Nuclear Orphan Receptor Nurr1 Directly Transactivates the Osteocalcin Gene in Osteoblasts*

Nurr1, an NGFI-B nuclear orphan receptor, which transactivates promoters through an NGFI-B response element (NBRE), is strongly induced by parathyroid hormone through the cAMP-protein kinase A signaling pathway in osteoblasts. Here, we demonstrate that multiple agents activating diverse signaling pathways in osteoblasts induce Nurr1. The strongest Nurr1 inducers were activators of cAMP-protein kinase A-coupled signaling, followed by protein kinase C- and calcium-coupled signaling activators. Receptor tyrosine kinase activators had minimal effect, whereas serine/threonine kinase activators had no effect on basal Nurr1 mRNA levels. Computer analysis of osteoblastic promoters indicated two potential NBREs in the rat osteocalcin (Ocn) promoter. Intriguingly, the proximal site maps to the cAMP-responsive cis-element. We tested whether Nurr1 induces Ocn expression through the NBRE-like site. Recombinant and endogenous Nurr1 protein from primary mouse osteoblasts bound to a consensus NBRE in EMSAs. Nurr1 induced a consensus 3×NBRE-luciferase reporter construct in mouse osteoblasts. Recombinant and endogenous Nurr1 protein bound to the proximal NBRE-like site in the Ocn promoter in EMSAs. Endogenous Nurr1 protein bound to this site as a monomer, because neither retinoid X receptor α nor retinoid X receptor β antibody supershifted the protein-DNA complex. Ocn promoter-luciferase constructs lacking or containing a mutated proximal NBRE-like site had markedly blunted responses to Nurr1 overexpression. Finally, adenovirally expressed Nurr1 protein bound to the proximal NBRE-like site in chromatin immunoprecipitation assays and induced Ocn mRNA in primary rat osteoblasts. We conclude that Ocn is a Nurr1 target gene, which positions Nurr1 in the core of transcriptional factors regulating osteoblastic gene expression.

osteoblasts, the precursor cell gives rise to chondroblasts, adipocytes, myoblasts, and marrow stromal cells. Differentiation along any of these lineages is determined by expression of cell type-and stage-specific markers. Precursors that differentiate along the osteoblast lineage express Runx1/cbfa1 and osteopontin at an early stage, whereas osteocalcin (Ocn) 1 is a late stage marker for mature osteoblasts (3). Identifying the regulators of osteoblastic gene expression is critical to understanding the control of precursor differentiation along specific lineages.
Nurr1 and LacZ Adenoviruses-The AdEasy TM Adenoviral Vector System (Invitrogen) was used to construct adenoviruses expressing Nurr1 or LacZ proteins according to the manufacturer's protocol. ROBs were plated at 4 ϫ 10 5 cells/well in 6-well tissue culture plates (Costar) in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 200 units/ml penicillin, and 100 g/ml streptomycin. The following day, the cells were transduced with 100 Nurr1-expressing virions/cell in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 200 units/ml penicillin, and 100 g/ml streptomycin. Experiments utilizing LacZ adenovirus and the same protocol demonstrated more than 90% transduction of the cells (data not shown). Two days after transduction, total RNA and nuclear extracts were collected for use for real-time PCR and EMSAs.
Immunocytochemistry-Cells were plated on culture slides (BD Falcon, San Jose, CA) at 5 ϫ 10 4 cells/chamber in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum, 200 units/ml penicillin, and 100 g/ml streptomycin. Two days later cells were transduced with Nurr1-expressing adenovirus. Two days after transduction, medium was aspirated, and cells were washed with PBS. Cells were fixed in cold 2% paraformaldehyde for 30 min at room temperature. After fixation, cells were washed in PBS and blocked for 1 h at room temperature with 3% normal goat serum (Jackson ImmunoResearch, West Grove, PA) in PBS containing 0.1% Triton. Cells were washed with PBS and incubated overnight at 4°C with Nurr1 antibody (Santa Cruz, sc-990x) diluted 1:200 in 1.5% normal goat serum in PBS containing 0.1% Triton. Cells were then washed in PBS and incubated for 2 h at room temperature with secondary antibody (Santa Cruz, sc-2004) diluted 1:200 in PBS containing 0.1% bovine serum albumin. Cells were washed with PBS and developed using 0.005% 3Ј3-diaminobenzidine/0.021% H 2 O 2 solution in PBS, pH 7.4. Cells were washed in PBS, rinsed through an increasing ethanol gradient, and immersed in xylene for 1 min.
Chromatin Immunoprecipitation Assays-Confluent ROB cells were transduced with adenoviruses as described above. Two days post-transduction, cells were subjected to chromatin immunoprecipitation (ChIP) assays utilizing the ChIP assay kit (Upstate) and the manufacturer's protocol. Briefly, 10 6 cells were treated with 1% formaldehyde solution for 10 min at 37°C to cross-link histones, were resuspended in lysis buffer containing 1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.1, and then were sonicated to shear DNA. DNA was recovered, and histone-DNA cross-links were reversed in an aliquot, which subsequently was used in PCR reactions to evaluate the amount of DNA present in various groups. The remaining DNA-histone complexes were used in immunoprecipitation reactions utilizing Nurr1 or E4bp4 primary antibodies (Santa Cruz) or a no primary antibody control and a Salmon Sperm DNA/Protein A-agarose slurry (Upstate). Histone-DNA crosslinks were reversed, and DNA was recovered and used in PCR reactions utilizing primers that bracket the proximal NBRE-like site in the osteocalcin promoter. The sequences of the primers that encompass the proximal NBRE-like element were Ϫ137/Ϫ120 5Ј-CCAACCACAG-CATCCTTT-3Ј and Ϫ2/Ϫ19 5Ј-GGACTTGTCTGTTCTGCA-3Ј in the proximal rat Ocn promoter.
PTH-induced Nurr1 Protein Binds to and Transactivates a Consensus NBRE-After demonstrating the induction of Nurr1 gene expression through multiple signaling pathways in osteoblasts, we tested potential function of Nurr1 protein. Because cAMP-PKA agonists yielded the strongest Nurr1 expression and PTH is an endogenous hormone, we utilized PTH to induce endogenous Nurr1 protein synthesis in subsequent studies. We have previously shown that 10 nM PTH causes maximum induction of Nurr1 gene expression (13). To determine if PTHinduced Nurr1 protein translocates to the nucleus and binds DNA target sequences, nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h were utilized in EMSAs. PTHinducible binding to a consensus WT NBRE probe peaked between 2 and 4 h (Fig. 3A). No binding to a mutated NBRE was detected. This binding was competed by a 50ϫ WT unlabeled competitor. To determine if Nurr1 is part of this PTH-induced protein complex, supershift assay was performed on the same proteins. Nurr1-specific antibody supershifted the majority of the PTH-induced complex (Fig. 3B), whereas nonspecific antibody had no effect (data not shown).
Having determined that Nurr1 is part of the PTH-inducible complex that binds to the WT NBRE, we next investigated if Nurr1 protein induces NBRE-containing promoters. MOBs were co-transfected with an empty vector or the Nurr1 expression vector and a chimeric WT 3ϫNBRE-luc construct. Nurr1 overexpression induced luciferase activity through the chimeric promoter (Fig. 3C).
Nurr1-transactivated Rat Ocn Promoter Activity Requires the Proximal 199 bp-Having determined that PTH-induced Nurr1 protein bound to and transactivated an NBRE-containing promoter, we used computer analysis to identify osteoblastic genes with NBRE-containing promoters. The Ϫ1050 bp rat Ocn promoter has two sites with close homology (7/8 bp) to the consensus NBRE at nt Ϫ934/Ϫ927 and nt Ϫ115/Ϫ108.
To determine whether Ocn is a potential Nurr1 target gene, we transiently co-transfected MOBs with the Nurr1 expression vector and luc-reporter constructs driven by 5Ј serial deletions of the rat (Ϫ1050) Ocn promoter. Nurr1 overexpression maximally induced Ocn promoter activity in constructs containing nt Ϫ1050/Ϫ199 (Fig. 4). However, deletion to nt Ϫ138 attenuated, whereas further deletion to nt Ϫ83 abolished, Nurr1induced Ocn promoter activity.
Nurr1 Is Part of a PTH-induced Nuclear Protein Complex That Binds the Rat Ocn Proximal NBRE-Our transfection data suggested that both nt Ϫ199/Ϫ138 and nt Ϫ138/Ϫ83 are important for Nurr1-mediated Ocn promoter induction (Fig. 4). Interestingly, computer analysis revealed an NBRE-like sequence within nt Ϫ138/Ϫ83 (Ϫ115/Ϫ108, ATAGGTCA in the antisense strand) but not within nt Ϫ199/Ϫ138. To determine potential direct Nurr1-DNA interaction accounting for the reduced Nurr1-mediated transcription seen in the (Ϫ199)Ocn-luc and (Ϫ138)Ocn-luc constructs, we performed EMSAs utilizing recombinant Nurr1 or luc proteins and nt Ϫ199/Ϫ138 or nt Ϫ142/Ϫ83 as probes. Nurr1 protein strongly bound to the nt Ϫ142/Ϫ83 probe but did not bind to the nt Ϫ199/Ϫ138 probe (Fig. 5A). Recombinant luc protein was used as control. Because computer analysis revealed an NBRE-like sequence in the distal Ocn promoter at nt Ϫ934/Ϫ927, we performed EM-SAs utilizing recombinant Nurr1 or PTH nuclear extracts and a probe corresponding to nt Ϫ942/Ϫ920 of the Ocn promoter. Neither recombinant Nurr1 nor PTH treated extracts bound to that site (data not shown).
Because the nt Ϫ142/Ϫ83 probe contains the putative NBRE at nt Ϫ115/Ϫ108, we tested the ability of endogenous Nurr1 protein to bind to this site. Nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h showed peak binding to the nt Ϫ120/Ϫ103 probe at 2 h. Furthermore, a 2-nt mutation in the Ϫ120/Ϫ103 probe eliminated PTH-induced Nurr1 protein binding (Fig. 5B). The PTH-induced binding completely supershifted with Nurr1 antibody, whereas a nonspecific antibody had no effect (Fig. 5C). To determine if Nurr1 binds to the NBRE-like sequence alone (17,18) or as an RXR-heterodimer (17,21,22), we performed supershift experiments on nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h utilizing the nt Ϫ130/Ϫ103 probe and Nurr1, RXR␣ or RXR␤ antibodies. Similar to Fig. 3B, Nurr1 antibody completely supershifted PTH-induced binding. No supershift was seen when the RXR antibodies were used (Fig. 5D).

FIG. 3. PTH induced nuclear protein binding to and transactivation of a consensus NBRE.
A, nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h show peak binding to the WT NBRE probe at 2-4 h. This binding is competed by 50ϫ unlabeled WT NBRE probe. The mut NBRE probe fails to bind PTH-induced nuclear proteins. P, probe alone. B, nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h show peak binding to the WT NBRE probe at 2-4 h that supershifts with Nurr1 antibody. C, MOBs were transiently co-transfected with empty or Nurr1 expression vector and 3ϫWT NBRE-luc reporter vector. Data are expressed as percent control. FIG. 5. Nurr1 protein binds nt ؊120/ ؊103 from the rat Ocn promoter. A, recombinant Nurr1 protein (N) bound to the nt Ϫ142/Ϫ83, but not to the Ϫ199/ Ϫ138, fragment from the rat Ocn promoter. Recombinant luciferase protein (L) was used as control. P, probe alone. B, nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h bound to the WT but not mut nt Ϫ120/Ϫ103 fragment from the rat Ocn promoter. C, nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h show peak binding to nt Ϫ120/ Ϫ103, which supershifts with Nurr1 antibody but not with nonspecific E4bp4 antibody (ϩN.S. Ab). D, nuclear proteins from MOBs treated with 10 nM PTH for 0 -6 h show peak binding to nt Ϫ120/ Ϫ103, which supershifts with Nurr1 but not RXR␣ or RXR␤ antibody.

The Proximal NBRE Is Critical for Nurr1-and PTH-induced
Ocn Promoter Activity-To evaluate the importance of the proximal NBRE-like element in Nurr1-mediated Ocn promoter activity, we introduced the same 2-nt mutation that abolished endogenous Nurr1 protein binding to the (Ϫ1050)Ocn promoter-luc construct. MOBs transiently co-transfected with an empty or Nurr1-expressing vector and the WT or mut (Ϫ1050)Ocn-luc construct showed that Nurr1 transactivated the WT, but not the mut (Ϫ1050)Ocn-luc construct (Fig. 6).

Adenovirally Expressed Nurr1 Protein Translocates to the Nucleus in Primary Osteoblasts and Recognizes a Consensus NBRE-To investigate the effect of Nurr1 in endogenous
Ocn gene expression, we generated Nurr1-expressing adenovirus constructs. These viruses at an multiplicity of infection of 100:1 transduced Nurr1 expression in the nucleus of more than 90% of osteoblasts, detected by immunocytochemistry utilizing a Nurr1 antibody (Fig. 7A). Similar treatment of ROBs with a LacZ-expressing adenovirus yielded no change in immunocytochemistry staining compared with control (data not shown).
Our data that recombinant and endogenous Nurr1 protein bound to and transactivated the rat Ocn promoter suggest that Ocn is a potential Nurr1 target gene in osteoblasts. To test this, we transduced ROBs with Nurr1-or LacZ-expressing adenovirus for 2 days. Non-transduced cells were also used as controls. EMSA demonstrates that nuclear proteins from Nurr1-transduced ROBs bound to a consensus NBRE (Fig. 7B).
Nurr1 Protein Binds to the Endogenous Ocn Promoter in Vivo-Our transient transfection and EMSA studies above indicate that Nurr1 protein recognized the proximal NBRE-like element in the rat Ocn promoter. To examine whether Nurr1 protein can recognize the endogenous Ocn promoter, we transduced ROB cells with Nurr1-expressing adenovirus and performed ChIP assays. Adenovirally expressed Nurr1 protein bound to an area of the Ocn promoter that encompasses the proximal NBRE-like site in primary ROB cells (Fig. 7C). LacZ control protein did not recognize this site.
Nurr1-expressing Adenovirus Induces Ocn mRNA Levels in Primary Fetal Rat Osteoblasts-To investigate whether Nurr1 binding to the endogenous Ocn promoter results in activation of Ocn gene expression, ROB cells were transduced with Nurr1 or LacZ adenoviruses. Non-transduced cells were also used as controls. Total RNA was extracted, reverse-transcribed, and subjected to real-time PCR. Nurr1 significantly induced mRNA levels compared with control. LacZ had no effect on Ocn mRNA expression compared with control. ␤-Actin expression was used to normalize the data (Fig. 7D). DISCUSSION We have previously reported the PTH-induced expression of the nuclear orphan receptor Nurr1 in osteoblasts (13). Here, we determined that multiple signaling pathways important in osteoblastic function regulate Nurr1 expression (Figs. 1 and 2). The strongest inducers were cAMP-PKA activators; PKC and calcium signaling, as well as selective agonists of tyrosine kinase receptors, caused a moderate increase of Nurr1 gene expression. Interestingly, TGF␤-BMP family members, which are potent osteoblast differentiation factors (32), had no effect on Nurr1 mRNA levels.
Induction of Nurr1 expression by various agonists and multiple signaling pathways has also been reported in other tissues. As in osteoblasts, cAMP signaling strongly induces Nurr1 expression in corticotrophs (33) and neuroblastoma and glial cells (34). Nurr1 is induced by luteinizing hormone in granulosa cells through PKC (16), vascular endothelial growth factor in human umbilical vein endothelial cells through the tyrosine kinase-coupled KDR receptor (15), and N-type calcium channel depolarization in primary sensory neurons (14).
The abundance of signaling regulators that induce Nurr1 raises the possibility that multiple signaling pathways may converge on Nurr1 to affect cell type specificity. Indeed, this seems to be the case for neuroblastoma and glial cells. Extracellular signal-regulated kinase, which mediates mitogen-activated protein kinase (MAPK) signaling, is required to mediate FSK-induced Nurr1 expression in N2A neuroblastoma cells but not in C6 glial cells (34). Likewise, MAPK signaling does not affect cAMP-induced Nurr1 expression in AtT-20 corticotrophs (33). In light of these data, the induction of Nurr1 by multiple agonists in osteoblasts stresses its potential importance in regulating target genes that affect osteoblast function.
To identify potential Nurr1 target genes in osteoblasts, we used a computer search to locate NBRE-containing promoters. NBRE-like elements were identified in the promoter/enhancer area of alkaline phosphatase, bone sialoprotein, osteopontin, type I(a)1 and type I(a)2 collagens, insulin-like growth factor I, collagenase, and osteocalcin genes. More specifically, we found two putative NBRE-like sites in the rat osteocalcin (Ocn) gene at nt Ϫ934/Ϫ927 and nt Ϫ120/Ϫ103. The downstream NBRElike site maps entirely within a previously identified cAMPresponsive region (26). Indeed, we found that maximum Ocn promoter activity required at least the proximal Ϫ199 nt of the promoter. Deletion of nt Ϫ199/Ϫ139 greatly reduced, and deletion of nt Ϫ138/Ϫ83 eliminated, Nurr1-induced Ocn promoter activity (Fig. 4). The latter response is easily explained by the loss of the proximal NBRE-like site at nt Ϫ120/Ϫ103. However, we were surprised when Nurr1-mediated Ocn promoter activity declined when the nt Ϫ198/Ϫ139 fragment containing an E box, AP-1, and cbfa1 sites was deleted.
Nurr1-mediated transcription is highly dependent upon Nurr1 stability on DNA (35). Loss of the E box, AP-1, and cbfa1 sites and any proteins bound to them may disrupt Nurr1 stability, thereby reducing Nurr1-mediated transcription. Yet, no direct interactions between HLH, AP-1, or cbfa1 and Nurr1 are known. It could be argued that AP-1 proteins are candidates for affecting Nurr1-mediated transcription, because AP-1 proteins regulate nuclear receptors through competition for the co-activator CREB-binding protein (CBP) (36). Although nuclear receptors compete for CBP, Nurr1 is unusual in that it does not bind CBP or any other classic nuclear receptor cofactor (35).
The lack of evidence for interactions between Nurr1 and AP-1 proteins in the regulation of Ocn promoter activity does not rule out the possibility that both transcription factors coordinately regulate induced Ocn promoter activity. AP-1 proteins are heterodimers formed from members of the Jun and Fos family of basic leucine zipper transcription factors. During osteoblast differentiation the composition of AP-1 dimers changes from a mixture of all the Jun and Fos family members in proliferating preosteoblasts to only JunD-Fra-2 heterodimers in differentiated osteoblasts (37). Interestingly, PTH maximally induces Fra-2 binding to a consensus AP-1 sequence at 3 h of treatment (38). This coincides with peak PTH-induced Nurr1 binding to nt Ϫ120/Ϫ103 of the Ocn promoter (Fig. 3, A  and B). This suggests that both AP-1 and Nurr1 proteins are present at the same time in PTH-treated osteoblasts and that they may coordinately induce Ocn promoter activity through the adjacent AP-1 and NBRE-like sites.
Our report here also adds Ocn to a short list of known Nurr1 target genes. CYP11B2 is induced through Nurr1 in angiotensin II-treated adrenocortical cells (39). Three Nurr1 targets are critical for neuronal differentiation and function. Ret is a protooncogene encoding a receptor tyrosine kinase (40 -43), tyrosine hydroxylase, catalyzes the rate-limiting step in catecholamine synthesis (24,44) and the dopamine transporter (DAT) regulates extracellular dopamine levels (25,45). Similar to Nurr1Ϫ/Ϫ mice (46), ret and tyrosine hydroxylase knockouts have severe neuronal agenesis that causes pre-or peri-natal death (47,48). DATϪ/Ϫ mice survive to adulthood and exhibit hyperlocomotion that mimics neurostimulant-induced activity (49). Interestingly, these mice are dwarfs due to anterior pituitary hypoplasia, which disrupts the growth hormone axis, and lactotroph hypoplasia, which decreases milk availability for nursing pups (50). The dwarfism correlates with decreased trabecular bone volume, cortical thickness, and cortical strength, although serum calcium, phosphorous, and PTH levels are unchanged compared with wild type littermates (51). Finally, very recently, osteopontin was shown to be a Nurr1 target in osteoblasts (22).
Interestingly, Nurr1 binds to the osteopontin promoter as a monomer, although it appears to be tethered to one of the RXR isoforms (22). In contrast, we found that Nurr1 bound to the Ocn promoter alone, because Nurr1 but neither RXR␣ nor RXR␤ antibody successfully supershifted PTH-induced nuclear proteins (Fig. 5). Indeed, the NBRE-1ike site in the Ocn promoter is a true monomeric site, suggesting that although Nurr1 can either homodimerize (19,20) or heterodimerize with RXR (17,21,22), there is no strict dimerization requirement for Nurr1 to transactivate osteoblastic gene expression.
Identification of both osteopontin and osteocalcin as Nurr1 target genes in osteoblasts is intriguing from a functional perspective. Neither osteopontin nor osteocalcin are required for normal skeletal development. Instead, both genes are needed for normal skeletal homoeostasis. OcnϪ/Ϫ mice undergo progressive osteopetrosis with age due to inappropriate skeletal remodeling (52), while osteopontin deletion impairs bone remodeling following ovariectomy (53) or PTH treatment (54). Interestingly, the two genes differ significantly in the timing of their expression during osteoblast development. Osteopontin is highly expressed in the earliest preosteoblast stage and remains high in fully mature osteoblasts (3). Conversely, Ocn is expressed only in fully mature osteoblasts (3). These data point to Nurr1 regulation of multiple genes along the osteoblastic lineage. Indeed, establishing Nurr1-induced Ocn gene expression points to Nurr1 as a potential transcriptional regulator influencing osteoblastic differentiation.