Functional Characterization of Heterogeneous Nuclear Ribonuclear Protein C1/C2 in Vitamin D Resistance

Clinically apparent hereditary vitamin D-resistant rickets (HVDRR) usually results from a loss of function mutation in the vitamin D receptor (VDR). We recently described a human with the classical HVDRR phenotype but normal VDR function. Hormone resistance resulted from constitutive overexpression of heterogeneous nuclear ribonucleoprotein (hnRNP) that competed with a normally functioning VDR-retinoid X receptor (RXR) dimer for binding to the vitamin D response element (VDRE). Here we describe the purification, molecular cloning, and expression of this vitamin D resistance-causing, competitive response element-binding protein (REBiP) hnRNP C1/C2. When overexpressed in vitamin D-responsive cells, cDNAs for both hnRNPC1 and hnRNPC2 inhibited VDR-VDRE-directed transactivation (28 and 43%, respectively; both p < 0.005). By contrast, transient expression of an hnRNP C1/C2 small interfering RNA increased VDR transactivation by 39% (p < 0.005). Chromatin immunoprecipitation of nucleoproteins bound to the transcriptionally active 1,25-dihydroxy vitamin D-driven CYP24 promoter revealed the presence of REBiP in vitamin D-responsive human cells and indicated that the normal pattern of 1,25-dihydroxy vitamin D-initiated cyclical movement of the VDR on and off the VDRE is legislated by competitive, reciprocal occupancy of the VDRE by hnRNP C1/C2. The temporal and reciprocal pattern of VDR and hnRNPC1/C2 interaction with the VDRE was lost in HVDRR cells overexpressing the hnRNP C1/C2 REBiP. These observations provide further evidence for the functional importance of REBiP as a component of the multiprotein complex involved in the regulation of vitamin D-mediated transcription. In particular, chromatin immunoprecipitation data suggest that, in addition to its RNA-processing functions, hnRNP C1/C2 may be a key determinant of the temporal patterns of VDRE occupancy.

The effects in vivo of vitamin D are largely mediated by a specific vitamin D receptor (VDR). 2 The VDR is a ligand-de-pendent transcription factor that recognizes and binds to Cisacting vitamin D response elements (VDREs) in the promoter regions of target genes to induce or repress transcription. In addition to VDR, there are many other factors that act to "fine tune" hormone responsiveness. First among these is the circulating vitamin D-binding protein that delivers active vitamin D metabolites to target tissues (1,2). Second are the various "acceptor" proteins, such as megalin and cubulin (3)(4)(5)(6), which are anchored in the plasma membrane of target cells and promote internalization of steroid hormones. Third are the so-called "co-integrator" chaperone proteins that direct the intracellular trafficking of vitamin D metabolites for metabolism, catabolism, and transactivation via the VDR (7,8). Fourth is the cognate VDR dimerization partner retinoid X receptor (RXR) (9,10). Fifth are receptor-associated coactivators and co-repressors that influence recruitment of other elements of the transcriptional machinery to the promoter (11,12). And sixth are the "co-modulator" Cis-acting proteins of the heterogeneous nuclear ribonucleoprotein (hnRNP) family that compete with the VDR-RXR for binding to the vitamin D response element (VDRE), thus altering hormone receptor-directed transactivation (13)(14)(15).
The hnRNPs, first recognized for their ability to bind single strand ribopolynucleotides (16,17), are a family of more than 20 proteins that contribute to the complex associated with nascent pre-mRNA and are thus able to modulate RNA processing, including the stabilization of pre-mRNAs for nuclear export and translation (18 -20). More recently, hnRNPs have been shown to be capable of binding DNA in both single- (16,17) and double-stranded formats (14,15). Acting in their capacity as double-stranded DNA-binding proteins, we have shown previously that individual hnRNPs may function as dominant-negative modulators of steroid hormone-mediated transcription by competing with the VDR-RXR (14,15) and the estrogen receptor (21)(22)(23) for VDRE and estrogen response element, respectively. Here we describe the isolation, purification, and cloning of the cDNA for the naturally occurring human retinoid X and vitamin D response element-binding protein (REBiP), which was shown to be hnRNP C1/C2. In further studies, we have demonstrated the ability of the REBiP to exert a dominantnegative effect on 1,25-dihydroxy vitamin D 3 (1,25(OH) 2 D 3 )induced transcription via the naturally occurring VDRE in the vitamin D-24-hydroxylase (CYP24) gene promoter. Finally, using chromatin immunoprecipitation (ChIP) we have revealed the temporal pre-existence of the REBiP on the VDRE in vivo in cells in advance of VDRE occupation by the liganded VDR, suggesting a possible role for the REBiP promoting cis sitespecific chromatin remodeling in the region of the human genome harboring VDREs.

EXPERIMENTAL PROCEDURES
Reagents and Cell Culture-Crystalline 1,25(OH) 2 D 3 (Biomol, Plymouth Meeting, PA) was solubilized in 100% ethanol for addition to reaction mixtures. An Epstein-Barr virus transformed B-lymphoblast cell line and primary dermal fibroblast cultures from a VDR-RXR-normal patient with classical symptoms of hereditary vitamin D-resistant rickets (HVDRR) were used as a source for REBiP (15). Cultures of cells from an unrelated female of the same age with normal skeletal development were used as control cells for comparison. Immortalized B-cells were routinely maintained in RPMI 1640 medium (Invitrogen) supplemented with 5% fetal calf serum (FCS) (Omega, Tarzana, CA). VDR-VDRE-reporter assays were performed in the VDRE ϩ ROS (rat osteosarcoma) 17/2.8 cells, a cell line stably transfected with luciferase under the control of the VDRE in the CYP24 promoter (a kind gift from Hector DeLuca, University of Wisconsin, Madison, WI) as previously described (24). VDRE ϩ ROS cells were propagated in Dulbecco's modified Eagle's medium (Invitrogen) containing 6 mM glutamine, 10 g/ml insulin, and 10% FCS. Vitamin D-responsive human kidney HKC-8 cells were cultured in Dulbecco's modified Eagle's medium/F-12 media (Invitrogen) supplemented with 5% FCS.
Purification and Molecular Cloning of REBiP-A DNA affinity resin was prepared as described by Kadonga and Tijan (25). Gel-purified oligonucleotides (30-mers) containing nucleotides of the complementary sequence to consensus RXRE (5Ј-GATCAGCTTCAGGTCAGAGGTCAGAGAGCT-3) with 4-bp cohesive ends were annealed with their complementary sequence, subjected to 5Ј-phosphorylation, and then concatamerized in reactions using DNA ligase. The concatamerized DNA was coupled to cyanogen bromide-activated Sepharose. Affinity-purified response element-binding proteins present in the nuclear extracts of HVDRR cells were sequentially eluted from the support in increasing concentrations of KCl as described previously (21). Selected fractions were then submitted to automated Edman degradation and amino acid sequencing using an Applied Biosystems 477A or Hewlett Packard G1005 protein sequencer (Harvard Microchemistry Facility, Cambridge, MA) as previously described (15).
Following the identification of hnRNP C1/C2 as the REBiP from amino acid sequencing, a cDNA expression construct, including the coding region of the hnRNP C1/C2 gene, was cloned. Approximately 200 ng of HVDRR total RNA, isolated with TRIzol reagent (Invitrogen), was used as a template. Successful amplification of the coding region cDNA sequence was achieved with 30 cycles of reverse transcription-PCR using the primers to hnRNP C1/C2 (5Ј-ACGATGGCCAGCAACGT-TACCAAC-3Ј and 5Ј-TCCTCCATTGGCGCTGTCTCT-3Ј). The resulting PCR product was ligated into the PCR 3.1/V5-His-TOPO vector (Invitrogen) and the cDNA sequence verified.
Antibodies and Immunoblotting Assays-Cell extracts from the HVDRR patient and the control subject were subjected to electrophoresis on 4 -20% SDS-PAGE and transferred to nitrocellulose membranes. Western blot analyses, using a panel of antibodies including anti-human hnRNP C1/C2, anti-human hnRNPC-like, anti-human hnRNPA, anti-human Ku protein, and anti-human actin antibody (all from Santa Cruz Biotechnology, Santa Cruz, CA) were performed as described previously (14,15,21,22).
Transient Transfection and Small Interfering RNA (siRNA) Assays-Cells were grown to 80 -90% confluence in 12-well plates. For transient expression analyses, each well received 0.8 g of REBiP cDNA expression plasmid or vector-alone plasmid. For siRNA experiments, each well received 10 M siRNP C1/C2 (Santa Cruz Biotechnology). Promoter-reporter assays carried out using VDRE ϩ ROS cells utilized the endogenous luciferase construct, whereas similar assays using HKC-8 cells required transfection of 0.8 g of the VDRE-luciferase plasmid. All transfections were carried out in 0.004% Lipofectamine-2000 in Opti-MEM (Invitrogen) followed by an overnight incubation. The next day, the medium was replaced by Opti-MEM containing 0.1% ethanol or 10 nM 1,25(OH) 2 D 3 . After an additional 24 h at 37°C, the cells were lysed and luciferase and ␤-galactosidase activities measured (Promega, Madison, WI).
Electrophoretic Mobility Shift Assay and ChIP Assays-Electrophoretic mobility shift assays were performed as described previously (15) using REBiP-enriched nuclear extract from DNA affinity chromatography as a source of response elementbinding proteins.
ChIP assays were performed as described previously (26). Briefly, HVDRR and control lymphocytes were cultured for 4 days in RPMI 1640 medium supplemented with 5% charcoalstripped, heat-inactivated FCS and treated with 1,25(OH) 2 D 3 for the indicated times. Following hormone treatment, the cells were washed twice with phosphate-buffered saline and crosslinked with 1% formaldehyde at 37°C for 10 min. After quenching of the cross-linking with 1.25 M glycine, cells were harvested, rinsed with phosphate-buffered saline, and the resulting pellets were resuspended in 1 ml of cell lysis buffer (5 mM Pipes, pH 8.0, 85 mM KCl, 0.5% Nonidet P-40, 1 mM dithiothreitol, 0.25 mM phenylmethylsulfonyl fluoride, and 1 g/ml each of pepstatin, leupeptin, and aprotinin). Nuclei were collected and resuspended in 500 l of nuclear lysis buffer (50 mM Tris-HCl, pH 8.1, 10 mM EDTA, 1% SDS, 1 mM dithiothreitol, 2.5 mM phenylmethylsulfonyl fluoride) and 1 g/ml each pepstatin, leupeptin, and aprotinin (all from Sigma). The resulting chromatin samples were sonicated to yield sheared DNA fragments of sizes between 300 and 1000 bp. For each immunoprecipitation, sheared chromatin was diluted with immunoprecipitation dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH 8.1, and 167 mM NaCl). The chromatin was collected and incubated at 4°C overnight with 5 g of anti-VDR-9A7 (Affinity Bioreagents Inc. Golden, CO) and anti-hnRNP antibodies (Santa Cruz Biotechnology). Rabbit IgG was used as a negative control. The immune complexes were precipitated with 60 l of protein A-Sepharose beads (source) at 4°C for 1 h. The beads were then subjected to serial 1-ml washes of the following: immunoprecipitation dilution buffer TSE-500 (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.1, 500 mM NaCl); LiCl/detergent buffer (100 mM Tris-HCl, pH 8.1; 500 mM LiCl; 1% Nonidet P-40; and 1% deoxycholic acid in TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA). The antibody-protein-DNA immunocomplexes were then eluted with 1% SDS in 50 mM NaHCO 3 . Formaldehyde cross-linking was reversed by heating at 65°C overnight with the addition of 5 M NaCl to a final concentration of 200 mM. All of the samples were then digested at 45°C for 1 h with 20 g of proteinase K. DNA was extracted by phenol-chloroform, ethanol-precipitated, and analyzed by PCR using amplifying primer sequences spanning the enhancer VDRE in the proximal CYP24 promoter, Ϫ334 to Ϫ22 bp relative to the start site of transcription. Amplifying primer pairs spanning a more distal section of the CYP24 promoter not harboring a VDRE, bp Ϫ1749 to Ϫ1424, were used as a negative control in ChIP assays.

RESULTS
Purification and Molecular Cloning of the REBiP-Previous work from this laboratory (14) identified two species of an anti-hnRNP-reactive protein from vitamin D-resistant New World primate nuclear extracts that bound specifically to a consensus human VDRE (VDRE-D3, AGGTCAGACAGGTCA) direct repeat motif in both double-and single-stranded format. Subsequently, we screened human HVDRR cell extracts for their ability to bind to the same response elements in vitro and deter-mined that, although both cis elements could specifically bind HVDRR extract, the most avid interaction appeared to be with the RXRE in double-stranded DNA format (15). Given these preliminary results, we used concatamers of the double-stranded RXRE as an affinity agent to purify the human REBiP. The affinity support bound a nuclear protein with a tryptic fragment bearing 100% sequence identity with the human hnRNPC1 and C2 proteins (Fig. 1A), the alternatively spliced, translated products of the hnRNPC1/C2 gene. Using tryptic peptide sequencing, a full-length REBiP cDNA was cloned that bore 99.5% nucleotide sequence identity and 99.3-99.7% deduced amino acid sequence identity with the human hnRNPC1 and C2 proteins. Western blot analyses using an anti-human hnRNP C1/C2 antibody confirmed that cells from the HVDRR patient overexpressed a pair of anti-hnRNP C1/C2-reactive proteins of 39 -40 kDa, compatible with the hnRNPC1 and the slightly larger hnRNPC2 (Fig.  1B). This contrasted with extracts from an age-and sexmatched normal, vitamin D-sensitive donor that expressed principally hnRNPC2. Data in Fig. 1C showed that the endogenous hnRNPC1/C2 gene product was not confined to the nucleus, as the REBiP was also readily detectable in the post-nuclear supernatant of both HVDRR and control B-lymphocytes.
REBiP Specifically Interacts with the VDRE and Blocks VDRE-directed Transactivation-Confirmation of the ability of the affinity-purified REBiP or hnRNPC to bind to a VDRE and affect VDRE-directed transactivation was then sought. As was the case with crude extracts from HVDRR cells (15), affinitypurified REBiP was specifically bound to the VDRE in doublestranded format (Fig. 2B) and competed with 100ϫ unlabeled oligonucleotide for occupation of a VDRE bearing a direct repeat of the AGGTCA (VDRE-D3) or RXRE (Fig. 2A). These data suggested that the affinity-purified hnRNPC could bind specifically to a direct repeat of AGGTCA separated by either one or three base pairs and might function in a dominant-negative mode to block VDRE-directed transactivation. Promoterreporter analyses confirmed that transient overexpression of the unspliced hnRNPC1/C2 cDNA in wild-type, vitamin D-sensitive cells blocked VDR-VDRE-directed reporter activity in the absence or presence of added 1,25(OH) 2 D 3 (57 and 44% decrease, respectively, compared with vector-only control; both p Ͻ 0.005) (Fig. 3A). Somewhat unexpectedly, co-expression of the siRNA for hnRNPC1/C2 increased transcription both in the absence and presence of stimulatory 1,25(OH) 2 D 3 (39 and 135% increase, respectively, compared with vector only control; both p Ͻ 0.005). These data suggested that an endogenous, functional, human hnRNPC1/C2 siRNA-reactive mRNA for a REBiP is normally expressed in this rat cell line; the inset Western blot (Fig. 3B) confirms this to be the case. Data in Fig. 3C shows that overexpression of hnRNPC1 and hnRNPC2 cDNAs alone or together suppressed VDR-VDRE-directed reporter activity in the absence or presence of added 1,25(OH) 2 D 3 . This inhibitory effect on transcription appeared to be modulated by a combination of the two splice variants, with hnRNPC2 being slightly more potent than hnRNPC1 and the inhibitory effect of the two cDNAs being additive in the absence of 1,25(OH) 2 D 3 .
REBiP Occupies the VDRE in Vivo-To determine whether the hnRNP C1/C2 REBiP could affect 1,25(OH) 2 D 3 -stimulated transcription in living cells, we performed ChIP assays looking for proteins from vitamin D-resistant HVDRR and vitamin D-responsive cells that bound to the CYP24 gene promoter. In initial experiments, both vitamin D-resistant HVDRR and vitamin D-responsive control cells were treated with or without a VDR-saturating concentration of 10 nM 1,25(OH) 2 D 3 for 15 min prior to isolation of cross-linked protein-chromatin complexes. The protein-DNA complexes were immunoprecipitated with anti-hnRNP C1/C2 and anti-VDR antibody and quantification of cross-linked DNA performed by reverse transcription-PCR using two sets of primer pairs, one set corresponding to a transcriptionally active enhancer VDRE (Ϫ334 to Ϫ22 bp of the CYP24 gene promoter) and a control, non-VDRE-containing region (Ϫ1749 to Ϫ1424 bp of the CYP24 A and B, lanes 4). gene promoter) (Fig. 4, top panels). Prior to and following treatment with 1,25(OH) 2 D 3 , neither the VDR nor the hnRNP C1/C2 REBiP were bound by the non-VDRE upstream region of the CYP24 promoter (Ϫ1794 to Ϫ1424 bp) in either control or HVDRR cells (Fig. 4A). By contrast, in vitamin D-responsive human B lymphoblasts, the proximal VDRE-containing promoter region of the CYP24 promoter (Ϫ334 to Ϫ22 bp) (Fig.  4B) was occupied mostly by anti-human hnRNP C1/C2-reactive protein and just minimally by the VDR in the basal state before 1,25(OH) 2 D 3 treatment. After exposure to a VDR-saturating concentration of 1,25(OH) 2 D 3 , the VDR was recruited to the VDRE with a coincident decrease in hnRNP C1/C2 interaction with the promoter. This suggested that there was competition between the two VDRE-binding proteins for the cis sequence. The rabbit IgG control did not detect any other proteins interacting with this region of the promoter, indicating that binding of both the hnRNP C1/C2 and VDR was specific. Surprisingly, in REBiP-overexpressing HVDRR cells, both the VDR and the hnRNPC1/C2 REBiP were present on the promoter in the absence of added hormone, and this pattern of relative occupancy of the VDRE by the two proteins changed little upon acute administration of 1,25(OH) 2 D 3 . Bearing in mind that this experiment afforded only a single 15-min "snapshot" of hormone-induced changes in VDR and hnRNP C1/C2, these data suggest that the two proteins remained as co-occupants of the VDRE and/or that one of the two was anchored to the VDRE with the other being bound to the other but not necessarily still binding to DNA.

Analysis of Temporal Changes in Association of VDR and hnRNPC1/C2 with the VDRE by
ChIP-To further clarify the dynamics of VDRE occupancy, we analyzed temporal changes in the interaction of the hnRNP C1/C2 REBiP with the Ϫ334 to Ϫ22 bp fragment of the CYP24 gene promoter (Fig. 5). It is now well recognized that occupancy of the VDRE by the VDR-RXR is cyclical in nature, eventually resulting in the relative stable association of the heterodimer with the promoter and sustained transcriptional effects (26). However, the molecular nature of the "on-off" behavior of the VDR-RXR at the VDRE remains a matter of conjecture. We postulated that these receptor-response element cycling events may be due to the competitive presence of the REBiP at the VDRE. As such, we examined the temporal sequence of proteins occupying the VDRE in control vitamin D-responsive and vitamin D-resistant HVDRR cells over a period of 165 min (Fig. 5). As was the case with the ChIP assay data presented in Fig. 4, before the addition of hormone, the VDR was relatively excluded from the VDRE in control vitamin D-responsive cells compared with vitamin D-resistant HVDRR cells in which both the VDR-and anti-hnRNP C1/C2-reactive REBiP were present on the VDRE. Following the addition of 1,25(OH) 2 D 3 to the control cells, the VDR was recruited to the VDRE and the REBiP content diminished. Further, as previously reported by others (26), the VDR cycled on and off the promoter at roughly 45-min intervals. Interestingly, REBiP displayed a reciprocal cycling pattern occupying the VDRE every 45 min when the VDR was off the cis element, with relatively less of the hnRNPC1/C2 gene product occupying the VDRE in favor of the VDR at the conclusion of the 165-min observation period. These results suggested that the VDR and REBiP were competing for occupancy of the VDRE. In contrast to the reciprocal on-off cycling of the VDR and hnRNP C1/C2 in vitamin-responsive control cells, the cycle time for the VDR with the VDRE in HVDRR cells was 60 min following treatment with 1,25(OH) 2 D 3 . Moreover, in HVDRR cells, the cycle "off" time for hnRNP C1/C2 was 45-60 min, not 15 min, after the addition of 1,25(OH) 2 D 3 , and the reciprocal association of the VDR and hnRNP C1/C2 with the VDRE was distorted compared with that in control vitamin D-responsive cells.  ChIP analysis of VDR, hnRNP C1/C2, or IgG (as the negative control) on the native rat CYP24 VDRE was carried out for lysates isolated from control or HVDRR cells treated with or without 1,25(OH) 2 D 3 (10 nM) for different periods from 0 to 165 min. The associated chromatin DNA fragments were amplified with 1,25(OH) 2 D 3 -responsive region (Ϫ334 to Ϫ22 bp).

DISCUSSION
The human hnRNPC1/C2 gene encodes two alternatively spliced translation products, hnRNPC1 and hnRNPC2 (16); as depicted in Fig. 1A, hnRNPC1 is the smaller of the two variants created by excision of 13 amino acids from the hnRNP C1/C2 coding sequence (27,28). Along with protein partners hnRNPA1 and -A2 and hnRNPB1 and -B2, hnRNPC1 and -C2 are the core elements of the hnRNP complex that is central to the handling of RNA transcripts for intracellular translation or destruction (19,29,30). The hnRNPC1/C2 gene products have also been shown to have additional actions in the nuclear compartment of cells, including splicing, telomere regulation, as well as nuclear retention of hnRNAs and nuclear matrix (31)(32)(33)(34)(35). As part of their action in RNA processing, the hnRNPA, -B, and -C proteins form heterotetrameric complexes that assemble as an anti-parallel 4-helix coiled coil (36) on nascent transcripts to regulate the splicing, polyadenylation, and turnover of that transcript (37). Among the core hnRNPA, -B, and -C proteins, hnRNPC1 and C2 were previously considered to be the only members of the grouping not found outside of the nucleus of the cell (38). However, our results (Fig. 1C), along with the work of others (37,39), demonstrate that hnRNPC1/C2 gene products can be recovered from the cytoplasm of the cell, with the export signal for hnRNPC1 and -C2 residing in a 40-amino-acid stretch of the C-terminal domain of the protein (37).
Recent work from this laboratory has also confirmed that hnRNP binding to nucleic acid is not confined to single strand mRNA but can be observed using either single or double strand DNA as a binding template (14,21). When overexpressed in vivo, as they are in certain steroid hormone-resistant New World primate genera (7), these proteins can compete with ligand-bound sterol/steroid receptor dimer pairs for binding to that receptor's cognate cis response element and block ligandmediated transactivation (23). These New World primate species represent a successful experiment of nature as they have evolved effective means of countering the dominant-negative actions of the hnRNPs as promoter-silencing factors by amplifying the endogenous production of the dominant positive-acting receptor-activating ligands (37). In 2003, we reported the first human example of clinical vitamin D resistance due to overexpression of an hnRNP (15). The homologue of the New World primate condition resulted from the constitutive overexpression of an hnRNP originally thought to reside in the hnRNPA family. Here we have purified this dominant-negative-acting, vitamin D resistance-causing human protein and shown it to be hnRNPC1/C2 (Fig. 1A); we now know that our characterization of the human REBiP as an hnRNPA in earlier studies (15) largely resulted from use of an anti-hnRNPA polyclonal antibody that cross-reacted with hnRNPC1/C2. However, because of the propensity of the hnRNPC1 or C2 to oligomerize with either hnRNPA1 or hnRNPA2 in native conditions (40), it is likely that there was indeed a mixture of hnRNPC and -A proteins present in unpurified nuclear extracts from our vitamin D-resistant patient cells. Nevertheless, the ability of a more specific anti-human hnRNP C1/C2 antibody to eliminate binding of a purified REBiP from complexes bound to either the VDRE or RXRE ( Fig. 2A) provides further evidence that hnRNP C1/C2 is overexpressed in HVDRR cells.
As anticipated, based on earlier work characterizing the hormone-resistant phenotype of our HVDRR patient and New World primates, overexpression of the hnRNPC1/C2 cDNA as well as the hnRNPC1 cDNA and the hnRNPC2 cDNA alone or together exerted a dominant-negative effect on 1,25(OH) 2 D 3driven VDRE-promoter activity (Fig. 3). More unexpected was the finding in the same experiments that a significant dominant-negative effect of hnRNP C1/C2 was exerted in the basal, no-hormone-added state in wild-type, vitamin D-sensitive osteosarcoma (ROS 17/2.8) and kidney (HKC-8) cells (Fig. 3, A and C), suggesting that "hormone-sensitive" promoters such as that found in the CYP24 gene can be extraordinarily susceptible to the dampening effects on transcription when only a small amount of 1,25(OH) 2 D 3 is available (i.e. as in diluted FCS) to the endogenous VDR. Even more unexpected was the significant stimulatory effect of hnRNP C1/C2 siRNA on VDRE-reporter activity in ROS 17/2.8 cells (Fig. 3A), suggesting that the "knockdown" of endogenously expressed hnRNP C1/C2 in vitamin D-responsive cells can relieve a naturally occurring brake on transcription. These results further indicate that the dominant-negative effects of endogenous hnRNP C1/C2 cannot be "competed out" completely by the addition of exogenous 1,25(OH) 2 D 3 . In fact, the presence of the hnRNP C1/C2 REBiP on the VDRE in vitamin D-responsive as well as HVDRR cells subjected to ChIP (Fig. 4B) confirms that the REBiP is bound to the VDRE in vivo.
Because of its relative abundance in the HVDRR compared with control nuclei (Fig. 1C), it is perhaps not surprising to find relatively more of the hnRNP C1/C2 REBiP occupying the VDRE in the basal state (Fig. 4B). However, it was somewhat unexpected to detect the presence of REBiP on the VDRE in vitamin D-responsive control cells in the basal state (Figs. 4 and 5). These data suggest that hnRNP C1/C2 is normally bound by VDRE-like cis elements preceding occupation of that cis sequence by its cognate receptor. This observation and the ability of the REBiP to bind single strand DNA with an AGGTCA motif in a direct repeat format (15) have led us to speculate that hnRNP C1/C2, and not the receptor protein itself as previously suggested (41), initiates the program of chromatin remodeling. Such hnRNP C1/C2 "priming" for chromatin remodeling has been demonstrated recently for an enhancer element in the ␤-globin promoter (42). If this is the case, then one would predict that the introduction of ligand 1,25(OH) 2 D 3 would promote more effective binding of the liganded VDR-RXR dimer pair to VDRE. This appears to be the case. In wild-type, vitamin D-responsive cells, the addition of 1,25(OH) 2 D 3 to the ChIP reactions results in recruitment of the VDR to the VDRE. There was no discernable change in hnRNP C1/C2 occupancy of the promoter in these non-quantitative experiments; a dynamic range in the number of rounds of PCR amplification to which ChIP fragments have been exposed has not yet been performed to assess the relative robustness of hnRNP C1/C2 binding to the VDRE before and after the addition of 1,25(OH) 2 D 3 .
Data presented in this study provide further evidence that, in addition to their established function as facilitators of posttranscriptional gene regulation, hnRNPs are able to influence Response Element-binding Protein DECEMBER 22, 2006 • VOLUME 281 • NUMBER 51 gene transcription itself by acting as binding proteins for hormone response elements (REBiP). We have shown that the REBiP associated with VDR-mediated transactivation is hnRNP C1/C2 and confirm that overexpression of this factor leads to suppression of vitamin D-mediated gene regulation in a similar fashion to that originally described for a patient with HVDRR (15). Significantly, the REBiP function of hnRNP C1/C2 was also evident in normal vitamin D-responsive cells, where its occupancy of the CYP24 VDRE was reciprocal to the VDR. It is therefore possible to hypothesize that REBiPs play an important role in directing the cyclical on-off equilibrium between the VDR and its target CYP24 response element (43). Moreover, the identification of an REBiP associated with estrogen receptor-mediated gene regulation suggests that this may be a common determinant of temporal variations in response element occupancy by steroid hormone receptors (22). As well as clarifying the general applicability of REBiP-receptor interaction, future studies will be required to elucidate the putative function of hnRNPs as initiators of chromatin remodeling and the potential link between this and the more established posttranscriptional function of hnRNPs.