Cloning of the human phospholipase C-gamma1 promoter and identification of a DR6-type vitamin D-responsive element.

The 5′-flanking region of the human phospholipase C-γ1 gene was isolated from a human P1 genomic DNA library. The S1-nuclease mapping and primer extension analysis revealed that there is a single transcriptional start site located at 135 bases upstream from the translation start codon in the human phospholipase C-γ1 gene. DNA sequence analysis showed that the sequence around the transcriptional start site is very GC-rich and has no TATA box. The fragment +135 to −877 in the 5′-flanking region of the human phospholipase C-γ1 gene was subcloned into a luciferase reporter vector. The chimeric gene produced a high level of luciferase activity and responded to 1,25-(OH)2D3 in transiently transfected human keratinocytes. Deletion and mutation studies of the fragment +135 to −877 demonstrated a vitamin D-responsive element that contains a motif arranged as two direct repeats separated by 6 bases (DR6), AGGTCAgaccacTGGACA, located between −786 and −803 base pairs. Incubation of the oligonucleotide containing the DR6 with keratinocyte nuclear extracts produced a specific protein-DNA complex that shifted to a higher molecular weight form upon the addition of an antibody specific to the 1,25-(OH)2D3 receptor. Therefore, the 5′-flanking region of the human phospholipase C-γ1 gene confers promoter activity and contains a DR6-type vitamin D-responsive element that mediates, at least in part, the enhanced expression of this gene in human keratinocytes by 1,25-(OH)2D3.

The 5-flanking region of the human phospholipase C-␥1 gene was isolated from a human P1 genomic DNA library. The S1-nuclease mapping and primer extension analysis revealed that there is a single transcriptional start site located at 135 bases upstream from the translation start codon in the human phospholipase C-␥1 gene. DNA sequence analysis showed that the sequence around the transcriptional start site is very GC-rich and has no TATA box. The fragment ؉135 to ؊877 in the 5-flanking region of the human phospholipase C-␥1 gene was subcloned into a luciferase reporter vector. The chimeric gene produced a high level of luciferase activity and responded to 1,25-(OH) 2 D 3 in transiently transfected human keratinocytes. Deletion and mutation studies of the fragment ؉135 to ؊877 demonstrated a vitamin D-responsive element that contains a motif arranged as two direct repeats separated by 6 bases (DR6), AGGTCAgaccacTGGACA, located between ؊786 and ؊803 base pairs. Incubation of the oligonucleotide containing the DR6 with keratinocyte nuclear extracts produced a specific protein-DNA complex that shifted to a higher molecular weight form upon the addition of an antibody specific to the 1,25-(OH) 2 D 3 receptor. Therefore, the 5-flanking region of the human phospholipase C-␥1 gene confers promoter activity and contains a DR6type vitamin D-responsive element that mediates, at least in part, the enhanced expression of this gene in human keratinocytes by 1,25-(OH) 2 D 3 .
Phospholipase C (PLC) 1 is a family of isoenzymes that cleave phosphatidyl inositol bisphosphate to two second messengers, inositol triphosphate and diacylglycerol, in response to a transmembrane signal (1,2). Diacylglycerol is the physiological activator of protein kinase C, and inositol triphosphate causes the release of calcium from the endoplasmic reticulum. PLCs can be divided into three types (PLC-␤, PLC-␥, and PLC-␦), and each type contains several subtypes (3,4). PLC-␥1, unlike the other PLC isoenzymes, contains a src homology 2 domain through which PLC-␥1 interacts with various tyrosine kinase growth factor receptors (5)(6)(7)(8). PLC-␥1 is overexpressed in primary human breast carcinoma (9), human colorectal cancer (10), familial adenomatous polyposis (11), and hyperprolifera-tive epidermal diseases (12). The amount of PLC-␥1 protein is higher in neoplastic keratinocyte cell lines than in normal keratinocytes (13). Calcium-induced differentiating keratinocytes express over 2-fold more PLC-␥1 protein than undifferentiated keratinocytes (14). These observations suggest that PLC-␥1 might be involved in the regulation of cell proliferation and differentiation.
The differentiation of normal human keratinocytes is induced by extracellular calcium and 1,25-(OH) 2 D 3 (15)(16)(17)(18)(19)(20). The mechanism underlying the regulation by 1,25-(OH) 2 D 3 is thought to include changes in intracellular calcium, PLC, and protein kinase C activation. PLC-␥1 is one of the major PLC isoenzymes that mediate cellular signal transduction. Treatment with 1,25-(OH) 2 D 3 dramatically up-regulates the protein and mRNA expression of PLC-␥1 (24). To understand the molecular mechanism of this regulation, we cloned the 5Ј-flanking region of the human PLC-␥1 gene that confers promoter activity and identified within the 5Ј-flanking region a DR6-type vitamin D-responsive element (VDRE).

MATERIALS AND METHODS
Isolation of Genomic Clone-The subclones containing phospholipase C-␥1 genomic DNA were obtained from a human P1 genomic DNA library using as probe a 5-kb PLC-␥1 cDNA (Genome Systems). To isolate the 5Ј-flanking region of the PLC-␥1 gene, the subclones were further screened by colony hybridization using the oligonucleotide (5Ј-CGTTGCGCTTGCTCCCGGGC-3Ј) from the 5Ј-untranslated region of PLC-␥1 cDNA as probe. From the selected subclone, a 1.1-kb XhoI fragment was resubcloned into a pBluscript SK(Ϫ) vector (Stratagene). The nucleotide sequence of the insert was sequenced using the dideoxy chain termination method. The sequence of each strand was confirmed by repeating the sequencing in both directions at least three times. The sequence of the GC-compressed region was confirmed using dITP instead of dGTP.
Construction of Plasmids-The XhoI-StylI fragment was subcloned into a pGL-3-basic vector (Promega). The PLC-␥1 gene was placed 2 bp upstream from the luciferase gene. Subsequent 5Ј deletion constructs were made with restriction enzyme digestion. The constructs containing the fragment Ϫ748 to Ϫ828 and the fragment Ϫ786 to Ϫ803 were made by ligating the fragments to the heterologous simian virus 40 (SV40) promoter in the pGL-3-promoter vector. Correct orientation of the inserts with respect to the luciferase sequence was verified by restriction enzyme analysis.
Cell Culture-Normal human keratinocytes were isolated from neonatal human foreskins and grown in serum free keratinocyte growth medium (Clonetics) (25). Briefly, keratinocytes were isolated from newborn human foreskins by trypsinization (0.25% trypsin, 4°C, overnight), and primary cultures were established in keratinocyte growth medium containing 0.07 mM calcium. Second passage keratinocytes were plated in 60-mm culture dishes with keratinocyte growth medium plus 0.03 mM calcium at 20 -30% confluency for the transfection experiments.
DNA Transfection and Luciferase Assay-PLC-␥1 luciferase chimeric plasmids were transfected into normal human keratinocytes using a polybrene method 24 h after plating cells in 60-mm culture dishes (26). Cells were co-transfected with 0.2 g of pRSV␤-gal (27), a ␤-galactosidase expression vector that contains a ␤-galactosidase gene that is driven by a Rous sarcoma virus promoter and enhancer, which was used as an internal control to normalize for transfection efficiency. 1,25-(OH) 2 D 3 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) U80983. The cells were lysed, and the cell extracts were assayed for luciferase activities using Luciferase Assay System (Promega). The ␤-galactosidase activities were assayed using Galacto-Light kit (TROPIX Inc.). A pGL-3-control vector (Promega) containing SV40 promoter and SV40 enhancer, which are known to be unresponsive to 1,25-(OH) 2 D 3 , was included in each transfection experiment as a control. Every experiment was done in triplicate and was repeated at least three times. S1 Nuclear Protection Assay-Total cellular RNA was isolated from the first passage of normal human keratinocytes by RNA STAT-60 kit (Tel-Test "B" Inc). The poly(A) RNA was obtained using a poly(A) mRNA isolation kit (Stratagene). The S1-nuclease Protection assay was carried out using the S1-Assay kit from Ambion. An antisense DNA probe was synthesized from a 212-bp Bsu36I-StylI fragment upstream of the translation start codon in the human PLC-␥1 5Ј-flanking region cloned into the pGL-3-basic vector, using Klenow and [␣-32 P]dCTP. This was accomplished by use of the antisense GL primer2 primer that bound the downstream sense GL primer2 primer in the vector such that the synthesized probe spanned the insert. The probe was coprecipitated with 1 g of the normal human keratinocyte poly(A) RNA. Hybridization was performed by dissolving the precipitate in 10 l of hybridization buffer at 42°C overnight. Unprotected DNA was digested with S1 nuclease at 37°C for 30 min. The resulting fragment was recovered by ethanol precipitation, denatured, and analyzed on an 8% sequencing gel with a sequencing ladder as a standard. The sequencing reaction was performed using the dideoxy chain termination method.
Primer Extension Analysis-The total RNA and poly(A) RNA from normal human keratinocytes were isolated in a same way as that for S1-nuclease protection assay. The primer extension analysis was performed using the Primer Extension System from Promega. 2 g of poly(A) RNA was hybridized with an end labeled primer corresponding to the region 40 -60 bp downstream from the translation start codon of the antisense strand of the human PLC-␥1 cDNA. The hybridization mixture was heated at 75°C for 15 min and then incubated at 42°C for 40 min. Actinomycin D was added to the mixture at a final concentration of 75 ng/l to inhibit secondary structure formation of the RNA. The extension products were analyzed on a denatured 6% polyacrylamide gel.
DNA Mobility Shift Assay-The nuclear extracts were made from normal human keratinocytes according to the method described by Abmayr and Workman (28). The recombinant vitamin D receptor was from Affinity Bioreagents Inc. Synthetic oligonucleotides used for the DNA mobility shift assay were end-labeled by T 4 polynucleotide kinase. The DNA-protein reactions were performed in a total of 17 l; nuclear extracts (12 g of protein) were incubated with 2 g of poly(dI⅐dC) (Pharmacia Biotech Inc.) and 10,000 cpm of 32 P-labeled probe in 10 ml of binding buffer (20 mM HEPES, pH 7.9, 20% glycerol, 50 mM KCl, 0.5 mM dithiothreitol) at 30°C for 25 min. Unlabeled competitors were added at the preincubation step. In the super gel shift reaction, a polyclonal anti-vitamin D receptor antibody (3 l from the original stock, Affinity Bioreagents Inc.) was added to the DNA-protein reaction and incubated for an additional 25 min. Protein-DNA complexes were electrophoresed in a 6% nondenaturing polyacrylamide gel in 1 ϫ gel shift running buffer (50 mM Tris, 380 mM glycerin, 2 mM EDTA, pH 8.5).

RESULTS
Three positive clones were obtained from the human P1 genomic DNA library screening. One of the positive clones was digested by HindIII, and the random fragments were subcloned into a pZErO vector (Invitrogen). Three independent subclones that contain the 5Ј-flanking region of the human PLC-␥1 gene were isolated from the random subclones by screening 96 subcultures using an oligonucleotide from the untranslated region of the human PLC-␥1 cDNA (see "Materials and Methods"). Restriction analysis indicated that all three positive subclones contained a 9-kb HindIII insert spanning more than 8 kb upstream from the translation start codon in the human PLC-␥1 gene. The restriction map for the 9-kb fragment is shown in Fig. 1A. The HindIII-StylI (9 kb), XbaI-StylI (2.5 kb), and XhoI-StylI (1 kb) fragments were individually subcloned in a pGL-3-basic vector and transfected into human keratinocytes. The results showed that the 1-kb XhoI-StylI fragment construct expressed the highest luciferase activity (data not shown). Therefore, we focused on the 1-kb XhoI-StylI fragment in the subsequent experiments. The sequence analysis revealed that the 1-kb XhoI-StylI fragment in the 5Ј-flanking region of the human PLC-␥1 gene was very GC-rich. 16 putative SP1 sites and 8 putative AP2 sites were clustered in the 1-kb XhoI-StylI fragment. No TATA box was found in this fragment. There was a putative CCAAT box located between Ϫ581 and Ϫ585 bp upstream from the transcriptional start site (Fig. 1B).
Both S1 nuclease protection assay and primer extension analysis were performed to determine the transcriptional start site for the human PLC-␥1 gene. The S1 nuclease protection assay was performed using an antisense probe spanning 212 bp upstream from the translation start codon. This probe hybridized to the poly(A) RNA isolated from human keratinocytes. After S1 nuclease digestion, a single protected fragment of 135 bp was detected ( Fig. 2A). The result suggested that the transcriptional start site is 135 bp upstream from the translation start codon. The primer extension analysis showed a 195-bp single extension fragment whose 5Ј end is 135 bp upstream from the translation start codon (Fig. 2B), confirming the result obtained with the S1 nuclease protection assay.
In an attempt to delineate the sequences essential for human PLC-␥1 gene transcription, nine deletional fragments spanning from ϩ135 to Ϫ877 bp in the 5Ј-flanking region were fused with the coding region of the luciferase gene in the luciferase vector and transfected into normal human keratinocytes (Fig. 3A). The construct containing the ϩ135 to Ϫ877 fragment construct expressed luciferase activity 50-fold higher than that from the vector alone (Fig. 3B). The data clearly indicated that the 5Ј-flanking region of the human PLC-␥1 gene contains a sequence that confers promoter activity. Deletional analysis to Ϫ200 bp showed little loss in basal activity. When the 5Ј deletions reached Ϫ39 bp, the luciferase activities were greatly reduced. The fragment ϩ13 to ϩ135, which did not contain the transcriptional start site, lost all activity (Fig. 3B). The data suggested that the most proximal 200 bp of the 5Ј-flanking region of the human PLC-␥1 gene are essential for transcriptional initiation.
To determine if the human PLC-␥1 gene transcriptionally responds to 1,25-(OH) 2 D 3 , the nine deletional constructs were transfected into human keratinocytes in the presence or the absence of 1,25-(OH) 2 D 3 . The results showed that the construct containing fragment ϩ135 to Ϫ877 was responsive to 1,25-(OH) 2 D 3 stimulation. The luciferase activity was increased over 3-fold after 24 h of exposure to 1,25-(OH) 2 D 3 . 5Ј deletion to Ϫ748 bp totally abolished the responsiveness to 1,25-(OH) 2 D 3

FIG. 3. The induction of human PLC-␥1 promoter by 1,25-(OH) 2 D 3 using transfection experiments with deletional and mutant constructs.
A, nine 5Ј deletional fragments spanning from ϩ135 bp to Ϫ877 bp were ligated to the luciferase gene in a pGL-3 basic vector. The fragments Ϫ748 to Ϫ828 and Ϫ786 to Ϫ803 were ligated to the SV40 promoter and luciferase gene in a pGL-3-promoter vector. The constructs were transfected into human keratinocytes as described under "Materials and Methods." B, the luciferase activities of the nine 5Ј deletional constructs were measured following 24 h of exposure to 1,25-(OH) 2 D 3 or vehicle, divided by ␤-galactosidase activity, and expressed as the percentage of activity of the ϩ135 to Ϫ877 construct in the absence of 1,25-(OH) 2 D 3 . The activity obtained with the pGL-3-basic vector showed the vector background. C, a similar experiment was performed with constructs containing Ϫ748 to Ϫ828, Ϫ786 to Ϫ803, and a mutant construct in which two random sequences replaced AGGTCA and TGGACA. The results are normalized to ␤-galactosidase activity. A construct containing the vitamin D-responsive region at Ϫ143 to Ϫ293 in the human 24-hydroxylase gene (24-hydroxylase) was used as a positive control. The activity obtained from the pGL-3 promoter vector showed the vector background. (Fig. 3B), indicating that the responsive region was located between Ϫ748 and Ϫ828 bp. To confirm that the Ϫ748 to Ϫ828 bp region contains a VDRE, this fragment was subcloned into the pGL-3-promoter vector. Transfection experiments showed that the promoter activity was induced over 2-fold in human keratinocytes by 1,25-(OH) 2 D 3 (Fig. 3C). The region Ϫ748 to Ϫ828 contains an SP1 site and two direct repeats separated by 6 bases (DR6), AGGTCAgaccacTGGACA (named PDR6) (Fig.  1). The PDR6 was located in the region Ϫ786 to Ϫ803. The construct containing only the PDR6 showed the same fold induction by 1,25-(OH) 2 D 3 as did the construct containing the region Ϫ748 to Ϫ828 (Fig. 3C). A mutant construct containing the sequence TAGGTAgaccacATGCAT (named MPDR6) gave no response to 1,25-(OH) 2 D 3 . The vitamin D-responsive region at Ϫ143 to Ϫ293 in the human 24-hydroxylase gene (29) subcloned into the pGL-3-promoter vector was used as a positive control in the transfection experiments. The 24-hydroxylase VDRE construct showed nearly 3-fold induction by 1,25-(OH) 2 D 3 . These results indicate that the human PLC-␥1 gene contains a DR6-type sequence in the region Ϫ786 to Ϫ803, which is of comparable responsiveness to 1,25-(OH) 3 D 3 as the VDRE in the 24-hydroxylase gene.
DR6-type VDREs in other genes have been reported to bind the VDR as a homodimer or heterodimer (30,31). To determine if the responsive region in the human PLC-␥1 gene binds to the VDR in human keratinocytes, an 80-bp synthetic oligonucleotide (named W1) representing the vitamin D-responsive region Ϫ748 to Ϫ828 bp was evaluated using the DNA mobility shift assay. Incubation of the oligonucleotide W1 with the nuclear extracts from the human keratinocytes yielded two specific DNA-protein binding complexes (Fig. 4A). The specificity of the binding was verified by competition with the same or mutant unlabeled oligonucleotides at 100 molar excess. The results showed that the two binding complexes were reduced by W1 but not by a mutant fragment (named M1) containing MPDR6 instead of PDR6 (Fig. 4A). These data suggest that the two bands are specific complexes of the sequence PDR6 with the nuclear factors in the human keratinocytes. The bands are not SP1 complexes because binding was not blocked by an SP1 consensus oligonucleotide even though there is a putative SP1 site in this region (Fig. 4A). However, the upper band was competed out by a 21-bp unlabeled oligonucleotide (named H) containing a DR3-type VDRE (AGGTGAgcgAGGGCG) found in the human 24-hydroxylase gene, suggesting that the upper band was a VDR-VDRE complex (Fig. 4A). To narrow down the vitamin D binding region, we repeated the experiment but used a 38-bp oligonucleotide (named W2) containing the sequence PDR6 with 10 flanking bases on each side. The results showed that a single main complex formed after the incubation of the fragment W2 with the nuclear extracts from the human keratinocytes (Fig. 4B). The binding complex was reduced by the unlabeled oligonucleotides W2 and H but not by a mutant fragment (named M2) containing MPDR6 instead of PDR6. The binding complex was shifted to a higher molecular weight form upon the addition of an antibody specific to the VDR. Incubation of the labeled fragment W2 with the recombinant vitamin D receptor yielded two shifted binds that were blocked by an unlabeled oligonucleotide W2 (Fig. 4B). The results indicate that the sequence PDR6 in the region Ϫ786 to Ϫ803 binds to the VDR in human keratinocytes. DISCUSSION We have cloned the 5Ј-flanking region of the human PLC-␥1 gene that confers promoter activity when transiently transfected into human keratinocytes. The sequence in the human PLC-␥1 flanking region is GC-rich and has no TATA box, similar to many genes that are important in the control of cell proliferation and differentiation such as transforming growth factor-␤1 (32), epidermal growth factor receptor (33), and nerve growth factor (34). The region between Ϫ200 and Ϫ39 bp contains several putative SP1 and AP2 binding sites and appears to contain the promoter because the deletion from Ϫ200 to Ϫ39 bp dramatically reduced promoter activity (Fig. 3B). The transcription initiation of the human PLC-␥1 gene could be similar to other GC-rich genes in which SP1 binding to the GC-boxes, rather than a TFIID-TATA complex, is able to activate gene transcription (35).
The pentanucleotide CCAAT, which is usually found within Ϫ50 to Ϫ100 bp upstream from the transcriptional start site in mammalian genes where it appears to have a role in mediating promoter function (36), was located between Ϫ581 bp and Ϫ585 bp in the human PLC-␥1 gene. However, this putative CCAAT  box has no clear function because the deletion from Ϫ613 to Ϫ551 bp did not remarkably reduce basal promoter activity. Therefore, basal transcription of the human PLC-␥1 gene does not appear to require a CCAAT-binding protein.
A DR-6 type VDRE, AGGTCAgaccacTGGACA, has been precisely localized within the 5Ј-flanking region of the human PLC-␥1 gene. The transfection experiments showed that the DR6 sequence was completely silent in the human PLC-␥1 gene in the absence of 1,25-(OH) 2 D 3 but was activated by the addition of 1,25-(OH) 2 D 3 . In the DNA mobility shift assays, the DR6 specifically bound to the vitamin D receptor in the human keratinocytes, as recognized by the vitamin D receptor antibody. Substitution mutation of the 6-base repeats in the DR6 sequence totally abolished the response to vitamin D as well as the DNA binding ability, indicating that 1,25-(OH) 2 D 3 activates the human PLC-␥1 gene through its vitamin D receptor interacting with the DR6-type VDRE localized in the 5Ј-flanking region. The sequences of the two repeats share some homology with the known DR6-type VDREs in the human osteocalcin gene (30) and rat 24-hydroxylase gene (31) (Fig. 5). Alignment of these DR6-type VDREs showed that one-third of the nucleotides are identical between each repeat. It seems that the second and the sixth nucleotides within each repeat are always G and A, respectively, suggesting that these bases are critical for DNA-receptor binding and confer transactivation upon vitamin D stimulation. Single base mutations will be required to define the precise nucleotides that are essential to mediate the responsiveness to 1,25-(OH) 2 D 3 .
Vitamin D receptor and nonreceptor transcriptional factors binding to distinct sites in a promoter or enhancer region is one mechanism by which the profound alteration in gene expression can occur from small changes in the concentration of trans-acting factors (37). A typical vitamin D receptor and nonreceptor transcriptional factor interaction model was reported in the human osteocalcin VDRE, which contains an AP1 site; AP1 binding proteins were shown to regulate VDRE function (38 -40). Although no AP1 site was found in the human PLC-␥1 gene, SP1 has also been reported to interact with the vitamin D receptor by independently binding to a different motif (37). We found 16 putative SP1 sites clustered downstream of the DR6 sequence in the 5Ј-flanking region of the human PLC-␥1 gene. However, the isolated human PLC-␥1 VDRE ligated to a heterologous SV40 promoter did not appear to differ in the degree of response to 1,25-(OH) 2 D 3 as the VDRE within its own gene context. The data suggest that the SP1 sites are not involved in the vitamin D-induced human PLC-␥1 transcription.
DR6-type VDREs of the human osteocalcin gene (41) and the rat 24-hydroxylase gene have been shown to bind VDR-RAR heterodimers, as well as VDR-VDR homodimers (31). In this report, we found that recombinant VDR was able to bind to the human PLC-␥1 VDRE, as shown by DNA mobility shift assay, implying that VDR might be binding to the human PLC-␥1 VDRE as a homodimer or monomer. However, the keratinocyte nuclear extracts give a different pattern of binding to the DR6 than the recombinant VDR, suggesting that other factors are also involved in VDR-VDRE binding. Further experiments are needed to identify these additional factors.