Suppression of the Human Parathyroid Hormone Promoter by Vitamin D Involves Displacement of NF-Y Binding to the Vitamin D Response Element*

An earlier report in the literature indicated the vitamin D response element (VDRE) in the human parathyroid hormone (hPTH) promoter could be specifically bound by an unidentified transcription factor in addition to the vitamin D receptor (VDR) complex. We confirmed that OK and HeLa cell nuclear extracts formed a specific complex with the hPTH VDRE that was insensitive to competition with other VDRE sequences. However, this factor could be competed for by a consensus NF-Y DNA-binding site, and an anti-NF-Y antibody was able to supershift the bound band. Mutational analysis indicated that the NF-Y-binding site partially overlapped the 3′ portion of the VDRE. Transfection studies using an hPTH promoter construct in Drosophila SL2 cells demonstrated strong synergistic transactivation by NF-Y interactions with both the VDRE site and a previously described distal NF-Y-binding site. Finally, mobility shift studies indicated that the VDR heterodimer competed with NF-Y for binding to the VDRE sequence, and NF-Y-stimulated activity of the hPTH promoter could be suppressed in a hormone-dependent manner when the VDR heterodimer complex was coexpressed in SL2 cells. In summary, these findings establish the presence of a proximal NF-Y-binding site in the hPTH promoter and highlight the potential for synergism between distal and proximal NF-Y DNA elements to strongly enhance transcription. Furthermore, findings suggest that the repressive effects of vitamin D on hPTH gene transcription may involve displacement of NF-Y binding to the proximal site by the VDR heterodimer, which subsequently attenuates synergistic transactivation.

Despite the clear importance of parathyroid hormone (PTH) 1 in the maintenance of calcium homeostasis, little is known about the activating and repressing factors that control the transcription of this peptide hormone. A conserved cyclic AMP response element in the promoters of the human, bovine, and more recently murine PTH genes has been described previously (1)(2)(3). Repressor DNA elements for the VDR have been identified in the human, bovine, rat, and chicken PTH promoters (4 -7), although an unknown transcription factor was reported to also bind to the human VDRE (8). Negative calcium response elements have been identified in the hPTH promoter and appear to involve interactions with apurinic/apyriminidic endonuclease/redox factor 1 (Ape1/Ref1) (9,10). In addition, a handful of proteins have recently been shown to be important in various stages of parathyroid gland (PTG) development (11)(12)(13). Glial cells missing 2 (Gcm2) is a transcription factor in which expression is largely restricted to parathyroid tissue (14), and mice lacking the Gcm2 fail to develop parathyroid glands (15). A homozygous mutation in the human homolog glial cells missing B (GCMB) was identified as the likely cause of a case of isolated familial hypoparathyroidism (16).
We initiated studies of the PTH gene by conducting a DNA sequence comparison of different mammalian promoters to identify conserved regions that might harbor binding sites for transcription factors involved in regulating promoter activity. In this analysis a highly conserved Sp1 DNA enhancer element was detected in conjunction with high levels of expression of Sp3 and Sp1 in bovine PTG cells (17). More recently we reported finding an NF-Y-binding site unique to the hPTH promoter that partially overlapped the aforementioned Sp1 element and functioned as a potent enhancer of PTH gene transcription (18). Results of the present study demonstrate that a second, proximal NF-Y-binding site is present in the hPTH promoter and partially overlaps with the repressor VDRE. Furthermore, NF-Y binding to both distal and proximal binding sites in the hPTH promoter results in synergistic enhancement of gene activity. The data suggest that one mechanism to account for repression of the hPTH promoter by vitamin D is through displacement of NF-Y binding to the proximal DNA element.
Preparation of Nuclear Extracts-Cultured HeLa and OK cells were maintained in Dulbecco's modified Eagle's medium/F-12 (1:1) solution with 10% charcoal-stripped fetal bovine serum containing penicillin (100 units/ml) and streptomycin (100 g/ml) at 37°C. The cells were detached with trypsin, washed with phosphate-buffered saline at ϳ4°C, incubated on ice in 3 volumes of a similarly cold low salt buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM EDTA, 2.0 mM dithiothreitol, 10% glycerol, and 1ϫ protease inhibitor mixture) for 20 min followed by cell disruption with a Teflon Dounce homogenizer. Following a 30-min spin at 100,000 ϫ g, the supernatants were removed, and the nuclear pellets were resuspended in 1 volume of cold high salt buffer (same as above with 400 mM KCl) and incubated on ice for 30 min with occasional gentle mixing. Samples were then spun at 100,000 ϫ g for 30 min. The supernatant fractions were collected, aliquoted into individual tubes, snap-frozen, and stored at Ϫ70°C prior to use. Extracts of recombinant human VDR (hVDR) and human RXR␣ (hRXR␣) were prepared as described previously (20).
Electrophoretic Mobility Shift Assay-The double-stranded oligonucleotide probe for the hPTH VDRE possessed overhanging BglII/XhoI ends and was radiolabeled using Klenow fragment (exo Ϫ ) and [␣-32 P]dATP (3000 Ci/mmol) (PerkinElmer Life Sciences). The radiolabeled DNA fragment was gel-purified prior to use in binding reactions, which were assembled as described previously (20). Briefly, indicated amounts of nuclear extracts were added to a binding solution (20 l final volume) (buffer components were 120 mM KCl, 20 mM Tris, pH 7.5, 1.5 mM EDTA, 2 mM dithiothreitol, 5% glycerol, 0.5% CHAPS, 10 mM NaF, 100 M Na 3 VO 4 , 1.0 g dI⅐dC, 100 M leupeptin) for 30 min at 4°C. Recombinant hVDR and hRXR␣ extracts were normally diluted 1:25 in KTEDG-400 buffer (400 mM KCl, 20 mM Tris, pH 7.5, 1.5 mM EDTA, 2 mM dithiothreitol, 5% glycerol) prior to use. Where indicated, samples were incubated with antiserum for 30 min prior to addition of the radiolabeled DNA probe. For DNA competition experiments, the excess (400-fold) unlabeled oligonucleotide was also allowed to incubate with the sample for 30 min prior to addition of the radiolabeled DNA probe. Vitamin D hormone (final concentration, 250 nM) was added to all samples that included the VDR heterodimer in the binding reaction. Following a 30-min incubation at 4°C with the radiolabeled DNA probe, samples were loaded onto prerun 5% polyacrylamide gels (29:1), and electrophoresis was performed at ϳ14 V/cm for ϳ2 h 20 min with buffer cooling. Gels were transferred and dried, and autoradiography was performed.
Interference Footprinting-The ethylation interference footprint experiments were performed as described previously (20). Briefly, the hPTH VDRE was subcloned into the BamHI/SalI sites of pTZ19R (Fermentas, Hanover, MD) and excised with the combination of HindIII/ EcoRI for the footprinting protocol. Singly, 32 P-end-labeled DNA probes in 50 mM sodium cacodylate buffer, pH 8.0, were treated with ethylnitrosourea-saturated ethanol for 20 min at 55°C. After precipitation with sodium acetate/ethanol and reprecipitation (3ϫ), the pellets were washed with 70% ethanol, dried, and resuspended in water. Ethylated probes were then used in the gel mobility shift assay as described above, except the amounts of probe were increased 3-4-fold. Following electrophoresis, the wet gels were exposed to x-ray film overnight at 4°C. Acrylamide sections corresponding to bound and free DNA were excised, and the DNA was recovered by electrochemical elution and precipitation. Cleavage of the modified DNA was accomplished by treating it with 100 mM NaOH, 0.1 mM EDTA in 10 mM phosphate buffer at 95°C for 30 min followed by neutralization with 3 M sodium acetate, pH 5.2, and precipitation with ethanol. Samples were separated through 8% sequencing gels and dried, and autoradiography was performed.
PCR and Preparation of hPTH Promoter Constructs-The hPTH promoter (Ϫ177 to ϩ21) luciferase reporter plasmid (hPTHp/luc) has been described previously (18). Selective mutation of NF-Y-binding sites within the context of the wild-type hPTH promoter was accomplished by two-step PCR using oligonucleotides containing mutant NF-Y elements and the primers hPTHPromF (5Ј-ATGGATCCAATTA-TCTGAAACTTAAGAAGA-3Ј) and hPTHPromR (5Ј-ATCTCGAGACA-ACTGATGAATTGGACTGCA-3Ј). For example, to inactivate the distal NF-Y site, two separate PCR reactions were performed to generate mutant promoter fragments with overlapping ends. One PCR reaction consisted of hPTHPromF primer and HP-R-NFY-dismut, (5Ј-ACACAC-ACCCACGGGGCGGTGCACACTCTT-3Ј); the second reaction consisted of hPTHPromR and HP-F-NFY-dismut (5Ј-GTGCACCGCCCCGTGGG-TGTGTGTATGTGC-3Ј) (mutations are italicized and underlined). PCR was performed by initial denaturation at 94°C for 3 min and then for 30 cycles as follows: 94°C for 30 s, 56°C for 15 s, and 72°C for 45 s. Following isolation of the two separate PCR products, aliquots of each were mixed, and PCR was performed again using the hPTHPromF and hPTHPromR primers under conditions described above to generate the distal mutant NF-Y hPTH promoter fragment. The mutant promoter was isolated, digested with BamHI/XhoI, and ligated into the same sites of the luciferase reporter. Mutation of the proximal NF-Y element was similarly prepared and involved the two mutant primers HP-F-NFYproxmut (5Ј-GTGCTGCTTTGAACCTAGCCTTGAGATCC-3Ј) and HP-R-NFY-proxmut (5Ј-GGATCTCAAGGCTAGGTTCAAAGCAGCAC-3Ј) (mutations are italicized and underlined). Simultaneous mutation of both NF-Y-binding sites was accomplished by using the NF-Y-distmut promoter construct as the template for PCR reactions using the NF-Yproxmut primers as outlined above. All mutant promoters were subjected to manual sequencing analysis to verify sequence identity.
Preparation of Vectors for Expression of VDR Heterodimer in Drosophila SL2 Cells-In previous work, the plasmid pTZ19R had been digested with BamHI/HindIII to remove the multiple cloning site and was replaced with a double-stranded oligonucleotide possessing compatible overhanging ends, 5Ј-GATCCCATATGATCGATATCCTCGAG-A-3Ј (top strand). The hVDR expression vector CMV-hVDR (generously provided by Dr. L. Freedman, Merck Research Laboratories, West Point, PA) had been digested with NdeI/EcoRV to release the hVDR coding region, which was isolated and subcloned into the same sites of the modified pTZ19R plasmid. For the present study, the modified pTZ19R/hVDR plasmid was digested with BamHI/HindIII, and the hVDR cDNA was cloned into the same sites of pBluescript II (Stratagene, La Jolla, CA). The hVDR cDNA was recovered following digestion with BamHI/XhoI. pPacSp1 (generously provided by Dr. G. Suske, Marburg, Germany) was digested with the same enzyme pair to excise the Sp1 cDNA and replace it in-frame with the hVDR sequence.
The hRXR␣ coding region was amplified by PCR (BD Advantage HF 2 PCR kit (BD Biosciences) using primers 5Ј-ATAGATCTGTCGCAGA-CATGGACACCAAAC-3Ј and 5Ј-TATCTCGAGCTAAGTCATTTGGTG-CGGCGC-3Ј and CMV-hRXR␣ as a template (21). Amplification was performed by initial denaturation at 94°C for 1 min followed by 30 cycles of 94°C for 15 s, 58°C for 15 s, and 68°C for 2 min. The amplified product was digested with BglII/XhoI, isolated, and ligated into the BamHI/XhoI sites of pPac as indicated above. DNA sequencing and Western blotting were performed on both pPacVDR and pPacRXR␣ to confirm in-frame connection and expression of receptor proteins (data not shown).
Transient Transfection-Drosophila SL2 cells were maintained in Drosophila SL2 medium supplemented with 10% fetal bovine serum at 27°C. Cells were distributed in 24-well plates on the day before transfection and transfected in triplicate with the indicated hPTH promoter luciferase reporter construct (100 ng) p97b-␤-galactosidase expression vector (50 ng), the indicated pPac expression vectors, and carrier plasmid DNA made up to 500 ng of DNA/well. SL2 cells were transfected using Cellfectin (4 l/well) (Invitrogen) for 3 h in medium lacking serum followed by supplementation to 7% serum. Where indicated, vitamin D hormone was also added to a final concentration of 100 nM. After 42 h, lysates were prepared by washing the cells twice with phosphatebuffered saline followed by overlaying with lysis buffer and two rounds of freeze thawing. Luciferase activities from individual wells were determined, normalized with respect to values for ␤-galactosidase enzymatic activity and average values calculated Ϯ S.E. from samples transfected in triplicate. Results are reflective of at least two independent experiments.

RESULTS
Previous work indicated that a transcription factor unrelated to the VDR could also specifically bind to the hPTH VDRE (8). This factor was reportedly present in a variety of cell lines and appeared to bind to the 3Ј half of the VDRE probe. To begin an analysis of this binding factor, an hPTH VDRE oligonucleotide probe was synthesized (Fig. 1A) that contained the previously described half-site sequence (4) but omitted much of the distal portion of a proposed direct repeat element (22). HeLa and OK cell nuclear extracts that both lack significant expression of the VDR were then assessed for binding activity with this sequence. A complex was observed with both of these extracts that exhibited self-competition with the addition of excess unlabeled hPTH VDRE oligonucleotide but was not competed for with an excess of either hOC or avian PTH VDREs or a consensus Sp1 element (Fig. 1, B and C). However, the addition of a consensus NF-Y DNA element completely prevented formation of the observed binding complex in both extracts (Fig. 1, B and C, lane 6). Accordingly, inclusion of an antibody against the B subunit of the NF-Y complex was able to supershift this complex, whereas normal serum had no effect (Fig. 1D). The same results were obtained using bovine PTG nuclear extracts (data not shown). Thus, we concluded that NF-Y could bind specifically to the hPTH VDRE.
To further characterize binding by NF-Y and VDR complexes to this DNA sequence and because of conflicting data surrounding the identity of the VDRE and nature of VDR interactions with the hPTH promoter (4,22,23), it was necessary to investigate VDR binding with our oligonucleotide sequence. As seen in Fig. 2A, the addition of increasing amounts of either recombinant hVDR-or hRXR␣-containing extracts to the hPTH VDRE probe failed to produce a bound complex in the mobility shift assay. However, incubation with a combination of the two nuclear receptor extracts produced a strong bound complex in this analysis. To assess the specificity of this interaction, a mixture of the receptor extracts was incubated with an excess of different unlabeled competitor oligonucleotides. An estrogen response element failed to displace the complex (Fig. 2B, lane  1), whereas other known VDREs were able to compete for the bound band (Fig. 2B, lanes 2-4), unlike the data obtained with the bound NF-Y complex (compare with Fig. 1). The presence of both the hVDR and hRXR␣ in the bound complex was confirmed when they were incubated with the individual receptor antisera, whereas normal serum did not disturb the bound band (Fig. 2B, lanes 5-7). Thus, binding by the VDR to the hPTH VDRE oligonucleotide was specific and as noted earlier (22) required the presence of RXR.
To localize the DNA-binding interaction, an ethylation interference experiment was performed using the recombinant heterodimer and hPTH VDRE fragment. As seen in Fig. 2C, strong interference was observed over the top strand sequence 5Ј-TGA, whereas on the lower strand the interference was limited to the region 5Ј-GGTT (Fig. 2D). Thus, the footprinting data support the previous observation that the repressor VDRE is localized to the half-site sequence (4) 5Ј-TGAACC (top strand), although additional weak contacts to other nucleotides outside of this sequence cannot be ruled out.
A series of mutant oligonucleotides were then used as unlabeled DNA competitors in binding experiments to compare and contrast the location of the VDR heterodimer and NF-Y-binding sites in the hPTH VDRE (Fig. 3). In Table I, Mutant 1, which disrupted the 3Ј-end of the VDRE, prevented competition for both VDR heterodimer and NF-Y binding, indicating that both factors were sensitive to changes in this part of the sequence. Mutations farther toward the 3Ј-end of the sequence (Mutant 2) prevented competition for NF-Y binding but resulted in a sequence that could now compete for VDR heterodimer binding. Mutant 3 at the 5Ј-end of the sequence was equally adept at displacing either VDR heterodimer or NF-Y proteins from binding, suggesting that neither complex was sensitive to this region of DNA. Finally, Mutant 4 within the VDRE could not compete for the VDR heterodimer complex but was still capable of displacing NF-Y. Thus, the different competitors localized the binding site for the NF-Y complex to the imperfect CAAT-like sequence, 5Ј-ACCTATAG, partially overlapping and extending 3Ј from the VDRE. Meanwhile, the VDR heterodimer was sensitive to changes in the previously recognized half-site VDRE (4), which was consistent with the footprinting data (see Fig. 2).
The above experiments suggested that competition between the VDR and NF-Y complexes might be occurring for binding to the hPTH VDRE. To assess this possibility, mobility shift assays were pursued using OK cell nuclear extracts containing the NF-Y complex together with increasing amounts of added recombinant heterodimer proteins. As seen in Fig. 4, control binding with the OK cell nuclear extract produced the expected strong NF-Y complex with the hPTH VDRE. When increasing amounts of recombinant VDR/RXR␣ extracts were added, two distinct sets of complexes were observed, NF-Y and the VDR heterodimer. As more of the heterodimer was added to the binding reactions, there was a corresponding increase in the observed bound VDR complex together with a concomitant decrease in the amount of bound NF-Y. There was no evidence of higher order structures that would be suggestive of simultaneous binding to the same DNA probe by both factors. Thus, the data are consistent with competitive binding by these factors for their respective overlapping binding sites contained within this DNA element.
Previous data from our laboratory (18) had identified an NF-Y-binding site, hereafter referred to as NF-Y dist , which is unique to the hPTH promoter and which partially overlaps the conserved Sp1 DNA element found in a variety of mammalian PTH genes. Strong activation of the hPTH promoter by NF-Y expression was observed in that study, which raised the possibility that this might result from interactions between NF-Y complexes bound to the two sites separated by ϳ30 bp. To explore this issue, NF-Y activation of the wild-type hPTH promoter (hPTHp/luc, see Ref. 18) was compared with analogous mutant promoter constructs that selectively inactivated binding to either NF-Y dist or the newly identified proximal NF-Ybinding site NF-Y prox . A third mutant promoter reporter that simultaneously inactivated binding to both distal and proximal NF-Y elements was also constructed. The transient transfections were carried out in Drosophila SL2 cells that lack endogenous expression of mammalian transcription factors (24) and thus provide a null background for assessing NF-Y interactions with various promoter elements (18,(25)(26)(27)(28). As seen in Fig. 5, inclusion of expression vectors for the NF-Y complex resulted in a 38-fold induction of wild-type hPTH promoter activity. When either of the two NF-Y-binding sites was individually mutated, there was an order of magnitude decrease (to Ͻ4-fold) in the capacity of NF-Y to stimulate the reporter gene. Furthermore, no enhancement in activity was observed for the promoter containing mutations in both NF-Y-binding sites in response to co-transfection of the NF-Y complex. Thus, the data are consistent with interactions between NF-Y complexes occupying the two DNA elements to drive strong, synergistic transactivation of the hPTH promoter, whereas occupancy of either site alone produced a much more muted enhancement of gene activity.
Based on the ability of the VDR heterodimer complex to compete with NF-Y for binding to the VDRE sequence (Fig. 4), the next set of experiments sought to evaluate whether the VDR complex could suppress NF-Y-stimulated transcriptional activity in SL2 cells. To test this possibility, vectors for expression of hVDR and hRXR␣ in Drosophila SL2 cells were prepared and simultaneously co-transfected in increasing amounts into cells expressing a constant amount of NF-Y complex and the hPTHp/luc reporter. As seen in Fig. 6A, a 1:1 ratio of transfected NF-Y complex and VDR heterodimer expression vectors resulted in a 69% decline in reporter activity in response to hormone. The repression grew to 86% at a 1:3 ratio (NF-Y/VDR heterodimer), which was maintained at the highest (1:10) amount tested. To confirm that this effect was dependent on expression of the VDR heterodimer, individual receptor vectors were analyzed in cells again transfected with the NF-Y complex and hPTHp/luc. As seen in Fig. 6B, expression of either hVDR or hRXR␣ alone had minimal effect on NF-Y-  stimulated gene activity from the hPTH promoter. However, when both VDR and RXR␣ were simultaneously expressed in SL2 cells, hormone treatment again caused a ϳ80% decrease in reporter activity.

DISCUSSION
The present data affirm an earlier study (8) of an unidentified transcription factor binding to the hPTH VDRE. The identification of this factor as NF-Y is consistent with the previous report describing the presence of the factor in a variety of mammalian cell lines, and there is substantial overlap in the binding sites identified in the former and present studies. However, the possibility cannot be excluded that the factor noted in the previous report is distinct from the NF-Y complex. For example, mutation of the first two nucleotides of the core binding site identified in the earlier report, 5Ј-TGAACCTAT, appeared to have a significant impact on DNA-binding by the unknown factor but had no effect on NF-Y binding in the mutant oligonucleotide analysis of the present study (see Table  I). Thus, although the data presented here clearly support the existence of an imperfect NF-Y-binding site, 5Ј-ACCTATAG, which partially overlaps with the repressor VDRE in the hPTH promoter, it remains to be seen whether NF-Y is the previously noted unidentified transcription factor.
We identified NF-Y dist as unique to the hPTH promoter in an earlier report (18), and the present study now extends those results to establish the existence of a second NF-Y-binding site ϳ30 bp downstream from the former DNA element. A number of mammalian promoters exhibit multiple NF-Y-binding sites, and many of these genes appear to be regulators of the cell cycle (29,30). It is noteworthy that strong transcriptional synergism with the wild-type hPTH promoter was observed for NF-Y expressed in SL2 cells relative to mutant promoters, inactivating either one or the other NF-Y-binding site (Fig. 4). However, in the earlier report we observed that co-transfection of a dominant negative isoform of the A subunit of NF-Y had a more modest effect on basal hPTH promoter activity in OK cells (18). Thus, the synergism observed in the SL2 cells may be an anomaly of the insect line, or conversely, NF-Y activity is restricted in some manner in OK cells, perhaps involving p300 (30), despite the ready presence of an NF-Y-binding complex in nuclear extracts (Fig. 1). Nevertheless, the presence of NF-Y in PTGs (18), together with the current SL2 cell data, highlights the potential for synergism between NF-Y complexes bound to the two DNA elements to strongly enhance transcription of the hPTH promoter.
Binding by the VDR to the hPTH VDRE (Fig. 2) was limited to the previously identified DNA half-site element (4), although binding also required heterodimerization with RXR (22). How this is reconciled with a three-dimensional arrangement of heterodimer proteins and DNA is not known at this time. It is well established that the VDR heterodimer functions to acti- vate transcription from an ever-growing number of gene promoters, often utilizing VDREs represented as direct repeats separated by three base pairs (31). However, we have not observed vitamin D-dependent transcriptional enhancement of reporter activity using the wild-type hPTH promoter construct in mammalian or insect cells (data not shown). Therefore, the strong interference footprint over the half-site repressor VDRE suggests that this may be a means of positioning the heterodimer complex in such a way as to assume the role of a ligand-dependent repressor as opposed to a transcriptional enhancer.
The present report also demonstrated competitive DNAbinding interactions between NF-Y and the VDR complex for the hPTH VDRE. Competition for binding with other transcription factors has been implicated in the repressive effects of vitamin D on other target gene promoters. For example, a negative VDRE capable of binding the heterodimer was identified in the rat bone sialoprotein gene promoter that overlapped a unique inverted TATA box (32). The authors suggested that competition between the VDR complex and TATA-binding proteins might be a plausible explanation for the repression observed with this promoter. Other transcription factors unrelated to the VDR complex were also observed specifically binding to the repressor VDRE from the Runtrelated transcription factor 2 promoter (33). Although the identity and transcriptional activities of the factors were not determined, they appeared to recognize some overlapping portion of the VDRE, which was distinct from competitor sensitivity exhibited by the VDR complex. The VDR also competed for binding with nuclear factor of activated T cells, an enhancer of granulocyte-macrophage colony-stimulating factor gene transcription (34). Additionally, when the VDR bound to this DNA element it also stabilized binding by the Jun/Fos complex to an adjoining AP-1 binding site. The VDR interaction was mediated through the Jun protein, and this led to a composite complex that attenuated the activated transcription of granulocyte-macrophage colony-stimulating factor. Fig. 7A outlines identified transcription factor binding sites in the region from Ϫ160 to Ϫ60 of the hPTH promoter. The VDRE/NF-Y-binding sites are highlighted and clearly indicate the overlapping nature of the two DNA elements. A potential mechanism to at least partially account for the repressive ef-fects of vitamin D on hPTH gene transcription is also shown (Fig. 7B). NF-Y binding to both distal and proximal elements would result in strong enhancement of hPTH promoter activity. Displacement of NF-Y from the proximal element by the liganded VDR heterodimer complex can be envisioned disrupting synergistic transactivation by NF-Y and strongly attenuating promoter activity. Although the model provides a basis for future studies, there are still many unanswered questions surrounding the role of NF-Y and the hPTH promoter. For example, does NF-Y contribute to basal activity in the PTG, or are there external factors, such as hypocalcemia, that selectively enhance NF-Y interactions with the promoter? NF-Y is known to synergize with Sp1 to enhance gene transcription (27,35,36), suggesting that the proximal NF-Y-binding site in the hPTH promoter may also be capable of interacting with Sp proteins bound to the upstream Sp1 DNA element. Finally, renal insufficiency is often accompanied by the development of secondary hyperparathyroidism, which has been associated with decreased vitamin D activity in the PTG (37)(38)(39)(40)(41)(42). Our data imply that unopposed NF-Y transactivation of the human promoter may be contributing to this condition; therefore drugs that attenuate NF-Y transcriptional activity may be candidates for use in treating this disorder (43,44). Additional studies are essential to assess these and other possible roles of NF-Y in regulating hPTH promoter activity. FIG. 7. Transcription factor interactions with the hPTH promoter. A, a schematic is shown of the hPTH promoter with identified DNA elements indicated. The sequence corresponding to the VDRE (overhead double arrow) and NF-Y sites (underline double arrow) is shown. CRE, cAMP response element. B, binding by NF-Y complexes to the distal and proximal elements results in strong, synergistic transactivation. Binding by the VDR heterodimer to the VDRE displaces the proximal NF-Y complex and disrupts the synergism between the two NF-Y molecules. ϩϩϩϩϩ, synergistic transactivation; ϩ, weak transactivation.