Transcriptional Regulation of Human Rev-erbα Gene Expression by the Orphan Nuclear Receptor Retinoic Acid-related Orphan Receptor α*

The Rev-erb and retinoic acid-related orphan receptors (ROR) are two related families of orphan nuclear receptors that recognize similar response elements but have opposite effects on transcription. Recently, the Rev-erbα gene promoter has been characterized and shown to harbor a functional Rev-erbα-binding site known as Rev-DR2, responsible for negative feedback down-regulation of promoter activity by Rev-erbα itself. The present study aimed to investigate whether Rev-erbα gene expression is regulated by RORα. Gel shift analysis demonstrated thatin vitro translated hRORα1 protein binds to the Rev-DR2 site, both as monomer and dimer. Chromatin immunoprecipitation assays demonstrated that binding of RORα to this site also occurred in vivo in human hepatoma HepG2 cells. The Rev-DR2 site was further shown to be functional as it conferred hRORα1 responsiveness to a heterologous promoter and to the natural human Rev-erbαgene promoter in these cells. Mutation of this site in the context of the natural Rev-erbα gene promoter abolished its activation by RORα, indicating that this site plays a key role in hRORα1 action. Finally, adenoviral overexpression of hRORα1 in HepG2 cells led to enhanced hRev-erbα mRNA accumulation, further confirming the physiological importance of RORα1 in the regulation of Rev-erbα expression.

The Rev-erb receptors form a subfamily of orphan nuclear receptors, consisting of two different genes, Rev-erb␣ (also termed ear1 or NR1D1) and Rev-erb␤ (also termed RVR, ear1␤, BD73, HZF-2, or NR1D2) (1). The Rev-erb␣ gene is located on human chromosome 17q21 and is encoded on the opposite strand of the thyroid hormone ␣2 receptor (2)(3)(4). Rev-erb␣ receptor was first reported to bind as monomers to response elements consisting of the half-core RGGTCA motif preceded by a 6-bp A/T-rich sequence (5,6). Later, it was furthermore shown to bind also as a homodimer on response elements consisting of a tandem repeat of two RGGTCA motifs spaced by two nucleotides preceded by a 6-bp A/T-rich sequence (7,8). The crystal structure of the Rev-erb␣ DNA-binding domain bound to its response element has been elucidated and revealed two major protein-DNA interfaces (9). Although the Rev-erb␣ receptor was initially reported to activate transcription (5), more recent data suggest that Rev-erb␣ actually acts as strong repressor of transcription (7). Transcriptional silencing by Rev-erb␣ requires the interaction with the corepressors N-CoR (or its variant RIP13a and RIP13⌬1) (10) and SUN-CoR (11) but not SMRT (12). This interaction was shown to rely on two conserved interaction domains (ID-I and ID-II) of N-CoR (13) and on at least two domains (CIR-1 and CIR-2) located in the E region of Rev-erb␣ (14). Rev-erb␣ is widely expressed, especially in muscle (6), liver (6,15), and brain (16). Expression of Rev-erb␣ is repressed during myocyte differentiation (17) or after exposure of liver to glucocorticoids (18). By contrast, its expression is induced during adipocyte differentiation (19) and in rat liver after chronic exposure to fibrates (15). Moreover, expression of Rev-erb␣ follows a circadian rhythm (18,20). A functional Rev-erb␣-binding site has been identified in the Rev-erb␣ gene promoter. Rev-erb␣ was shown to negatively regulate the activity of its own promoter via this site (8). This site is also essential for the fibrate response of the Rev-erb␣ gene via PPAR␣ 1 (15). Based on the presence of putative response elements in their promoter or in vitro data, several target genes for Rev-erb family members were proposed (8,17,(21)(22)(23)(24). A role for Rev-erb␣ has also been proposed in myocyte (17) and adipocyte differentiation (19). A transgenic mouse line that carries a deleted Rev-erb␣ gene has been shown to present transient alterations in cerebellar development (25).
The retinoic acid-related orphan receptors (ROR; also termed RZR) form another subfamily of orphan nuclear receptors consisting of three different genes ROR␣, -␤, and -␥ (NR1F1, NR1F2, and NR1F3) (26). Despite early controversial evidence that melatonin could be a ROR ligand (27), no natural ligands have been identified so far for this class of receptors. A subclass of synthetic thiazolidinediones was reported to bind and activate ROR␣ (28). Recently, Ca 2ϩ /calmodulin-dependent protein kinase type IV (CaM-KIV) has been shown to activate indi-rectly ROR␣ probably via a novel unidentified class of regulated co-activator molecules (29,30). The ROR␣ gene is located on human chromosome 15q21-q22 (31). Due to alternative splicing and promoter usage, it gives rise to four isoforms, ␣1, ␣2, ␣3, and ␣4 (also called RZR␣) (32)(33)(34), that differ in their N-terminal domains and display distinct DNA recognition and transactivation properties (32). RORs were initially reported to bind as monomers to response elements consisting of the halfcore RGGTCA motif preceded by a 6-bp A/T-rich sequence (6, 32, 34 -38). Later, binding to direct repeats of the RGGTCA motifs spaced by two nucleotides and preceded by a 6-bp A/Trich sequence has also been described (39,40). ROR␣ is widely expressed in peripheral tissues especially liver and muscle (6,26,33). The expression of ROR␣ is modulated by interleukin-1␤ in cartilage (41) and by thyroid hormone during Purkinje cell development (42). Based on the presence of putative response elements in their promoters, several target genes for ROR subfamily members have been proposed and analyzed in vitro (21, 22, 36 -38, 43-48). A role for ROR␣1 has been proposed in myocyte differentiation (49), lipid metabolism (50,51), bone and cartilage metabolism (52), ischemia-induced angiogenesis (53), small resistance arteries smooth muscle cell function (54), and inflammation control (48,55). Two mouse lines have been generated that carry a deleted ROR␣ gene (56,57). The phenotype of these strains is similar to the one of staggerer mice that carry a natural deletion in the ROR␣ gene that prevents the translation of its putative ligand-binding domain, thereby presumably disrupting its function (58). These mice exhibit cerebellar defects (58), abnormal pro-inflammatory cytokine production (59), as well as deficient intestinal apolipoprotein A-I and liver apolipoprotein C-III expression (50,51).
Because Rev-erb␣ and ROR␣ display similar expression patterns and regulate shared target genes in an opposite manner, the present study aimed to investigate whether Rev-erb␣ gene expression is under the control of the orphan nuclear receptor ROR␣ in HepG2 hepatoma cells. Our results show that ROR␣ binds in vitro and in vivo to the Rev-DR2 site in the Rev-erb␣ gene promoter. The Rev-DR2 site conferred specific hROR␣ responsiveness to a heterologous promoter. A striking elevation of Rev-erb␣ promoter activity was observed in HepG2 cells upon hROR␣1 expression vector co-transfection. Rev-erb␣ gene expression was increased when HepG2 cells were infected with an adenoviral construct allowing hROR␣1 overexpression. Taken together, our results identify ROR␣ as a positive regulator of Rev-erb␣ gene transcription and further delineate the complex cross-talks between the two receptors.

MATERIALS AND METHODS
RNA Analysis-RNA was extracted with Trizol reagent (Invitrogen) as indicated by the manufacturer. Reverse transcription was performed with MMLV-reverse transcriptase starting with 1 g of total RNA following the manufacturer's instructions (Invitrogen). 2.5 l of cDNA were used for semiquantitative PCR (annealing temperature: 55°C, 30 cycles) with the primers listed in Table I. Quantitative RT-PCR was performed (annealing temperature: 55°C, 40 cycles) on a Light Cycler apparatus (Hoffmann-La Roche) with the Faststart DNA SYBRGreen Master Mix I (Hoffmann-La Roche) quantification kit according to the manufacturer's instructions, using a 10-fold dilution of reverse transcription products, the specific primers indicated in Table I (200 nM), and 3 mM MgCl 2 for hRev-erb␣ or 4 mM MgCl 2 for ␤-actin. Considering C t as the cycle number in which the SYBRGreen fluorescence exceeds a constant threshold value and ⌬C t as the value corresponding to the difference C t (hRev-erb␣) Ϫ C t (reference), where the ␤-actin mRNA is used as reference, the relative hRev-erb␣/␤-actin relative level L is determined by L ϭ 2 Ϫ⌬Ct . The values presented are means Ϯ S.D. of triplicates.
Cell Culture, Viral Infection, and Transient Transfection Assays-Human hepatoma HepG2 cells were obtained from E.C.A.C.C. (Portondown, Salisbury, UK). Cell lines were maintained in standard culture conditions (Dulbecco's modified Eagle's minimal essential medium, supplemented with 10% fetal calf serum, 5% non-essential amino acids, and 5% sodium pyruvate (Invitrogen)) at 37°C in a humidified atmosphere of 5% CO 2 , 95% air. Medium was changed every 2 days. For infection experiments, HepG2 cells were seeded in 100-mm Petri dishes at a density of 10 7 cells/dish and incubated at 37°C for 16 h prior to viral infection. After seeding, viral particles were added at a multiplicity of infection of 100 and incubated for 3 h. Thereafter, cells were washed three times with PBS (0.15 M NaCl, 0.01 M sodium phosphate buffer, pH 7.2) and incubated in culture medium for the indicated times. At the end of the experiment, cells were washed once with ice-cold PBS, lysed, and scraped in 2 ml of ice-cold Trizol reagent.
For transfection experiments, HepG2 cells were seeded in 24-well plates at a density of 6 ϫ 10 4 cells/well and incubated at 37°C for 16 h prior to transfection using the cationic lipid RPR 120535B as described previously (50) with reporter plasmids (at 50 ng/well), expression vectors (pCDNA3 or pCDNA3-hROR␣1 at 100 ng/well), and the transfection efficiency control plasmid pRenContϩ at 1 ng/well. At the end of the experiment, the cells were washed once with ice-cold PBS, and the luciferase activity was measured with the Dual-Luciferase TM Reporter Assay System (Promega, Madison, WI) according to the manufacturer's instructions. All transfection experiments were performed at least 3 times. Protein content of the extract was evaluated by the Bradford assay using the kit from Bio-Rad.
Gel Retardation Assays-hROR␣1 and hRev-erb␣ were transcribed in vitro from the pCDNA3-hROR␣1 and pSG5-hRev-erb␣ plasmids, respectively, using T7 polymerase and subsequently translated using the TNT-coupled transcription/translation system (Promega, Madison, WI) following the manufacturer's instructions. DNA-protein binding assays were conducted as described (60) using the following binding buffer: Hepes 10 mM, KCl 50 mM, glycerol 1%, MgCl 2 2.5 mM, dithiothreitol 1.25 mM, poly(dI-dC) 0.1 g/l, herring sperm DNA 50 ng/l, bovine serum albumin 1 g/l containing 10% of programmed or unprogrammed reticulocyte lysate. Double-stranded oligonucleotides corresponding to the wild type Rev-DR2-response element present in position Ϫ45/Ϫ22 (with respect to the S1 transcription initiation site as defined previously (8)) of the human Rev-erb␣ promoter were endlabeled using T4 polynucleotide kinase and [␥-32 P]ATP and used as probe. For competition experiments, 10-, 50-, and 100-fold excess of cold wild type or mutated double-stranded oligonucleotides corresponding to the wild type Rev-DR2-response element present in position Ϫ45/Ϫ22 human Rev-erb␣ gene promoter (Table I) or a consensus DR2 oligonucleotide (Table I) was included 15 min before adding labeled oligonucleotides corresponding to the wild type Rev-DR2-response element. DNA-protein complexes were finally resolved by non-denaturating PAGE.
Chromatin Immunoprecipitation (ChIP) Assays-ChIP experiments were performed according to the method of Shang et al. (61), as modified by Giraud et al. (62). Briefly, 200 ϫ 10 6 HepG2 cells were grown to 60% confluence. Cell lysates were sonicated on ice, 15 times for 15 s and separated by 45 s. A volume of lysate equivalent to 20 ϫ 10 6 cells was immunoprecipitated using 4 g of an anti-ROR␣ antibody (48) or of an anti-hemagglutinin antibody (Santa Cruz Biotechnology, Santa Cruz, CA) as negative control. The same lysate volume was kept without immunoprecipitation for subsequent purification of input genomic DNA. One-tenth of the immunoprecipitated DNA was PCR-amplified twice for 35 cycles (30 s at 95°C, 30 s at 55°C, and 30 s at 72°C) using primers reported in Table I. An equal volume of non-precipitated (input) genomic DNA was amplified as positive control. One-fifteenth (input) or one-fifth (precipitated DNA) of PCR products was separated on an ethidium bromide-stained 2% agarose gel.

RESULTS
In Vitro Translated hROR␣1 Protein Specifically Binds to the Rev-DR2 Site of the Human Rev-erb␣ Gene Promoter-Rev-erb␣ as well as the PPAR␣/RXR␣ heterodimer bind to the Rev-DR2 site present in the human Rev-erb␣ gene promoter (8,15). As this site presents structural homologies with a binding site for the orphan nuclear receptor hROR␣1, we evaluated whether hROR␣1 protein translated in vitro could bind to this Rev-DR2 site by electromobility shift assay. As shown in Fig. 1, in vitro translated hRev-erb␣ protein used as a positive control binds as a monomer (noted M) and a dimer (noted D) on the Rev-DR2 site, as described previously (8). In addition, in vitro translated hROR␣1 protein binds as a monomer (noted M) on the Rev-DR2 site. Surprisingly, in vitro translated hROR␣1 protein binds also as a dimer on the Rev-DR2 site (noted D in Fig. 1), although to a lesser extent than as a monomer. To verify the specificity of the in vitro translated hROR␣1 protein binding to the Rev-DR2 site, competition with unlabeled double-stranded nucleotides corresponding to wild type or mutated Rev-DR2 sites was performed. As shown in Fig. 2, wild type double-stranded nucleotide corresponding to the Rev-DR2 site or a double-stranded nucleotide corresponding to a consensus DR2 (DR2Cons) site efficiently competed with the binding of the in vitro translated hROR␣1 protein to the Rev-DR2 site. Competition was still observed with a double-stranded nucleotide corresponding to the Rev-DR2 site with a mutated 3Ј-RGGTCA half-site (noted Rev-DR2M3Ј) leaving intact the A/Trich region and the first RGGTCA half-site. By contrast, double-stranded nucleotides corresponding to the Rev-DR2 site with a mutated 5Ј-RGGTCA half-site (noted Rev-DR2M5Ј) or with a mutated (A/T)-rich region (noted Rev-DR2CCC) failed to compete with the binding of the in vitro translated hROR␣1 protein to the Rev-DR2 site. Taken together, these data indicate that the Rev-DR2 site present in the human Rev-erb␣ gene promoter is indeed a specific binding site for the orphan nuclear receptor hROR␣1.
Endogenous hROR␣1 Protein Binds to the Rev-DR2 Site of the Human Rev-erb␣ Gene Promoter in HepG2 Cells-In order to evaluate whether hROR␣ binds to the human Rev-erb␣ gene promoter in vivo, in vivo occupancy of the Rev-erb␣ promoter by ROR␣ was analyzed using ChIP assays performed on DNA from HepG2 cells using an anti-ROR␣ antibody. The DNA encompassing the Rev-DR2 was precipitated by the anti-ROR␣ antibody (Fig. 3, row 1, lane 4). By contrast, no amplification was observed when the same DNA samples were PCR-amplified using primers covering either a region containing a RG-GTCA half-site preceded by a degenerated A/T-rich region previously designated as the Rd site (8) (Fig. 3, row 2, lane 4) or a region 1400 bp upstream of the Rev-DR2 site (Fig. 3, row 3,  lane 4). PCR amplification using oligonucleotides for ␤-actin, as negative control for the immunoprecipitation, did not result in any signal (Fig. 3, row 4, lanes 4). Taken together, these results indicate that hROR␣ binds specifically the Rev-DR2 site of the human Rev-erb␣ gene promoter in vivo.

Overexpression of hROR␣1 Enhances the Activity of Reporter Constructs Containing the Rev-DR2 Site in the Context of a Heterologous Promoter or of the Natural Human Rev-erb␣ Gene
Promoter-To determine whether the Rev-DR2 site is functional in the context of a heterologous promoter, the wild type or mutated Ϫ66/ϩ6 region of the human Rev-erb␣ gene promoter was cloned in front of the SV40 promoter driving the luciferase reporter gene in the pGL2-prom vector. This reporter construct was cotransfected in HepG2 cells along with a hROR␣1 expression vector. As shown in Fig. 4, hROR␣1 overexpression enhanced the luciferase activity in cellular extracts from HepG2 cells transfected with the wild type human Rev-erb␣ gene promoter Ϫ66/ϩ6 fragment-driven pGL2-prom vec-tor (noted Rev73wt). This effect was lost when the 5Ј-RGGTCA half-site of the Rev-DR2 site (noted Rev73M5Ј) or when the A/T-rich region preceding the RevDR2 site (noted Rev73CCC) was mutated. Moreover, hROR␣1 overexpression enhanced the luciferase activity of cellular extracts from HepG2 cells transfected with a construct harboring two copies of the wild type RevDR2 site cloned in front of the SV40 promoter of the pGL2prom vector (noted RevDR2wt 2X ). This effect was dose-dependent (data not shown). The cells transfected with a construct harboring two copies of the RevDR2 site with a mutated 3Ј-RGGTCA half-site (noted Rev-DR2M3Ј 2X ) cloned in front of the SV40 promoter were still responsive to hROR␣1, although to a much lesser extent than the wild type construct. By contrast, cells transfected with constructs harboring two copies of the RevDR2 site with a mutated 5Ј-RGGTCA half-site (noted Rev-DR2M5Ј 2X ) or the empty pGL2-prom vector remained insensitive to hROR␣1 action. In addition, consistently with the ChIP data (Fig. 3), the luciferase activity of cells transfected with a construct harboring three copies of the Rd site cloned in front of the same heterologous promoter was also not enhanced by hROR␣1 overexpression (data not shown). These data indicate that the Rev-DR2 site of the human Rev-erb␣ gene promoter confers hROR␣1 responsiveness to a heterologous promoter.
Next, we investigated whether the Rev-DR2 site could confer hROR␣1 responsiveness in the context of the Ϫ1517/ϩ216 fragment of the human Rev-erb␣ gene promoter. As shown in Fig.  5, hROR␣1 enhanced the luciferase activity of cellular extracts from HepG2 cells transfected with the wild type Ϫ1517/ϩ216 fragment-driven pGL2 vector (noted Rev1.7). This effect remained in two deletion mutants that harbored an intact Rev-DR2 site (noted Rev␦2 (Ϫ481/ϩ216) and Rev␦9 (Ϫ99/ϩ216)). Mutation of the 5Ј-RGGTCA half-site of the Rev-DR2 site in the Rev1.7 construct (noted Rev1.7M5Ј) or of the A/T-rich region preceding the Rev-DR2 site (noted Rev1.7CCC) severely impaired the hROR␣1 effect on the Ϫ1517/ϩ216 fragment of the human Rev-erb␣ gene promoter. By contrast, mutation of the Rd site did not significantly affect human Rev-erb␣ gene promoter activity, nor its responsiveness to hROR␣1 (data not shown). These results indicate that hROR␣1 enhances the ac- FIG. 2. hROR␣1 binding to the ؊45/؊22 region (RevDR2) of the human Rev-erb␣ gene promoter is specific. A double-stranded oligonucleotide corresponding to the Ϫ45/Ϫ22 fragment of the human Rev-erb␣ gene promoter was prepared and labeled as described under "Materials and Methods." In vitro translated hROR␣1 protein or unprogrammed lysate were incubated without or with 10-, 50-, and 100fold excess of the indicated unlabeled double-stranded oligonucleotides for 15 min at 4°C before labeled probes were added for 5 min at room temperature. These oligonucleotides correspond to the wild type (Rev-DR2wt) or mutated Rev-DR2 sites with either a mutated 5Ј-RGGTCA motif (Rev-DR2M5Ј), a mutated 3Ј-RGGTCA motif (Rev-DR2M3Ј), or with a mutated (A/T)-rich region (Rev-DR2CCC). The consensus DR2 (DR2cons) oligonucleotide was previously described (7,40). DNA-protein complexes were resolved by non-denaturating PAGE as described under "Materials and Methods." Specific complexes not observed with unprogrammed lysate and corresponding to nuclear receptor monomers (M) or dimers (D) are indicated by arrows. HepG2 cells were co-transfected with the pCDNA3-hROR␣1 expression vector (hROR␣1) (100 ng) or the empty pCDNA3 vector as control (Cont.) and reporter constructs (50 ng) containing the wild type or mutated Ϫ66/ϩ6 region of the human Rev-erb␣ gene promoter cloned in front of the minimal SV40 promoter of the pGL2-prom construct noted Rev73 WT, Rev73M5Ј, or Rev73CCC, respectively, or two copies of the wild type or mutated Ϫ45/Ϫ22 region of the human Rev-erb␣ gene promoter cloned in front of the minimal SV40 promoter of the pGL2-prom construct noted (Rev-DR2WT) 2x , (Rev-DR2M5Ј) 2x , or (Rev-DR2M3Ј) 2x , respectively, as described under "Materials and Methods." The empty pGL2-prom construct was used as negative control. Cells were transfected and luciferase activity measured and expressed as described under "Materials and Methods." tivity of the human Rev-erb␣ gene promoter and that the Rev-DR2 site plays a key role in this response.
Modification of hROR␣1 Expression Leads to Changes in hRev-erb␣ mRNA Levels-To confirm that hROR␣1 overexpression could lead to increased accumulation of hRev-erb␣ mRNA, HepG2 cells were infected with the Ad-ROR␣1 adenoviral expression vector (48). As control, HepG2 cells were infected with the same vector allowing expression of GFP. After adenoviral infection, cells were washed and lysed. Total mRNA was extracted and analyzed by semiquantitative and quantitative RT-PCR using hRev-erb␣-specific oligonucleotides. As shown in Fig. 6A, an increase in hRev-erb␣ mRNA was observed by semiquantitative RT-PCR in cells infected for 24 h with the hROR␣1 adenoviral expression vector compared with cells infected with the GFP adenoviral expression vector. No variation was observed in the level of cyclophilin mRNA used as control. In a separate experiment, the adenovirus-induced hRev-erb␣ mRNA accumulation was shown to be time-dependent by quantitative RT-PCR using SyberGreen detection and ␤-actin as reference gene (Fig. 6B). Up to a 20-fold increase in relative hRev-erb␣ mRNA level was reached after 48 h upon infection with the Ad-ROR␣1 adenovirus as compared with cells infected with the negative control Ad-GFP adenovirus. Taken together, these results confirm that hROR␣1 plays a key role in the physiological control of the expression of the Rev-erb␣ gene in human HepG2 cells.

DISCUSSION
The Rev-erb and ROR receptors are two subfamilies of orphan nuclear receptors that recognize similar response elements consisting of RGGTCA half-sites preceded by an (A/T)rich region but that have opposite effects on gene transcription (6). Because both receptors are co-expressed in certain tissues (e.g. liver and muscle) (6, 63), they were suggested to participate in a network of cross-talking receptors. Our observations showing that ROR␣ overexpression enhances Rev-erb␣ promoter activity and increases Rev-erb␣ gene expression further extend the complexity of this cross-talk. Moreover, we show that the Rev-DR2 site that mediates the transcriptional repression of Rev-erb␣ promoter activity by Rev-erb␣ itself (8) and its activation by PPAR␣/RXR␣ heterodimers (15) is also a major ROR␣-response element involved in the control of Rev-erb␣ gene promoter activity. This observation confirms the central role of this site in the regulation of Rev-erb␣ gene promoter activity and indicates that Rev-erb␣ gene expression is under the dynamic control of both receptor subfamilies acting through this site.
Our data indicate that wild type hROR␣1 binds both in vitro and in vivo to the natural Rev-DR2 site present in human Rev-erb␣ gene promoter. Binding in vitro occurred not only as a monomer but also as a dimer. Yet binding as a dimer to this site is weaker than binding as a monomer. A similar difference in binding intensity of Rev-erb␣ monomers and homodimers was observed on a DR2-response element (see Fig. 1) (8). Our data obtained in vitro with a natural response element confirms previous observations analyzing hROR␣1 binding on a consensus DR2-response element (39,40). Furthermore, the Rev-DR2 site confers hROR␣1 responsiveness to a reporter construct when cloned in front of a heterologous promoter. As the activation of a reporter construct containing the wild type RevDR2 is more than 2-fold greater than the activation of a reporter construct containing the RevDR2 with a mutated 3Ј-RGGTCA half-site leaving intact the A/T-rich region and the first RGGTCA half-site to which hROR␣1 only binds as monomer, the binding of hROR␣1 to the Rev-DR2 site as a dimer appears functionally important. Our data suggest that the two half-sites cooperate to ensure maximal activation by hROR␣. However, our data contrast with previous observations (64) that cooperative binding of hROR␣ as a dimer on a consensus DR2-response elements requires mutation of four amino acids in its DNA binding domain. Hence, our results rather suggest that functional dimer formation relies both on the nuclear receptor structure and on the actual sequence of the response element. Further work will be needed to clarify the relationship between response element sequence, dimer formation, and hROR␣ structure.
Phenotypic analysis of staggerer mice, which carry a natural deletion in the ROR␣ gene that disrupts its function, suggests that ROR␣ plays a key role in cerebellar development (58), as well as in bone (52) and lipid metabolism (50,51). A similar cerebellar ataxia due to deficient Purkinje cell development was observed in transgenic mice that carry a deleted ROR␣ gene (56,57). Interestingly, mice deficient for the Rev-erb␣ gene display a similar albeit transient neurological defect (25). The similarity between the phenotype of ROR␣-and Rev-erb␣deficient mice, despite the fact that both receptors have opposite actions, is intriguing. It is therefore tempting to speculate that the activation of Rev-erb␣ gene expression by ROR␣ is a crucial early event in Purkinje cell development leading to the FIG. 5. The ؊1517/؉216 region of the human Rev-erb␣ gene promoter is activated by hROR␣1 via its Rev-DR2-response elements. HepG2 cells were co-transfected with the pCDNA3-hROR␣1 expression vector (hROR␣1) (100 ng) or the empty pCDNA3 vector as control (Cont.) and reporter constructs (50 ng) containing the wild type full-length (noted as Rev1. 7) or deletion (noted as Rev␦2 or Rev␦9) and point mutants (noted as Rev1.7CCC or Rev1.7M5Ј) of the Ϫ1517/ϩ216 region of the human Rev-erb␣ gene promoter cloned in front of the luciferase reporter gene of the pGL2 construct as described under "Materials and Methods." The empty pGL2 construct was used as negative control. Cells were transfected and luciferase activity measured and expressed as described under "Materials and Methods." early repression of a specific set of genes by Rev-erb␣ and the specific positive regulation by ROR␣ of another set of target genes. Analysis of the spectrum of action of both receptors in these cells will be required to verify this hypothesis.
staggerer mice display alterations in the expression of apolipoprotein A-I (51), a major constituent of high density lipoproteins involved in reverse cholesterol transport. Recently, we have shown (24,51) that the expression of this gene is controlled by both ROR␣ and Rev-erb␣. Furthermore, the human apolipoprotein C-III, a major constituent of very low density lipoprotein involved in triglyceride transport, was identified as a ROR␣ target gene (50). Preliminary evidence suggests that apolipoprotein C-III is also repressed by Rev-erb␣ (65). Hence, the cross-talk between ROR␣ and Rev-erb␣ might be physiologically important for the control of the cholesterol and triglyceride metabolism. Unbalanced action of any of these receptors could therefore play a role in the pathogenesis of the dyslipidemia predisposing to atherosclerosis as already observed with ROR␣ (66). Overexpression of a dominant negative ROR␣ protein in myogenic cells was shown to delay the expression of mRNAs encoding MyoD and myogenin, two proteins involved in muscle development, and of p21 Waf-1/Cip-1 , a CDK inhibitor involved in cell cycle regulation (49). As a similar phenotype is observed when Rev-erb␣ is overexpressed (17), it seems likely that the ROR␣/Rev-erb␣ cross-talk is also important in muscle development. A role for ROR␣ is also suspected in bone (52) and cartilage metabolism (41) as well as in the inflammatory response (48,59). A role for Rev-erb␣ in these processes has not yet been described. It will be of interest to evaluate whether Rev-erb␣ antagonizes these effects of ROR␣. On the other hand, Rev-erb␣ has been implicated in the control of adipocyte differentiation (19), and its expression has been shown to follow a circadian rhythm (18,20). A possible role of ROR␣ as well as its cross-talk with Rev-erb␣ in these processes deserves further investigations.
Because ROR␣ and Rev-erb␣ have opposite activities and because these receptors have been implicated in several diseases, it appears important to study the factors that affect the balance between the two receptor subfamilies. So far, no natural ligand has been identified for Rev-erb␣. The lack of an AF2 transactivation domain in the hRev-erb␣ ligand binding domain rather suggests that it is unlikely that such ligand exists (1). Hence, its activity will probably be defined mainly by its expression level. Further characterization of the functional elements of its promoter is therefore of great interest. On the other hand, the promoter structure of the ROR␣ gene is not yet described, and little is know about the factors that determine its expression. Some cytokines like interleukin-1␤ (41) or hormones like T3 (42) may control ROR␣ gene expression. The repression of endogenous ROR␣ expression in myocytes forced to express an exogenous dominant negative ROR␣ suggests that ROR␣ could directly or indirectly drive its own expression (49). Hence, a role for hRev-erb␣ in the control of the expression of this receptor deserves attention and could further extend the cross-talk between the two receptor families.
In the present study, we have provided evidence that the Rev-erb␣ gene is a target for the orphan nuclear receptor ROR␣. Moreover, our results demonstrate that ROR␣ acts mainly by binding as a dimer on the Rev-DR2 site previously identified in the Rev-erb␣ gene promoter. Our results therefore further delineate the complex cross-talk between ROR␣ and Rev-erb␣.