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

J. Biol. Chem., Vol. 277, Issue 38, 35088-35096, September 20, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/38/35088    most recent M205816200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sacchetti, P. Articles by Lefebvre, P. Search for Related Content PubMed PubMed Citation Articles by Sacchetti, P. Articles by Lefebvre, P. Social Bookmarking                      What's this?

Requirements for Heterodimerization between the Orphan Nuclear Receptor Nurr1 and Retinoid X Receptors^*

Paola Sacchetti, Hélène Dwornik, Pierre Formstecher, Christophe Rachez, and Philippe Lefebvre^§

From the INSERM Unité 459, Faculté de Medecine Henri Warembourg, 1 Place de Verdun, Lille 59045, France

Received for publication, June 12, 2002, and in revised form, July 12, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The nuclear receptor nurr1 is a transcription factor involved in the development and maintenance of neurons synthesizing the neurotransmitter dopamine. Although the lack of nurr1 expression has dramatic consequences for these cells either in terms of differentiation or survival, the mechanisms by which nurr1 controls gene transcription still remain unclear. In the intent to understand better the modalities of action of this nuclear receptor, we have undertaken a systematic analysis of the transcriptional effects and DNA binding properties of nurr1 as a monomer or when forming dimers with the different isotypes of the retinoic X receptor (RXR). Here, we show that nurr1 acts as a gene activator independently of RXR and through an AF2-independent mechanism. In addition, heterodimerization with RXR is isotype-specific, involves multiple domains in the C-terminal region of nurr1, and requires RXR binding to DNA. RXR-nurr1 and RXR-nurr1 heterodimers bind direct repeat response elements and display no specific requirements with respect to half-site spacing. However, the retinoid responsiveness of DNA-bound heterodimers requires the reiteration of at least three nurr1 binding sites, thereby limiting retinoid-induced nurr1 transcriptional activity to specific direct response elements.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The nuclear receptor nurr1 (NR4A2) is a brain-specific transcription factor (1), member of the superfamily of nuclear receptors (NR)¹ that plays a major role in the development and maintenance of a specific subset of neuronal cells. Knockout experiments have shown that absence of nurr1 causes an abnormal development of dopamine-synthesizing neurons of the midbrain region (2-4). Interestingly, further analysis of the knockout mice demonstrated that early neuronal differentiation is not affected by absence of nurr1 (5-7), whereas this factor is fundamental for the final differentiation of mesencephalic cells and expression of dopamine-specific marker genes like tyrosine hydroxylase, dopamine transporter, and L-aromatic amino acid decarboxylase (3, 4). Recently, it has been shown that the human dopamine transporter gene (8) is activated in vitro by nurr1 (9, 10) and that nurr1 might be playing an important transcription regulatory role in vivo as well (11).

To date, only the human dopamine transporter and tyrosine hydroxylase genes have been shown to be regulated by nurr1 (12). However, in light of its important role in dopaminergic phenotype maintenance and its gene expression in several brain regions in adult tissue (13, 14), a wider nurr1 target gene pool is conceivable. In addition, it is necessary to characterize further the transcriptional mechanisms by which this orphan nuclear receptor controls gene expression to identify nurr1-regulated genes. Nurr1, like its related factors nur77 and nor1 (15), binds as a monomer to the nerve growth factor inducible- DNA binding sequence AAAGGTCA (NBRE (16)) and strongly activates the expression of reporter genes bearing this sequence (10, 17). Furthermore, nurr1 also activates transcription by interacting with the 9-cis- retinoic acid receptor (RXR (18)) and forms heterodimers permissive to retinoids on multimerized AGGTCA response elements (17, 19, 20). However, retinoid-mediated transcription is not observed for natural nurr1-regulated genes containing functional NBREs in their promoter (10, 21).

RXR acts as a critical partner for several other NRs and forms either silent or 9-cis-retinoic acid-responsive heterodimers, depending on the dimer partner (22). Characterization of the RXR-nurr1 DNA binding properties showed that RXR-nurr1 dimers bind preferentially to DR5 response element, in a manner reminiscent of the RXR-all-trans-retinoic acid receptor (RAR) dimer (20). However, the nurr1 C-terminal transactivation domain AF2 is required for RXR-nurr1 dimer formation (17), unlike other RXR-containing heterodimers. The exact interaction modalities among nurr1, RXR, and their potential target recognition sequences need further analysis. Notably, the recent identification of a brain-specific ligand for RXR-nurr1 heterodimers (23) suggests that RXR ligands may play a crucial role in regulating nurr1 transcriptional activity and prompts the identification of functional target genes for this RXR-nurr1 regulatory complex.

In this study, we characterized the dimerization process between nurr1 and RXR by analyzing transcriptional activities and DNA binding properties of these receptors in PC12 cells, a pheochromocytoma cell line with neuronal characteristics and the ability to express nurr1 (1). Nurr1 interacted specifically with the RXR isotypes  and , but not , and two regions in the C-terminal region of nurr1 are necessary for dimerization with RXR. RXR-nurr1 heterodimers displayed low stringency with respect to DR half-site spacing. Formation of DNA-bound heterodimers on DR with spacing ranging from 10 to 27 bases was detected, but these dimers were refractory to retinoid stimulation in transcription assays. In addition, our studies reiterate the function of nurr1 as a monomeric protein because preventing nurr1 interaction with RXR through modification of putative dimerization interfaces did not alter its transcription activating function.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmids-- The luciferase reporter plasmids used in transient transfection studies were cloned by inserting one, two, or three copies of NBRE (see Fig. 1) upstream of the herpes simplex virus thymidine kinase (tk) gene minimal promoter in a pGL3-based vector (obtained from K. Ozato). The pCMX-nurr1 and pCMX-RXR encoding full-length mouse nurr1 and RXR cDNAs were obtained from R. M. Evans. cDNAs coding for human wild type (RXR), AF2-deleted RXR (RXRAF2; amino acids 1-419), and mutant RXGR subcloned into pSG5 (Stratagene, Amsterdam) have been described elsewhere (24-26). The (DR5)3xtkLuc contains three repeats of a retinoic acid receptor response element spaced by five base pairs. The RXR plasmid encoding full-length human RXR cDNA cloned into pcDNA3 was a kind gift of R. Polakowska. The dimerization-negative mutants, pCMX-nurr1^dim and pSG5-RXR^dim, and the truncated form of nurr1, pCMX-nurr1AF2 (amino acids 1-584), were constructed by site-directed mutagenesis (QuikChange, Stratagene) following the manufacturer's protocol. Specific oligonucleotides were used to mutate nurr1 residues Lys⁵⁵⁴, Leu⁵⁵⁵, and Leu⁵⁵⁶ into Ala (nurr1^dim (27)), RXR residues Glu³⁹⁰ and Glu³⁹⁴ into Ala (RXR^dim), and to insert a stop codon at nurr1 residue Pro⁵⁸⁴ (nurr1AF2). All constructs were verified by automated sequencing and restriction analysis.

Cell Culture and Transfections-- Transient transfection studies were performed in PC12 cells, grown at 37 °C in a 5% CO₂ humidified atmosphere in Dulbecco's modified Eagle's medium high glucose (Invitrogen) containing 5% heat-inactivated fetal calf serum (BioWhittaker), 5% heat-inactivated horse serum (HyQ, Logan, UT), and supplemented with 2 mM L-glutamine, 1,000 units/ml penicillin, and 100 µg/ml streptomycin (Invitrogen). Cells were plated in 24-well plates (2.5 × 10⁵ cells/well) 24 h before transfection and transfected with 0.5 µg of plasmid DNA/well complexed with 3 µl of LipofectAMINE (Invitrogen). Typically, cells were transfected with 50 ng of the reporter construct and 100 ng of receptor expression vectors. A reporter gene expressing the -galactosidase cDNA driven by the cytomegalovirus promoter was cotransfected (10 ng) in all experiments as an internal control for normalization of transfection efficiency. After a 5-h incubation, the lipid/DNA mix was replaced with fresh 2.5% serum medium containing dimethyl sulfoxide or 1 µM CD2624. Luciferase and -galactosidase activities were assayed 24 h later using Bright-Glo (Promega) and Galacto-Star (Tropix, Bedford, MA) reagents, respectively, and a LumiCount luminometer (Packard).

Western Blotting and Antibodies-- The antibody directed against RXR (sc-553) was obtained from Santa Cruz (Santa Cruz, CA). The monoclonal -actin antibody (A5441) and peroxidase-coupled anti-rabbit IgG were purchased from Sigma and the alkaline phosphatase-conjugated anti-mouse IgG from Promega. Proteins were resolved on a 10% SDS-polyacrylamide gel and transferred onto nitrocellulose membrane (Hybond-C, Amersham Biosciences). Immunodetections were carried out using the ECL Plus (RXR; Amersham Biosciences) and the ECF (actin; Promega) detection systems.

Reverse Transcription-PCR-- Total RNA from cells or tissue was isolated (RNeasy Minikit, Qiagen, Courtaboeuf, France), and 1-µg aliquots were reverse transcribed in the presence of random primers by Moloney murine leukemia virus reverse transcriptase (Promega) following the manufacturer's protocol. Conditions for PCR amplifications were as follows: 94 °C for 5 min, 30 cycles at 94 °C for 30 s, 50 °C (58 °C for nurr1 and actin) for 1 min, 72 °C for 1 min, and a final extension at 72 °C for 7 min. Reactions were carried out using specific primers designed based on sequences published previously (nurr1 and actin primers as described elsewhere (10, 26), hRXR (bp 582-898), mRXR (bp 249-592)).

Electrophoretic Mobility Shift Assays (EMSAs)-- The following oligonucleotides containing different NBRE repetition elements were end labeled with [-³²P]dATP (3,000 Ci/mmol) and T4 polynucleotide kinase and used as probes:

NBRE1x, gtaccctcgagctAAAGGTCAcgctagca;

NBRE2xDR10, gtaccgcagcagccctcgagctAAAGGTCAcgctagctAAAGGTCAa;

NBRE2xDR11, gtaccAAAGGTCAcctcgagctAAAGGTCAcgctagctgcagcagca;

NBRE2xDR27, gtaccAAAGGTCAcctcgagctgcagcagccgctagctAAAGGTCAa;

NBRE3x, gtaccAAAGGTCAcctcgagctAAAGGTCAcgctagctAAAGGTCAa.

Proteins were synthesized and labeled by coupled in vitro transcription/translation in rabbit reticulocyte lysates (TNT^TM T7-coupled transcription/translation system, Promega) and incubated in a binding buffer containing 10 mM Tris-HCl, pH 8.0, 40 mM KCl, 7% glycerol, and 1 mM dithiothreitol. The probe and 1 µg of poly(dI·dC) were added to the reaction, and the mix was incubated on ice for 20 min. For supershift assays, 0.3 µg of nurr1-specific antibody (sc-991 X; Santa Cruz), RXR antibody (sc-553; Santa Cruz), or anti-glutathione S-transferase IgG (sc-138; Santa Cruz) were preincubated with the in vitro translated proteins for 15 min at room temperature before addition of the probe. Electrophoresis was run in 0.5 × TBE on a 5.5% nondenaturing polyacrylamide gel at 4 °C. After electrophoresis, gels were dried for autoradiography or PhosphorImager (Molecular Dynamics) quantification.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Nurr1 Activates Transcription in PC12 Cells-- The transcriptional activity of the nuclear receptor nurr1 has been previously shown to be cell type-specific and mostly restricted to neuronal cell types (17). In the intent to characterize the transcriptional activity of nurr1 acting as monomer or as a heterodimer with RXR, we used the rat pheochromocytoma PC12 cells because they exhibit neuronal-like characteristics and can be stimulated to induce expression of nurr1 (1). Thus, this cell line provides an appropriate background to study the transcriptional mechanisms of this brain-specific receptor.

Because the functional activity of nurr1 in these cells has not been described previously, we tested the ability of nurr1 to induce the expression of a luciferase reporter gene containing three canonical nurr1 binding sites (NBRE) upstream of a tk minimal promoter (NBRE3xtkLuc). The three NBREs introduced in the reporter vector form direct repeats spaced by 9 (DR11) and 8 (DR10) nucleotides, respectively (Fig. 1B). Fig. 1 shows that increasing doses of nurr1 strongly activated the NBRE3x plasmid in a dose-dependent manner, whereas nurr1 did not affect the activity of the parental plasmid, pGL3tkLuc.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Dose-dependent activation of a NBRE3x reporter gene by the transcription factor nurr1. A, analysis of luciferase activity in PC12 cell extracts transfected with 50 ng of pGL3tkLuc or a luciferase reporter plasmid containing three NBRE binding sites (NBRE3xtkLuc) and the indicated doses of nurr1. Luciferase activity was normalized to -galactosidase activity, and the values are expressed as -fold induction over the normalized basal NBRE3xtkLuc activity set to 1. Data are the means ± S.E. (bars) of a representative experiment (n = 3), and the experiment was performed three times with similar results. B, sequence and geometry of the NBRE3x element.

Nurr1 Interacts Functionally with RXR Isotypes  and  but Not -- Nurr1 can interact with RXR and form heterodimers responsive to retinoids (17, 19). Three forms of RXR have been identified (28) which exhibit different tissue distributions in the brain (29, 30) and could therefore potentially interact with nurr1 to form distinct functional regulatory complexes in different cell types. Thus, we were interested in testing the functional activities of the different RXR-nurr1 dimers and compared them with the ability of nurr1 to induce reporter gene expression. Nurr1 significantly increased the basal activity of the NBRE3x plasmid (Fig. 2A), and this activation was enhanced further upon addition of the RXR-specific ligand (rexinoid) CD2624, reflecting a contribution of endogenous RXRs. Overexpression of RXR alone had no effect on the luciferase activity of the reporter gene, even after stimulation with CD2624. Upon cotransfection of nurr1 and RXR, the luciferase level was only slightly enhanced compared with nurr1 alone; however, stimulation with CD2624 strongly induced (approximately 12-fold) the activity of the NBRE3x reporter gene (Fig. 2A). We then studied the ability of RXR-nurr1 and RXR-nurr1 to activate the NBRE3xtkLuc construct. As expected, RXR (Fig. 2A) and RXR (data not shown) alone, in the presence or absence of the rexinoid, did not induce luciferase activity above the NBRE3x basal level. Coexpression of RXR and nurr1 did not significantly alter the luciferase activity level compared with nurr1 alone. Surprisingly, luciferase activity was not further enhanced upon addition of the rexinoid, suggesting a lack of functional interaction between nurr1 and RXR (Fig. 2A). On the contrary, cotransfection of nurr1 and RXR enhanced luciferase activity in a ligand-dependent manner (34-fold; Fig. 2A, right panel). Similar results were obtained using a different rexinoid (CD2425) and 9-cis-retinoic acid (data not shown), confirming that the effects observed are the result of isotype-specific heterodimer formations and not ligand-specific properties. Thus, RXR-nurr1 and RXR-nurr1 heterodimers have similar transcriptional properties in this system, suggesting that the isotypes  and  are potent partners of nurr1 and potential functional relays of retinoid action on nurr1-controlled pathways.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Formation of functional heterodimers between nurr1 and isoforms  and  of RXR. A, analysis of luciferase activity of PC12 cell extracts after stimulation with 1 µM CD2624, a RXR-specific ligand. Cells were transfected with 50 ng of NBRE3xtkLuc reporter plasmid and 100 ng of the different receptor expression vectors, pCMX-nurr1, pSG5-hRXR, pSG5-hRXR AF2, pcDNA3-hRXR, and pCMX-mRXR, as indicated. Results are expressed as described in Fig. 1. DMSO, dimethyl sulfoxide. B and C, expression of nuclear receptors used in PC12 cells. B, 25 µg of whole cell extracts transfected with an empty pSG5 (control) and a pSG5-hRXR expression vector was analyzed by Western blotting with an antibody to RXR. In vitro transcribed and translated pSG5-hRXR (TnT) and actin were used as controls. C, total RNAs from rat midbrain and PC12 cells were analyzed by reverse transcription-PCR for endogenous RXR isoforms  and  and nurr1 transcripts. Actin was used as an internal control, and RNA from PC12 cells transfected with the different receptor expression vectors was used as control for receptor overexpression.

We were interested in defining the contribution of the RXR AF2 domain in heterodimer formation between nurr1 and RXR. To this end, we used a RXRAF2 mutant (25) that conserves the ability to homo- and heterodimerize with RAR (31) but lacks the terminal 19 residues that allow for ligand-dependent transcriptional activity. Overexpression of this mutant alone had no effect on the reporter basal activity, and cotransfection with nurr1 evidenced a wild type-like sensitivity to nurr1 overexpression (Fig. 2A). However, addition of the RXR ligand did not enhance luciferase expression as observed in the wild type RXR-nurr1. Thus, the RXRAF2-nurr1 heterodimers are insensitive to retinoids, showing that the presence of the AF2 domain of RXR is required for RXR-nurr1 heterodimer response to these ligands.

We suspected that the enhancing effects induced by the RXR ligand CD2624 on nurr1 activity in the absence of overexpressed RXR (Fig. 2A) were because of heterodimerization of nurr1 with endogenous RXR/ expressed in PC12 cells. Thus, we tested the presence of RXR in these cells by Western blotting assays. As shown in Fig. 2B, PC12 cells expressed a significant amount of RXR, and this level could be enhanced strongly upon transfection of the RXR expression vector. We similarly tested the levels of nurr1, RXR, and RXR expression in PC12 cells by reverse transcription-PCR because of unavailability of antibodies or unsatisfying results obtained by Western blotting using available antibodies. PC12 cells expressed detectable levels of RXR mRNA (Fig. 2C) but not of nurr1 and RXR. After transfection with the appropriate cDNA-encoding plasmids, we were able to detect nurr1 as well as RXR mRNAs and an increased RXR expression.

Nurr1 Forms DNA-bound Heterodimers with RXR and -- To understand whether the different transcriptional effects observed on the NBRE3x reporter were the result of different receptor complexes binding to the same DNA element, we tested the DNA binding properties of these heterodimers by EMSA (Fig. 3). Nurr1 was able to bind efficiently to the NBRE sites present in the 3x element, producing a single band shift (Fig. 3A). This interaction was specific because addition of a nurr1-specific antibody prevented the formation of the receptor-DNA complex (compare lane 2 with lane 3). The antibody used was raised against the N-terminal domain of nurr1 and probably interfered with DNA binding. As expected, because of the absence of its specific DNA response element, RXR was not able to bind to the NBRE3x probe (Fig. 3A, lane 4). Coincubation of nurr1 and RXR resulted in the formation of a complex of lower mobility (Fig. 3A, lane 5) indicative of the presence of RXR-nurr1 dimers. Once again, addition of a nurr1-specific antibody abolished the formation of the higher mobility complex, confirming the presence of nurr1 in this complex (lane 6).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   DNA binding activities of RXR-nurr1 dimers on an NBRE3x motif. A, DNA binding of nurr1, RXR, and nurr1-RXR in the presence or absence of a nurr1-specific antibody. Receptors were obtained by in vitro translation and incubated for 15 min at room temperature with the antibody. The labeled NBRE3x probe was then added for 20 min, and complexes were resolved on nondenaturing 5.5% PAGE. Complexes were then visualized by autoradiography of dried gels. B, dimerization between nurr1 and , , and  forms of RXR. Receptor-DNA complexes were obtained and visualized as in A. Lower panel, control for RXR protein loadings used in EMSA. RXR receptors were obtained by in vitro translation in the presence of [35S]methionine and resolved on 10% SDS-PAGE. The gel was dried and visualized by PhosphorImager exposure. C, RXR AF2 is not required for dimerization with nurr1. Complexes between wt RXR or RXRAF2 and nurr1 were preincubated with a RXR-specific antibody for 15 min. The labeled probe was then added, and complexes were resolved on a polyacrylamide gel. The positions of the different complexes are indicated. The dark arrow indicates the position of supershifted complexes.

Functional data shown in Fig. 2 suggested the formation of heterodimers between nurr1 and RXR and , but not RXR. Thus, we tested this hypothesis by comparing the ability of RXR isotypes to form heterodimers on this NBRE3x probe. Fig. 3B shows that coincubation of nurr1 and RXR or nurr1 and RXR yielded two complexes, one having migration properties identical to nurr1 alone (lanes 2, 3, and 5) and the other of lower mobility, indicative of dimer formation (lanes 3 and 5). However, when nurr1 and RXR were coincubated with the labeled NBRE3x, only a single band was detectable which migrated similarly to nurr1 alone (compare lane 2 with lane 4). Thus, RXR is unable to form dimers with nurr1, providing a molecular basis for the lack of nurr1 responsiveness to rexinoid in the presence of RXR.

The lack of rexinoid sensitivity of the RXRAF2-nurr1 heterodimers (Fig. 2A) prompted us to test the ability of the two receptors to interact and bind DNA. Coincubation of nurr1 and RXRAF2 (Fig. 3C) resulted in a low mobility complex that was supershifted in the presence of RXR antibody (lanes 5 and 6), suggesting that deletion of the RXR AF2 domain does not prevent RXR-nurr1 dimerization, but interferes only with RXR-mediated transcription.

Nurr1 and RXR Interact through Two Domains Present in the Ligand Binding Domain (LBD)-- The fact that the C-terminal truncation of RXR did not affect the heterodimerization process with nurr1 was reminiscent of the results obtained with a truncated RXR (387-429) and RAR (32). Because one of the dimerization interface domains has been localized in the LBD of receptors such as RXR and RAR (32), we tested whether RXR-nurr1 dimers interacted through the same region. Based on the RXR-RAR dimer crystallographic structure (33), amino acids in helix 9 of the RXR LBD (Glu³⁹⁰ and Glu³⁹⁴), potentially involved in RXR-nurr1 dimer interaction, were mutated into Ala residues. We hypothesized that mutations of these cardinal residues would disrupt the dimer interface and prevent the complex from binding to NBREs. To verify our hypothesis, we performed EMSA with in vitro translated wt and mutant RXR. We also used a dimerization-deficient nurr1 (nurr1^dim) whose residues Lys⁵⁵⁴-Leu⁵⁵⁵-Leu⁵⁵⁶, present in the putative consensus dimerization interface, had been mutated to Ala residues (27). As expected, wt nurr1 and nurr1^dim bound the NBRE3x with similar affinities and formed a single mobility complex (Fig. 4A, lanes 2 and 3), whereas neither wt RXR nor the RXR^dim mutant bound to DNA (lanes 4 and 5). Coincubation of wt nurr1 and wt RXR resulted in the formation of RXR-nurr1 heterodimers (lane 6) as confirmed by supershift assays (see Fig. 3). On the contrary, coincubation of nurr1 with RXR^dim did not yield RXR-nurr1 heterodimers (compare lane 7 with lane 2). Similar results were obtained using the nurr1 derivative defective for dimerization, nurr1^dim (Fig. 4A). These results thus identify RXR Glu³⁹⁰, Glu³⁹⁴ and nurr1 Lys⁵⁵⁴-Leu⁵⁵⁵-Leu⁵⁵⁶ as critical contributors to heterodimer formation.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Mutations within the dimerization interface or the DNA binding domain prevent the formation of stable heterodimers on the NBRE3x motif. A, DNA binding of wt nurr1 and nurr1dim mutant in the presence and absence of wt RXR and RXRdim on the NBRE3x probe. Receptors were obtained by in vitro translation and incubated for 20 min with labeled NBRE3x probe. Receptor-DNA complexes were resolved on nondenaturing 5.5% PAGE and visualized by autoradiography of dried gels. Lanes 10 and 11, DNA binding of nurr1AF2 mutant to the probe in the presence or absence of wt RXR. B, absence of dimerization between nurr1 and RXGR mutant, unable to bind NBRE motifs. Supershifting with RXR-specific antibody did not affect band migration. Receptor-DNA complexes were obtained and visualized as in Fig. 3. Lanes 2 and 3, nurr1 alone; lanes 4 and 5, RXR alone; lanes 6-9, RXR-nurr1 complexes.

Previously published work (17) suggested that the AF2 domain of nurr1 is necessary for RXR-nurr1 dimerization. Thus, we tested the ability of a truncated form of nurr1 (nurr1AF2) to heterodimerize with RXR and bind to the NBRE3x response element (Fig. 4A, lanes 10 and 11). The nurr1AF2 mutant was able to bind to DNA, and the addition of RXR did not induce heterodimer formation, suggesting that the nurr1 AF2 domain is not necessary for DNA binding of nurr1 but is required for its heterodimerization with RXR.

We were interested in determining whether RXR-nurr1 heterodimer formation required RXR binding to the NBRE3x element. We therefore tested this potential interaction using an RXR mutant whose P box had been substituted for a glucocorticoid receptor P box (RXGR (26)). The RXGR mutant can only recognize the half-site AGAACA and thus is unable to bind to AGGTCA motifs. As predicted, the RXGR protein alone did not bind to the NBRE3x probe (Fig. 4B, lane 1), and coincubation of RXGR with wt nurr1 resulted in the formation of a unique complex migrating similarly to nurr1 alone (Fig. 4B, lane 2). Supershift experiments showed the absence of RXR in this complex (Fig. 4B, lane 3). Thus, inactivation of RXR-specific interaction with the AGGTCA motif suppressed the formation of stable RXR-nurr1 dimers.

Mutations in the LBD of Nurr1 Affect Only Heterodimerization-- We then tested the functionality of the different complexes observed by EMSA using transient transfection studies. As predicted, wt RXR and the RXR^dim mutant did not increase the activity of the NBRE3x reporter gene (Fig. 5) even after stimulation with CD2624. Overexpression of nurr1 or nurr1 and RXR induced luciferase expression (5-fold), and, as expected, stimulation with the rexinoid CD2624 (and others, data not shown) further enhanced gene activity (Fig. 5; see also Fig. 2A). However, coexpression of nurr1 and the RXR^dim mutant did not confer rexinoid responsiveness, indicating a lack of functional interaction between these receptors. We then tested the functional interaction between wt RXR and nurr1^dim. The nurr1^dim construct was able to activate luciferase expression in the NBRE3x plasmid (7-fold), but its activity was not enhanced further by CD2624. Coexpression of this mutant with wt RXR or RXR^dim did not alter the basal activity of the NBRE3x, which was also not responsive to the rexinoid (Fig. 5). Thus, the nurr1^dim mutant could interact functionally neither with endogenous (RXR,) nor overexpressed RXR to form retinoid-responsive heterodimers. The stronger basal activation observed with nurr1^dim compared with wt nurr1 remains unexplained at the moment. Nurr1AF2 was able to bind to the NBRE3x probe (see Fig. 4, lane 10) and activated luciferase expression of the NBRE3xtkLuc plasmid (Fig. 5B), although to slightly lower levels than the wt receptor. Coexpression of this mutant with RXR did not reveal any functional interaction with RXR in agreement with EMSAs showing no RXR-nurr1AF2 interaction (Fig. 4A, lane 11). This suggests that deletion of nurr1 AF2 domain prevents the formation of functional RXR-nurr1 dimers without altering the functional activity of nurr1 alone.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   The canonical dimerization interface of nurr1 is required for rexinoid inductibility. A, analysis of luciferase activity of PC12 cell extracts transfected with 50 ng of NBRE3xtkLuc reporter plasmid and 100 ng of the different wt and dim receptor expression vectors, pCMX-nurr1, pCMX-nurr1dim, pSG5-hRXR, and pSG5-hRXRdim. Cells were stimulated as described under "Experimental Procedures" with dimethyl sulfoxide (DMSO) or 1 µM CD2624. B, comparison of transcriptional activity between wt nurr1 and nurr1AF2. Luciferase activity was normalized to -galactosidase activity, and results are expressed as described in Fig. 1.

The Presence of a DR Element Is Necessary for the Formation of Functional Dimers-- We were then interested in comparing the interaction of nurr1 plus or minus RXR with a single NBRE element (Fig. 6A). Nurr1 bound the NBRE1x element, producing a single complex (lane 2), and addition of RXR did not alter the mobility of this complex (lane 4). We observed that the complex formation was inhibited by the addition of the anti-nurr1 antibody (lane 5) but not affected by the presence of anti-RXR or nonspecific antibodies (lanes 6 and 7). From this, we concluded that nurr1 is likely to bind stably to a single response element as a monomeric protein.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   A single NBRE element does not favor binding of functional RXR-nurr1 heterodimers. A, DNA binding of nurr1, RXR, and RXR-nurr1 to a single NBRE. Receptors were obtained by in vitro translation and incubated for 15 min at room temperature with the indicated antibodies. The labeled NBRE1x probe was then added for 20 min, and complexes were resolved on nondenaturing 5.5% PAGE. Complexes were then visualized by autoradiography of dried gels. B, analysis of luciferase activity of PC12 cell extracts transfected with 50 ng of NBRE1xtkLuc reporter plasmid and 200 ng of pCMX-nurr1 ± 200 ng of pSG5-hRXR. Cells were stimulated as described under "Experimental Procedures" with dimethyl sulfoxide (DMSO) or 1 µM CD2624. Luciferase activity was normalized to -galactosidase activity, and results are expressed as described in Fig. 1.

The functional activity of a luciferase reporter plasmid containing a single AAAGGTCA upstream of the tk minimal promoter (NBRE1x) was examined. Like the NBRE3xtkLuc construct, this plasmid responded in a dose-dependent manner to nurr1 (data not shown), albeit to a lower extent (maximal induction ~ 4-fold). The 2-fold induction by 200 ng of nurr1 was not enhanced further upon addition of the ligand CD2624, unlike the NBRE3x element (see Figs. 2 and 5). Coexpression of nurr1 and RXR did not result in increased luciferase expression of the NBRE1x plasmid, and the addition of the RXR ligand did not further alter luciferase level (Fig. 6B), suggesting the absence of functional interaction between nurr1 and RXR on this response element. Such inhibition has been reported previously on the NurRE proopiomelanocortin element (21) and may be a potential squelching phenomenon because of the formation of RXR-nurr1 dimers in solution.

Low Stringency of DNA Recognition by RXR-Nurr1 Heterodimers-- It has been shown previously that RXR binds to the 5'-half-site when dimerizing with nurr1 and that these dimers exhibit strong binding to a direct repeat spaced by 5 nucleotides (DR5 (20)). Because our data suggested that RXR-nurr1 heterodimers could bind DNA only in the presence of multiple NBRE sites, we were interested in identifying which two sites were bound by the receptors on the NBRE3x probe. We used different oligonucleotides in which a single NBRE was mutated at a time and originated response elements of DR11, DR27, and DR10 type, respectively. The comparison between binding properties of nurr1 to the different NBRE2x probes gave rise to surprising results. In vitro translated nurr1 was able to bind the three probes as monomeric protein (Fig. 7A), but unexpectedly, coincubation of nurr1 and RXR yielded a complex of slower mobility than nurr1 alone on all three DR elements (lanes 3, 6, and 9), suggesting that the heterodimer can recognize and form DNA-bound heterodimers on DR10, 11, and 27 elements. To confirm this hypothesis, we compared the band shift pattern obtained with NBRE3x with that obtained with NBRE2xDR27. Binding of nurr1 or nurr1 and RXR to the NBRE3x probe yielded DNA-protein complexes of distinct mobilities (Fig. 7B, lanes 6 and 7, and Fig. 3). The slower migrating complex was supershifted by the RXR antibody (lane 8), confirming the presence of RXR in this complex. Identical results were obtained using the NBRE2xDR27 element as probe (compare lane 3 with 7 and lane 4 with 8) and the DR10 as well as the DR11 (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   RXR-nurr1 heterodimers recognize widely spaced DR elements. A, DNA binding of nurr1 and RXR-nurr1 to probes containing differently spaced DR elements. Receptors were obtained by in vitro translation, incubated for 20 min with the appropriate labeled probe (as indicated), and complexes were then resolved on nondenaturing 5.5% PAGE. Complexes were visualized by autoradiography of dried gels. B, supershift assays with RXR antibody on RXR-nurr1 bound to DR27 and NBRE3x elements. Receptor-DNA complexes were obtained and visualized as in Fig. 3. The positions of the different complexes are indicated.

The NBRE2x sequences were introduced in the pGL3tkLuc backbone, and their functional activities were tested (Fig. 8). All three DR-containing elements were activated by nurr1 in a dose-dependent manner (data not shown) and reached maximal inductions similar to that obtained with the NBRE3xtkLuc plasmid (Fig. 8). Overexpression of nurr1 alone, or nurr1 and RXR, activated luciferase expression of the NBRE2x plasmids to similar extents. Surprisingly, treatment with the rexinoid CD2624 did not further enhance the activity of any of the three NBRE2xtkLuc plasmids (Fig. 8), suggesting that the heterodimers formed on these elements are not functional and are unable to transduce the retinoid signal. In similar conditions, the (DR5) 3xtkLuc reporter plasmid responded to retinoid stimulation similarly to the NBRE3xtkLuc, as reported previously (21).

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Widely spaced DRs are unresponsive to rexinoid stimulation. Analysis of luciferase activity of PC12 cell extracts transfected with 50 ng of different response element-tkLuc reporter plasmids and 100 ng of wt nurr1 ± 100 ng of RXR. Cells were stimulated as described under "Experimental Procedures" with dimethyl sulfoxide (DMSO) or 1 µM CD2624. Luciferase activity was normalized to -galactosidase activity, and results were expressed as described in Fig. 1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Our data show for the first time that nurr1 heterodimerizes selectively with RXR and RXR, but not RXR on the NBRE3x (Figs. 2 and 3B) as well as on a DR5 (data not shown). This isotype-specific interaction is an important element to bear in mind when searching for target genes for these putative transcriptional regulatory complexes. In fact, data available on the expression of the RXR isotypes show a different spatial distribution of these receptors in adult brain (29, 30), with high protein expression levels of RXR concentrated in the corpus striatum, the hypothalamus, and the anterior pituitary and a lower but more ubiquitous expression of RXR throughout the brain (29). Consequently, based on nurr1 mRNA localization (13, 14), the potential genes regulated by RXR-nurr1 heterodimers would be located in the olfactory bulb, specific subregions of the cortex, and the hippocampus for RXR and the hypothalamus for RXR. These predictions take into account the important differences that have been reported between localization of mRNAs encoding for RXR receptors and actual protein expression (29). Because such differences could be applicable to other species, clearly further studies should be performed to characterize the protein distributions of nurr1 and RXR isotypes in human brain, data not yet available, to identify which brain pathways could be regulated by retinoids. In light of the importance of nurr1 for dopaminergic cell survival and phenotypic expression (5) and high level of expression of the nurr1 protein in human substantia nigra (11), the absence of RXR proteins in the mesencephalic region (29) seems an important feature to define the functional role of nurr1 in dopaminergic pathways. This would suggest that nurr1 regulates mesencephalic dopaminergic-specific genes via a RXR- and retinoid-independent mechanism. This conclusion seems supported by the fact that expression of the human dopamine transporter gene is enhanced in the presence of nurr1, but not of RXR (10).

The response to retinoids of RXR-nurr1 heterodimers requires the AF2 domain of RXR (Fig. 2), suggesting that retinoid control of nurr1 pathways will require the presence of RXR. Our data show that the AF2 domain of RXR is a requirement only for rexinoid-induced transcription and that it is not a prerequisite for RXR-nurr1 heterodimer formation, as observed by EMSA (Fig. 3C). Thus, our data confirm that RXR is not silent in these heterodimers and that retinoids are potential nurr1 modulators in RXR-expressing cell types.

The role of the RXR AF2 domain is different from what we and others (17) observed for the nurr1 AF2 domain. Deletion of the nurr1 AF2 domain prevented dimerization with RXR (Fig. 4, lane 11), suggesting that this domain has a unique function in this receptor, unusual for NRs and for RXR-containing dimers. This conclusion had been suggested previously but then dismissed (17) by hypothesizing that this deletion altered the nurr1 LBD structural integrity, which in turn affected other receptor functions, such as DNA binding. However, truncation of nurr1 AF2 did not alter the DNA binding capacities of nurr1 alone (Fig. 4) or its transcriptional activity (Fig. 5). In addition, another interaction domain is located within the LBD of nurr1, in a region described previously for other NRs as the dimerization interface domain (32). We were able to prevent heterodimer formation by mutating few amino acids in the LBD of nurr1 and RXR (Figs. 4 and 5), suggesting that RXR and nurr1 interact with each other similarly to other RXR heterodimers, such as RAR-RXR. In our hands, neither point mutations in the LBD nor the absence of the AF2 domain altered functional and DNA binding properties of nurr1 alone, confirming that this receptor is a strong activator on its own, which can act independently of its dimerization partner.

Our data demonstrate that the simultaneous binding of nurr1 and RXR to DNA is required for the formation of stable heterodimers on DR response elements because no RXR-nurr1 complex is detected when the RXR P box is altered (Fig. 4B). This conclusion is strengthened further by the lack of rexinoid responsiveness of the NBRE1x reporter gene and the inability of RXR to associate to nurr1 bound to this monovalent response element. Based on two-hybrid experiments in mammalian cells, nurr1 has been reported to form rexinoid-responsive units with RXR and thus to interact with RXR in a DNA-independent manner_. (19, 20). However, in vitro protein-protein interaction assays demonstrated that structural requirements for RXR-nurr1 interaction in the absence of DNA are distinct from those involved in RXR-nurr1-DNA complex formation. Indeed, glutathione S-transferase pulldown assays showed that nurr1AF2, as well as RXR^dim and nurr1^dim mutants, formed heterodimers as efficiently as their wt counterpart (data not shown). Our data therefore suggest the inability of these receptors to bind as dimers on a single AAAGGTCA element and favor a classical nuclear receptor heterodimerization mechanism between these two receptors. Thus, the target genes must contain DR elements to be regulated by RXR-nurr1 dimer complexes.

Our studies further document that DNA binding of RXR-nurr1 dimers is more permissive than other nuclear receptor dimers in terms of spacing between repeated elements. Previous data showed a preference for DNA-bound heterodimers on DR5 and DR6 elements (20); however, we observe here the formation of heterodimers exhibiting similar binding affinities on DR10, 11, and as far as DR27 DNA elements (Fig. 7). These results underline a significant plasticity of LBD dimerization interface(s) between nurr1 and RXR which allow stabilization of the dimers on these unusual response elements. The use of two strong dimerization interfaces makes it conceivable that the strong protein-protein interaction would allow cooperative binding on widely spaced elements by inducing DNA bending in the long spacer between repeat sites.

Interestingly, rexinoid responsiveness of RXR-nurr1 heterodimers is limited to a subset of response elements because only multimerized DR5 response elements (our data and (20)) or NBRE3x (our data and (19)) are activated by rexinoids. Despite the interaction between these receptors and DNA, it is plausible that RXR undergoes such important conformational changes imposed by the geometry of the response element that they would prevent ligand binding. This could be viewed as another mechanism to control unwanted dimer activation, but it implicates a different method of selection of correctly spaced response elements, at least for these types of dimers. It remains still unclear how nurr1 and RXR can bind to the NBRE3x element forming stable and functional dimers, whereas binding to only two of the three repeats does not allow transduction of the retinoid signal. Quite intriguingly, the reiteration of DNA binding sites has been shown to be required for optimal RXR-mediated transcription on the CRBP2 gene promoter and other artificial reporter genes. Moreover, the ability of RXR to bind cooperatively as tetramers to these reiterated RXREs is isotype-selective, with RXR being significantly less prone to form such tetramers (35, 36). The ability of RXR to form dimers of dimers is mediated notably by the AF2 activating domain (helix 11), and deletion or mutation of this region prevents RXR tetramer formation (37). Thus analogies of the RXR system with the RXR-nurr1 system which we describe in this paper are striking and warrant further investigations to elucidate the interplay of nurr1 and RXR on reiterated response elements.

In conclusion, we have determined important new features of the mechanism of dimerization between nurr1 and its potential transcriptional regulator partner RXR. Nurr1 interacts exclusively with RXR and , and nurr1 mutants unable to heterodimerize with RXR are still strong transactivators, suggesting that this nuclear receptor functions through two distinct mechanisms of action which are probably implicated in the regulation of different brain signaling pathways. The conclusion that both receptors need to interact with DNA to form functional complexes and that only specific DR elements can transduce retinoid signaling will help identification of potential target genes for these complexes in regions coexpressing the different receptors. Because our functional data were collected in a single cell type, our conclusions need to be extended to other neuronal cell lines. The search for target genes warrants further studies because nurr1 is a tissue-specific factor that could be targeted in the development of drugs to cure severe neurodegenerative conditions, such as Parkinson's disease.

	ACKNOWLEDGEMENTS

We are grateful to Pascaline Ségard for technical assistance in reverse transcription-PCR and to Bruno Lefebvre and Céline Brand for a critical reading of the manuscript and important discussions. We thank K. Ozato, R. M. Evans, and R. Polakowska for providing DNAs, and U. Reichert and S. Michel for retinoids.

	FOOTNOTES

^* This work was supported in part by a Marie Curie Fellowship of the European Community Program "Improving Human Research Potential and the Socio-economic Knowledge Base" under Contract HPMF-CT-2001-01362 (to P. S.) and by the INSERM and Association France-Parkinson.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Recipient of an INSERM poste vert.

^§ To whom correspondence should be addressed. Tel.: 33-3-2062-6876; Fax: 33-3-2062-6884; E-mail: p.lefebvre@lille.inserm.fr.

Published, JBC Papers in Press, July 16, 2002, DOI 10.1074/jbc.M205816200

	ABBREVIATIONS

The abbreviations used are: NR, nuclear receptor(s); DRn, direct repeats with a spacing of n bases; EMSA(s), electrophoretic mobility shift assay(s); LBD, ligand binding domain; Luc, luciferase; NBRE, nerve growth factor inducible--binding response element; RAR, all-trans-retinoic acid receptor; RXR, 9-cis-retinoic acid receptor; tk, thymidine kinase; wt, wild type.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us