A single amino acid change in the first zinc finger of the DNA binding domain of the glucocorticoid receptor regulates differential promoter selectivity.

Mammalian species are well known to differ in their sensitivity to glucocorticoids, but the molecular basis for this difference remains largely uncharacterized. To address this issue, the transcriptional activity of the mouse and human glucocorticoid receptor (GR) was analyzed on two model glucocorticoid-responsive promoters. Mouse GR (mGR) displayed unique promoter discrimination in response to a range of glucocorticoids, with enhanced activity on a simple glucocorticoid response element (GRE)-based promoter and diminished activity on the complex mouse mammary tumor virus promoter compared with human GR (hGR). Promoter discrimination between mGR and hGR was mapped to a single amino acid change at residue 437 (glycine to valine) of mGR and to sequence differences within individual GREs of the different promoters. Mouse GR displayed higher activation on GREs with a guanine rather than a thymine at the -6 position. Binding studies indicated mGR (mGR437V) displayed a weaker affinity for GREs containing a thymine at the -6 position than a mGR mutant containing a glycine at residue 437 (mGR437G). Despite distinct transcriptional activities, both receptors had similar affinities for response elements that contain a guanine at the -6 position. Our findings support a model by which the presence of a valine residue at position 437 of mGR induces a conformational change that leads to alterations in affinity and/or transcriptional activation in a promoter-dependent context.

Mammalian species are well known to differ in their sensitivity to glucocorticoids, but the molecular basis for this difference remains largely uncharacterized. To address this issue, the transcriptional activity of the mouse and human glucocorticoid receptor (GR) was analyzed on two model glucocorticoid-responsive promoters. Mouse GR (mGR) displayed unique promoter discrimination in response to a range of glucocorticoids, with enhanced activity on a simple glucocorticoid response element (GRE)-based promoter and diminished activity on the complex mouse mammary tumor virus promoter compared with human GR (hGR). Promoter discrimination between mGR and hGR was mapped to a single amino acid change at residue 437 (glycine to valine) of mGR and to sequence differences within individual GREs of the different promoters. Mouse GR displayed higher activation on GREs with a guanine rather than a thymine at the ؊6 position. Binding studies indicated mGR (mGR437V) displayed a weaker affinity for GREs containing a thymine at the ؊6 position than a mGR mutant containing a glycine at residue 437 (mGR437G). Despite distinct transcriptional activities, both receptors had similar affinities for response elements that contain a guanine at the ؊6 position. Our findings support a model by which the presence of a valine residue at position 437 of mGR induces a conformational change that leads to alterations in affinity and/or transcriptional activation in a promoter-dependent context.
The biological actions of glucocorticoid hormones are mediated through the intracellular glucocorticoid receptor (GR), 1 which belongs to the nuclear receptor subfamily that includes receptors for mineralocorticoids, estrogen and thyroid hormones, retinoic acid, and vitamin D (1). In the absence of hormone, GR resides in the cytoplasm as a multipro-tein complex with the chaperone proteins Hsp90 and Hsp70, the immunophilin p59, and the phosphoprotein p23 (2). Upon binding hormone, GR dissociates from this complex and undergoes a conformational change that unmasks nuclear localization signals found within the receptor (3). The GR translocates to the nucleus and binds glucocorticoid response elements (GREs) as a dimer in the promoter regions of target genes (4 -6). The association of the glucocorticoid receptor dimer with the response element results in an allosteric induced conformational change within the receptor and subsequent recruitment of coactivator complexes critical for chromatin remodeling and transcription (4 -7). The functional result of the GR-GRE interaction can be largely cell type-, promoter-, and ligand-specific. In some instances, the GR binds negative GREs in promoters of genes and inhibits gene transcription (8). Alternatively, GR represses gene transcription by physically interacting with other transcription factors such as AP-1 and nuclear factor-B (8 -11).
As with other members of the nuclear receptor family, the GR contains a modular structure consisting of three major domains: an N-terminal domain, a central DNA binding domain (DBD), and a C-terminal ligand binding domain (LBD) (12,13). The N-terminal domain represents the most variable region among the species of the GR and the nuclear receptor family. The N-terminal domain contains the AF-1 transcriptional activation domain required for transcriptional enhancement and association with basal transcription factors (12)(13)(14). The central DBD is composed of two highly conserved zinc finger regions and is most conserved region among nuclear receptors (15). The first zinc finger is primarily thought to be responsible for target site recognition. Three residues in the carboxyl half of the first zinc finger termed the "P box" are thought to be responsible for response element discrimination (16,17). The second zinc finger stabilizes protein-DNA interactions and contains the "D box" region critical for receptor dimerization (5,6). The central DBD is also required for the repression of other transcription factors such as nuclear factor-B and AP-1 (9,11,18). The C-terminal LBD serves as the binding site for hormones, chaperone Hsp90, and coactivators (12,14). In addition, the DBD and LBD are important determinants in receptor mobility (19).
The GR is an evolutionary conserved transcription factor found in such species as human, rat, mouse, frog, rainbow trout, flounder, and others. The highest sequence homology exists among mammalian species, particularly in the highly conserved DBD and LBD. Nonmammalian species display less homology overall, both in conserved (DBD, LBD) and variable domains (N terminus) of the GR. Despite relatively high sequence homology, striking differences in glucocorticoid sensitivity have been observed among mammalian species. In some * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  instances, species-specific differences in glucocorticoid action result from amino acid changes within functional domains of the GR protein. For example, the guinea pig displays resistance to cortisol because of the presence of several amino acid residues in the LBD which differs from the human GR (hGR) (20). Similarly, sequence differences in the LBD are responsible for humans and other cortisol-secreting species (rabbit, pig) displaying a higher affinity for cortisol than rodents (rat, mouse), which bind corticosterone with a higher affinity (21). In addition, phosphorylation patterns vary among species, providing a mechanism for species-specific differences in phosphorylationinfluenced functions of GR which include protein stability, transcriptional activation, and subcellular localization (22)(23)(24). For example, glycogen synthase kinase-3 inhibits the transcriptional activity of rat GR through phosphorylation but has no effect on hGR (24). Alternatively, species-specific differences in GR signaling may result from the elevated expression of proteins that alter GR function. For example, the elevated expression of the FK506-immunophilin FKBP51 causes glucocorticoid resistance in several New World primates by decreasing binding of glucocorticoids to GR (25). In addition, the dominant-negative form of the GR, hGR␤, is expressed in humans but not other species (26).
In the present study, we investigated the molecular basis for species-specific differences in the transcriptional activation of simple and complex promoters by mouse GR (mGR) and hGR in response to glucocorticoid agonists and antagonists. Our studies reveal that mGR displays distinct promoter-specific transcriptional activities compared with hGR. A comprehensive analysis of GR domains and hormone-responsive promoters indicated that promoter discrimination mapped to a single amino acid at residue 437 of the DBD of mGR and to the identity of the Ϫ6 position of the GRE.
DBD-VP16 fusion constructs were constructed by PCR. The nucleotide sequence corresponding to the DBD of mGR (amino acids 414 -524) was amplified from the pCMVmGR using the primers 5Јtataggtaccggcaccatgtcagtgttttc-3Ј and 5Ј-tatagatatccacctccagcagtgacacca-3Ј. The product was digested with KpnI and EcoRI enzymes and ligated into digested pcDNA3.1 (Invitrogen). The VP16 activation domain was amplified from pAct (Clontech) with primers 5Ј-tatagatatctcgacggcccccccgacc and 5Ј-tatactcgagttatccccgacccgggaat-3Ј. The product, digested with EcoRI and XhoI, was then inserted downstream of the DBD to generate the pcDBD437V-VP16 fusion expression vector. The expression vector pcDBD437G-VP16 was generated by mutagenesis using mGR437G primers as described above.
Construction of Reporter Gene Constructs-The TAT2-LUC reporter gene was generated by excising a BamHI-HpaI fragment from GRE2-CAT (28) which contained the TAT GREs and TATA box sequence. The fragment was subcloned into BamHI-HpaI sites of pXP2 plasmid (29). The reporter gene MMTV-LUC (pHH-LUC) has been described previously (29). To generate single GRE luciferase reporter genes, the tandem GREs were excised from the plasmid TAT2-LUC by restriction analysis with BamHI and XhoI enzymes. Double-stranded synthetic oligonucleotides were generated which correspond to the GREs from the mouse mammary tumor virus (MMTV; 30), rat tyrosine aminotransferase (TAT; 31), human metallothionein (MTIIA; 32), and tryptophan oxygenase (TO; 33) genes (34). The double-stranded oligonucleotide contained BamHI and XhoI sites and was subsequently ligated into previously digested plasmid. The sequences of oligonucleotides used were: TAT, tataggatccctctgctgtacaggatgttctagctacctcgagatata; MMTVI, tataggatcctttttggttacaaactgttcttaaaacctcgagatata; MMTVII, tataggatcctggtttggatcaaaggttctgatctgc tcgagatat; MTIIA, tataggatccgcacccggtacactgtgtcctcccgctctcgagatat; and TO, tataggatcctgctccctttcatgatgcctggcccactcgagatat. The chimeric TAT2-MMTV reporter gene was generated by excising a Sal-EcoRV fragment from TAT2-LUC, which contains the TATA box sequence and luciferase gene. A fragment containing the TATA box, NF-1, Oct-1, and luciferase gene was then amplified by PCR from MMTV-LUC with primers 5Ј-ggtactcgagtttggaattttatccaa-3Ј and 5Ј-ggggccacctgatatcctt-3. The PCR product was digested with XhoI-EcoRV and inserted into the SalI-EcoRV-digested TAT2 plasmid to generate the full-length TAT2-MMTV reporter gene.
Cells and Growth Conditions-COS-1 cells were purchased from American Type Culture Collection (Manassas, VA). E8.2 cells were a gift from Dr. Paul Housley (Department of Pharmacology, Physiology, and Neurobiology, University of South Carolina). All cells were propagated in Dulbecco's modified Eagle's medium supplemented with 2 mM glutamine and 10% fetal calf serum. All cultures were maintained in a 5% CO 2 humidified atmosphere at 37°C and passaged every 3-4 days and used in experiments during a 2-month period.
Eukaryotic Transfections and Reporter Gene Assays-Approximately 0.75 ϫ 10 5 cells/20-mm dish or 2 ϫ 10 5 cells/35-mm dish were plated and incubated at 37°C for 16 -24 h. Cell culture medium was then removed and replaced with Opti-MEM (Invitrogen). The appropriate plasmid vector and reporter gene were transfected with TransIT as detailed by the manufacturer (Mirus Corp., Madison, WI). The transfection complexes were incubated with cells for 8 h, removed, and replaced with Dulbecco's modified Eagle's medium containing 10% charcoal-stripped fetal calf serum. After a 24-h recovery, the appropriate steroid or vehicle control was added to the cells and incubated for an additional 16 -24 h. Cells were then scraped in 1ϫ reporter gene assay buffer (Roche Applied Science) supplemented with protease inhibitors (0.1 nM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1 M pepstatin, 1 M leupeptin). Cell lysates were vortexed and incubated at room temperature for 15 min. After a brief centrifugation at 15,000 ϫ g, luciferase activities were determined as detailed by the manufacturer (Roche Applied Science) using a Dynex Technologies (Chantilly, VA) luminometer microplate reader. For determination of protein expression, cell lysates were processed for Western blot analysis as detailed below.
Western Blot Analysis-Total cell lysates were prepared by briefly sonicating cell lysates prepared in 1ϫ reporter lysis buffer as detailed above. Protein concentrations were determined by a Bio-Rad protein assay kit using bovine serum albumin as the standard. Protein samples were resolved on 8% precast Tris-glycine gels (Invitrogen) and were transferred electrophoretically to nitrocellulose using the Hoefer semiphor semidry electrophoretic system (Amersham Biosciences) as detailed by the manufacturer. Blots were stained with a 1:2,000 dilution of anti-GR antibody (57) and a 1:10,000 dilution of anti-␤-actin (Sigma) in TTBS (50 mM Tris, 150 mM NaCl, 0.15% Tween, pH 7.5) supplemented with 5% dry milk for 1 h at 22°C, according to the manufacturer's protocol. Blots were washed with three changes of TTBS for a total of 45 min. Blots were then incubated in TTBS buffer with 5% dry milk containing a 1:10,000 dilution of GAR-HRP for 1 h at room temperature. After three washes with TTBS, bands were visualized with the enhanced chemiluminescence (ECL) kit as specified by the manufacturer (Amersham Biosciences). Multiple exposures of each set of samples were produced and were within the linear range established for each antibody.
Competitive DNA Binding ELISA-The GREs utilized in the ELISAs were generated by annealing biotinylated oligonucleotides (IDT, Coralville, IA). The sequences of the GREs used were: TAT GRE, 5Ј-aagctttgtacaggatgttctaagctt-3Ј; and MMTVI GRE, 5Ј-tttttggttacaaactgttcttaaaac-3Ј. Double-stranded oligonucleotides were quantified with the PicoGreen double-stranded DNA quantitation kit as described by the manufacturer (Molecular Probes, Eugene, OR). A DNA binding ELISA was performed essentially as described previously (35,36). Briefly, 1 pM TAT or MMTVI was incubated in streptavidin-coated 96-well plates (Pierce) for 1-2 h.
Cytosol was prepared from COS-1 cells transfected with the appropriate recombinant GR plasmid by homogenization in MENG buffer (25 mM MOPS, pH 7.1, 1 mM EDTA, 10% glycerol) supplemented with protease inhibitors. Homogenates were centrifuged at 22,000 ϫ g for 10 min at 4°C, and aliquots were stored at Ϫ70°C prior to use. Protein concentrations were determined by a Bio-Rad protein assay kit. The relative expression of each recombinant GR protein expressed in the cytosol was determined by Western blot analysis as described above. GR protein was activated by incubating the cytosol with 100 nM dexamethasone for 2 h at room temperature. The in vitro activated cytosol was then incubated at 22°C for 15 min in 1ϫ DNA ELISA buffer (20 mM HEPES, 80 mM KCl, 5 mM MgCl 2 , 2 mM dithiothreitol, 10% glycerol) containing 0.1 mg/ml poly(dI⅐dC) for a total of 100 l. Samples were then added to 96-well plates coated previously with the biotinylated TAT GRE and incubated for 1 h at room temperature. To generate competition curves, excess cold nonbiotinylated oligonucleotides corresponding to the TAT GRE or MMTVI GRE were added to wells prior to the addition of samples. Wells were washed 10 times with PBST (1ϫ phosphate-buffered saline, 0.05% Tween) to remove nonbound complexes. Anti-GR antibody was diluted 1:2,000 in 1ϫ DNA-binding buffer supplemented with 1% dry milk. 100 l of the solution was added to the wells and incubated for 1 h. Wells were washed 10 times with 1ϫ PBST and then incubated for additional 1 h with a 1:5,000 dilution of GAR-HRP antibody in 1ϫ DNA-binding buffer containing 3% dry milk. After a final wash, immunocomplexes were detected with the SuperSignal ELISA Femto Maximum Sensitivity Substrate (Pierce) and luminescence measured on a microplate reader (Dynex Technologies). Competitor displacement curves were generated based on the LUC values. EC 50 values were calculated using one-site competition curve analysis by GraphPad Prism software (San Diego).
Amino Acid Analysis-Protein and nucleotide sequences of members of mouse and hGR were obtained using GenBank (37, www.ncbi.nlm. nih.gov) and the Nuclear Receptors data base (38, www.receptors.org/ NR). GenBank TM accession numbers for mGR and hGR are X04435 and X03225, respectively.

Comparison of Promoter Activation by mGR and hGR-
The transcriptional activities of the mGR and hGR were evaluated on two model glucocorticoid responsive promoters: 1) a simple GRE promoter, and 2) the complex MMTV promoter (Fig. 1A). The simple reporter gene, TAT2-LUC, contains two copies of the GRE from the rat TAT gene upstream of a TATA box sequence and the LUC reporter gene (Fig. 1A). The complex MMTV-LUC reporter gene is comprised of two dissimilar GREs, two half-site GREs (TGTTCT), and sites for the transcription factors sites NF-1 and Oct-1. Mouse GR and hGR were evaluated for their ability to activate the simple and complex glucocorticoid responsive reporter genes in COS-1 cells in response to natural and synthetic glucocorticoids (Fig. 1B). Mouse GR displayed enhanced activity on the simple TAT2-LUC reporter gene over the complex MMTV promoter (Fig. 1B). In contrast, hGR activated transcription better in the context of the complex MMTV promoter than the simple TAT2-LUC reporter gene. Mouse GR displayed ϳ5-fold higher activation over hGR on TAT2-LUC and 1.6-fold less activity on the MMTV promoter in response to dexamethasone. Importantly, the promoter-specific activities were independent of the agonist tested and were observed with corticosterone, cortisol, and triamcinolone, ligands that display differing affinities for GR (27,39,(41)(42)(43). The agonists represent a structurally similar class of compounds that are thought to induce similar conformational changes within the GR. The expression levels of each protein in the reporter gene assays were nearly identical as determined by Western blot analysis (Fig. 1C). Collectively, these data revealed that mGR displayed distinct promoter-specific transcriptional activities compared with hGR in response to glucocorticoid agonists.
A Single Amino Acid Change in the DBD of mGR Exerts Promoter Discrimination-We next focused on identifying the functional domains that contributed to the species-specific promoter discrimination by mGR. In preliminary studies, we examined the role of the N terminus of GR and found that the domain was not responsible for the distinct promoter activities between mGR and hGR in response to glucocorticoids (data not shown). This led us to investigate the consequence of a single amino acid change at the tip of the first zinc finger of the DBD of the mGR (Fig. 2). The mGR utilized in this study, originally reported by Danielsen et al. (44), contains a valine at position 437 instead of a glycine found in hGR. The amino acid change results from codon change of GGA to GTA (44). The cDNA sequence (GenBank TM accession number X04435) containing this nucleotide change acts as the reference standard for the mGR by genomic data bases such as NCBI SwissProt, and MGI. Residue 437 of mGR also appears to be polymorphic among mouse cell lines (45,46). Similar polymorphisms have been found in the vitamin D receptor (G30D) and the androgen receptor (G551V) and have been associated with hypocalcemic rickets and partial androgen insensitivity syndrome, respectively (47)(48)(49). The mutations in androgen receptor and vitamin D receptor appear to disrupt DNA binding and transcriptional activation properties of the receptor.
Amino acid 437 of mGR is located outside the P box (Fig. 2), a region thought to act as the primary determinant of sequence recognition among glucocorticoid and estrogen receptors (16,17). Response element specificity depends largely on the identity of the amino acids within the P box. Residues outside the P box region have been previously reported to make contacts with phosphate backbone of DNA and are thought to help provide structural information (5,6,15). The functional importance the conserved glycine at positions corresponding to residue 437 of mGR has not been investigated extensively. Danielsen et al. (44) reported the glycine to valine change in mGR results in a functional protein capable of activating the MMTV promoter in COS-1 cells. However, examination of their data suggests that the mutation did cause a decrease in transcriptional activity. In contrast, Zandi et al. (50) reported that the glycine to valine amino acid change had no effect on activity of mGR in a MMTVbased assay. In both studies, analysis was limited to the MMTV promoter and in response to a single concentration of dexamethasone. The effect of this mutation on the GR activity has not been examined with other glucocorticoid-responsive promoters.
The valine residue at position 437 of mGR (mGR437V) was mutated to the conserved glycine residue (mGR437G) and the mutants assayed for their ability to activate the simple and complex reporter genes in response to dexamethasone in both COS-1 and E8.2 cells. In both cell types, mutation of the valine to glycine reduced the transcriptional activity of mGR on the TAT2 promoter and enhanced the activity on the MMTV reporter gene (Fig. 3A). Dose-response curves performed in COS-1 cells confirmed that the effect of the mutation on promoter specificity was independent of concentration of dexamethasone (Fig. 3B). Western blot analysis revealed mGR437V expressed slightly less than mGR437G (Fig. 3A). To confirm that the promoter-specific activities were not caused by slight differences in expression levels, activation of the TAT2 reporter gene was measured in response to increasing concentrations of transfected receptors. Fig. 3C shows that mGR437V has enhanced activity over mGR437G at each GR concentration in response to glucocorticoids. In addition, mGR437G failed to reach the maximum level of transcriptional activation achieved by mGR437V. To determine whether the functional consequence of the mutation is conserved among species, the glycine at the corresponding position in human (amino acid 430) and rat (amino acid 449) GRs was mutated to a valine residue. Consistent with mGR, the valine mutants of human and rat GR displayed enhanced activity on the simple TAT2-LUC reporter and decreased activity on the complex MMTV promoter (Fig.  4A). Collectively, these findings provide evidence that the promoter-specific activity of mGR results from a single amino acid change of glycine to valine at residue 437 in the DBD. Further, these findings suggest that a residue outside the P box region also provides important structural information for response element recognition.
Promoter Specificity Depends on the Identity of the Amino Acid at Position 437-To investigate further how promoterspecific activity of mGR could be modified by the identity of the amino acid at position 437, the residue was replaced by a leucine (mGR437L), alanine (mGR437A), histidine (mGR437H), or an aspartic acid (mGR437D). The range of substitutions represents various classes of amino acids with different sizes of side chains. Alanine, valine, and leucine are nonpolar amino acids with increasingly larger aliphatic side chains, respectively. Histidine contains a basic side chain composed of an imidazole ring. The mGR437 mutants were assayed for their ability to activate the simple TAT2 and complex MMTV reporter genes in response to a concentration range of dexamethasone (Fig. 4B). On the TAT2 promoter, potency of activation occurred in the order of mGR437L, mGR437A, mGR437V, mGR437H, mGR437D and mGR437G. All mutants displayed higher activation than mGR437G on the TAT2 promoter. Mutants containing residues at position 437 with smaller side chains (Leu, Ala, Val) showed the strongest activation. The activation profile on the complex MMTV promoter differed drastically with no direct correlation to the identity of side chain within amino acid 437. The mGR437L mutant displayed the highest activation on the simple promoter but the lowest on the complex MMTV promoter. In contrast, the mGR437A mutant maintained strong activation on both promoters. Aspartic acid was the least favorable substitution, with mGR437D exhibiting diminished activity on both promoters. In addition, the dose-response curve was shifted to the right for mGR437D on the TAT2 reporter gene. The diminished activity of mGR437D may relate to its stability as a major degradation product of the protein as detected by Western blot analysis (Fig. 4B). The expression levels of the other mutants were comparable. Together, these findings suggest that the identity of the amino acid at position 437 influences the transcriptional properties of mGR on the simple and complex promoters, demonstrating its role in promoter discrimination.
Molecular Basis of Promoter Specificity-Studies thus far have identified residue 437 as a primary determinant within the mGR responsible for promoter discrimination. We next analyzed the composition of the simple and complex promoters to determine the role of the DNA elements in the differential transcriptional activity of mGR437V. The simple TAT2 promoter differs from the MMTV promoter in several aspects: the number of GREs, the absence of NF-1 and OCT-1 sites of the MMTV promoter, and the individual response element sequence (Figs. 1A, and 5B). We first investigated whether the number of GREs could be a determinant of promoter-specific activation because the number of GREs in the context of a simple promoter has been reported to lead to cooperatively of GR activity (51). The TAT2 promoter contains two full-length GREs, whereas the MMTV promoters contain two full-length and two half-site GREs (Figs. 1A and 5B). The half-site GREs correspond to the conserved 3Ј-site (TGTTCT), and each has been shown to be functionally occupied by one subunit of the GR (52). To evaluate whether the number of GREs within the simple promoter influenced the activity of mGR437V, the mutant was analyzed on a simple promoter comprised of one, two, or three copies of the TAT GRE. In response to dexamethasone, mGR437V displayed enhanced activity over mGR437G regardless of the number of copies of GRE (Fig. 5A). We next investigated the consequence of the NF-1 and Oct-1 sites present in the MMTV promoter but absent in the simple TAT2 promoter.
A chimeric reporter gene was generated which contained the tandem GRE elements of the TAT2 promoter fused with the NF-1 and Oc-1 sites of the MMTV promoter (Fig. 5B). The presence of the NF-1 and Oct-1 sites in the context of TAT GREs had no effect on the transcriptional activities of the mGR437 mutants (Fig. 5C). The mGR437V mutant displayed enhanced activity over mGR437G on simple GRE and chimeric TAT2-MMTV promoters.
The findings suggest that a second determinant of promoterspecific activity of mGR437V resides in the response element sequences, which differ among the promoters. To test this hypothesis, reporter genes were constructed which contained a single copy of the rat TAT, human MTIIA, rat TO, MMTVI or MMTVII GREs fused upstream of a minimal TATA promoter and luciferase reporter gene. The MMTV response elements represent the first (MMTVI) and second (MMTVII) GREs from the complex MMTV promoter. The sequences of the various GREs differ mainly in the 5Ј-variable half-site and spacer regions (Fig. 6A). The mGR437 mutants were analyzed for their ability to activate the individual reporter genes in response to a concentration range of dexamethasone. The mGR437V mutant displayed higher transcriptional activity on the TAT, MMTVII, and MTIIA response elements and lower activity on the MMTVI and TO GREs compared with mGR437G (Fig. 6B). The transcriptional activity of the mutants correlated directly with the identity of the nucleotide present at the Ϫ6 position of the response element. Higher activity was observed with mGR437V on individual response elements (TAT, MTIIA, MMTVII) that contain a guanine at the Ϫ6 position of the GRE and lower activity on GREs (MMTVI, TO) that contained a thymine at this position. A guanine at the Ϫ6 position is the preferred nucleotide in the consensus response element of nuclear receptors (Fig. 7A). The GR has been reported to display a higher affinity for a GRE with a guanine at Ϫ6 than any other base (53,54). To determine whether the identity of the Ϫ6 base was actually responsible for influencing the activity of mGR437, the 5Ј-half-site of MMTVI was mutated to that of the TAT GRE. The TAT and MMTVI GREs differ at the Ϫ7 and Ϫ6 positions and the spacer region (Fig. 7B). The Ϫ7G and Ϫ6T of the MMTVI GRE were mutated alone and in combination to the Ϫ7T and Ϫ7G of the TAT response element. Mutation of the guanine at the Ϫ7 position of MMTVI GRE to a thymine caused a minimal increase in mGR437V and a decrease in mGR437G transcriptional activities (MMTVI versus MMTVI G Ϫ7 T, Fig. 7B). The minimal decrease in transcriptional activity of mGR437G is consistent with previous DNA binding specificity studies in which the GR displayed a lower affinity for GREs with a Ϫ7 thymine instead of a guanine (53,54). In contrast, mutation of the Ϫ6 thymine to guanine in MMTVI strongly increased transactivation by mGR437V and mGR437G by 8-and 1.5-fold, respectively (MMTVI versus MMTVI T Ϫ6 G, Fig. 7B). At least for mGR437G, the enhanced activity is consistent with a reported increased affinity of wildtype GR for GREs containing a guanine at the Ϫ6 position. The Western blot to the right of each graph represents 20 g of total cell lysate from the reporter gene assays stained with anti-GR (1:2,000) and anti-␤-actin (1:10,000) and detected with GAR-HRP (1:10,000) by ECL. B, response element discrimination by substitution mutants at residue 437. Transactivation activity of mGR mutants that contain a valine, glycine, leucine, histidine, alanine, or aspartic acid at residue 437 was measured in COS-1 cells on the TAT2 and MMTV promoters in response to 16 h of treatment with increasing concentrations of dexamethasone (0.1-1,000 nM). Each data point represents the mean Ϯ S.E. of triplicate samples from a representative experiment. Differences in relative activities were reproducible in three independent experiments. The Western blot to the right shows the expression of the individual mutants in the reporter gene assay. Staining for ␤-actin was included as a control for loading. An arrow indicates possible degradation products of the mGR437D mutant.
More importantly, the substitution of a guanine at the Ϫ6 position restored the enhanced activity of mGR437V over mGR437G, confirming the role of this residue in the observed promoter discrimination. Mutation of both Ϫ7 and Ϫ6 positions in the MMTVI GRE essentially converted the response element into the TAT GRE (TAT versus MMTVI G Ϫ7 T, T Ϫ6 G). Transcriptional activation by the mGR mutants was similar on both response elements. Collectively, these findings indicate that a major determinant of the response element discrimination of mGR resides with the identity of Ϫ6 position of the GRE. The Ϫ7 position of the GRE can further regulate the potential of transcriptional activation depending on the identity of Ϫ6 base.
Basis of Enhanced Transcriptional Activity on Promoters-Our studies have identified the key positions in mGR and response elements responsible for mGR promoter discrimination. We next focused on determining whether the differential promoter activities of mGR437V resulted from alterations in DNA binding affinity and/or transcriptional activity. DNA binding was first measured with isolated DBDs of the receptor mutants by a reporter gene approach. The DBDs of mGR437V and mGR437G were fused to the VP16 activation domain (Fig.  8A). This approach allows the determination of DNA binding properties without ligand or influences by the N-and C-terminal domains of GR. DNA binding was measured by determining activation on the simple TAT2 and complex MMTV promoters and reporter genes containing a single copy of the TAT or MMTVI GREs (Fig. 8B). The mGR437V and mGR437G mutants showed similar DNA binding on reporter genes containing the TAT GRE (TAT1, TAT2). In contrast, the DNA binding of mGR437V was reduced dramatically on MMTVI and fulllength MMTV promoters compared with mGR437G. The overall DNA binding on the MMTVI response element was also lower for both mutants compared with the TAT GRE, consistent with the reported decreased affinity of GR for GREs containing a thymine at the Ϫ6 position (53,54). The DNA binding activities were also confirmed by determining the relative affinity of the full-length mGR437 proteins for the TAT and MMTVI GREs using competitor DNA binding ELISAs. Consistent with the DBD studies, the full-length mGR437V and mGR437G proteins displayed similar relative affinities for the TAT GRE (Fig. 8C). Similarly, a lower affinity was observed for the MMTVI GRE for mGR437V with respect to mGR437G. Collectively, these data suggest that the diminished transcriptional activities of mGR437V on MMTV GRE and the complex MMTV promoter may in part come from a decrease in DNA binding affinity. The decreased DNA affinity of mGR437V for MMTVI GRE correlates with the presence of a thymine at the Ϫ6 position of the response element. However, the basis for the decreased DNA binding affinity for the full-length MMTV promoter is complicated by the additional presence of the MMTVII and half-site GREs. The contribution and cooperativity of the individual GREs of the MMTV promoter to GR-DNA binding and transactivation remain to be fully characterized. In contrast, mGR437 displayed enhanced activity on the TAT GREbased promoters despite having DNA binding similar to that of mGR437G. These findings suggest that the glycine to valine change at position 437 can enhance the transcriptional activation function of GR without affecting DNA affinity. DISCUSSION Promoter specificity among nuclear receptors has largely been attributed to the identity of P box residues located at the base of the first zinc finger (16,17). In this paper, we illustrate that residues outside the P box are critically important for maintaining the functional integrity of promoter discrimination by GR. By investigating species-specific differences in promoter activation, we identified a single amino acid change (G437V) at the tip of the first zinger finger domain of mGR which has a profound impact on target site recognition, transcriptional activation, and receptor mobility. A glycine residue is critically important at this position as demonstrated by its conservation among the nuclear receptor family and conse- quence of its mutation. For example, patients with the corresponding mutation at this position (G551V) in the human androgen receptor gene exhibit the partial androgen insensitivity syndrome (47,48). Similarly, the mutation (G30D) in the human vitamin D receptor gene has been associated with hypocalcemic rickets (49). In both androgen receptor and vitamin D receptor studies, the corresponding Gly to Val mutation appears to influence DNA binding and transactivation negatively, transiently transfected with pCMVmGR437V or pCMVmGR437G together with the indicated reporter gene. After transfection, cells were treated with increasing concentrations of dexamethasone (0.1-1,000 nM) for 16 h, harvested, and analyzed for luciferase activity. Each data point represents the mean Ϯ S.E. of triplicate samples from a representative experiment. Differences in relative activities were reproducible in three independent experiments. although transcriptional analysis was only limited to a few promoters. Our studies indicate that the mutation in the GR also has a profound effect on DNA binding and transcriptional activation in a promoter-specific fashion and the mobility of GR within the nucleus. The functional outcome of the GR mutation depends largely on two determinants: the identity of the amino acid at position 437 of GR, and the nucleotide at the Ϫ6 position of the GRE. Mouse GR with a valine at position 437 (mGR437V) displayed reduced DNA affinity and transcriptional activity on GREs that contain a thymine at the Ϫ6 position compared with wild-type GR (mGR437G). In contrast, mGR437V showed enhanced activation over mGR437G on GRE sequences with a Ϫ6 guanine despite both mutants having similar affinities for the GRE sequence.
Importantly, these data suggest a nonlinear relationship between DNA binding and glucocorticoid-mediated transactivation. Indeed, a linear correlation between DNA binding and transactivation does not appear to exist always. The GR can activate or repress gene expression to different extents depending on the context of the response element and the ligand involved. For example, the GR can bind composite GREs such as in the proliferin, proopiomelanocortin, and human corticotropin-releasing hormone genes and repress transcription (55,56). In addition, GR shows similar affinities for response elements in response to dexamethasone and the partial agonist/ antagonist RU486 but displays strikingly different transcriptional activation properties (43,58). The uncoupled relationship between DNA binding and transactivation appears to be especially apparent for GRs with mutations in the DBD. For example, the C438G mutation within the first zinc finger domain of hGR leads to 10% of DNA binding and 80% of transcriptional activity of wild-type GR on the MMTV promoter (47). Similarly, a K442G mutation in hGR results in less than 1% transactivation but 60% DNA binding activity of wild-type GR (12). One major reason for the discord between DNA binding and transactivation is the sequence of the response element. In addition to influencing receptor affinity, the response element acts as an allosteric regulator of receptor conformation (7, 59 -62). Binding of nuclear receptors to DNA is thought to transduce structural changes to other functional domains and expose transcriptional activation surfaces required for efficient transcription (63). The sequence of the response element influences the transcriptional potential of the receptor by modulating receptor conformation. In support of this, Yamamato and co-workers (7) identified a lysine residue (Lys-461) in rat GR which acts as an allosteric lock and restricts receptors to inactive configurations except in favorable response element contexts (7). Replacing Lys-461 with an alanine in rat GR yields a receptor capable of activating transcription from most response elements and promoters. In addition, mutations of Ser-459 and Pro-493R in GR mimic the allosteric effect of DNA and produce a receptor that is in DNA-bound conformation state in the absence of GRE binding (61). Recently, numerous studies have also shown that estrogen response element sequence alters the conformation of ER and influences coactivator recruitment and transcriptional regulation (59,60).
The findings in this paper support the role of response element sequence as a determinant of both receptor affinity and in the allosteric regulation of GR transactivation. We propose that promoter discrimination involves two interdependent mechanisms of action. First, the DNA binding affinity of GR is influenced by the response element sequence and the conformation of the receptor. Distinct changes in receptor conformation induced by mutation or different ligands influence the ability of the receptor to recognize different response element sequences. Second, the response element sequence allosterically regulates GR conformation to affect transactivation. The outcome of the DNA-induced conformation change is determined primarily by the relationship between the GRE sequence and receptor conformation prior to DNA binding. Understanding how the response element sequence induces changes in receptor conformation is limited primarily by the availability of solved crystal structures of GR bound to different response element sequences. Multiple structures are required for proper analysis of hydrogen bonding patterns between GR and the GRE that differ depending on the nucleotide composition of the response element. Currently, the only available crystal structure is the rat GR DBD⅐GRE complex solved by Luisi et al. (6). The 5Ј-halfsite including the Ϫ6 position of the GRE within the complex differs from the response elements utilized in the present FIG. 8. DNA binding analysis of the mGR437 mutants. A, schematic diagram of the DBD-VP16 chimeras. The VP16 activation domain was fused to the 3Ј-end of the DBD (amino acids 415-524) of mGR437V (DBD437V-VP16) and mGR437G (DBD437G-VP16). B and C, COS-1 cells were transfected with the indicated DBD-VP16 construct and MMTV-LUC, TAT2-LUC, or reporter genes containing one copy of the TAT or MMTVI GRE. After a 48-h recovery, cells were harvested for Western blot analysis (B) and determination of luciferase activities (C). B, total cell lysates (20 g) from reporter gene assays were analyzed by Western blotting with anti-VP16 (1 g/ ml) and anti-␤-actin (1:10,000) and visualized by ECL with GAR-HRP (1:10,000). D, relative affinities of full-length mGR437 mutants for TAT and MMTVI response elements. Competitive DNA binding ELISA was used to measure DNA binding and calculate relative affinities of full-length mGR mutants for the TAT and MMTVI response elements essentially as described under "Experimental Procedures." EC 50 values were calculated based on the one-site competition curves using GraphPad Prism software. study. Furthermore, the mechanism by which the DBD of GR could transduce structural changes to other functional domains remains unresolved because of the absence of a crystal structure of the full-length GR complexed to DNA. However, it has been proposed that site-specific GRE binding influences the folding of the AF-1 transactivation domain and alters the ability to bind cofactors crucial for the transactivation process (40). Merit for this hypothesis is supported by recent studies demonstrating estrogen receptor structure, function, and coactivator recruitment can be modulated by response element sequence (59,60). Collectively, our studies demonstrate that residues outside the P box are critically important for maintaining the functional integrity of promoter discrimination by GR and in communicating with the response element sequences for the regulation of transcriptional activation.