV DNA. Extracts of COS-7 cells transfected with GR treated with or without ligand bind with similar affinity to the MMTV duplex. C, the binding of extracts containing GR, GR-Dex, and GR-RU486 to MMTV DNA is highly cooperative with respect to extract concentration. The presence of Dex or RU486 ligand does not significantly alter the binding affinity or cooperativeness of GR to MMTV DNA.

Cell extracts of GR prepared with agonist dexamethasone or antagonist RU486 bind GRE1 (containing the MMTV site) with similar affinity as ligand-free GR (Fig. 1, B and C). Interestingly, the binding of all three extracts to the dimeric MMTV-DNA is highly cooperative as a function of GR concentration (Fig. 1C). Similarly, studies carried out with the DNA binding domain of GR (dbd-GR) indicate cooperativeness in DNA binding, suggesting that the first monomer molecule of dbd-GR to DNA may significantly enhance binding of the second monomer. Although others have reported a specific monomer dbd-GR-DNA band in gel-shift studies, we detect no complex containing monomer full-length GR bound to DNA by gel-shift analyses (32, 34-36). Moreover, experiments carried out with cell extracts of GR and a duplex containing a single MMTV half-site indicate no stable and specific GR-DNA band shift (data not shown). The cooperative binding data for dimeric GR bound to DNA (Fig. 1C) yields half-maximal concentrations or EC₅₀ values, which are estimations of the dissociation constants (K_d) of all three extracts ranging from 1.8 × 10⁹ M (for both GR-Dex and GR-RU486) to 2.0 × 10⁹ M (for GR). As demonstrated by these studies, ligands dexamethasone or RU486 do not significantly alter the affinity of GR for MMTV DNA (Fig. 1), a result also observed with duplexes GRE2 and GRE3, which contain GREs from other gene promoters, suggesting a more universal effect. However, Dex-GR extracts induce subtle effects on the mobility of the GR-DNA band containing the MMTV (GRE1) substrate (Fig. 1B). The Dex-GR-DNA gel-shift band appears as a sharper doublet suggesting that this complex has a conformation that is distinct from the GR-DNA and RU486-GR-DNA complexes (Figs. 1B and 3B).

GR-DNA Kinetics-- Given the lack of effect of ligand on the affinity of GR for GRE-containing DNA substrates, attention is shifted to investigate a potential ligand-dependent effect on the rate of association or dissociation of GR from DNA. The cooperative binding data indicate that the association rate for dimer GR binding to DNA is likely dependent on the binding of monomer GR to DNA; however, we detect no stable monomer GR-DNA band in our gel-shift studies (Fig. 1). Instead, in our attempts to measure association kinetics we detect transient intermediates that migrate with gel mobilities similar to the stable GR-DNA band and that associate and dissociate too fast to measure by gel-shift methods (data not shown). Moreover, association of the stable dimer Dex-GR-DNA complex is significantly faster than the GR-DNA and RU486-GR-DNA complexes. Because direct measurement and comparison of the association kinetics for the GR-DNA, Dex-GR-DNA, and RU486-GR-DNA are not feasible, we focus our attention on measuring the dissociation kinetics of these complexes.

Studies were conducted to characterize the dissociation of dimeric GR from DNA containing the MMTV GRE (duplex GRE1). Gel-shift experiments indicate that the half-life (dissociation kinetics) of the GR-DNA complex is not a simple first order reaction but is dependent on the concentration of GRE-containing duplex DNA. An increase in the concentration of unlabeled GRE1 from 625 to 3800 nM promotes a nonlinear increase in the dissociation kinetics of GR (Fig. 2). However, an increase in the concentration of nonspecific DNA (containing no GREs) in the reaction does not have a measurable effect on the dissociation kinetics of the GR-DNA complex (data not shown). Dissociation data are fit to a simple exponential decay curve to extract the half-life (t_1/2) and dissociation constants (k_d) of the GR-DNA complex at each GRE1 DNA concentration. The dissociation constant (k_d) increases linearly with the square of duplex DNA concentration (Fig. 2B), indicating that GR-GRE dissociation is second order in excess GRE-containing DNA. The limiting dissociation rate (k₀ = 0.98 × 10⁴ s¹) is obtained by extrapolating the plot of k_d versus [DNA]² to [DNA]² = 0 (Fig. 2B). Further validation of second order dissociation kinetics is demonstrated in a linear regression plot of ln(k_d  k₀) versus ln[DNA], which has a slope of 2.3. A similar second order dependence on DNA concentration is observed with sf9 cell extracts of GR, obtained from baculoviral infection of recombinant human GR (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Relationship between GR-DNA dissociation rates (kd) and the concentration of specific GRE-containing DNA. A, dissociation of GR is dependent on the concentration of excess unlabeled MMTV DNA. Dissociation data at each concentration of DNA (0.63-3.8 µM) are an average of three independent experiments, with the error represented as standard deviation. The data are fit to a single exponential decay curve to obtain the dissociation constants (kd) and half-lives (t1/2) of the receptor-DNA complex. B, the dissociation rates increase linearly with the square of DNA concentration. The linear plot is extrapolated to obtain the intrinsic kd (1.0 × 104 s1) at [DNA] = 0 µM.

Agonist dexamethasone and antagonist RU486 have opposite effects on GR-DNA dissociation kinetics (Fig. 3 and Table I). At a GRE1 duplex concentration of 1.25 µM, the Dex-GR-DNA complex exhibits faster dissociation kinetics (t_1/2 = 220 s; k_d = 2.9 × 10³ s¹) whereas the RU486-GR-DNA complex shows modestly slower dissociation kinetics (t_1/2 = 5600 s; k_d = 1.2 × 10⁴ s¹) than the unliganded GR-DNA complex (t_1/2 = 2300 s; k_d = 3.0 × 10⁴ s¹). Importantly, dissociation kinetics of the Dex-GR-DNA varies as a nonlinear function of duplex DNA concentration. The dependence of k_d for the Dex-GR-DNA complex on duplex GRE1 concentration is analogous to the GR-DNA complex. A 2-fold increase in the DNA concentration promotes a 4-fold increase in k_d for both the Dex-GR-DNA and the GR-DNA complex (Table I). For the comparison of DNA dissociation, extracts of GR, GR-Dex, and GR-RU486 all show ~20-50% DNA binding at the initial time point prior to addition of excess GRE. In this range, the dissociation rates are independent of the fraction of DNA bound.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Effect of ligands dexamethasone and RU486 on GR-DNA dissociation. A and B, gel-shifts showing dissociation of GR-RU486 (A) and GR-Dex (B) from MMTV DNA at 1250 nM excess unlabeled MMTV DNA. The lane marked 20* represents a control lane with DNA minus GR extract loaded at the 20-min time point. C, analyses of the gel dissociation data in A and B by single exponential decay curves demonstrate that dissociation of GR-Dex (t1/2 = 220 s; kd = 2.9 × 103 s1) is significantly faster than GR-RU486 (t1/2 = 5600 s; kd = 1.2 × 104 s1) from DNA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The mechanism by which ligands modulate GR-dependent transcriptional regulation is complex; several laboratories are currently investigating this problem from different perspectives and using distinct methodologies. One piece of this puzzle is the question of whether or not ligands modulate the interactions between GR and specific DNA recognition sites. Methods including gel-shifts, gradient sedimentation, and footprinting have been used to measure the allosteric effects of ligands on the interactions between GR and DNA in vitro; however, these studies have led to contradictory results (12-17). Importantly, each observation is distinct with respect to the choice of ligand, DNA promoter site, and method of detection. To resolve these discrepancies, we have utilized in vitro gel-shift methods to systematically measure the allosteric effects of an agonist dexamethasone and an antagonist RU486 on the affinity and dissociation kinetics of GR for natural and artificial GRE sequences.

Naturally occurring GREs identified in gene promoters are characterized by the nonpalindromic consensus sequence 5'-XXTACAXXXTGTTCT-3' containing two hexamer binding sites for the GR dimer. However, GR has been shown to bind to and promote transactivation from specific mutants of the GRE consensus sequence to varying degrees (12, 37). Our studies indicate that extracts of GR-transfected cells treated with or without dexamethasone or RU486 bind cooperatively to consensus GREs from the MMTV and TAT promoters with similar affinities (Fig. 1). Similarly, others have reported on the cooperative binding of the DNA binding domain of GR to DNA, suggesting that binding of the first monomer to DNA greatly enhances binding of the second monomer (32, 34-36). Furthermore, extracts of ligand-free GR bind with similar affinity to the artificial palindromic GRE 5'-AGAACAXXXTGTTCT-3', compared with naturally occurring GREs from the MMTV or TAT promoters. Again, ligands induce no significant effect on the affinity or cooperativeness of GR-DNA binding. However, GR binding to a duplex containing a palindromic GRE is ablated if the site is positioned at the ends of a duplex. The importance of nonspecific interactions between GR and base pairs that flank the GRE site is consistent with footprinting experiments of a GR-DNA complex described by DNaseI protection of base pairs on either side of the consensus GR binding site (38).

Ligands have no effect on GR-DNA binding affinity and cooperativeness; however, they have distinct and measurable allosteric effects on GR-DNA dissociation. Most significantly, the Dex-GR complex dissociates rapidly from DNA with a rate that is 10-fold greater than ligand-free GR. Moreover, the RU486-GR complex dissociates at a rate 2-fold slower than ligand-free GR and 20-fold slower than the Dex-GR complex. Because the binding affinity of GR for DNA in the presence and absence of ligand is equivalent, it is likely that the association rate of Dex-GR binding to DNA is also greatly enhanced compared with ligand-free GR or RU486-GR. However, given the fast association kinetics and cooperativeness of DNA binding by all three extracts, a direct measurement of association rates for dimer GR binding to DNA is not feasible. This finding is consistent with reports that describe the limitations of gel-shift methods in measuring the association kinetics of dbd-GR binding to DNA (32).

The observation that the dissociation of the Dex-GR-DNA complex is significantly faster than the GR-DNA or RU486-DNA complex is consistent with an emerging model in the literature that points to the importance of a dynamic GR-DNA complex in transcriptional activation (27-31). An agonist such as dexamethasone may enhance the rate of dissociation of GR from DNA, thereby promoting the entry of other DNA remodeling factors that prepare the site for transcription by RNA polymerase. By way of comparison an antagonist like RU486, which slows down the dissociation of GR from DNA, may prevent the access of DNA remodeling factors to the site of transcription. A similar but less striking finding was reported for dissociation kinetics of the steroid progesterone receptor (PR) from a DNA substrate containing the identical MMTV GRE site. PR bound to an agonist R5020 dissociates at a modest rate, 2-fold faster than PR bound to an antagonist RU486 (39) from MMTV DNA. Therefore, although PR and GR bind to the identical functional GRE from the MMTV promoter, the active role that DNA plays in ligand-dependent allosteric effects appears to be more significant for GR than PR. These results are consistent with reports that highlight the importance of the allosteric role that DNA plays in transcriptional activation by GR (40-42).

This work directly demonstrates the differential effects of an agonist and an antagonist on the dissociation of GR from simple (non-chromatin) GRE-containing DNA substrates in vitro. Importantly, our experiments are carried out with extracts of cells transfected with GR and treated with agonist or antagonist. Therefore, it is likely that other components present in these cell extracts induce the effects on GR-DNA dissociation in a ligand-dependent manner. The involvement of other cofactors in affecting dissociation of GR from chromatin has been described recently by other laboratories (43, 44). First, a combination of HeLa nuclear extracts and ATP was shown to promote restriction enzyme access to GR-bound chromatin DNA sites in vitro, implicating a role for cofactors in the enhancement of GR-DNA dissociation (43). Second, it was recently demonstrated that the molecular chaperone protein p23 forms a ternary complex with hormone-bound GR and chromatin DNA and enhances disassembly of the transcriptional complex in vivo in an ATP-dependent manner.² Furthermore, p23 was reported to enhance agonist-dependent GR transcription and to promote the dissociation of the thyroid receptor from its response elements in vitro (44, 45). These observations suggest that p23 and other DNA remodeling factors present in nuclear extracts may contribute to the in vitro agonist-dependent rapid GR-DNA dissociation observed in this work.

In addition to the effects of ligands our studies point to the importance of the concentration of unbound GREs in facilitating the dissociation of GR from DNA. Previous reports have shown that the dissociation of the DNA binding domain of GR (dbd-GR) from DNA is dependent on the concentration of competitor DNA (32); however, our studies indicate that this effect is exaggerated in the dissociation of full-length GR from GREs. A 2-fold increase in DNA concentration is reported to promote a modest 1.4-fold increase in the dissociation rate of dbd-GR from DNA (32), whereas a 2-fold increase in DNA concentration promotes a 4-fold increase in the dissociation rate of full-length GR from DNA. The observation that full-length GR-DNA dissociation is second order in specific DNA indicates that dissociation is catalyzed by two additional molecules of MMTV DNA. A similar observation was reported for the Escherichia coli transcriptional factor protein CAP (catabolite activator protein), in which dissociation of CAP from promoter DNA is second order with respect to nonspecific DNA concentration. In general, our studies emphasize that dissociation of GR from DNA is clearly not a simple first order reaction but is more complex and is facilitated by DNA and potentially other proteins and cofactors present in cell extracts that are associated with the GR-DNA complex.

The effect of DNA concentration on GR-DNA dissociation kinetics may have important physiological implications with respect to how GR searches for specific GRE sites. A central question in gene regulation by transcriptional factors is how do intracellular DNA-binding proteins rapidly search for specific binding sites from billions of DNA base pairs? A "direct transfer" model has been used to explain faster than diffusion-controlled kinetics of site searching by some DNA-binding proteins including the prokaryotic transcriptional factor CAP (46). This two-dimensional search strategy is distinct from proposed one-dimensional search mechanisms, including protein sliding along a DNA chain or repeated cycles of dissociation/association of proteins from DNA (32, 46).

The direct transfer model proposes that a protein forms a bridge between two or more DNA segments; dissociation from one DNA promotes transfer to another DNA molecule. In contrast to a simple first order dissociation/reassociation model, this model predicts that increasing concentrations of DNA will increase the apparent dissociation rate of the protein as observed with CAP, dbd-GR, and GR (32, 46). An important distinction in the results of these studies is that CAP-DNA dissociation can be enhanced by nonspecific DNA (46) and dbd-GR-DNA dissociation can be enhanced more significantly by specific than nonspecific DNA (32), whereas this work demonstrates that GR-DNA dissociation can only be enhanced by specific DNA. This dependence on specific DNA in GR-DNA dissociation suggests that the DNA binding region of one monomer dissociates from the first DNA molecule (while still dimerized to the first monomer) and binds specifically to a second DNA molecule to form a bridged intermediate.

The simultaneous involvement of two additional molecules of DNA in GR-DNA dissociation, as indicated by the second order dependence of k_d on [DNA] (also observed with the E. coli transcriptional factor protein CAP), would suggest that the mechanism of dissociation is a complex version of the direct transfer model. The involvement of two molecules of DNA suggests that monomer subunits of GR may dissociate individually from each half-site of a GRE or through some bridged intermediate direct transfer complex requiring two equivalents of GRE-containing DNA. The report that dissociation of the DNA binding domain of GR from DNA is approximately first order in DNA concentration (32) suggests that the second order dependence on DNA for dissociation of the full-length receptor involves regions outside of the DNA binding domain of the receptor that may make specific contacts with MMTV DNA to accelerate dissociation.

Interestingly, our estimation of the intrinsic DNA dissociation rate k₀ (independent of unbound DNA) for full-length GR is 1.0 × 10⁴ s¹ and compares well with the rates reported for dbd-GR (2.2 × 10⁴ s¹) and CAP (1.2 × 10⁴ s¹). DNA dissociation by dbd-GR and CAP has been explained by a direct transfer model for fast kinetics of DNA site searching, and our results with full-length GR further support this model. At concentrations of DNA estimated (32) in a eukaryotic cell (~10-50 mg/ml), one would predict that GR is in fast dissociation and association with DNA. This prediction is consistent with fast kinetics of cortisol-GR binding to MMTV chromatin in a cell, as observed by photobleaching experiments (29). Moreover, this work shows that although dexamethasone does not alter the concentration dependence of GR dissociation for DNA, it does enhance the intrinsic k_d of GR. We propose that an agonist like dexamethasone may not only promote the access of DNA remodeling factors to DNA but may also indirectly enhance the rate at which GR searches for specific sites in a cell containing several thousands of potential binding sites.

	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: Merck & Co., Inc., Dept. of Metabolic Disorders, P.O. Box 2000, Mail Code: R80Y-300, Rahway, NJ 07065. Tel.: 732-594-0788; Fax: 732-594-5468; E-mail: ayesha_sitlani@merck.com.

Published, JBC Papers in Press, October 26, 2001, DOI 10.1074/jbc.M105438200

² Yamamoto, K. R., EMBO Workshop: Nuclear Receptor Structure and Function, Erice, Italy, 2001.

	ABBREVIATIONS

The abbreviations used are: GR, glucocorticoid receptor; GRE, glucocorticoid response element; MMTV, mouse mammary tumor virus; TAT, tyrosine amino transferase; EMSA, electrophoretic mobility shift assay; dbd-GR, DNA binding domain of GR; Dex, dexamethasone; PR, progesterone receptor; CAP, catabolite activator protein.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				S. Salemi, S. Yousefi, D. Lochmatter, A. Eble, J. Deladoey, I. C. A. F. Robinson, H.-U. Simon, and P. E. Mullis Isolated Autosomal Dominant Growth Hormone Deficiency: Stimulating Mutant GH-1 Gene Expression Drives GH-1 Splice-Site Selection, Cell Proliferation, and Apoptosis Endocrinology, January 1, 2007; 148(1): 45 - 53. [Abstract] [Full Text] [PDF]

				T. C. Voss, S. John, and G. L. Hager Single-Cell Analysis of Glucocorticoid Receptor Action Reveals that Stochastic Post-Chromatin Association Mechanisms Regulate Ligand-Specific Transcription Mol. Endocrinol., November 1, 2006; 20(11): 2641 - 2655. [Abstract] [Full Text] [PDF]

				N. Yoshikawa, K. Yamamoto, N. Shimizu, S. Yamada, C. Morimoto, and H. Tanaka The Distinct Agonistic Properties of the Phenylpyrazolosteroid Cortivazol Reveal Interdomain Communication within the Glucocorticoid Receptor Mol. Endocrinol., May 1, 2005; 19(5): 1110 - 1124. [Abstract] [Full Text] [PDF]

				H. Garside, A. Stevens, S. Farrow, C. Normand, B. Houle, A. Berry, B. Maschera, and D. Ray Glucocorticoid Ligands Specify Different Interactions with NF-{kappa}B by Allosteric Effects on the Glucocorticoid Receptor DNA Binding Domain J. Biol. Chem., November 26, 2004; 279(48): 50050 - 50059. [Abstract] [Full Text] [PDF]

				B. M. Necela and J. A. Cidlowski A Single Amino Acid Change in the First Zinc Finger of the DNA Binding Domain of the Glucocorticoid Receptor Regulates Differential Promoter Selectivity J. Biol. Chem., September 17, 2004; 279(38): 39279 - 39288. [Abstract] [Full Text] [PDF]

				D Jung, A C Fantin, U Scheurer, M Fried, and G A Kullak-Ublick Human ileal bile acid transporter gene ASBT (SLC10A2) is transactivated by the glucocorticoid receptor Gut, January 1, 2004; 53(1): 78 - 84. [Abstract] [Full Text] [PDF]

				K. De Bosscher, W. Vanden Berghe, and G. Haegeman The Interplay between the Glucocorticoid Receptor and Nuclear Factor-{kappa}B or Activator Protein-1: Molecular Mechanisms for Gene Repression Endocr. Rev., August 1, 2003; 24(4): 488 - 522. [Abstract] [Full Text] [PDF]

				B. Kauppi, C. Jakob, M. Farnegardh, J. Yang, H. Ahola, M. Alarcon, K. Calles, O. Engstrom, J. Harlan, S. Muchmore, et al. The Three-dimensional Structures of Antagonistic and Agonistic Forms of the Glucocorticoid Receptor Ligand-binding Domain: RU-486 INDUCES A TRANSCONFORMATION THAT LEADS TO ACTIVE ANTAGONISM J. Biol. Chem., June 13, 2003; 278(25): 22748 - 22754. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/2/1538    most recent M105438200v1 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 Pandit, S. Articles by Sitlani, A. Search for Related Content PubMed PubMed Citation Articles by Pandit, S. Articles by Sitlani, A. Social Bookmarking