0013-7227/02/$15.00/0 Printed in U.S.A.

Endocrinology 143(9):3482–3489 Copyright © 2002 by The Endocrine Society doi: 10.1210/en.2002-220135

Potentiation of Glucocorticoid Receptor Transcriptional Activity by Sumoylation YVES LE DREAN, NATHALIE MINCHENEAU, PASCALE LE GOFF,

AND

DENIS MICHEL

Information et Programmation Cellulaire, Unite´ Mixte de Recherche 6026 Centre National de la Recherche Scientifique, University of Rennes I, Rennes 35 042, France The glucocorticoid receptor (GR) is a transcription factor, subject to several types of posttranslational modifications including phosphorylation and ubiquitination. We showed that the GR is covalently modified by the small ubiquitin-related modifier-1 (SUMO-1) peptide in mammalian cells. We demonstrated that GR sumoylation is not dependent on the presence of the ligand and regulates the stability of the protein as well as its transcriptional activity. SUMO-1 overexpression induces dramatic GR degradation, abolished by proteasome inhibition. We also found that SUMO-1 stimulates the transactivation capacity of GRs to an extent largely exceeding those observed so far for other sumoylated transcription factors.

G

LUCOCORTICOID HORMONES ARE involved in the basal and stress-related regulation of carbohydrate, protein, and fat metabolism and play key roles in the modulation of immune and inflammatory responses (1). Most of these glucocorticoid hormone effects are mediated by a specific intracellular receptor protein from the nuclear receptor superfamily: the glucocorticoid receptor (GR). The GR is a ubiquitously expressed transcription factor that presents a modular structure consisting of a DNA-binding domain, a ligand-binding domain, and two independent activating functions in the N- and the C-terminal parts of the protein, respectively (2). In the absence of glucocorticoid hormones, the GR is retained in the cytoplasm in an inactive state in association with heat shock proteins. Upon ligand binding, the heat shock protein/GR complex dissociates and the receptor translocates into the nucleus (2). The nuclear action of GR can be divided in two ways: The GR can exert its action via protein-protein interaction with other transcription factors such as activator protein-1 or nuclear factor ␬B, or the GR directly binds to specific DNA sequences termed glucocorticoid-response element (GRE) in the regulatory region of target genes (3, 4). In that context, GRs can bind to one or several GREs and transcription is activated by interaction of the receptor with coregulators or with the basal transcriptional machinery (5, 6). If ligand binding is the main inducer of the GR transcriptional activity, posttranslational modifications such as phosphorylation have also been shown to play an important role in GR regulation (7). We report here Abbreviations: AR, Androgen receptor; CMV, cytomegalovirus; Dex, dexamethasone; EGFP, enhanced green fluorescent protein; ER, estrogen receptor; GR, glucocorticoid receptor; GRE, glucocorticoid-response element; HSF-1, heat shock factor 1; MMTV, mouse mammary tumor virus; PIAS, protein inhibitor of activated STAT; rtGR, rainbow trout glucocorticoid receptor; SC, synergy control; SCF, synergy control factor; SUMO-1, small ubiquitin-related modifier-1.

Overexpression of SUMO-1 specifically enhances the ligandinduced transactivation of GR up to 8-fold. However, this hyperactivation occurs only in the context of a synergy between multiple molecules of GRs. It requires more than one receptor DNA-binding site in promoter and becomes more prominent as the number of sites increases. Interestingly, these observations may be related to the transcriptional properties of the synergy control region of GRs, which precisely contains two evolutionary conserved sumoylation sites. We propose a model in which SUMO-1 regulates the synergy control function of GR and serves as a unique signal for activation and destruction. (Endocrinology 143: 3482–3489, 2002)

that GR function is dramatically regulated by another posttranslational mechanism: the covalent addition of small ubiquitin-related modifier-1 (SUMO-1) peptide. SUMO-1, also known as PIC-1, sentrin, or GMP-1, is a ubiquitin-like conjugation peptide (8) with only 18% identity to ubiquitin but with a remarkably similar secondary structure. SUMO-1 differs from ubiquitin in its surface-charge distribution, which can explain its specificity (9). Like ubiquitin, SUMO-1 was found covalently conjugated to various cellular proteins, but contrary to ubiquitin, which targets its protein substrates to the proteasome, SUMO-1 does not signal proteolysis. Instead, sumoylation appears to play multiple roles in subcellular protein translocation, nuclear body formation, protein stabilization, and modulation of the transcriptional activity of several transcription factors (10, 11). One third of SUMO-1 cellular substrates are transcription factors, but, depending on the nature of these factors, the effect of SUMO-1 modification is very diverse. SUMO-1 induces a nuclear relocalization of p53, which results in increased transcriptional activation of p53-regulated proapoptotic genes and affects cell survival (12, 13). In the opposite, SUMO-1 conjugation to p73 does not affect its transcriptional activity (14), but SUMO-1 modification attenuates c-jun (15) and androgen receptor transactivation (16). Sumoylation is a dynamic and reversible process, requiring a multiple-step reaction catalyzed by specific enzymes related to, but distinct from, enzymes involved in ubiquitination (17). The specific SUMO-1-conjugating enzyme is called Ubc9 and is necessary for protein substrate recognition and SUMO-1 linkage, which explains why numerous Ubc9 interacting proteins turned out to be true substrates for SUMO-1. However, a large number of proteins shown to interact with Ubc9 in the yeast two-hybrid system have not been found to be sumoylated (18). They may be regulated by

3482

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

Le Drean et al. • SUMO-1 Potentiates GR Action

Ubc9 via a mechanism that does not involve direct covalent modification, as proposed for the androgen receptor (19). The glucocorticoid receptor has been shown to interact with Ubc9 (20) and Ubc9 can catalyze sumoylation of GR in an in vitro assay (16). The purpose of this study was to verify whether such a direct SUMO-1 modification occurs in vivo and is able to alter GR functions. We show here that GR may be modified by SUMO-1 and that conjugation dramatically affects GR stability and GR-mediated transcription in the context of promoters containing multiple GREs. Materials and Methods Plasmid constructions The mouse SUMO-1, SUMO-3, and ubiquitin cDNAs were isolated by RT-PCR and ligated in-frame with a DNA sequence encoding the sequence MDYKDDDDK (FLAG epitope-tag). The inserts were cloned downstream of the cytomegalovirus (CMV) promoter in the pXJ41 plasmid, generating the X-flag-SUMO-1, the X-flag-SUMO-3, or the X-flagubiquitin expression vectors. The human wild-type GR cDNA vector (pCMV-GR), the human androgen receptor (AR), the rat estrogen receptor (ER), and the reporter construct (p(GRE)x4-tk-luc) containing four GREs upstream a thymidine kinase promoter driving luciferase expression were obtained from Dr. Thieulant (University of Rennes I, Rennes, France). The rainbow trout GR (rtGR) expression vector was a gift from Dr. Ducouret (University of Rennes I). The empty mammalian expression vector pCDNA3 was purchased from Invitrogen (Paisley, UK). The control vectors encoding the green fluorescent protein (pEGFP-C3) and the ␤-galactosidase (pCMV-␤gal) were from CLONTECH Laboratories, Inc. (Palo Alto, CA). The pEGFP-GR vector containing the wild-type rat GR fused to the green fluorescent protein was a gift from Dr. Lefe`vre from University of Lilles (Lilles, France). The mouse heat shock factor 1 (HSF-1) was a gift from Dr. Morimoto (Northwestern University, Evanston, IL). To construct pEGFP-HSF-1, a cDNA fragment encoding the mouse HSF-1 open reading frame was PCR amplified and cloned between XhoI and EcoRI into the pEGFP-N1 vector.

Cell culture and transfections COS-7 and Chinese hamster ovary-CH3S mammalian cell lines were maintained in DMEM containing 8% fetal bovine serum and 25 U/ml streptomycin and penicillin. For luciferase assay, 3 ⫻ 104 cells were seeded on 24-well plates 24 h before transfection. As described previously (21), a total of 1 ␮g plasmid DNA was cotransfected using the calcium-phosphate/DNA coprecipitation technique. Cells were treated with steroid ligands [dexamethasone (Dex), dihydrotestosterone, estradiol, purchased from Sigma, St. Louis, MO] or with vehicle (ethanol) and harvested after 24 h of transient expression. The ␤-galactosidase and the luciferase activities were measured as previously described (21).

Western blotting A total of 5 ⫻ 105 cells were spread into 60-mm dishes and were transfected with a total of 5 ␮g DNA as described above. In addition to the expression vectors, all transfections contained Bluescript as a carrier and 500 ng ␤-galactosidase expression vector as internal control of transfection efficiency. Cells were treated or not with Dex for 48 h and/or with MG132 (Z-Leu-Leu-Leu-CHO; BIOMOL, Plymouth Meeting, PA) for the last 24 h of the culture. After 48 h of transient expression, cells were collected. Ten percent of cell culture were used for ␤-galactosidase assay, and the rest of the cells were lysed in a denaturing buffer (60 mm Tris-HCl, pH 6.8; 25% glycerol; 2% sodium dodecyl sulfate; 0.01% bromophenol blue; 5% ␤-mercaptoethanol). Proteins from total cell lysate were fractionated by electrophoresis on 7.5% polyacrylamide gel and transferred to a nitrocellulose membrane (Pharmacia Biotech Europe GmbH, Saclay, France). Western blotting was performed according to standard procedures. Primary antibodies [polyclonal anti-GFP and polyclonal anti-SUMO-1 purchased from Abcam (Cambridge, UK) and Tebu International (Le-Perray-en-Yvelines, France) respectively] were revealed by horseradish peroxidase-conjugated antirabbit IgG (Phar-

Endocrinology, September 2002, 143(9):3482–3489 3483

macia Biotech Europe GmbH) as secondary antibodies, followed by enhanced chemiluminescence (Pharmacia Biotech Europe GmbH) detection as recommended by the manufacturer’s instructions.

Results Sumoylation of GR occurs in vivo and induces downregulation of GR protein

To determine whether GRs can be conjugated to the SUMO-1 peptide in a cellular context, we cotransfected COS-7 cells with pEGFP-GR fusion-protein expression vector and an increasing quantity of SUMO-1 expression vector. As presented in Fig. 1A, Western blotting failed to reveal any high molecular SUMO-1-modified enhanced green fluorescent protein (EGFP)-GR but clearly showed that the level of EGFP-GR decreases in proportion with overexpressed SUMO-1 (compare lane 1 with lanes 2– 4, Fig. 1A). Because SUMO-1 is known to control the subcellular distribution of certain protein substrates (22), care was taken to use wholecell extracts to perform our Western blotting in denaturing buffer. This avoids to misinterpret as a degradation, the passage of EGFP-GR from soluble to insoluble fraction, which can lead to an apparent decrease in soluble fraction. Moreover, we found that the SUMO-1-dependent decrease of protein level is specific of GR because the same experiment performed with EGFP-HSF-1 expression vector (lanes 5– 8, Fig. 1A) or with EGFP alone (lanes 9 –12, Fig. 1A) showed that the cellular concentration of these proteins is not affected when augmenting cellular SUMO-1 levels. Furthermore, a slowly migrating form of EGFP-HSF-1 corresponding to the already described SUMO-1-modification form of HSF-1 (23) is observed when SUMO-1 is overexpressed (upper migrating band, lanes 6 – 8, Fig. 1A), indicating the SUMO-1 conjugation can be detected in our experimental conditions. Because protein stability is mainly regulated by the ubiquitin/proteasome pathway, we next tested the possible involvement of this system in the SUMO-1-induced GR instability. Transfections were performed into COS-7 and CHO-CH3S cells with EGFP-GR and SUMO-1 expression vectors, in presence or absence of the proteasome inhibitor MG132. Figure 1B shows that intensity of the EGFP-GR band increased on MG132 treatment in both cell lines. The ligand and the SUMO-1 dependence of GR degradation were obvious in COS-7 cells, whereas in CHOCH3S, MG132 led to a stronger effect on GR stabilization, allowing the detection of a new and minor slow migrating EGFP-GR band. Data presented in Fig. 2 clearly show that presence of this unique EGFP-GR upper band was mainly dependent on the overexpression of SUMO-1 but not on the presence of MG132 alone (compare lanes 1 and 2 with lanes 3 and 4, Fig. 2). This high-molecular-weight band is consistent with the formation of a Flag-SUMO-1conjugated form of EGFP-GR, which is expected to yield a unique complex and left out the possibility of ubiquitinated GR, which should result in formation of characteristic laddering pattern of high-molecular-weight species, containing increasing numbers of ubiquitin moieties. Identity of this complex was analyzed by Western blot, using anti-GFP and anti-SUMO-1 antibodies (Fig. 2, upper and lower panel, respectively). The anti-GFP antibodies allowed detection of two bands: the major 120-kDa unmod-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

3484

Endocrinology, September 2002, 143(9):3482–3489

Le Drean et al. • SUMO-1 Potentiates GR Action

FIG. 2. GR is covalently modified by SUMO-1 in vivo. CHO-CH3S cells were transfected with 2 ␮g pEGFP-GR and 500 ng pCMV-␤gal in association with 2 ␮g SUMO-1 expression vector (lanes 3 and 4) or its empty vector (lanes 1 and 2). Cells were treated for 48 h with either 0.1 ␮M Dex (lanes 2 and 4) or the vehicle (lanes 1 and 3). For the last 12 h of transient expression, cells were treated with the inhibitor of the proteasome MG132 (5 ␮M). After treatment, whole-cell extracts were prepared and normalized for ␤-galactosidase, and protein levels were determined by Western analysis using anti-GFP antibodies (upper panel) or anti-SUMO-1 antibodies (lower panel). The fusion protein EGFP-GR is indicated by an arrow and its corresponding conjugated form by an arrowhead.

is subject to SUMO-1 modification, which can subsequently affect its stability. Sumoylation of GR potentiates its transcriptional activity

FIG. 1. Overexpression of SUMO-1 induces down-regulation of GR protein. A, COS-7 cells were transfected with 1 ␮g pEGFP-GR fusionprotein expression vector (lanes 1– 4), 1 ␮g pEGFP-HSF-1 expression vector (lanes 5– 8), or 1 ␮g pEGFP-C3 vector (lanes 9 –12) in presence of pCMV-␤gal plasmid as internal control and an increasing quantity of SUMO-1 expression vector (0 ␮g: lanes 1, 5, and 9; 0.3 ␮g: lanes 2, 6, and 10; 1 ␮g: lanes 3, 7, and 11; 3 ␮g: lanes 4, 8, and 12). Whole-cell extracts were prepared and aliquots normalized for ␤-galactosidase were resolved by SDS-PAGE and immunoblotted with anti-GFP antibodies. B, COS-7 or CHO-CH3S cells were transfected as described above. As indicated, cells were treated or not with 5 ␮M MG132 and whole-cell extracts were normalized for ␤-galactosidase, resolved by SDS-PAGE and immunoblotted with anti-GFP antibodies. The fusion protein EGFP-GR is indicated by an arrow and its corresponding conjugated form by an arrowhead.

ified EGFP-GR fusion protein as well as the slowly migrating 150-kDa band. Of these two major GFP-antibodyreacting bands, only the high 150-kDa protein was also recognized by SUMO-1 specific antibody (Fig. 2, lanes 3 and 4, lower panel). Altogether, our results indicate that GR

To determine the role of SUMO-1 conjugation on GR transcriptional activity, additional transfection assays were performed using a glucocorticoid responsive reporter gene (p(GRE)x4-tk-Luc) and GR in conjunction or not with a SUMO-1 expression vector. In the presence of GR alone, treatment of cells with Dex enhances the glucocorticoid responsive reporter gene activity up to a 5-fold (compare histograms 5 and 7, Fig. 3A). The presence of SUMO-1 further increased in a dose-dependent manner the GR-mediated transactivation (Fig. 3A, histograms 8 –12). Classically, transactivation by GR requires the presence of ligand; interestingly, we found that the dramatic enhancement of GR transactivation by SUMO-1 also needs the GR stimulation by Dex. As shown by our control experiments, in absence of ligand, activation of the promoter activity by SUMO-1 was weak and not significant (Fig. 3A, histograms 5 and 6). Furthermore, the action of SUMO-1 required the presence of a GRE within the responsive gene promoter (data not shown) as well as the presence of GR into the cells, as indicated by control transfections with the empty pCDNA3 vector, showing that the basal activity of our reporter gene was not affected by the presence of either SUMO-1, SUMO-3, or ubiquitin (Fig. 3A, histograms 1– 4). We then tested the specificity of SUMO-1 for this hyperactivation of GR. Overexpression of another mammalian SUMO isoform (SUMO-3) also induces GR hyperactivity, whereas the same experiment performed with the related protein ubiquitin, which can also conjugate to GR (24), failed to show any activation (Fig. 3A, histograms 13 and 14).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

Le Drean et al. • SUMO-1 Potentiates GR Action

Endocrinology, September 2002, 143(9):3482–3489 3485

FIG. 3. Sumoylation of GR potentiates its transcriptional activity. A, COS-7 cells were transfected with 400 ng p(GRE)x4-TK-Luc reporter plasmid in presence of pCMV␤gal as internal control and 100 ng GR expression vector (histograms 5–14) or the empty vector pCDNA3 (histograms 1– 4). Then 200 ng X-flag empty vector (black histograms), X-flag-SUMO-1 (hatched histograms), X-flagSUMO-3 (gray histograms), or X-flag-ubiquitin (dotted histograms) expression vectors were added to the plasmid mixture. The SUMO-1 dose effect was performed with 0.5, 2, 10, 50, and 200 ng X-flag-SUMO-1 expression vector (histograms 8 –12, respectively). Cells were treated with 10⫺7 M Dex for 24 h before being assayed for luciferase activity. B, Cells were transfected as previously with an androgen-responsive reporter plasmid (p(ARE)x4-TK-Luc) or an estradiol-responsive reporter plasmid (pERE-TKLuc) in association with an AR or an ER expression vector as indicated. Cells were treated for 24 h with 10⫺7 M dihydroxytestosterone, 10⫺7 M estradiol, or vehicle. The values are an average ⫾ SEM of three to eight independent experiments.

Transactivation assays using reporter plasmids responsive to other steroid receptors were done to examine whether the dramatic effect of SUMO-1 on GR function could be generalized to other related transcription factors. The AR was previously described as a target of SUMO-1; Poukka et al. (16) showed that mutation of the target lysine of AR sumoylation abolished SUMO-1 addition and modestly enhanced transactivation, compared with wild-type AR. Intriguingly, we failed to obtain any effect of SUMO-1 or SUMO-3 overexpression on AR activity (Fig. 3B, histograms 1– 6), and overexpression of ubiquitin inhibited the androgen transcriptional induction, which was consistent with a degradation of AR through the ubiquitin/proteasome pathway (25). Besides, we also found that SUMO-1 had no effect on the ER transactivation of a pERE-TK-Luc reporter plasmid (Fig. 3B, histograms 7–10). Together, our data demonstrate SUMO-1 has a strong and specific effect on GR transcriptional activity. SUMO-1 hyperactivation of GR depends on the promoter context

SUMO-1 conjugation occurs at the level of lysine residues in the context of recognizable consensus sequences (⌽KxE,

where ⌽ is an aliphatic residue) (26). GR contains three such sumoylation consensus sequences. Two of them are located in the N-terminal domain of GR and are identical with the AR sumoylated sites (16). Intriguingly, this motif was independently identified as a region important for GR transactivation (27). This region, termed synergy control (SC) motif, enhances the ability of multiple DNA-bound molecules of GR to activate transcription in a synergistic manner. The function of this SC motif thus depends on the promoter context. It requires more than one binding site but not cooperative DNA binding, which is the case in the context of our glucocorticoid responsive reporter gene p(GRE)x4-TK-Luc (27). To determine whether SUMO-1 activation of GR is somehow related to the SC motif, we performed a transfection assay with either a reporter gene containing four consensus GREs (p(GRE)x4-TK-Luc) or a reporter gene containing the mouse mammary tumor virus (MMTV) promoter, in which the GR activity was shown independent of the SC motif (27). Using the MMTV-Luc reporter plasmid, GR alone (in absence of SUMO-1), triggered an up to 10-fold ligand induction of transcription. More interestingly, in this promoter context, overexpression of SUMO-1 had no effect on GR transacti-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

3486

Endocrinology, September 2002, 143(9):3482–3489

vation (Fig. 4A, compare histograms 10 and 11), but the hyperactivation of GR induced by SUMO-1 was observed in the context of a p(GRE)x4-TK-Luc promoter (Fig. 4A, histograms 2– 4). Transfection assays performed with other reporter plasmids containing one or two GREs (Fig. 4B), clearly showed that hyperactivation of the ligand-activated GR by SUMO-1 depended on the number of GR-binding sites. It required more than one binding site per promoter and became more prominent as the number of sites was increased. To further assess the biological significance of this powerful regulation of GR by SUMO-1, we tested concurrently the action of SUMO-1 on the rtGR because the putative SUMO-1-target/SC site is also found conserved in the GR from this distant species (Fig. 5A). Data presented in Fig. 4A (histograms 5–7) show that SUMO-1 also potentiated the activation of rtGR, indicating that the effect of SUMO-1 on GR was conserved during evolution, suggesting an important biological function. Together, these results indicate that sumoylation is an important mechanism determining GR activity. It is conserved through vertebrate evolution, specific of GR nuclear receptor, and can regulate the induction of a subset of GR target genes containing multiple GREs.

Le Drean et al. • SUMO-1 Potentiates GR Action

Discussion

In the present study, we show physical and functional interactions between the steroid receptor GR and the conjugation peptide SUMO-1 in intact cells. The GR was already found to bind to Ubc9 (20), the specific SUMO-conjugating enzyme. Furthermore, Poukka et al. (16) demonstrated, in an in vitro assay, that Ubc9 can catalyze the sumoylation of GR, but no direct evidence of such a GR modification in a cellular context was so far presented. By transfection, we were able to detect, on SUMO-1 overexpression, a high-molecularweight form of GR. The steady-state amount of SUMO-1 in cells was shown to be limiting so that de novo sumoylation requires SUMO-1 recycling after desumoylation of other substrates (28). This situation implies that the intracellular concentration of SUMO-1 must be increased to reveal sumoylation events that would have been otherwise undetectable. After SUMO-1 overexpression, the steady-state sumoylated GR represents a minority form in GR total population (about 7%) but leads to important changes of GR properties. Conjugated forms of proteins are considered as very transient structures resulting from a dynamic equilibrium between the opposite activities of conjugating and deconjugating enzymes (17). Protein substrates are conjugated during a short

FIG. 4. SUMO-1 activation of GR depends on the promoter context. A, Cells were transfected as described in Materials and Methods. Two different glucocorticoid-responsive reporter plasmids (p(GRE)x4-TK-Luc and MMTV-Luc) were used as indicated. Cells were cotransfected with a human or a trout GR expression vector in association or not with the SUMO-1 expression vector. Cells were treated 24 h with Dex (10⫺7 M) or vehicle and harvested for luciferase analysis. Histograms present means ⫾ SEM of four to six independents experiments. B, Several glucocorticoid responsive reporter genes were used to determine by transfection the fold of ligand induction triggered with GR alone or the level of hyperactivation induced by overexpression of SUMO-1 in conjunction with the ligand-activated GR. Data present the means ⫾ SEM of different ratios: luciferase activity obtained in presence of Dex (10⫺7 M) divided by luciferase activity obtained in absence of ligand or luciferase activity obtained in presence of Dex and SUMO-1 divided by luciferase activity obtained in presence of ligand and absence of SUMO-1.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

Le Drean et al. • SUMO-1 Potentiates GR Action

FIG. 5. Model of SUMO-1 action on GR. A, Alignment of human, rat, and trout GR residues of the SC region and comparison with the consensus sequence of sumoylation signal. B, Speculative model described by Iniguez-Lluhi and Pearce (27), where the SC motif of GR is negatively regulated by a hypothetical SCF. C, Conjugation of SUMO-1 to the SC motif of the GR might interfere with the SCF binding and permit the multiple DNA-bound molecules of the GR to activate transcription in a synergistic manner. GR is then destroyed by ubiquitin-mediated proteolysis.

period of time, which is nevertheless sufficient to allow their biological activity. As a consequence, although the proportion of conjugated proteins at the very same instant may appear limited, the corresponding biological effect can be obvious. As an example, sumoylation has been shown to induce global stabilization of I␬B, but the cellular fraction of SUMO-modified I␬B appeared barely detectable in cells (29). We observed that sumoylation induces specific instability of GR protein. This degradation is so important that SUMO1-conjugated GR can be detected only in the presence of the protease inhibitor MG132. This observation was surprising because SUMO-1 conjugation has not been frequently reported to target protein degradation as ubiquitin does. On the contrary, SUMO-1 ligation was sometimes described as

Endocrinology, September 2002, 143(9):3482–3489 3487

a stabilizing mechanism by competing with attachment of ubiquitin, as in the case of I␬B (29), in which SUMO-1 and ubiquitin ligation compete for the same lysine residue on protein substrate. Nevertheless, this competition mechanism is not general, and our observation of a SUMO-1-induced protein degradation is not unique because SUMO-1 was also reported to enhance p73 instability (14). The dramatic enhancement of GR transactivation by SUMO-1 also raises interesting questions. Some reports already showed that SUMO-1 can potentiate transcriptional activities of other transcription factors (12, 23), but these effects were generally modest, compared with those presented here. The SUMO-1-driven hyperactivation of GR is linked to its transcriptional properties, it requires presence of the ligand, and seems specific of GR because we found no effect of SUMO-1 overexpression on AR activity. Interestingly, Poukka et al. (16) showed that AR mutants no longer modifiable by SUMO-1 have enhanced transcriptional activity, which suggested that SUMO-1 had a negative effect on AR transactivation. However, we did not use the same ARsensitive reporter gene, and it is possible that action of SUMO-1 on ARs might also depend on the promoter context. Our study is in agreement with very recently reported stimulating influence of SUMO-1 pathway enzymes (Ubc9 and PIAS) on GR activation (30, 31). PIAS, an E3 SUMO-1-ligase enzyme (32), was already described as a coactivator of ARs and GRs but at different degrees, depending on the receptor, cell type, and promoter context (33). This suggests that the biological role of AR and GR sumoylation might be different and may explain the dichotomy that we observed in our system between the action of SUMO-1 on ARs and GRs. Modification by SUMO-1 has already been shown to have quite different biological effects, even on closely related protein substrates. Sumoylation activates p53 transcriptional activity (12), and sumoylation of p73 regulates protein stability and localization but not transcriptional activity (14). Interestingly, two of the three potential sumoylation sites (⌽KxE) in GRs are precisely located in an SC region within the N-terminal domain of the GR (26, 27). This potential SC/sumoylation site was found to be functional for SUMO-1 ligation in the AR and c-Myb context (16, 34). It is worth noting that this SC motif is precisely one of the rare regions of the GR N-terminal domain conserved among vertebrate classes from fish to mammals. The possibility of an implication of the SC motif in SUMO-1 activation is consistent with the observation that the SUMO-1 effect, like that of the SC motif, depends on the promoter context (27). The MMTV promoter is induced by ligand-activated GRs through a mechanism involving cooperative binding of GRs but that does not implicate the SC motif. Similarly, we did not find any effect of SUMO-1 on GR-mediated activity of the MMTV promoter. Moreover, we found that, as described for the SC motif, the SUMO-1-mediated hyperactivation of GR requires more than one GRE. The SC region was shown by IniguezLluhi and Pearce (27) to control the ability of multiple DNAbound molecules of GR to activate transcription in a synergistic manner. The authors proposed that this region restrains GR synergy by interacting with a hypothetical SC factor (SCF). This factor could be recruited by recognizing several SC motifs when several GRs are present in an ap-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

3488

Endocrinology, September 2002, 143(9):3482–3489

propriate context of multiple GREs and may then interfere with an active transcription complex (cf. model presented in Fig. 5B). According to this model, full transcriptional activation by GRs in such promoter contexts requires SCF release, allowing the activation function domains of multiple GRs to act in synergism. Based on our observations, we propose a new model (Fig. 5C), in which ligation of SUMO-1 in the SC motif can prevent the interaction between the GR and SCF, thus allowing GRs to fully activate target genes containing several GREs. Our observations that SUMO-1 can simultaneously enhance both GR degradation and GR transcriptional activation may appear contradictory but can be reconciled by observations from the literature. Several reports (35, 36) have already shown that transcriptional activation and active degradation are closely coupled events. Moreover, it has been shown that the GR undergoes continuous exchange between its DNA-binding elements and the nucleoplasmic compartment (37). It is possible that the GR is targeted for degradation once transcriptionally activated, which provides an efficient way to limit in time and intensity the glucocorticoid action. Interestingly, the activation domain ␶1 present in the N-terminal domain of GRs was described as an acidic region (38), and Salghetti et al. (36) showed that such domains specifically signal ubiquitin-mediated proteolysis of activated transcription factors. It was postulated that transcription factor proteolysis may be a direct consequence of the recruitment of the basal transcriptional machinery to a promoter. Consistently, some coregulators interacting with GRs were described as ubiquitin-ligase proteins or associated with the 26S proteasome (39, 40). In conclusion, our findings suggest that SUMO-1 conjugation plays a pivotal role in regulating GR transcriptional synergy in the context of the subset of target genes containing multiple GREs. Because the rate of SUMO-1 conjugation can be influenced by environmental factors such as cellular stress, the sumoylation of GRs may provide a means to regulate a subgroup of glucocorticoid-responsive genes in response to changes in the environment. It will be important to determine exactly when and where GR sumoylation occurs, how it is regulated, and how it contributes to the physiological glucocorticoid response. Acknowledgments We thank Drs. Ducouret, Thieulant, Lefe`vre, and Morimoto for plasmids. The critical support by Dr. Salbert and Dr. Saulier is also gratefully acknowledged. Received February 4, 2002. Accepted May 21, 2002. Address all correspondence and requests for reprints to: Dr. Yves Le Drean, Information et Programation Cellulaire, UMR CNRS 6026, Bat. 13, Campus de Beaulieu, Universite´ de Rennes I, 35 042 Rennes cedex, France. E-mail: [email protected]. This work was supported by “la re´ gion Bretagne” (op. A1CAA8) and by grants from “Association pour la Recherche contre le Cancer” (ARC 4479) and the “Comite´ Grand Ouest de la Ligue contre le Cancer.”

References 1. Newton R 2000 Molecular mechanisms of glucocorticoid action: what is important? Thorax 55:603– 613 2. Beato M, Herrlich P, Schutz G 1995 Steroid hormone receptors: many actors in search of a plot. Cell 83:851– 857

Le Drean et al. • SUMO-1 Potentiates GR Action

3. Reichardt HM, Schutz G 1998 Glucocorticoid signalling—multiple variations of a common theme. Mol Cell Endocrinol 146:1– 6 4. Adcock IM 2000 Molecular mechanisms of glucocorticosteroid actions. Pulm Pharmacol Ther 13:115–126 5. Glass CK, Rosenfeld MG 2000 The coregulator exchange in transcriptional functions of nuclear receptors. Genes Dev 14:121–141 6. Beato M, Chavez S, Truss M 1996 Transcriptional regulation by steroid hormones. Steroids 61:240 –251 7. Bodwell JE, Webster JC, Jewell CM, Cidlowski JA, Hu JM, Munck A 1998 Glucocorticoid receptor phosphorylation: overview, function and cell cycledependence. J Steroid Biochem Mol Biol 65:91–99 8. Saitoh H, Pu RT, Dasso M 1997 SUMO-1: wrestling with a new ubiquitinrelated modifier. Trends Biochem Sci 22:374 –376 9. Jin C, Shiyanova T, Shen Z, Liao X 2001 Heteronuclear nuclear magnetic resonance assignments, structure and dynamics of SUMO-1, a human ubiquitin-like protein. Int J Biol Macromol 28:227–234 10. Melchior F 2000 SUMO—nonclassical ubiquitin. Annu Rev Cell Dev Biol 16:591– 626 11. Muller S, Hoege C, Pyrowolakis G, Jentsch S 2001 SUMO, ubiquitin’s mysterious cousin. Nat Rev Mol Cell Biol 2:202–210 12. Rodriguez MS, Desterro JM, Lain S, Midgley CA, Lane DP, Hay RT 1999 SUMO-1 modification activates the transcriptional response of p53. EMBO J 18:6455– 6461 13. Fogal V, Gostissa M, Sandy P, Zacchi P, Sternsdorf T, Jensen K, Pandolfi PP, Will H, Schneider C, Del Sal G 2000 Regulation of p53 activity in nuclear bodies by a specific PML isoform. EMBO J 19:6185– 6195 14. Minty A, Dumont X, Kaghad M, Caput D 2000 Covalent modification of p73alpha by SUMO-1. Two-hybrid screening with p73 identifies novel SUMO1-interacting proteins and a SUMO-1 interaction motif. J Biol Chem 275:36316 – 36323 15. Muller S, Berger M, Lehembre F, Seeler JS, Haupt Y, Dejean A 2000 c-Jun and p53 activity is modulated by SUMO-1 modification. J Biol Chem 275: 13321–13329 16. Poukka H, Karvonen U, Janne OA, Palvimo JJ 2000 Covalent modification of the androgen receptor by small ubiquitin-like modifier 1 (SUMO-1). Proc Natl Acad Sci USA 97:14145–14150 17. Schwienhorst I, Johnson ES, Dohmen RJ 2000 SUMO conjugation and deconjugation. Mol Gen Genet 263:771–786 18. Yeh ET, Gong L, Kamitani T 2000 Ubiquitin-like proteins: new wines in new bottles. Gene 248:1–14 19. Poukka H, Aarnisalo P, Karvonen U, Palvimo JJ, Janne OA 1999 Ubc9 interacts with the androgen receptor and activates receptor-dependent transcription. J Biol Chem 274:19441–19446 20. Gottlicher M, Heck S, Doucas V, Wade E, Kullmann M, Cato AC, Evans RM, Herrlich P 1996 Interaction of the Ubc9 human homologue with c-Jun and with the glucocorticoid receptor. Steroids 61:257–262 21. Le Drean Y, Kern L, Pakdel F, Valotaire Y 1995 Rainbow trout estrogen receptor presents an equal specificity but a differential sensitivity for estrogens than human estrogen receptor. Mol Cell Endocrinol 109:27–35 22. Wilson VG, Rangasamy D 2001 Intracellular targeting of proteins by sumoylation. Exp Cell Res 271:57– 65 23. Hong Y, Rogers R, Matunis MJ, Mayhew CN, Goodson M, Park-Sarge OK, Sarge KD 2001 Regulation of heat shock transcription factor 1 by stressinduced SUMO-1 modification. J Biol Chem 276:40263– 40267 24. Wallace AD, Cidlowski JA 2001 Proteasome-mediated glucocorticoid receptor degradation restricts transcriptional signaling by glucocorticoids. J Biol Chem 276:42714 – 42721 25. Sheflin L, Keegan B, Zhang W, Spaulding SW 2000 Inhibiting proteasomes in human HepG2 and LNCaP cells increases endogenous androgen receptor levels. Biochem Biophys Res Commun 276:144 –150 26. Rodriguez MS, Dargemont C, Hay RT 2001 SUMO-1 conjugation in vivo requires both a consensus modification motif and nuclear targeting. J Biol Chem 276:12654 –12659 27. Iniguez-Lluhi JA, Pearce D 2000 A common motif within the negative regulatory regions of multiple factors inhibits their transcriptional synergy. Mol Cell Biol 20:6040 – 6050 28. Saitoh H, Hinchey J 2000 Functional heterogeneity of small ubiquitin-related protein modifiers SUMO-1 versus SUMO-2/3. J Biol Chem 275:6252– 6258 29. Desterro JM, Rodriguez MS, Hay RT 1998 SUMO-1 modification of I␬B␣ inhibits NF-␬B activation. Mol Cell 2:233–239 30. Kotaja N, Vihinen M, Palvimo JJ, Janne OA 2002 Androgen receptor-interacting protein 3 and other PIAS proteins cooperate with glucocorticoid receptor-interacting protein 1 in steroid receptor-dependent signaling. J Biol Chem 277:17781–17788 31. Kaul S, Blackford Jr JA, Cho S, Simons Jr SS 2002 Ubc9 is a novel modulator of the induction properties of glucocorticoid receptors. J Biol Chem 277:12541– 12549 32. Schmidt D, Muller S 2002 Members of the PIAS family act as SUMO ligases for c-Jun and p53 and repress p53 activity. Proc Natl Acad Sci USA 99:2872– 2877 33. Kotaja N, Aittomaki S, Silvennoinen O, Palvimo JJ, Janne OA 2000 ARIP3

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.

Le Drean et al. • SUMO-1 Potentiates GR Action

34. 35. 36. 37.

(androgen receptor-interacting protein 3) and other PIAS (protein inhibitor of activated STAT) proteins differ in their ability to modulate steroid receptordependent transcriptional activation. Mol Endocrinol 14:1986 –2000 Bies J, Markus J, Wolff L 2002 Covalent attachment of the SUMO-1 protein to the negative regulatory domain of the c-Myb transcription factor modifies its stability and transactivation capacity. J Biol Chem 277:8999 –9009 Thomas D, Tyers M 2000 Transcriptional regulation: kamikaze activators. Curr Biol 10:R341–R343 Salghetti SE, Muratani M, Wijnen H, Futcher B, Tansey WP 2000 Functional overlap of sequences that activate transcription and signal ubiquitin-mediated proteolysis. Proc Natl Acad Sci USA 97:3118 –3723 McNally JG, Muller WG, Walker D, Wolford R, Hager GL 2000 The glu-

Endocrinology, September 2002, 143(9):3482–3489 3489

cocorticoid receptor: rapid exchange with regulatory sites in living cells. Science 287:1262–1265 38. Almlof T, Wright AP, Gustafsson JA 1995 Role of acidic and phosphorylated residues in gene activation by the glucocorticoid receptor. J Biol Chem 270: 17535–17540 39. Baumann CT, Ma H, Wolford R, Reyes JC, Maruvada P, Lim C, Yen PM, Stallcup MR, Hager GL 2001 The glucocorticoid receptor interacting protein 1 (GRIP1) localizes in discrete nuclear foci that associate with ND10 bodies and are enriched in components of the 26S proteasome. Mol Endocrinol 15:485–500 40. McKenna NJ, Lanz RB, O’Malley BW 1999 Nuclear receptor coregulators: cellular and molecular biology. Endocr Rev 20:321–344

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 January 2017. at 11:20 For personal use only. No other uses without permission. . All rights reserved.