Phosphorylation and Inhibition of Rat Glucocorticoid Receptor Transcriptional Activation by Glycogen Synthase Kinase-3 (GSK-3)

Transcriptional activation by the glucocorticoid receptor (GR) is regulated by both glucocorticoid binding and phosphorylation. The rat GR N-terminal transcriptional regulatory domain contains four major phosphorylation sites: threonine 171 (Thr171), serine 224 (Ser224), serine 232 (Ser232), and serine 246 (Ser246). We have previously demonstrated that Ser224 and Ser232are phosphorylated by cyclin-dependent kinases, while Ser246 is phosphorylated by the c-Jun N-terminal kinase. We report here that the remaining GR phosphorylation site, Thr171, is a target for glycogen synthase kinase-3 (GSK-3)in vitro and in cultured mammalian cells. Increasing GSK-3 activity through its overexpression in cultured cells inhibits GR transcriptional enhancement, an effect dependent upon Thr171. Correspondingly, overexpression of a constitutively active form of the GSK-3 inhibitor, protein kinase B/Akt, increases GR transcriptional enhancement. Overexpression of GSK-3 had no effect on GR-mediated transcriptional repression of AP1-dependent gene expression. Importantly, transcriptional activation by the human GR (hGR), which contains an alanine (Ala150) at the position equivalent to Thr171 in rat GR, is not affected by GSK-3 overexpression. Introduction of a threonine residue at this position (A150T) establishes GSK-3-mediated inhibition of hGR transcriptional activation. These findings demonstrate species-specific differences in GR signaling, as revealed through GSK-3 phosphorylation, which suggests that GR function in rodents may not fully recapitulate receptor action in humans and that hGR is capable of adopting the GSK-3 signaling pathway through a somatic mutation.

Transcriptional activation by the glucocorticoid receptor (GR) is regulated by both glucocorticoid binding and phosphorylation. The rat GR N-terminal transcriptional regulatory domain contains four major phosphorylation sites: threonine 171 (Thr 171 ), serine 224 (Ser 224 ), serine 232 (Ser 232 ), and serine 246 (Ser 246 ). We have previously demonstrated that Ser 224 and Ser 232 are phosphorylated by cyclin-dependent kinases, while Ser 246 is phosphorylated by the c-Jun N-terminal kinase. We report here that the remaining GR phosphorylation site, Thr 171 , is a target for glycogen synthase kinase-3 (GSK-3) in vitro and in cultured mammalian cells. Increasing GSK-3 activity through its overexpression in cultured cells inhibits GR transcriptional enhancement, an effect dependent upon Thr 171 . Correspondingly, overexpression of a constitutively active form of the GSK-3 inhibitor, protein kinase B/Akt, increases GR transcriptional enhancement. Overexpression of GSK-3 had no effect on GR-mediated transcriptional repression of AP1dependent gene expression. Importantly, transcriptional activation by the human GR (hGR), which contains an alanine (Ala 150 ) at the position equivalent to Thr 171 in rat GR, is not affected by GSK-3 overexpression. Introduction of a threonine residue at this position (A150T) establishes GSK-3-mediated inhibition of hGR transcriptional activation. These findings demonstrate species-specific differences in GR signaling, as revealed through GSK-3 phosphorylation, which suggests that GR function in rodents may not fully recapitulate receptor action in humans and that hGR is capable of adopting the GSK-3 signaling pathway through a somatic mutation.
Glucocorticoid hormones control cellular proliferation and metabolism through their association with the glucocorticoid receptor (GR), 1 a member of the intracellular receptor super-family of transcriptional regulatory proteins (1). Upon glucocorticoid binding, GR enters the nucleus, associates with specific DNA sequences termed glucocorticoid response elements (GREs), and increases transcriptional initiation from nearby promoters. GR can also repress transcription mediated by the heterodimeric AP1 transcription factor complex (c-Jun and c-Fos) (2). Although glucocorticoids act as the primary signal in activating GR's transcriptional regulatory functions, GR-mediated transcriptional activation is also modulated by phosphorylation (3)(4)(5).
Rat GR isolated from cultured mammalian cells or ectopically expressed in yeast (Saccharomyces cerevisiae) is phosphorylated on four major residues (6). These sites cluster to the N-terminal transcriptional regulatory domain and include threonine 171 (Thr 171 ), serine 224 (Ser 224 ), serine 232 (Ser 232 ), and serine 246 (Ser 246 ) (Fig. 1A). Each of these residues is followed by a proline, thereby forming a motif phosphorylated by a family of serine/threonine-proline-directed kinases that includes the cyclin-dependent kinases (Cdk), the mitogen-activated protein kinases, and glycogen synthase kinase-3 (GSK- 3). Differential phosphorylation at these sites both positively and negatively regulate GR transcriptional activation. Positive regulation is accomplished by cyclin-Cdk complexes: cyclin E-Cdk2 phosphorylates Ser 224 , while cyclin A-Cdk2 phosphorylates both Ser 224 and Ser 232 . Mutations at these sites, or of particular Cdk genes in yeast, reduce GR-dependent transcriptional activation, suggesting that phosphorylation of Ser 224 and Ser 232 is required for full GR transcriptional enhancement (7). In contrast, phosphorylation of Ser 246 by c-Jun N-terminal kinase, a member of the mitogen-activated protein kinases family, inhibits GR transcriptional activation (8).
The remaining GR phosphorylation site, Thr 171 , also resides in a motif recognized by serine/threonine-proline-directed kinases. However, our previous studies indicate that neither the Cdks, nor c-Jun N-terminal kinase efficiently phosphorylate Thr 171 in vitro. Furthermore, phosphorylation of Thr 171 is evident in both serum-deprived quiescent and serum-stimulated proliferating cells (8), suggesting that Cdks and c-Jun N-terminal kinase are unlikely to phosphorylate Thr 171 in vivo, since these kinases are largely inactive in serum-starved, nonproliferating cells. GSK-3, on the other hand, is active throughout the cell cycle, as well as in serum-deprived cells (9). Thus, GSK-3 may represent the GR kinase that phosphorylates Thr 171 .
Recent studies in Dictyostelium, Xenopus, and Drosophila have implicated GSK-3 in pathways other than glycogen metabolism. GSK-3 has been implicated in cell fate determination and differentiation through its ability to phosphorylate and regulate factors involved in cellular proliferation including CREB, c-Myc, c-Jun, and ␤-catenin (10, 20 -23). Although GSK-3 has no known activators, its activity in cultured cells can be increased through overexpression. GSK-3 enzymatic activity is, however, negatively regulated by protein kinase B/Akt, an enzyme that phosphorylates and inhibits GSK-3 (24). Akt is, in turn, activated through an association with lipid products generated by phosphatidylinositol-3 kinase at the cell membrane and through phosphorylation (25)(26)(27). The phosphatidylinositol-3 kinase-Akt pathway is induced in response to insulin, insulin-like growth factor, epidermal growth factor, and other mitogens (28,29). Recently, the phosphatidylinositol-3 kinase-Akt pathway has been implicated in cell survival, with a constitutively activated form of Akt leading to a reduction in apoptosis in neuronal cells (30,31). GSK-3 activity is also inhibited by the Wnt signaling pathway through an unknown mechanism, involving the Dishevelled protein (32,33). Here we examine whether rat GR is a substrate for GSK-3 in vitro and investigate the consequences of GSK-3 activation and inhibition on GR transcriptional regulation in cultured mammalian cells.

Purification of Receptor Derivatives and in Vitro Kinase
Assays-A wild type rat GR derivative containing amino acids 106 -318 or receptor mutant with a single amino acid substitution T171A, and a wild type human ER derivative containing estrogen receptor (ER) N-terminal amino acids 1-121, were expressed in Escherichia coli as glutathione S-transferase (GST)-fusion proteins (GST-GR 106 -318 and GST-ER 1-121 ) exactly as described previously (7). The most concentrated fractions (1 mg/ml) were used as substrates for the in vitro kinase assays.
GST-GR 106 -318 substrate (2 g) was bound to 100 l of a 50% slurry of glutathione beads for 20 min on ice and washed twice with 1 ml of DK buffer (50 mM potassium phosphate, pH 7.15, 10 mM MgCl 2 , 5 mM NaF, 5 mM dithiothreitol, supplemented with protease inhibitors, 1 mM phenylmethylsulfonyl fluoride, and 1 g/ml each of aprotinin, pepstatin A, and leupeptin). Approximately 50 milliunits of rabbit purified GSK-3 (Upstate Biotechnology) were added to the immobilized receptor along with 25 M ATP, 10 mM MgCl 2 , 1 mM dithiothreitol, and [␥-32 P]ATP (100 Ci) in a total volume of 80 l. The kinase reactions were allowed to proceed for 30 min at room temperature with continuous shaking. The beads were washed five times with 1 ml of phosphate-buffered saline to remove unincorporated radioisotope. GST-c-Jun fusion protein containing c-Jun amino acids 1-223 (GST-c-Jun 1-223 , kindly provided by P. Schwenger) and GST-ER 1-121 were phosphorylated under identical conditions and used as an independent measure of kinase activity. The labeled GST-GR 106 -318 protein, GST-ER 1-121 , and the GST-c-Jun 1-223 control substrates were released by boiling for 2 min in an equal volume of 2ϫ SDS sample buffer and run on 10% SDS-PAGE. The gels were stained with Coomassie Blue to visualize the substrate proteins, dried, and incorporation of isotope was detected by autoradiography.
Phosphopeptide Mapping and Phosphoamino Acid Analysis-For phosphopeptide mapping, polyacrylamide gels containing the labeled receptor were washed three times for 10 min each in 500 ml of water and dried between cellophane sheets. Following autoradiography, the GR band was excised and the rehydrated gel slice was placed into a microcentrifuge tube in 50 mM ammonium acetate (pH 4.0), 1 mM dithiothreitol, and 50 g of V8 protease (endoproteinase Glu-C, Boehringer Mannheim). After 10 h at room temperature, an additional 50 g of V8 protease was added and incubation was continued for 5 h at room temperature. Samples was centrifuged for 5 min at 12,000 ϫ g and the supernatant containing the digested peptides was evaporated to dryness. Peptides were resuspended in 0.5 ml of water, dried, washed again, and dissolved in 10 l of electrophoresis buffer I, pH 1.9 (15% acetic acid, 5% formic acid). Peptides were electrophoresed in the same buffer on a thin layer chromatography plate (microcrystalline cellulose adsorbent without fluorescent indicator; Kodak) at 1000 V for 50 min. Plates were then dried, subjected to ascending chromatography in the second dimension for 3.5 h with 37.5% butanol, 25% pyridine, and 7.5% acetic acid, air-dried, and exposed to film (34).
For phosphoamino acid analysis, 32 P-labeled receptor was transferred to Immobilon paper (Millipore Corp.), and the GR band was visualized by autoradiography, excised from the membrane, and hydrolyzed in 100 l of 6 N HCl (Pierce) by heating to 110°C for 60 min. Samples were washed twice in 0.5 ml of water, dried, and resuspended in 8 l of electrophoresis buffer I. The hydrolysates were spotted onto a TLC plate, along with phosphoamino acid standards (1 l of mixture of phosphoserine, phosphothreonine, and phosphotyrosine (Sigma), 1 mg/ml each), and resolved in the first dimension by electrophoresis at 1500 V for 20 min in electrophoresis buffer I, and in the second dimension by electrophoresis at 1300 V for 16 min in buffer II, pH 3.4 (5% acetic acid, 0.5% pyridine). After drying, plates were sprayed with 0.25% w/v ninhydrin in acetone and developed at 70°C for 10 min to visualize the phosphoamino acid standards, and autoradiography was performed.
Site-directed Mutagenesis-Site-directed mutagenesis of the human GR alanine 150 to threonine was performed using Stratagene's Quick Change site-directed mutagenesis procedure and high fidelity Pfu DNA polymerase, according to the manufacturer's instructions, with the following oligos: 5Ј-GCTGTGTCTGCTACCCCCACAGAGAAG-3Ј and 5Ј-CTTCTCTGTGGGGGTAGCAGACACAGC-3Ј.
Plasmids-pCMV-wt GR and pCMV-GR T171A expression plasmids were used to produce rat GR, and XG 46 TL reporter plasmid, containing two consensus GREs upstream of thymidine kinase promoter (Ϫ109) linked to a luciferase gene was used to assay GR transcriptional activity. An XAP1TL reporter plasmid, containing a single AP1 binding site upstream of the thymidine kinase promoter fused to a luciferase gene, was used to assay transcriptional repression. pcDNA3-hGR and pcDNA3-hGR A150T plasmids expressed the human wild type and the alanine to threonine mutant GRs, respectively. pCMV5-HA-GSK-3␤ expressed HA-tagged GSK-3 and pCMV6-HA-Akt plasmid expressed a constitutively active myristylated HA-tagged form of Akt (28). A pCMV5 empty vector was used to equalize the total amount of DNA transfected in each experiment. pCMV-LacZ plasmid produced ␤-galactosidase.
Transient Transfections and Reporter Activity Assays-U-2 OS cells were plated on 60-mm dishes in DMEM, 10% FBS. One hour prior to transfection cells were refed with fresh medium and transfected with the indicated plasmids via the calcium phosphate precipitation method as described elsewhere (35). Eight hours later, cells were washed three times with prewarmed phosphate-buffered saline to remove calcium phosphate precipitates, allowed to recover overnight in DMEM, 10% FBS and incubated with fresh medium containing 100 nM dexamethasone where indicated, for an additional 8 h.
HeLa and PC-12 cells were plated in 60-mm dishes, washed once with serum-free medium and transfected with the indicated plasmids using 20 l of LipofectAMINE reagent (Life Technologies, Inc.) in a total volume of 2.5 ml of serum-free phenol red-free DMEM per 60-mm dish according to the manufacturer's instructions. Three hours posttransfection 2.5 ml of DMEM, 20% FBS was added to each dish and cells were incubated for another 12 h. The next day cells were refed with fresh DMEM, 10% FBS with 100 nM dexamethasone or identical volume of 100% ethanol and incubated for an additional 8 h.
Western Blotting-To make protein extracts from transfected cells, U-2 OS cells were washed twice with phosphate-buffered saline and lysed directly on the dishes in 200 l of ice-cold lysis buffer (150 mM NaCl, 50 mM Hepes, pH 7.5, 1 mM EDTA, 1 mM EGTA, 10% glycerol, 1% Triton X-100, 1 mM NaF, 25 M ZnCl 2 , supplemented with protease inhibitors (described above) and phosphatase inhibitor 1 mM sodium orthovanadate). The lysates were collected and precleared by centrifugation (10,000 ϫ g for 10 min at 4°C). The protein concentration in all samples was adjusted with the lysis buffer, and 200 l of the whole cell extracts was boiled for 2 min with 50 l of 5ϫ SDS sample buffer. For immunoblotting, protein extracts were fractionated by 10% SDS-PAGE, transferred to Immobilon paper, and probed with mouse monoclonal antibodies against GSK-3 (Upstate Biotechnology), HA-tagged Akt (Boehringer Mannheim), or with anti-GR rabbit polyclonal antiserum (Santa Cruz Biotechnology, Inc.). The blots were developed using horseradish peroxidase-coupled sheep anti-mouse or donkey anti-rabbit antibodies and the enhanced chemiluminescence (ECL) substrate per the manufacturer's instructions (Amersham Corp.).

GSK-3 Phosphorylates the Rat Glucocorticoid Receptor in
Vitro at Threonine 171-To examine whether GSK-3 can utilize the rat GR as a substrate in vitro, we tested purified rabbit GSK-3␤ for its ability to phosphorylate a GST-GR fusion protein containing the receptor residues 106 through 318 (GST-GR 106 -318 ). We compared GR phosphorylation by GSK-3 to that of the established GSK-3 substrate, the transcription factor c-Jun. Fig. 1B demonstrates that GSK-3 phosphorylates the GST-GR 106 -318 and GST-c-Jun 1-223 in vitro with similar effi-ciency. In contrast, under the same experimental conditions GSK-3 failed to phosphorylate an ER derivative encompassing residues 1 through 121 (GST-ER 1-121 ), which contains three serine-proline phosphorylation sites at Ser 104 , Ser 106 , and Ser 118 and has been shown previously to be substrate for cyclin A-Cdk2 complex (37) (Fig. 1B). Thus, it appears that the Nterminal transcriptional regulatory domain of rat GR is a substrate for GSK-3 in vitro.
Two-dimensional phosphopeptide mapping was used to identify specific residues that are phosphorylated by GSK-3 in vitro based on the known positions of each GR phosphopeptide when digested with V8 (Glu-C) protease ( Fig. 2A). The GSK-3-phosphorylated GST-GR 106 -318 fusion protein was digested with V8 protease and the resulting peptides were resolved by electrophoresis and chromatography, yielding a single phosphopeptide that corresponds to a GR peptide containing phospho-Thr 171 (Fig. 2B). Since Thr 171 is the only threonine residue phosphorylated in the rat GR, we would expect phosphoamino acid analysis to detect only phosphothreonine if Thr 171 is indeed the target for GSK-3 phosphorylation. As shown in Fig.  2C, phosphothreonine is the only phosphoamino acid present in GR when phosphorylated by GSK-3 in vitro. To further confirm that GR Thr 171 is the primary target for GSK-3 phosphoryla-

FIG. 2. GSK-3 phosphorylates threonine 171 of the rat GR in vitro.
A, a two-dimensional V8 phosphopeptide map of the rat GR isolated from dexamethasone (Dex)-treated rat hepatoma cells metabolically labeled with [ 32 P]orthophosphate. Phosphorylated residues contained within each GR phosphopeptide are indicated. B, GR phosphopeptides generated in vitro by GSK-3 phosphorylation. A GST-GR 106 -318 fusion protein was phosphorylated by GSK-3 in vitro as described in Fig. 1, the labeled receptor band excised from the gel, and eluted GR was digested with V8 protease. Resulting peptides were resolved by electrophoresis and chromatography, and the phosphorylated peptides were visualized by autoradiography. Note spot 4, corresponding to receptor phosphothreonine-171. C, phosphoamino acid analysis of rat GR phosphorylated by GSK-3 in vitro. A GST-GR 106 -318 was phosphorylated by GSK-3 in vitro, the reaction products were fractionated by SDS-PAGE, transferred to Immobilon paper and hydrolyzed in 6 N HCl for 1 h at 110°C. Resulting phosphoamino acids were mixed with the nonlabeled phosphoamino acid standards (phosphoserine, phosphothreonine, and phosphotyrosine) and electrophoretically separated in two dimensions. The plates were stained with ninhydrin to visualize the standards (shown by dashed circles) and autoradiography was performed. Note selective incorporation of 32 P into GR threonine residue. D, GR phosphorylation by GSK-3 in vitro is abolished by T171A mutation. GST-GR, either the wild type (wt) or with an alanine substitution at Thr 171 (T171A), was phosphorylated in vitro by GSK-3. The reaction products were separated by SDS-PAGE, stained with Coomassie Blue (bottom panel) and visualized by autoradiography (top panel). Note the lack of 32 P incorporation into GST-GR T171A mutant, relative to the wt GR. tion in vitro, we have generated a GST-GR 106 -318 fusion protein with a single threonine to alanine substitution at Thr 171 (T171A) and subjected it and the wild type (wt) GR to an in vitro kinase assay with GSK-3. Fig. 2D demonstrates that the T171A amino acid substitution abolished GR phosphorylation. Taken together, these finding indicate that Thr 171 is the site of phosphorylation by GSK-3 in the rat GR.

Ectopic Expression of GSK-3 Inhibits GR Transcriptional Activation, but Not Repression, in Cultured Mammalian
Cells-Since GSK-3 is expressed ubiquitously and is constitutively active in virtually all mammalian cell lines, we examined whether GSK-3 overexpression alters GR-dependent transcriptional activation. We transiently co-expressed GSK-3, rat GR, and a GR-responsive reporter plasmid containing two consensus GREs upstream of the luciferase gene, in U-2 OS human osteosarcoma cells, which lack endogenous GR. Fig. 3A demonstrates that increasing amounts of transfected GSK-3 inhibit GR hormone-dependent transcriptional enhancement in a dosedependent manner, with the maximal dose of GSK-3 used reducing GR transcriptional activity by over 50% compared with control cells transfected with the vector alone. This effect likely represents an underestimate of the total impact of GSK-3 on GR activity since the results are obtained in a cell line expressing endogenous GSK-3. The reduction in GR activity was not due to the inhibition of GR protein expression, since the steady-state level of GR was not affected by the GSK-3 overexpression (Fig. 3B). Inhibition of GR activity by ectopically expressed GSK-3 was also observed in rat PC-12 cells (Fig. 3C), which express GR endogenously, suggesting that the effect of GSK-3 on GR transcriptional activity extends to multiple cell types.
Since hormone-activated GR can also repress transcription (2, 38, 39), we examined the effect of GSK-3 on GR-mediated transcriptional repression of AP1-dependent transcription by transiently co-expressing GSK-3, the rat GR, and a reporter gene containing a single AP1 site upstream of the luciferase gene in U-2 OS cells. In the absence of GR activation by the glucocorticoid dexamethasone, AP1 activity was reduced as a result of GSK-3 overexpression (Fig. 3D), consistent with the inhibitory phosphorylation of c-Jun by GSK-3 (20,23). In contrast, no significant changes in either wt (Fig. 3D) or T171A (not shown) GR-mediated repression of AP1 activity in the presence of dexamethasone was observed upon GSK-3 overexpression, suggesting that phosphorylation of Thr 171 by GSK-3 does not alter GR's ability to repress transcription.
On the basis of our in vitro findings indicating that Thr 171 is the site of GSK-3 phosphorylation in rat GR, we anticipate that GSK-3-mediated reduction in GR transcriptional activity should be dependent on the presence of Thr 171 target site. To examine this hypothesis, we transiently transfected either the wt GR or GR T171A mutant into U-2 OS cells and assessed the effects of GSK-3 overexpression on GR transcriptional enhancement. As shown in Fig. 4A, the T171A substitution relieved the inhibitory effect of GSK-3 on GR activity. Furthermore, the T171A mutation increased GR transcriptional enhancement over the wt GR by 26 and 40% in the absence and

FIG. 3. Overexpression of GSK-3 inhibits rat GR transcriptional enhancement but not repression in vivo.
A, overexpression of GSK-3 inhibits GR transcriptional activity in vivo in a dose-dependent manner. U-2 OS human osteosarcoma cells were transiently transfected via calcium phosphate precipitation method with the full-length rat GR (pCMV-GR, 0.75 g per 60-mm dish), an XG 46 TL reporter plasmid containing two consensus GREs upstream of a luciferase gene (2 g per 60-mm dish), a pCMV-LacZ plasmid (0.5 g per 60-mm dish), and indicated amounts of pCMV5-HA-GSK-3 plasmid. The total amount of DNA transfected per dish was equalized with a pCMV5 empty expression vector. Receptor transcriptional activity in the presence of glucocorticoid dexamethasone (Dex) was measured via luciferase assay 8 h after the addition of dexamethasone to the medium, normalized to ␤-galactosidase activity and expressed as relative luminescence units (RLU). B, GSK-3 overexpression does not alter the level of the receptor protein. Whole cell extracts were prepared as described under "Experimental Procedures" from a set of dishes transfected under identical conditions. The expression of GR and GSK-3 was analyzed by immunoblotting. C, GSK-3 inhibits the transcriptional activity of endogenous GR in rat PC-12 cells. PC-12 rat pheochromocytoma cells were transiently transfected with an XG 46 TL reporter plasmid (2 g per 60-mm dish) and indicated amounts of pCMV5-HA-GSK-3 plasmid (or an empty pCMV5 expression vector) using LipofectAMINE reagent as described under "Experimental Procedures." Transcriptional activity of the endogenous GR in the presence of dexamethasone was measured via luciferase assay, normalized to the total protein concentration and expressed as RLU. D, GSK-3 does not affect GR repression of AP1 transcriptional activity. U-2 OS cells were transiently transfected with the full length rat GR (pCMV-GR, 0.75 g per 60-mm dish), a reporter plasmid containing a single consensus AP1 site upstream of a luciferase gene (XAP1TL, 2 g per 60-mm dish), a pCMV-LacZ plasmid (0.5 g per 60-mm dish), and 4 g of pCMV5-HA-GSK-3 or the empty pCMV5 expression vector and transcriptional activity was measured as above.
presence of co-transfected GSK-3, respectively, consistent with the inhibitory effect of GSK-3 on rat GR transcriptional activation and reflecting the loss of inhibition by endogenous GSK-3 (Fig. 4A). The steady state level of GR protein was unaffected for both the mutant and wt GR in both the presence and absence of ectopic GSK-3 expression (Fig. 4B). Thus, rat GR is phosphorylated by GSK-3 at Thr 171 in vitro, and mutation of this site to alanine eliminates the effect of GSK-3 overexpression on rat GR transcriptional activation in cultured mammalian cells.
Akt, an Inhibitor of GSK-3, Increases GR Transcriptional Enhancement-We next asked whether a decrease in GSK-3 activity would increase GR transcriptional activation. Since no GSK-3-deficient cell line is available, we chose to inhibit GSK-3 activity by using Akt, a protein kinase that phosphorylates GSK-3 and inhibits its catalytic activity. We overexpressed a constitutively active membrane-targeted myristylated form of Akt (28) in U-2 OS cells and measured GR-dependent transcriptional activation. GR transcriptional enhancement in the presence of ectopically expressed Akt was increased to nearly 300% compared with that of control cells receiving an empty expression vector (Fig. 5A). Akt also increased transcriptional activation of endogenous GR in rat PC-12 cells, and the rat GR introduced into GR-negative SAOS2 human osteosarcoma cells (data not shown), suggesting that this effect is not confined to a single cell type. Thus, activation of Akt increases GR transcriptional activation.
Our hypothesis that Akt acts by phosphorylating and inactivating GSK-3, and thus, relieving the inhibitory effect of GSK-3 on GR transcriptional enhancement, predicts that the effect of Akt should also be dependent on Thr 171 , the target of GSK-3 phosphorylation. Fig. 5A demonstrates that transcriptional enhancement by the rat GR T171A mutant was largely insensitive to Akt overexpression, suggesting that Akt's affect on GR transcriptional activation is accomplished principally through inhibition of GSK-3-mediated phosphorylation of Thr 171 . There is, however, a small (ϳ25%) effect of Akt on GR transcriptional activity that is independent of Thr 171 , which suggests that Akt may also be modifying accessory factors involved in GR transcriptional regulation. The Akt-mediated increase of GR transcriptional activity was not a result of the changes in steady-state GR protein levels (Fig. 5B). Thus, overexpression of Akt increases GR transcriptional activation, in a Thr 171 -dependent manner, suggesting that Akt's effect is accomplished primarily through inhibition of GSK-3.
Inhibition of GR Transcriptional Activity by GSK-3 Is Spe-

FIG. 4. Reduction of GR transcriptional enhancement by GSK-3 is dependent on receptor Thr 171 .
A, T171A amino acid substitution abolishes the GSK-3-mediated inhibition of GR transcriptional activation. U-2 OS cells were transfected as in Fig. 3 with pCMV-GR (wt GR or GR T171A mutant), an XG 46 TL reporter plasmid and a pCMV-HA-GSK-3 (4 g per 60-mm dish), and GR transcriptional activity in the presence and absence of dexamethasone (Dex) was assessed. Luciferase activity in each sample was normalized to the total protein concentration and expressed as relative luminescence units (RLU). B, GR expression level is not affected by GSK-3 overexpression or T171A mutation. Whole cell extracts were prepared from transfected cells as described under "Experimental Procedures" and the expression of GR and GSK-3 was analyzed by Western blotting.

FIG. 5. Inhibition of GSK-3 by Akt prevents GR Thr 171 phosphorylation and rescues GR transcriptional activity.
A, overexpression of Akt increases GR transcriptional enhancement in a Thr 171dependent manner. GR-negative U-2 OS cells were transfected with pCMV-GR (wt or T171A), a GSK-3-expressing vector (pCMV5-HA-GSK-3, 0.5 g per 60-mm dish), an XG 46 TL luciferase-linked reporter plasmid to monitor receptor transcriptional activity, and either pCMV6-HA-Akt (0.75 g per 60-mm dish), or an empty pCMV5 expression vector, as indicated. Cells were treated with 100 nM dexamethasone (Dex) for 8 h, where indicated, harvested, the luciferase activity was measured, normalized to the total protein concentration in each sample and expressed as relative luminescence units (RLU). The experiment was performed in duplicates five times with similar results. B, the level of GR protein is not altered by the T171A mutation or ectopic expression of Akt. Cells were transfected and treated exactly as above, and the whole cell extracts were analyzed by immunoblotting for the expression of myristylated HA-tagged Akt (anti-HA mouse monoclonal antibodies, top panel) and GR (GR-specific rabbit polyclonal antiserum, bottom panel).
cies-restricted-The analysis of GRs isolated and sequenced from different species demonstrates that the majority of GR phosphorylation sites are evolutionarily conserved. For example, serine residues 224, 232, and 246 in the rat GR are conserved among human, primates, mouse, guinea pig and Xenopus receptors (40 -44). In contrast, Thr 171 is conserved between rat, mouse, and guinea pig GR, but corresponds to an alanine (Ala 150 ) residue in the hGR, suggesting that hGR may be insensitive to signaling by GSK-3. To examine this possibility, we tested the effects of GSK-3 overexpression on the transcriptional activation of endogenous hGR in HeLa cells, and in U-2 OS cells, where hGR was introduced ectopically. Transient overexpression of GSK-3 failed to inhibit hGR transcriptional activation in HeLa cells (Fig. 6A). Since our transfection studies with the rat GR were performed in U-2 OS cells, we had to eliminate the possibility that these cells contain a specific co-factor necessary for GSK-3 action on GR. Fig. 6B demonstrates that in U-2 OS cells, no decrease in the hGR transcriptional activity was observed upon GSK-3 overexpression. Thus, under conditions identical to those used for the experiments with the rat GR (Fig. 3A), GSK-3 did not inhibit transcriptional enhancement by hGR, revealing species-specific differences in GR signaling.
We next asked whether replacing the alanine residue at position 150 with a threonine (A150T) in the hGR would confer sensitivity to GSK-3. To test this hypothesis, we have constructed a hGR A150T mutant and assessed the effects of GSK-3 overexpression on the transcriptional activation by this hGR mutant in U-2 OS cells. As shown in Fig. 6C, the transcriptional activity of the hGR A150T mutant decreases as a function of increasing amounts of ectopically expressed GSK-3, whereas no such inhibition was noted with wt hGR. Thus, introduction of a threonine residue into hGR at a position equivalent to Thr 171 in the rat GR, results in a GSK-3-dependent inhibition of the hGR transcriptional activation. DISCUSSION We have demonstrated that GSK-3 phosphorylates the rat GR at Thr 171 in vitro. In cultured mammalian cells, overexpression of GSK-3 inhibits GR transcriptional activation, while decreasing GSK-3 activity, through expression of the GSK-3 inhibitor, Akt, increases GR transcriptional enhancement. GRmediated repression of AP1-dependent transcriptional activity, however, was not affected by GSK-3 overexpression. A threonine to alanine mutation at Thr 171 , the site of rat GR phosphorylation by GSK-3 in vitro, eliminates the effect of GSK-3 on rat GR transcriptional enhancement. Although the effect of GSK-3 overexpression on GR-mediated transcriptional activation was relatively modest (ϳ50%), this likely represents an underestimate of the impact of GSK-3 on GR, since the studies were performed in cell lines containing active endogenous GSK-3. Our in vitro phosphorylation and mapping studies, coupled with activity assays using GR mutants, strongly suggest that GSK-3 phosphorylates rat GR at Thr 171 , and as a consequence, reduces GR transcriptional activation.
The mechanism by which GSK-3 phosphorylation of Thr 171 decreases GR transcriptional activity is unclear. GSK-3 phosphorylates and inactivates other regulatory factors including c-Myc, c-Jun, NF-ATc, and ␤-catenin. Although the mechanism of c-Myc and c-Jun inactivation by GSK-3 phosphorylation is unknown, GSK-3 phosphorylation of ␤-catenin targets it for degradation (45,46), while GSK-3 phosphorylation of NF-ATc promotes its export from the nucleus (47). It is doubtful, however, that either of these established mechanisms explain GSK-3 regulation of GR, since 1) neither GSK-3 nor Akt overexpression alter steady state GR protein levels, and 2) increased export of GR from the nucleus would also affect GR-dependent transcriptional repression, which has not been observed in our experiments. Alternatively, GSK-3-mediated phosphorylation of rat GR at Thr 171 may disrupt protein-protein interactions that favor GR transcriptional enhancement, or recruit inhibitory proteins that antagonize GR-dependent transcriptional activation, hypotheses that are currently being tested.
Given the high degree of conservation between GRs from FIG. 6. GSK-3 overexpression does not inhibit transcriptional response of the human GR. HeLa (A) and U-2 OS (B and C) cells were transfected with indicated amounts of pCMV5-HA-GSK-3 or an empty pCMV5 vector using LipofectAMINE reagent or calcium phosphate precipitation, respectively, as described under "Experimental Procedures." In addition, each 60-mm dish of HeLa cells received 2 g of XT 46 TL reporter, whereas each 60-mm dish of U-2 OS cells received 0.75 g of pcDNA3-hGR (B) or pcDNA3-hGR A150T (C) expression plasmid and 2 g of XT 46 TL reporter. Glucocorticoid-dependent transcriptional activity of the human GR was measured by luciferase assay, normalized to total protein concentration and expressed as relative luminescence units (RLU). different species, it is particularly striking that the hGR does not contain a site of GSK-3-mediated phosphorylation, thereby making hGR insensitive to GSK-3 overexpression. However, when an alanine residue at the position homologous to Thr 171 in rat GR is replaced with a threonine, transcriptional activation by the hGR A150T mutant becomes sensitive to GSK-3 overexpression. Sequence comparison between GRs isolated from different species shows that the primary amino acid sequence surrounding and including rat GR Thr 171 (residues 164 through 173) is conserved among rodents, including rat, mouse, and guinea pig (40,41,43). The equivalent region from human, squirrel monkey, owl monkey, and cotton-top tamarin GR remains conserved among primates, but has diverged from rodents (42,48,49). Why this region of GR has diverged between primates and rodents, while the other major phosphorylation sites (Ser 224 , Ser 232 , and Ser 246 ) are conserved remains unclear, but likely reflects alternative strategies adopted by each species to regulate GR action.
The differences in GR primary amino acid structure and signaling between rodents and humans may contribute to the greater sensitivity of murine lymphocytes to glucocorticoidinduced apoptosis relative to human cells. It is conceivable that GSK-3-mediated inhibition of GR transcriptional activation in rodents results in the reduced expression of a putative survival factor induced by GR. Recently, the cyclin-dependent kinase inhibitor p21 Cip1 has been shown to be a GR-responsive gene (49 -51) and forced expression of p21 can block apoptosis (52). Thus, p21 expression protects cells from apoptosis, and as such, can be considered a survival factor. It is tempting to speculate that inhibition of GR by GSK-3, and the subsequent lack of p21 induction, may facilitate apoptosis in murine but not human lymphocytes. It would be interesting to replace mouse GR with that of the human GR in vivo and examine whether the glucocorticoid-induced apoptosis of murine lymphocytes expressing hGR still occurs. We speculate further that a threonine at position 150 in hGR would result in greater glucocorticoid sensitivity compared with an alanine at this position.
Our findings demonstrate species-specific differences in human and rat GR signaling, which suggest that studies on GR function in mice and rats may not fully translate into hGR activity. In addition, our results indicate that hGR is capable of adopting the GSK-3 signaling pathway through a somatic mutation, which antagonizes hGR-dependent transcriptional activation. It would be informative to examine whether alanine to threonine substitutions at residue 150 in hGR are present in glucocorticoid-sensitive, but absent in glucocorticoid-resistant, malignancies.