The Vitamin D Response Element-binding Protein

Vitamin D resistance in certain primate genera is associated with the constitutive overexpression of a non-vitamin D receptor (VDR)-related, vitamin D response element-binding protein (VDRE-BP) and squelching of vitamin d-directed transactivation. We used DNA affinity chromatography to purify proteins associated with non-VDR-VDRE binding activity from vitamind-resistant New World primate cells. In electrophoretic mobility shift assays, these proteins bound specifically to either single-strand or double-strand oligonucleotides harboring the VDRE. Amino acid sequencing of tryptic peptides from a 34-kDa (VDRE-BP1) and 38-kDa species (VDRE-BP-2) possessed sequence homology with human heterogeneous nuclear ribonucleoprotein (hnRNP) A1 and hnRNPA2, respectively. cDNAs bearing the open reading frame for both VDRE-BPs were cloned and used to transfect wild-type, hormone-responsive primate cells. Transient and stable overexpression of the VDRE-BP2 cDNA, but not the VDRE-BP1 cDNA, in wild-type cells with a VDRE-luciferase reporter resulted in significant reduction in reporter activity. These data suggest that the hnRNPA2-related VDRE-BP2 is a dominant-negative regulator of vitamin D action.

Dominant-negative control of steroid hormone-regulated gene transcription has been traditionally attributed to: 1) expression of transcriptionally silent receptor molecules, which compete with transactivating receptor dimer pairs for binding to enhancer elements (1,2); and 2) expression of non-DNAbinding co-repressor molecules, which interact with other constituents of the transcriptional complex to halt or slow transcription (3). More recently, another family of dominantnegative-acting proteins which squelch transcription have been identified. These proteins are in the heterogeneous nuclear ribonuclear protein (hnRNP) 1 superfamily (4 -8). hnRNPs were initially recognized for their ability to bind to single-strand pre-mRNA (9,10) and control the modification and movement of mRNA in the cell. The fact that hnRNPs could function as dominant-negative regulators of transcription was discovered in New World primates, a nonhuman primate suborder that is characterized phenotypically by relative targettissue resistance to adrenal, gonadal, and vitamin D sterol/ steroid hormones (11)(12)(13)(14)(15)(16)(17)(18)(19)(20).
We recently cloned and characterized the first member of the hnRNP family capable of modifying steroid hormone-directed transactivation (21,22), the estrogen response element-binding protein (ERE-BP). The 42-kDa ERE-BP is highly homologous to proteins in the hnRNPC subfamily (23). It acts to squelch estrogen receptor (ER)-estrogen response element (ERE)-directed transactivation by competing with the ER homodimer for binding to the ERE. Although discovered because of its overexpression in estrogen-resistant New World primate species, the ERE-BP is also expressed in Old World primate species including man (21). The vitamin D response elementbinding protein or VDRE-BP described here is in the second set of non-receptor sterol/steroid hormone response element-binding protein to be discovered. VDRE-BP was also initially identified in New World primates resistant to the vitamin D prehormone, 25-hydroxyvitamin D, and vitamin D hormone, 1,25dihydroxyvitamin D (24). Here we report the purification, cloning, and initial functional characterization of two VDRE-BPs as a members of the hnRNPA family. One of these VDRE-BPs, VDRE-BP2, is capable of squelching sterol/steroid hormone-mediated transactivation.

EXPERIMENTAL PROCEDURES
Cells-All primate cell lines were obtained from the American Type Culture Collection (Manassas, VA) and cultured as described previously (22). Preparation of nuclear extracts was according to a slight modification (22) of the method of Zerivitz and Akusjarvi (25).
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF192348.
ʈ To whom correspondence should be addressed: Div. of Endocrinology, Diabetes and Metabolism, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Rm. B-131, Los Angeles, CA 90048. Tel.: 310-855-8970; Fax: 310-652-0578; E-mail: adamsj@cshs.org. 1 The abbreviations used are: hnRNP, heterogeneous nuclear ribonucleoprotein; VDRE-BP, vitamin response element-binding protein; ER, estrogen receptor ␣; VDR, vitamin D receptor; RXR, retinoid X receptor; ERE, estrogen response element; VDRE, vitamin D response element; RXRE, retinoid X response element; YY1, yin-yang-1 protein; RACE, rapid amplification of cDNA ends; EMSA, electromobility shift assay; extracts from B95-8 or Vero cells or affinity-purified nuclear extracts from B95-8 cells, in the presence or absence of antibodies, 9A7␥ anti-VDR antibody (Ref. 26; generously provided by J. W. Pike), anti-hnRNPA1, anti-hnRNPA2 (a generous gift from Dr. G. Dreyfuss), and anti-RXR␣ antibody (27) were incubated with 2 g of poly(dI-dC), 20 mM Hepes (pH 7.9), 100 mM KC1, 5 mM MgCl 2 , and 10% glycerol on ice for 15 min. In some experiments transfected yeast (Saccharomyces cerevisiae) extract enriched in human VDR (28) and human RXR (29) were added to the reaction mixture, either in the presence or absence of primate cell nuclear extract. The double-strand oligonucleotides used as specific competitors in EMSAs were prepared by annealing complementary sequence. The above-mentioned single-strand consensus oligonucleotides were used to prepare both single-and double-strand labeled probes. The VDRE oligonucleotides, either annealed with complementary sequence or not, were radiolabeled with [ 32 P]ATP (PerkinElmer Life Sciences) by T4 kinase (Life Technologies, Inc.) to a specific activity of 10 8 cpm/g DNA. Radiolabeled probe (50 fmol), with or without antibodies and/or a 100-fold molar excess of radioinert, competitive, single-or double-stranded oligonucleotide, was then added and the incubation continued at room temperature for 15 min. Aliquots of the reaction were subjected to electrophoresis in a 6% polyacrylamide, 0.5ϫ TBE gel in 0.5ϫ TBE running buffer at 100 V. The gels were dried and exposed to Kodak X-Omat AR film.
Amino Acid Sequencing of Tryptic Peptides-Nuclear extracts were prepared as described previously by us (23). Extract fractions that eluted from the DNA affinity column were desalted and microconcentrated with a Micron-30 filter (30-kDa molecular mass cut-off; Amicon, Beverly, MA). Samples were loaded and electrophoresed through 10% SDS-polyacrylamide gels. Coomassie Blue-staining bands at 34 and 38 kDa in the 0.4 M KC1 fraction of the affinity column was excised from gels. The protein in the 34-and 38-kDa pools was subjected to tryptic digestion and amino-terminal amino acid sequence determination at the Harvard Microchemistry Facility as described previously (32).
Western Blot Analysis-Denatured cell extracts were subjected to electrophoresis using 12% SDS-polyacrylamide gels and transferred to nitrocellulose membranes as described previously (8). The membranes were blocked with 5% nonfat dry milk for 1 h and then incubated with monoclonal anti-VDR antibody, anti-human hnRNP-A1 antibody, or anti-human hnRNP-A2 antibody for 2 h and with horseradish peroxidase-conjugated secondary antibody for another 1 h prior to detection of antibody-reactive proteins with chemiluminescence reagent (ECL, Amersham Pharmacia Biotech).
RT-PCR and Northern Blot Analyses-Approximately 200 ng of total cellular RNA from B95-8, OMK, and Vero cells was extracted with RNA isolating reagent (Life Technologies, Inc.) and used as template for RT-PCR reactions and Northern blots. Successful amplification of a 435-bp cDNA sequence was achieved with 30 cycles of RT-PCR using B95-8 cell RNA as template and four primers, based on the internal amino acid sequence of the four tryptic peptides (IGGLSFET and FD-DHDSV for VDRE-BP1, and GGLSFET and GFVTFDDHD for VDRE-BP2, respectively), that were recovered in digests of both the 34-kDa VDRE-BP1 and 38-kDa VDRE-BP2. This PCR product was ligated into the PCR 2.1 vector (Invitrogen, Carlsbad, CA) and cDNA sequence verified. The PCR product was subsequently used as probe to screen equal amounts (15 g) of total cellular RNA extracted from both New World primate (B95-8 and OMK) and Old World primate (Vero) cells by Northern blot analysis.
Molecular Cloning of the VDRE-BPs-B95-8 poly(A) ϩ RNA (2.5 g) was used as template to generate the 5Ј and 3Ј ends of the VDRE-BP cDNA with the Marathon cDNA amplification kit (CLONTECH Laboratories Inc., Palo Alto, CA). Second-strand cDNA synthesis and adapter ligation were performed as instructed in the enclosed manual. The adapter-ligated cDNA was then used as template for annealing adapter-and VDRE-BP-specific primers for the RACE reaction: 5Ј-CTTTGACGACCATGACTCCGTGGA-3Ј and 5Ј-TTTGTTACTTTTGAT-GACCATGATC-3Ј with their complementary sequence for the 5Ј-and 3Ј-RACE of VDRE-BP1 and VDRE-BP2, respectively. A cDNA for the VDRE-BP1 and VDRE-BP2 was generated by end-to-end amplification using specific 5Ј and 3Ј primers. The amplified products were then separately subcloned into the PCR 3.1 expression vector (identified as pCR3.1-VDRE-BP1 and pCR3.1-VDRE-BP2) and sequenced.
Creation of Cell Lines Overexpressing the VDRE-BPs and ERE-BP-Vitamin D-responsive Old World primate Vero or COS-7 cells were incubated with 5.5 g of pCR3.1-VDRE-BP1, pCR3.1-VDRE-BP2, and pCR3.1 ERE-BP1 in LipoTAXI solution for 5.5 h followed by the addition of an equal volume of 20% fetal calf serum-supplemented, antibiotic-free medium. After incubation overnight, cells were split 1:10 and exposed to fresh medium containing 500 g/ml Geneticin selective antibiotic (G418 sulfate; Life Technologies, Inc.). Antibiotic-resistant colonies were harvested and expanded to confluence prior to study.

New World Primate Cells Overexpress a Non-VDR-related
Protein That Binds to the VDRE-Interaction of the VDR-RXR␣ heterodimer with the VDRE is necessary for transcriptional control of vitamin D-regulated genes, such as the osteopontin gene (33). We initially employed an oligonucleotide containing the consensus sequence for the mouse osteopontin VDRE (VDRE-op) as a probe of protein extracts from New World primate cells, Old World primate cells, or S. cerevisiae cells transformed with the human VDR and RXR␣ in EMSAs. , respectively. Addition of New World primate nuclear extract also retarded movement of the VDRE probe through the gel (lane 6). The major retarded complex migrated slower than the VDR-RXR-VDRE complex, presumably due to the highly charged nature of hnRNP-like proteins (34,35), and was competed away with the addition of a 100-fold excess of radioinert VDRE (Fig. 1A, lane 7). New World primate B95-8 cell nuclear extract also bound to the VDRE-oc (Fig. 1B, lane 1), and VDRE-BP-oc-VDRE binding was competed away by addition of excess radioinert VDRE-op (Fig. 1B, lane 2) or VDRE-oc (Fig. 1B, lane 3). VDRE-BP-VDRE binding was not challenged by addition anti-VDR antibody (Fig. 1B, lane 4) or a 100-fold excess of unlabeled double-strand RXRE, ERE, or the irrelevant cis element in the consensus CTF/NF1 oligonucleotide (Fig. 1B, lanes 5-7, respectively), confirming the specific nature of the VDRE-BP-VDRE interaction.
VDRE-BP Is Distinct from the Dominant-negative-acting Response Element-binding Protein, yin-yang-1 (YY1)-The YY1 protein is a ubiquitous nuclear factor (36) with the potential to bind DNA and act as either an enhancer (37) or repressor (38) of transcription. There are YY1 recognition sequences within the 5Ј regulatory region of the osteocalcin gene that are known to bind the YY1 protein and squelch VDR-RXR␣-directed transcription (38). Fig. 1B also demonstrates that the VDRE-BP and the YY1 protein are functionally distinguishable from one another; an excess of an oligonucleotide harboring the consensus and a mutated YY1 recognition sequence (Fig. 1B, lanes 8 and 9, respectively), as well as anti-YY1 antibody (data not shown), were without effect on the VDRE-BP-VDRE complex; similar results were obtained when the VDRE-op was used as probe (data not shown).

Endogenous Squelching of VDR-VDRE-directed Transactivation in Vitamin D-resistant New World Primate Cells-If
VDRE-BP overexpression is wholly or partially responsible for vitamin D prehormone and hormone resistance in New World primates in vivo, then evidence of VDRE-BP's ability to blunt VDR-VDRE-directed transactivation of a vitamin-D regulated gene should be present in vitro in cells from hormone-resistant primates. In the representative experiments depicted in Fig. 2, cells from a vitamin D-resistant New World primate host (B95-8), but not cells from an Old World primate host (Vero), transfected with the VDRE-op-luciferase reporter construct were resistant to a 1,25-dihydroxyvitamin D-induced increase in reporter activity. The 7-fold increase in 1,25-dihydroxyvitamin D-directed reporter activity in vitamin D-responsive Old World primate Vero cells was absent in hormone-resistant B95-8 New World primate cells. Squelching of reporter activity in vitamin D-resistant B95-8 cells overexpressing VDRE-BP activity was also present in the basal state (i.e. without hormone induction); basal luciferase activity was at least a full order of magnitude less in New World primate cells than in Old World primate cells. In order to be certain that differences in endogenous VDR content between Vero and B95-8 cells was not a factor determining reporter activity, these experiments were repeated in the presence of co-transfected VDR (Fig. 2B). Co-transfection of the VDR did not alter the profound quelching of luciferase activity observed in vitamin D-resistant New World primate B95-8 cells.
Affinity Purification of VDRE-BP-Although the experiments described above suggested that VDRE-BP participated in the state of 1,25-dihydroxyvitamin D resistance in New World primate cells, proof that VDRE-BP is a participant in the vitamin D-resistant state in New World primate cells required demonstration that overexpression of the VDRE-BP would convert wild-type hormone-responsive Old World primate cells to a vitamin D-resistant phenotype. As a first step in the recapitulation of the resistant phenotype, we created a double-strand DNA affinity support bearing concatemers of the VDRE-op for purification of VDRE-binding proteins from New World primate cells. Coomassie-stained bands at 34 and 38 kDa were reactive in immunoblots probed with anti-hnRNPA1 (39) and hnRNPA2 antibody, respectively (Fig. 3A, middle panels). The elution positions of these proteins were distinct from the 55-kDa VDR present in the nuclear extracts of B95-8 New World primate cells (Fig. 3A, lower panel). The hnRNPA1-reactive 34-kDa species (henceforth referred to as VDRE-BP1) and the hnRNPA2-reactive 38-kDa species (henceforth referred to as VDRE-BP2) were excised from gels loaded with the 0.4 M KC1 affinity column fractions. The pooled fractions subjected to tryptic digestion and amino-terminal amino acid sequencing.
Three distinct internal tryptic peptides fragments were recovered from the 34-kDa VDRE-BP1 pool and four from the 38-kDa VDRE-BP2 pool (Fig. 3B). All seven tryptic peptides bore sequence similarity with proteins in the hnRNPA family. The sequence of two peptides each in the 34-kDa VDRE-BP1 pool and in 38-kDa VDRE-BP2 pool were closely related; this set of related peptides is found in the highly conserved RNA binding domain of hnRNPs (40). Collectively, the tryptic peptides in the 34-kDa VDRE-BP1 pool possessed a high degree of sequence similarity with proteins of the human hnRNPA1 subfamily, while peptides from the 38-kDa VDRE-BP2 pool bore the greatest sequence similarity with proteins in the hnRNPA2 subfamily (Fig. 3B).
VDRE-BP mRNA Is Overexpressed in Hormone-resistant New World Primate Cells-Using a panel of degenerate oligonucleotides corresponding to the two tryptic peptide sequences common to both VDRE-BP1 and VDRE-BP2 (see Fig. 3B), we amplified by RT-PCR a 435-bp cDNA that demonstrated 94% sequence identity to the cDNA for human hnRNPA2. This cDNA was used as probe in Northern analysis of RNA extracted from hormone-resistant New World primate B95-8 cells, from OMK New World primate cells with the intermediate, partially hormone-responsive phenotype, and from hormone-responsive Old World primate Vero cells (Fig. 3C). The probe hybridized to a 1.8-kilobase mRNA species that was plentiful in hormone-resistant B95-8 New World primate cells, was less plentiful in New World primate cells with the intermediate phenotype of partial hormone-responsiveness (OMK cells), and was expressed at a much lower level in Old World primate Vero cells with a hormone-responsive phenotype.
Affinity-purified VDRE-BP Binds Either Double-or Single-strand DNA Bearing the VDRE Recognition Sequence-hnRNPs are traditionally recognized as single-strand pre-mRNA-binding proteins (40,41). We theorized that VDRE-BPs would also be capable of binding single-strand DNA. Proof was provided by the experiments shown in Fig. 4. EMSA was performed with the 0.4 M KC1 affinity column fractions containing VDRE-BP1 and VDRE-BP2, with either single-strand VDRE-op or double-strand VDRE-op as probe and with either single-strand or double-strand radioinert response elementcontaining oligonucleotides as competitors for probe binding.
Using the top strand of VDRE-op as probe (Fig. 4A) (Fig. 4B). As expected, irrelevant sequences (i.e. CTF/NF, lanes 6 and 9), whether single-or double-strand in nature, had no effect on VDRE-BP binding to the VDRE. Molecular Cloning of VDRE-BP-Confirmation that the VDRE-BPs were members of the hnRNPA family was obtained by cloning cDNAs that encoded the entire open reading frames of VDRE-BP1 and VDRE-BP2. The 435-bp cDNA (see above) was employed to develop 5Ј-and 3Ј-nested primers for use in cloning by 5Ј-and 3Ј-RACE cDNAs for both VDRE-BPs that extended 5Ј through the translation start site and 3Ј through the untranslated region. The 1751-nucleotide VDRE-BP1 cDNA bore 61% sequence identity to the human hnRNPA1 cDNA, and the 1788-nucleotide VDRE-BP2 cDNA possessed 71% sequence identity to the human hnRNPA2 cDNA. Nucleotide sequence identity with human hnRNPA1 and hnRNPA2 was 96% through the open reading frame of the two cDNAs; most sequence divergence was accounted for in the extensive 3Ј-untranslated region of the two cDNAs. The translated portions of the cDNAs predicted proteins at 34 and 38 kDa, the same size of the anti-hnRNP-reactive proteins observed in immunoblots of eluents from VDRE affinity support (see Fig. 3).
Overexpression  wild-type cells that have been induced to overexpress the protein. In order to investigate this postulate, hormone-responsive, Old World primate Vero cells bearing a fully functional VDR were transiently co-transfected with the pCR3.1-VDRE-BP1, pCR3.1-VDRE-BP2, and VDRE-reporter constructs (Fig.  5). VDRE-reporter activity was unaffected by expression of pCR3.1-VDRE-BP1. On the other hand, basal VDRE-reporter activity was reduced 71% in pCR3.1-VDRE-BP2-transfected Vero cells compared with that observed in the same cells cotransfected with the empty PCR3.1 vector and VDRE-reporter constructs. In fact, basal reporter activity in pCR3.1-VDRE-BP2-bearing cells approached levels observed in hormone-resistant B95-8 cells also co-transfected with the empty pCR3.1 vector and the luciferase reporter construct. Stimulation of hormone-responsive Old World primate Vero cells carrying the empty pCR3.1 vector with 1,25-dihydroxyvitamin D 3 increased reporter activity 2.6-fold. Treatment of Vero cells bearing pRS-VDRE-BP2 with the vitamin D hormone elicited only a 1.3-fold increase in reporter activity to a level that was only 19% of that observed in hormone-stimulated, empty pCR3.1-bearing Vero cells and similar to that observed in hormone-resistant B95-8 cells, which overexpress VDRE-BP2 endogenously, confirming the squelching potential VDRE-BP2.

FIG. 2. VDRE-directed transactivation is squelched in the hormone-resistant New World primate cell line B95-8 either in the absence (panel A) or presence (panel B) of co-transfected VDR (0.5 g) and VDRE-op
Additional confirmatory evidence that VDRE-BP2 was the naturally occurring dominant VDR-VDRE squelching species in vitamin D-resistant New World primate cells was obtained by Western blot detection of an abundance of the anti-hnRNPA2-reactive VDRE-BP2 only in the B95-8 cell line derived from a vitamin-D-resistant New World primate (Fig. 6A) and by demonstration in EMSA of a VDRE-binding protein that was supershifted upon interaction with the VDRE-BP2cross-reactive anti-hnRNPA2 antibody (Fig. 6B). Recapitulation of the hormone-resistant phenotype in vitro was further demonstrated in VDR-VDRE-directed luciferase activity in wild-type Vero cells stably transfected with VDRE-BP2 (Fig. 7). Six VDRE-BP2-overexpressing cell lines were examined (Fig.  7A); compared with wide type cells without transfection, the six clones overexpressed VDRE-BP2 6 -17-fold. Compared with extracts from wild-type Vero cells, protein extracted from the nuclei of VDRE-BP2-overexpressing cells lines bound to VDRE probe in EMSA (Fig. 7B); VDRE-BP-probe binding in the stably transfected cell lines was similar to that observed in the vitamin D-resistant New World primate cell line from which the VDRE-BPs were initially cloned. VDRE-BP2-overexpressing clones were also assessed for their ability to squelch VDR-VDRE-directed luciferase activity (Fig. 7C). Among the four cell lines examined, luciferase activity was suppressed to 18 -27% of wild-type levels with the clone overexpressing VDRE-BP2 to the greatest degree (clone 4) exhibiting the most suppression of reporter activity.
Specificity of the VDRE-BP for Transactivational Regulation through the VDRE-We showed that ERE does not compete with the VDRE for binding VDRE-BPs (see Fig. 4). If this is the case, then constitutive overexpression of the VDRE-BPs in cells should exhibit no squelching of estrogen-ER-ERE-directed transactivation. Panel A of Fig. 8 confirms this postulate. Vero cells transiently transfected with either VDRE-BP1 or VDRE-BP2 did not inhibit luciferase activity driven off a promoter bearing an enhancing ERE, either in the absence or presence of an ER-saturating concentration of estradiol; in fact, VDRE-BP2 induced a modest increase in reporter activity. Specificity of VDRE-BP2 for squelching transactivation from a VDRE was also suggested by experiments performed in wild-type Old World primate COS-7 cells stably overexpressing another dominant-negative acting hnRNP, the estrogen-response elementbinding protein or ERE-BP ( Ref. 22; Fig. 8B). Compared with wild-type COS-7 cells, COS-7 cells stably overexpressing the ERE-BP and transfected with the VDRE-op-LUC reporter construct demonstrated no change in VDR-VDRE-directed luciferase activity. On the contrary, significant squelching of ER-ERE-directed transactivation was observed in COS-7 cells stably overexpressing ERE-BP1. These data 1) reinforce the concept that the VDRE-BPs do not normally interact with the ERE and 2) imply that modulatory effects of a response element-binding protein are specific for a single hnRNP-related species interacting with a specific response element. DISCUSSION Our knowledge of the rheostatic control of sterol/steroid hormone action has evolved considerably in the recent past. The earliest concepts of control focused on hormone synthesis, hormone metabolism (or catabolism), and hormone-receptor interaction in the target cell. In the vitamin D hormone system, target cell responsiveness is known to be governed by 1) the cutaneous, nonenzymatic photosynthesis of vitamin D (42); 2) the serial enzymatic modification of vitamin D to its active hormone, 1,25-dihydroxyvitamin D (43); and 3) interaction of the 1,25-dihydroxyvitamin D hormone with its nuclear receptor protein, the VDR (43). For example, without adequate sunlight exposure mammalian species, including Homo sapiens, fail to synthesize enough of the prohormone vitamin D to provide a sufficient supply of substrates to the vitamin D-25-hydroxylase and vitamin D-1-hydroxylase enzymes required for 1,25-dihydroxyvitamin D production. The resultant phenotype is failure to properly mineralize the skeleton. The same or similar skeletal phenotype is also seen in humans with inactivating mutations of the vitamin D-1-hydroxylase (44,45) or the VDR (46). Because they elicit the most dramatic changes in hormone effect and the most definitive phenotype, vitamin D prohormone synthesis, metabolism, and interaction of the hormone with VDR can be considered sites of "macroregulation" of vitamin D action.
However, there are additional factors in the hormone action schema that produce a much less subtle change in the phenotype. These factors serve as "microregulators" to fine tune hormone responses. In the vitamin D system, such microregulators include 1) the circulating vitamin D-binding protein, the shuttle protein in the serum that serves to deliver active vitamin D metabolites to target tissues for metabolism or action and carry away inactive metabolites for catabolism and excretion (47)(48)(49); and 2) the dimerization partner of the VDR (i.e. the RXR; Refs. 43 and 50), as well as the receptor-associated co-activators, co-repressors, and co-modulators of transcription (51). Another proposed microregulator in the vitamin D metabolism/action system is the newly discovered family of hsp70related intracellular vitamin D-binding proteins. These proteins, initially identified in vitamin D-resistant New World primate cells (52,53), have a high capacity and relatively high affinity (2-20 nM) for 25-hydroxylated vitamin D metabolites (54). Recent work 2 indicates that these proteins have the ability to facilitate hormone action by delivery of 1,25-dihydroxyvitamin D hormone to the nucleus and VDR, which has a higher affinity for hormone than the intracellular vitamin D-binding proteins.
Although the protein-protein interactions that occur among receptor proteins and various co-regulators may provide the most complex means of control of transcription, these interactions are not the most direct mode of transcriptional regulation. For example, even if assembled for maximal transcriptional effectiveness, the transactivating complex must be able to interact with a specific promoter element in order to achieve its transcriptional potential (55). If the VDR-RXR-directed tethering the transcriptional machinery to promoter elements via the VDRE is not possible because the cis element is occupied by another protein, then transcription will be thwarted although the complex is optimally suited to enhance transcription. The YY1 protein is one such protein with the potential to exert dominant-negative modulatory control over VDR-RXR-directed transcription through the VDRE (38). Guo et al. (38) have shown that YY1 protein can compete with VDR-RXR heterodimer for occupation of the VDRE in the osteocalcin promoter. We found that neither antibodies to the YY1 protein nor exposure to a 100-fold excess of the YY1 recognition sequence altered binding of New World primate cell VDRE-BPs to either the osteocalcin or the osteopontin VDRE (Fig. 1B). These results suggested that the VDRE-BPs in crude extracts of New World primate cells were not related to the YY1 protein family, 2 S. Wu, unpublished data. We have not yet determined independent affinity constants of the two known VDRE-BPs for the VDRE, but preliminary results 3 indicate that VDRE-BP2 possesses an avidity for the VDRE that rivals that of the VDR-RXR complex and harbors the same transcriptional squelching potential of YY1 (see Fig.  1B). Hence, when present in relatively small quantities in the nuclear compartment compared with the VDR and RXR, we propose that VDRE-BP2 has the potential to compete with VDR-RXR dimer for binding to the VDRE. When present in relatively large quantities as they are vitamin D-resistant New World primate cells, it is possible that VDRE-BP2 can substantially limit hormone-stimulated gene expression (Fig. 2). Consequently, we theorize that depending on their relative abundance in the target cell the VDRE-BPs have the potential to function as either a microregulator (i.e. when present in relative low abundance compared with VDR-RXR) or as a macroregulator when the VDRE-BP(s) overwhelm the VDR-RXR and its access to the VDRE.
hnRNPs are classically recognized as single-strand pre-mRNA-binding proteins (10). Recently proteins in this class have also been shown to possess the ability to bind DNA including double-strand DNA and to alter, principally by silencing, transcription (4 -8). The hnRNPA2-related VDRE-BP2 falls into this category. Affinity-purified VDRE-BP(s) will bind to either double-strand (Figs. 1 and 4) or single-strand (Fig. 4) DNA bearing a VDRE. The effectiveness of a single-strand oligonucleotide bearing the consensus RXRE (consisting of a direct repeat of the AGGTCA motif separated by only a single, not three, nucleotides) in competing for VDRE-BP binding suggests that the specificity for VDRE-BP binding resides in the 6-bp half-site motif (G/A)GGT(G/C)A; variations of this sixnucleotide motif encompasses nucleotide sequences present in the VDRE-op (GGGTCA), VDRE-oc (GGGTGA), and RXRE (AGGTCA) used in our DNA binding analyses. Binding specificity was confirmed by failure of the complementary (opposite) strand of the VDRE-op, VDRE-oc, and RXRE to compete for VDRE-BP binding when either single-strand or double-strand VDRE was employed as probe and by the ability of anti-hnRNPA1 antibody to disrupt VDRE-BP-VDRE binding (data not shown).
We recently described an hnRNPC-like protein, which competes with the ER for binding to the ERE (21,22). This protein, termed the ERE-BP, exerts a profound dominant-negative influence on ER-ERE-directed transactivation in vitro and in-duces estrogen resistance in vivo when overexpressed (22). The VDRE-BP2 described here appears to act in similar fashion with overexpression of this protein clearly squelching VDR-RXR-directed transactivation (Figs. 2, 5, and 7) via its ability to bind to the VDRE (Figs. 6 and 7). The question remains whether VDRE-BP2 is absolutely specific for the VDRE-BP or whether it might interact with other hormone response elements as well. To address this question, we examined the ability of VDRE-BP2, to alter transactivation directed by a ERE (Fig. 8A). It did not. We also queried whether transient overexpression of the ERE-BP acted to squelch VDRE-directed transactivation. Although a number of separate experiments showed no substantial or consistent effect with transient overexpression of the ERE-BP (data not shown). Stable overexression of the ERE-BP in wild-type cells had no effect on VDREdirected reporter activity while ERE-BP significantly squelched ERE-directed transactivation. These findings and the observation that squelching of VDRE-directed transactivation in cells transiently (Fig. 5) and in cell lines stably overexpressing VDRE-BP2 (Fig. 7) approximated that observed in vitamin D-resistant New World primate cells, which endogenously overexpress VDRE-BP2 and VDRE-BP1, indicates that VDRE-BP2 is the principal dominant-negative agent active at the VDRE.
The fact that the hnRNPA1-related VDRE-BP1 identified here had VDRE binding potential (Fig. 3A) but was without significant squelching potential (Fig. 5) was surprising. It is possible that the affinity of VDRE-BP1 for the VDRE is much lower, hence its ability to remain on the cis element and interfere with VDR-RXR binding is diminished compared with VDRE-BP2. The reduced but still existing affinity of VDRE-BP1 for the VDRE may also be an explanation why VDREdirected transactivation of a reporter cell lines overexpressing VDRE-BP2 at levels observed in hormone-resistant New World primate cells was not completely squelched (Fig. 7C). If present in sufficient quantity, as VDRE-BP1 appears to be in vitamin D-resistant New World primates cells (Fig. 3A), VDRE-BP1 may compete with VDRE-BP2 for binding to the VDRE. It is also likely that squelching of transcription requires interactions with other proteins in the transcriptional complex. As a consequence it is possible that the reason VDRE-BP1 is ineffective as regulator of transcription (Fig. 5) is because VDRE-BP1 in incapable of interacting with the transcriptional machinery, while VDRE-BP2 is capable of interacting with those proteins.
Are there in vivo equivalents to the microregulatory and macroregulatory phenotypes induced by the VDRE-BP(s) among primate genera? In subhuman primates the answer is yes. The partially hormone-responsive New World primate ge-3 H. Chen, unpublished data. FIG. 8. VDRE-BP2 is a specific inhibitor of transativation initiated through the VDRE. Panel A exhibits luciferase activity in wild-type, vitamin D-responsive Vero cells before (Ϫ) and after transient co-transfected with pCR3.1-VDRE-BP1 (BP-1) or pCR3.1-VDRE-BP2 (BP-2) and the ERE-LUC reporter plasmid and not exposed (left panel) or exposed (right panel) to 100 nM estradiol. Each value is the mean Ϯ S.D. of triplicate determinations of reporter activity; significant differences assessed by Student's t test are shown. Panel B shows hormone stimulated reported activity in COS-7 cell stably transfected with pCR3.1-ERE-BP1 and transiently co-transfected with either the VDRE-op-LUC (left panel) or the ERE-LUC (right panel) reporter plasmid. The asterisks denote a significant difference (p Յ 0.01) in reporter activity (mean Ϯ S.D. of triplicate values) compared with control COS-7 cells not bearing the ERE-BP.
nus Aotus and the hormone-resistant New World primate genus Callithrix (52) are respective examples of the micro-and macroregulatory phenotype. Because Aotus cells express relatively little mRNA (Fig. 3C) and even less protein for the VDRE-BPs, this genus possesses circulating 1,25-dihydroxyvitamin D levels similar to those of Old World primates and a clinical phenotype that is difficult to detect even when challenged with sunlight deprivation (19,20). On the other hand, Callithrix, which expresses high levels of the VDRE-BP mRNA and protein (Figs. 3C and 6A), possesses very high serum levels of 1,25-dihydroxyvitamin D and an extreme propensity to develop rickets when cutaneous vitamin D synthesis is challenged (19,20,52). Are there human equivalents to the microregulatory and macroregulatory phenotypes caused by the VDRE-BPs in subhuman primates? Because the phenotype that would likely result from a relatively low level VDRE-BP expression is likely to be subtle or nonexistent in the basal (unchallenged) state, it is highly unlikely that we have yet identified the human counterpart of the microregulatory phenotype. However, this may not be the case with the macroregulatory phenotype as exemplified by Callithrix. There are human families described with the classical phenotype of vitamin D-resistant rickets that harbor high serum concentrations of the active vitamin D metabolite 1,25-dihydroxyvitamin D but no recognized mutation in the VDR (56). It is possible that these subjects may suffer from a disturbance resulting in the constitutive overexpression of a protein(s) that interacts with the VDRE. 3