HCaRG, a Novel Calcium-regulated Gene Coding for a Nuclear Protein, Is Potentially Involved in the Regulation of Cell Proliferation*

Since a negative calcium balance is present in spontaneously hypertensive rats, we searched for the gene(s) involved in this dysregulation. A cDNA library was constructed from the spontaneously hypertensive rat parathyroid gland, which is a key regulator of serum-ionized calcium. From seven overlapping DNA fragments, a 1100-base pair novel cDNA containing an open reading frame of 224 codons was reconstituted. This novel gene, named HCaRG (hypertension-related,calcium-regulated gene), was negatively regulated by extracellular calcium concentration, and its basal mRNA levels were higher in hypertensive animals. The deduced protein showed no transmembrane domain, 67% α-helix content, a mutated calcium-binding site (EF-hand motif), four putative “leucine zipper” motifs, and a nuclear receptor-binding domain. At the subcellular level, HCaRG had a nuclear localization. We cloned the human homolog of this gene. Sequence comparison revealed 80% homology between rats and humans at the nucleotide and amino acid sequences. Tissue distribution showed a preponderance in the heart, stomach, jejunum, kidney (tubular fraction), liver, and adrenal gland (mainly in the medulla). HCaRG mRNA was significantly more expressed in adult than in fetal organs, and its levels were decreased in tumors and cancerous cell lines. We observed that after 60-min ischemia followed by reperfusion, HCaRG mRNA declined rapidly in contrast with an increase in c-mycmRNA. Its levels then rose steadily to exceed base line at 48 h of reperfusion. HEK293 cells stably transfected withHCaRG exhibited much lower proliferation, as shown by cell count and [3H]thymidine incorporation. Taken together, our results suggest that HCaRG is a nuclear protein potentially involved in the control of cell proliferation.

Ionized calcium concentration in plasma is maintained within a very narrow range. The major players maintaining extracellular calcium homeostasis are calciotropic hormones, parathyroid hormone (PTH), 1 1,25-dihydroxyvitamin D, calcitonin, and calcium itself. Indeed, extracellular calcium regulates its own concentration as an extracellular messenger by acting on calcium receptors or calcium sensors. The calciumsensing receptor is linked to several intracellular second messenger systems via guanylyl nucleotide-regulating G proteins and activates phosphoinositide-specific phospholipase C, leading to accumulation of inositol 1,4,5-trisphosphate and diacylglycerol (1)(2)(3)(4)(5).
Cells of the parathyroid gland possess such a calcium sensor (6). Even slight reductions in extracellular ionized calcium concentration (on the order of 1-2% or less) elicit prompt increases in the rate of PTH secretion and mRNA levels. Historically, research on the parathyroid gland has focused on the chemistry, regulation, synthesis, and secretion of PTH. There is growing interest in other calcium-regulating proteins of this gland that are also negatively regulated by extracellular calcium, such as chromogranin A and secretory protein-I (7), as well as a hypertensive factor of parathyroid origin (PHF) (8,9).
Arterial hypertension is associated with numerous disturbances of calcium metabolism manifested not only in humans but also in genetic as well as acquired models of hypertension (10 -14). Disturbances in renal and intestinal handling of calcium in hypertension have been reported by several investigators (15). Urinary calcium has generally been shown to be increased (so-called urinary leak) and intestinal calcium absorption diminished in genetically hypertensive or spontaneously hypertensive rats (SHR) (15,16). Cytoplasmic free calcium concentration has most often been found to be elevated in circulating platelets, lymphocytes, erythrocytes, and vascular smooth muscle cells (VSMC) from hypertensive animals and humans (for a review, see Ref. 17). In SHR as well as in low renin hypertensive patients, there seems to be an inverse relationship between extracellular and intracellular calcium (18). It has been hypothesized that certain genetic abnormalities might be responsible for the link between some forms of hypertension, calcium homeostasis, and the parathyroid gland. To identify new genes that might be abnormally regulated by extracellular calcium in the parathyroid gland of genetically hypertensive rats, we prepared a cDNA library from the parathyroids of SHR. In this study, we describe the isolation and characterization of a novel gene, designated HCaRG (for hypertension-related, calcium-regulated gene), negatively regulated by extracellular calcium with higher mRNA levels in SHR. HCaRG is a nuclear protein with putative "leucine zipper" motifs and is potentially involved in the regulation of cell proliferation.

EXPERIMENTAL PROCEDURES
Cell Cultures-Parathyroid cells (PTC) were isolated from SHR and Wistar-Kyoto rats (WKY). Primary cultures were passaged in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal calf serum, as described previously (9). They were then maintained in Ham's F-12 medium containing a low (0.3 mM) or normal (2.0 mM) total calcium concentration for 2 or 48 h. COS-7 or HEK293 cells were cultured in DMEM containing 10% fetal calf serum. All cell types were maintained in 5% CO 2 at 37°C.
Ischemia-Reperfusion-SHR were anesthetized lightly with flurane, and the right kidney was removed through a mid-abdominal incision. The rats were kept at ambient temperature during the surgery. Their core temperature, monitored by radio-telemetry, was 38°C. The left kidney was subjected to warm transient ischemia by occlusion of the left renal artery and vein with a microclip, as described previously (19). The skin incision was closed during the 60-min renal ischemia period. It was then reopened, and the clip was removed. The wound was closed with a 2"O" suture. The rats had access to water immediately after surgery.
SHR Parathyroid cDNA Library-Parathyroid glands were removed from 100 12-week-old SHR and frozen immediately in liquid nitrogen. The glands were added to a guanidinium thiocyanate solution and homogenized. Poly(A) RNA was isolated on an oligo(dT) column. The cDNA library was constructed with poly(A) RNA as template and the ZAP-cDNA synthesis kit (Stratagene, La Jolla, CA). A summary of the protocol is as follows. mRNA was reverse-transcribed from an XhoI linker oligo(dT) primer using Moloney murine leukemia virus reverse transcriptase. Second strand synthesis was then produced with DNA polymerase I in the presence of RNase H. cDNA termini were blunted by incubation with the Klenow fragment of DNA polymerase I and dNTPs. EcoRI adaptors were added using T4 ligase, and the ends were phosphorylated with T4 polynucleotide kinase. This mixture was then digested with XhoI to release adaptors from the 3Ј-end of the cDNA. The resulting mixture was separated on a Sephacryl S-400 column. cDNAs were ligated into the Uni-ZAP XR vector using T4 DNA ligase and packaged into Gigapack II Gold packaging extract. The packaged products were plated onto XL1-Blue MRFЈ. To screen the cDNA library, phages were plated onto bacterial host plates (XL1-Blue MRFЈ) and incubated overnight. After chilling at 4°C for 2 h, a nitrocellulose filter was overlaid for 2 min. The filter was then denatured and neutralized, and DNA was cross-linked to it with UV light. Hybridization was performed with digoxigenin-dUTP-labeled probes (Roche Molecular Biochemicals) derived from 3Ј-and 5Ј-rapid amplification of cDNA ends (RACE) products described below.
RNA and cDNA Preparation-Total RNAs were prepared from rat cells and organs according to the standard guanidinium thiocyanatephenol-chloroform method (20) and kept at Ϫ70°C until used. mRNA was extracted from total RNA with the Poly(A)Ttract system (Promega, Nepean, Canada). cDNAs, unless stated, were synthesized with random hexamers for first strand synthesis and reverse-transcribed. Radiolabeled DNA probes were prepared by the random priming technique or polymerase chain reaction (PCR) amplification with [ 32 P]dCTP.
3Ј-or 5Ј-RACE-Four mixtures of degenerate oligonucleotide primers were initially designed according to the putative amino acid sequence of PHF with the following degenerate sequence: 5Ј-TA(T/C) TCI GTI TCI CA(T/C) TT(T/C) (A/C)G-3Ј. From initial RACE experiments (described below), one unique sequence primer TAC TCC GTG TCC CAC TTC CG was selected for its ability to generate reverse transcription (RT)-PCR DNA fragments from PTC total RNA and used subsequently as candidate primer for 3Ј-RACE. In brief, for 3Ј-RACE, total RNA from PTC was reverse-transcribed with a hybrid primer consisting of oligo(dT) (17-mer) extended by a unique 17-base oligonucleotide (adaptor). PCR amplification was subsequently performed with the adapter, which bound to cDNA at its 3Ј-ends, and the candidate primer mentioned above (21). For 5Ј-RACE, RT was undertaken with an internal primer derived from the sequence of the cDNA fragment generated by 3Ј-RACE. A dA homopolymer tail was then appended to the first strand reaction products using terminal deoxynucleotidyl transferase. Finally, PCR amplification was accomplished with the hybrid primer described previously and a second internal primer upstream to the first one (21).
Subcloning-The DNA fragments generated from the RACE experiments were separated by electrophoresis, isolated from agarose gel, and extracted by the phenol-chloroform method (20). pSP72 plasmid (Promega) was digested at the SmaI site and ligated to blunt DNA fragments with T4 DNA ligase. Transformed DH5Ϯ Escherichia coli bacteria were grown, and recombinant bacteria were selected by PCR. Similarly, HCaRG was subcloned in pcDNA1/Neo (Invitrogen, Carlsbad, CA).
To determine the subcellular localization of HCaRG protein in mammalian cells, the coding region of HCaRG was fused to green fluorescent protein (GFP) cDNA and was transfected in the cells. Briefly, the entire coding region of HCaRG was amplified by PCR with the primers ATG TCT GCT TTG GGG GCT GCA GCT CCA TAC TTG CAC CAT CCC and TAA TAC GAC TCA CTA TAG GGA GAC, gel-purified, and fused in frame to GFP in pEGFP-C1 (CLONTECH, Palo Alto, CA) through a blunt HindIII site. pEGFP-HCaRG was then sequenced. Similarly, the coding sequence of HCaRG was fused in frame to glutathione S-transferase in pGEX-3X (Amersham Pharmacia Biotech) through a SmaI site and a blunt EcoRI site.
Sequencing-Double-stranded sequencing of cloned cDNA inserts was performed with Sequenase version 2.0 (U.S. Biochemical Corp.). 5 g of recombinant plasmid template were denatured, annealed with T7 or SP6 primers, and labeled with [ 35 S]dATP by extension, using the chain termination method of Sanger according to the manufacturer's protocol.
Cloning of Human HCaRG-A 439-bp cDNA fragment of rat HCaRG was 32 P-labeled and served as a probe for screening a human VSMC cDNA library. DNA from positive phages was purified, and the fragments were cloned in pBluescript. All fragments were sequenced. We obtained a 1355-bp fragment containing the coding region of HCaRG.
Northern Blot Hybridization, Dot Blot Hydridization, and Competitive RT-PCR-2 g of poly(A) RNA from PTC or 10 g of total RNA from kidneys were denatured at 68°C and separated on denaturing formamide 1% agarose gel. After transfer onto nitrocellulose by vacuum, hybridization was performed overnight using 32 P-labeled probes generated from cDNA clone(s) by PCR or random labeling. 1 g of total RNA was used in dot blot experiments. A human multiple tissue expression (MTE) array (CLONTECH) and human fetal and tumor panel Northern Territory RNA blots (Invitrogen) were hybridized with 32 P-labeled human HCaRG cDNA according to the manufacturer's specifications. For quantitative determinations of HCaRG mRNA, total RNA was extracted from PTC and reverse-transcribed. A HCaRG competitor was constructed using the PCR Mimic Construction Kit (CLONTECH) with the following composite primers: GCA CGA GCC ACA GCC AGC TAC CCC AGC CAC CCA TTT GTA CC (sense) and TGT GAC TGT CAG CGG GAT GGA GTC CGA GAT GTA GAG GGC (antisense). The 344-bp DNA obtained was cloned into pSP72 and transcribed with SP6 RNA polymerase. The resulting RNA was quantified by photometry and subsequently used in competitive RT-PCR. The competitive reaction contained 1 or 2 g of total RNA with increasing amounts of competitor cRNA along with 32 P-labeled nucleotide. Two primers, TGT GAC TGT CAG CGG GAT GG and GCA CGA GCC ACA GCC AGC TACC, flanking the HCaRG intron were employed to amplify a 186-bp cDNA fragment. PCR was performed as follows: 15 s at 95°C, 20 s at 68°C, 30 s at 72°C for 30 cycles, followed by a 5-min elongation step at 72°C. 10 l of the PCR were loaded on 1.8% agarose gel and then dried and exposed in a PhosphorImager cassette for quantification.
In Situ mRNA Hybridization-Tissues from SHR and WKY were rinsed in phosphate buffer, fixed in 4% paraformaldehyde, and embedded in paraffin. 3-5-m sections were cut and mounted on microscope slides pretreated with aminopropylthiethoxysilane. The slides were first dried at 37°C and then at 60°C for 10 min prior to use. The probe applied was a unique 300-bp fragment (3r 290 in Fig. 1A) that had been subcloned into the BamHI site of a pSP72 vector. The DNA was transcribed using T7 or SP6 polymerases to create sense and antisense riboprobes, which were labeled with digoxigenin-UTP. They were validated by dot blot hybridization with template DNA. Prehybridization of slides was undertaken after dewaxing in xylene, followed by progressive ethanol-water hydration (from 95 to 50%). The slides were rinsed in phosphate-buffered saline (PBS) and incubated with proteinase K (20 g/ml) for 20 min at room temperature. After this digestion, they were rinsed successively in glycine buffer plus PBS and then dehydrated in ethanol. Actual prehybridization was done with 50% formamide, 0.2% SDS, 0.1% Sarcosyl, 5ϫ standard sodium citrate (SSC: NaCl (0.15 M), sodium citrate (0.015 M, pH 7.0)), and 2% blocking reagent (Roche Molecular Biochemicals) for 1 h at 60°C. Hybridization was performed by adding the probe (200 ng/ml) to 50 l of 4ϫ SSC and 50% formamide per section. The slides were incubated overnight at 60°C in a humidified chamber. During hybridization, a coverslip was placed over the tissue section. After hybridization, it was removed, and the sections were rinsed with 4ϫ SSC and then washed with 4ϫ SSC for 15 min and in 2ϫ SSC for 15 min at room temperature. Finally, the sections were washed with 0.1% SSC for 30 min at 60°C. For coloration, they were washed with buffers 1 and 2 of the DIG Luminescent Detection Kit (Roche Molecular Biochemicals). They were then incubated with anti-DIG alkaline phosphatase antibody (1:500) in buffer 2 for 40 min and washed twice in buffer 1 for 15 min and in buffer 3 for 2 min. Incubation in the color solution nitro blue tetrazolium/5-bromo-4-chloro-3-indoyl phosphate (NBT/x-phos) was carried out for 45 min, after which the slides were washed in distilled water and dry-mounted with Geltol.
In Vitro Translation-The full length of the HCaRG coding sequence was synthesized by RT-PCR with specific primers and inserted downstream of the T7 promoter into the pSP72 vector. In vitro transcription and translation were performed using a TNT-T7-coupled reticulocyte lysate system (Promega) in the presence of [ 35 S]methionine. A plasmid containing the luciferase gene supplied by the manufacturer was used as a control. The synthesized proteins were analyzed by 15% SDSpolyacrylamide gel electrophoresis in the absence or presence of ␤-mercaptoethanol. Radioactive protein bands were detected by scanning with a PhosphorImager.
Antibody Production-E. coli cells transformed with pGEX-3X were grown in LB medium containing 50 g/ml ampicillin at 37°C until A 595 ϭ 0.5. Isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.1 mM, and the cells were cultured for 2 h. Purification of glutathione S-transferase-HCaRG was performed according to the manufacturer's protocol. Polyclonal antisera with antibodies recognizing HCaRG were produced by immunization of rabbits with glutathione S-transferase-HCaRG protein.
Immunocytological Reaction at the Electron Microscopic Level-Rat tissues (liver, anterior pituitary, spleen, heart, and adrenal gland) were quickly removed and fixed in 4% paraformaldehyde with 0.05% glutaraldehyde in phosphate buffer solution for 90 min. A part of the specimens was cryoprotected in 0.4 M sucrose phosphate buffer solution for 30 min at 4°C and then frozen in a cold gradient of fuming nitrogen (Biogel, CFPO, Saint Priest, France) to Ϫ4°C and immersed in liquid nitrogen, as described previously (22). Ultrathin frozen sections of 80-nm thickness were obtained using a dry sectioning method at Ϫ120°C with an Ultracut S microtome (Leica, Lyon, France). The other part of the specimens was dehydrated before embedding in Lowicryl K4M with the AFS system (Leica) (23). Sections were mounted on 400-mesh collodion-carbon-coated nickel grids. For ultrastructural localization of HCaRG protein, the grids were first placed in buffer containing 0.1 M phosphate buffer, 0.15 M NaCl, and 1% albumin, pH 7.4, for 10 min. They were then incubated for 1 h with polyclonal IgG raised against HCaRG protein at concentrations of 1:1000 and 1:50 for ultrathin frozen sections and Lowicryl sections, respectively. After 10min washing in the same buffer, antigen-antibody complexes were revealed with anti-rabbit IgG conjugated with 10-nm gold particles in buffer containing 0.05 M Tris, 0.15 M NaCl, 1% albumin, pH 7.6, for 1 h. The grids were washed in the same buffer and fixed with 2.5% glutaraldehyde. The specificity of the immunocytological reaction was tested on sections with omission of primary antibody and incubation of the primary antibody with particle-adsorbed antigen. No signal was observed on these tissue sections. Before observation in a Philips CM 120 electron microscope at 80 kV, the cryosections were contrasted in 2% uranyl acetate and embedded in 8% methylcellulose, and the Lowicryl sections were contrasted for 20 min in 5% uranyl acetate.
Transfection and Subcellular Localization-COS-7 cells were plated at ϳ30 -50% confluency 1 day prior to transfection, which was performed with 5 g/well of pEGFP-HCaRG or pcDNA1/Neo-HCaRG, according to the calcium phosphate method. After 24 h, the cells were fixed with 4% paraformaldehyde in PBS for 30 min at room temperature. Following three washes with PBS, cells transfected with pEGFP-HCaRG or pcDNA1/Neo-HCaRG were mounted on coverslips. The cells were permeabilized with 0.3% Triton X-100 for 12 min, blocked with 1% bovine serum albumin, 1% gelatin for 15 min, incubated with HCaRG antibodies at 37°C for 1 h, washed in 0.5% bovine serum albumin, incubated with anti-rabbit fluorescein isothiocyanate-labeled antibodies, and washed again. Fluorescence and immunofluorescence were detected with a Zeiss fluorescence microscope.
Stable Transfection-HEK293 cells were plated in a 100-mm plate at a density of 0.5 ϫ 10 6 cells/plate. They were transfected with the control plasmid pcDNA1/Neo (Invitrogen) or with the plasmid containing rat HCaRG using a standard calcium phosphate coprecipitation method. 48 h after transfection, the cells were plated in 150-mm plates in the presence of 400 g/ml G418 (Life Technologies, Inc.). After 2 weeks, the clones were picked, and the level of ectopic HCaRG expression was determined by Northern hybridization.
Cell Counting and [ 3 H]Thymidine Incorporation-The rate of stable clone cell proliferation was measured by counting the number of cells after plating. Cells were seeded at a density of 0.1 ϫ 10 6 cells/six-well plate, with triplicate plates for each cell line. Every 24 h, the cells were trypsinized and counted in a hemocytometer. HEK293 cells that stably expressed either Neo control plasmid or HCaRG were used for the estimation of DNA synthesis by [ 3 H]thymidine incorporation. The clones were trypsinized at 90% confluency, counted in a standard hemocytometer, and inoculated at an identical initial cell density of 40,000 cells/ml in DMEM containing 10% fetal bovine serum and G418 at 400 g/ml. All cells were inoculated in poly-D-lysine-pretreated 24well plates in a volume of 1 ml/well (40,000 cells/well). They were allowed to attach and grow for a period of 24 -48 h. The growth media were then replaced by DMEM containing 0.2% fetal bovine serum and G418 (400 g/ml) for a period of 48 h to synchronize the cells. After the synchronization period, the cells were supplied with fresh medium containing 10% fetal bovine serum and allowed to grow for 48 h.

Isolation of a Novel cDNA Whose Expression Is Negatively Regulated by Extracellular Calcium in the SHR Parathyroid
Gland-Using sense candidate primers (from a putative amino acid sequence of PHF (24)) and a hybrid oligo(dT) primer, 3Ј-RACE experiments, performed on total RNA extracted from SHR PTC cultured in low calcium medium, generated one major 700-bp fragment that was digested and cloned in the BamHI site of pSP72. Since a BamHI site was present in the 700-bp fragment, a recombinant plasmid containing a 300-bp insert was isolated and sequenced. This fragment was used to screen the PTC library and to generate new oligonucleotide primers to extend the cDNA toward the 5Ј-and 3Ј-ends by RACE. From seven overlapping DNA fragments isolated in the above experiments and from SHR PTC cDNA library screening, a 1100-bp cDNA was reconstituted (Fig. 1A). The rat 1100-bp reconstituted cDNA sequence contained an open reading frame of 224 codons preceded by two in-frame stop codons and followed by the most frequent variant of the poly(A) tail (Fig. 1B). A 342-bp intron was localized at position Ϫ52 from the translation initiation site.
Poly(A) RNA was isolated as described and analyzed by Northern hybridization with the 32 P-labeled 300-bp fragment ( Fig. 2A). Two bands were detected with this probe, with approximate lengths of 1.2 and 1.4 kilobase pairs. These results suggest either the existence of two genes or differential splicing. Furthermore, they indicate that the reconstituted 1100-bp cDNA is almost full-length cDNA, estimated at 1.2 kilobase pairs by the major band in the Northern hydridization experiments.
Regulation of the expression of this novel gene was investigated by competitive RT-PCR assay in PTC from WKY and SHR. Cells between 5 and 12 passages were tested in these studies. In WKY PTC, lowering of ambient calcium from 2.0 to 0.3 mM induced a rapid 2-fold increase in the mRNA levels of this novel gene at 2 h, which lasted up to 48 h (Fig. 2B). This calcium regulation was detected in WKY PTC up to about 12 passages but disappeared in long term cultures. Lowering of calcium concentrations in the cell medium also increased the mRNA levels of this novel gene in SHR PTC but to a lesser extent than in WKY cells (data not shown). We then compared its mRNA levels between two normotensive rat strains (Brown Norway, BN.lx, or WKY) and hypertensive animals (SHR). We observed that the mRNA levels of this novel gene were significantly higher in PTC derived from SHR (Fig. 2C, left panel) compared with normotensive WKY at normal calcium. Similarly, when we extracted RNA (Fig. 2C, right panel) or proteins (Fig. 2D) directly from the kidneys, we found significantly higher levels of this novel gene in hypertensive rats. These results clearly show that this novel gene is negatively regu-lated by extracellular calcium concentrations and that its levels are significantly higher in genetically hypertensive rats compared with two normotensive strains. We therefore named this gene HCaRG (hypertension-related, calcium-regulated gene).
Sequence and Structure of HCaRG cDNA-The deduced protein contained 224 amino acids with a calculated molecular mass of 22,456 Da. The estimated pI of the protein was 6.0. It comprised no known membrane-spanning motif but had an estimated 67% ␣-helix content. The absence of a putative signal peptide sequence suggested an intracellular protein. There were two cysteines in the sequence, indicating possible intra-or intermolecular disulfide bridges (Cys 64 -Cys 218 ). The protein had several putative phosphorylation sites for protein kinase C and protein kinase A and one potential Asn-glycosylation site (Asn 76 ). To confirm that HCaRG mRNA encodes a peptide of expected size, the HCaRG cDNA inserted into pSP72 was incubated in vitro in a coupled transcription/translation labeling system. It was transcribed by T7 RNA polymerase and translated in rabbit reticulocyte lysate. As shown in Fig. 3 (lane 4), HCaRG mRNA directed the synthesis of a peptide with a molecular mass of 27 kDa, which closely corresponded to the molecular mass calculated from the amino acid sequence. Polyacrylamide gel electrophoresis analysis of the reaction product in the absence of the reducing agent ␤-mercaptoethanol showed bands of 27 and 43 kDa (Fig. 3, lane 5). These results suggest possible intramolecular or intermolecular disulfide bridges and the formation of homodimers or heterodimers with other protein(s) present in the lysate.
Cloning of Human HCaRG-We then used a 439-bp cDNA fragment of rat HCaRG (ϩ1 to ϩ440 in Fig. 1) to screen a human VSMC cDNA library. We identified several positive clones that were purified, subcloned in pBluescript vector, and sequenced. We obtained a 1355-bp sequence containing fulllength human cDNA, while all other clones contained only partial sequences. A recent sequence search in GenBank TM revealed a region with complete DNA sequence homology within three cosmids containing the zinc finger protein 7 gene (accession numbers AF124523, AF146367, and AF118808). Although the nucleotide sequence of human HCaRG could be found in these cosmids, we are the first to assign an expressed gene sequence to this DNA region.
Sequence comparison between human HCaRG and rat HCaRG showed 80% identity at the nucleotide level (data not presented) and, similarly, 80% homology at the amino acid level (Fig. 4). Analysis of protein structure with the PROSE-ARCH data base revealed four overlapping putative leucine zipper consensus motifs (Fig. 4, underlined). Further analysis revealed homology to the EF-hand calcium-binding motif (eight of the 10 most conserved amino acids) (Fig. 4, dashed box). We also identified a nuclear receptor-binding motif (Fig. 4, boldface and italic type). All of these motifs were conserved in the rat and human amino acid sequence.
Subcellular Localization of HCaRG-We expressed GFP-HCaRG in COS-7 cells. Fluorescence study showed that GFP-HCaRG localized in the nucleus, while cytoplasmic fluorescence was very faint (Fig. 5B). GFP, on the other hand, had a very diffuse localization (Fig. 5A). This result was confirmed by immunofluorescence using antibodies specific to HCaRG (Fig.  5C) and by electron microscopy (Fig. 5D). Electron microscopy was also performed on different tissues. In all tissues studied, HCaRG was found in the nucleus with some labeling in protein synthesis sites.
HCaRG Expression in Various Human Tissues-A human MTE array was hybridized with human 32 P-labeled HCaRG cDNA as a probe. The array contained 76 poly(A) RNAs from various adult tissues, cell lines, fetal tissues, and cancerous cell lines. These arrays were normalized to eight different housekeeping genes. Analysis of the array showed that HCaRG was expressed preponderantly in the heart, stomach, jejunum, kidney, liver, and adrenal glands. Comparison of HCaRG expression in fetal organs with expression in adult organs revealed that HCaRG mRNA was less expressed in all fetal tissues compared (Fig. 6A), particularly in the heart, kidney, and liver. Northern blots confirmed the lower abundance of HCaRG in the fetal heart compared with all regions of the adult heart (Fig. 6B). We also compared HCaRG mRNA levels in various cancerous cell lines to normal tissues (Fig. 6C). HCaRG mRNA levels were decreased in all cancerous cell lines studied. They were also much lower in a glioblastoma, a partly differentiated renal cell carcinoma, and a moderately differentiated hepatocellular tumor compared with the same amount of normal RNA of adjacent tissues excised from the same operational site (Fig. 6D).
In Situ Hybridization of HCaRG mRNA in the Kidney and Adrenal Gland-HCaRG expression was determined in SHR tissues by in situ hybridization. The labeled antisense riboprobe hybridized to the medulla and zona fasciculata of the adrenal cortex (Fig. 7). In the kidney, labeling was almost exclusively located in the cortex and concentrated in the tubular component, contrasting with the virtual absence of the signal in glomeruli (Fig. 7). In these organs, the signal was clearly greater in hypertensive rats compared with their normotensive controls. 2 The sense probe was used as a negative control and appropriately revealed a low signal under our hybridization conditions, demonstrating the specificity of the reaction (Fig. 7, lower panels).
HCaRG mRNA Levels after Ischemia-Reperfusion-The process of kidney injury and repair recapitulates many aspect of development. It involves dedifferentiation and regeneration of epithelial cells, followed by differentiation (25)(26)(27). Since we observed that HCaRG mRNA levels are lower in fetal than in adult organs, we evaluated HCaRG expression after unilateral renal ischemia in uninephrectomized rats (19), since contralateral nephrectomy has been shown to stimulate cell regeneration (28 -31). We noted that HCaRG mRNA declined rapidly to its lowest levels at 3 and 6 h of reperfusion (Fig. 8A). These values then increased steadily to higher than base line at 48 h of reperfusion. This was observed in both the kidney medulla (Fig. 8A) and cortex (Fig. 8B). In contrast to the decline in HCaRG mRNA levels, the proto-oncogene c-myc expression, which is correlated with hyperplastic response in mammalian cells, was rapidly increased following renal ischemia and reperfusion (31). c-myc mRNA levels were low in control kidneys and increased dramatically in the postischemic kidney at 3 h of reperfusion, at a time when HCaRG mRNA levels were already reduced (Fig. 8, A and C).
Overexpression of HCaRG Inhibits Cell Proliferation-HEK293 cells were stably transfected with either plasmid alone or with plasmid containing rat HCaRG. After transfection, several clones were examined for the determination of rat HCaRG mRNA levels. Four clones (HCaRG clones 1, 5, 8, and 9) expressed variable amounts of rat HCaRG mRNA, as detected by Northern blots, while no HCaRG mRNA levels were 2 R. Lewanczuk and J. Tremblay, unpublished data. The deduced amino acid sequences of rat HCaRG (rHCaRG) and of human HCaRG (hHCaRG) are aligned. Identical amino acids are boxed, while homologous amino acids are shaded. We calculated 80% homology between these two sequences. Analysis revealed homology to the EF-hand motif, with eight out of the 10 most conserved amino acids (dashed box). Further analysis using the PROSEARCH data base revealed four overlapping putative leucine zipper consensus motifs (underlined). We also identified a nuclear receptor-binding domain (boldface and italic type).
found in clones transfected with the plasmid alone (Fig. 9). Clones expressing the highest levels of HCaRG (clones 8 and 9) were selected for cell proliferation studies. For these studies, cells that were transfected with the vector alone or polyclonal HCaRG-transfected cells served as controls. The proliferation rates of the HCaRG-transfected cell lines and vector control cells were examined under normal growth conditions (10% fetal calf serum and G-418) by counting cell numbers every day for a period of 8 days after plating. Cell lines transfected with the vector alone (Neo clones 1 and 6) showed a similar growth rate as nontransfected cells (not presented). Clones 8 and 9 expressing high levels of rat HCaRG revealed a much lower proliferation rate than vector control cells, while polyclonal cells expressing intermediate values of HCaRG fell in between (Fig.  10A). Consistent with a lower proliferation rate, stable HCaRG transfection clones 8 and 9 showed much lower [ 3 H]thymidine incorporation than clones transfected with the empty vector (Fig. 10B). DISCUSSION The cloning of a novel extracellular calcium-responsive gene (HCaRG) in the rat parathyroid gland from SHR is described here. HCaRG mRNA and protein levels were higher in cultured PTC and in several organs of SHR, compared with their normotensive counterparts. They were negatively regulated by extracellular calcium; i.e. lowering extracellular calcium led to increases in HCaRG mRNA. The identification of an extracellular calcium-sensing receptor from the parathyroid gland has provided novel insights into the mechanisms of direct action of extracellular calcium on several cell types. The calcium sensor has also been localized in the cerebral cortex and cerebellum, in the tubular region of the kidney cortex, the thyroid, adrenal medulla, lung, and blood vessels (1,32,33). As shown here, HCaRG mRNA levels are also detected in several of these tissues. The calcium receptor is a member of the superfamily of G protein-coupled receptors activating phospholipase C (34,35). In the parathyroid gland, it is a key mediator of inhibition of PTH expression by high calcium (36). The calcium sensor has been shown, in the kidney, to be directly related to inhibition of tubular reabsorption of calcium and magnesium in the thick ascending loop (for a review, see Ref. 34). In PTC cultures prepared from human or bovine parathyroids, low extracellular calcium (0.3 mM) has been demonstrated to increase PTH secretion and mRNA levels, whereas augmentation of calcium in the incubation medium reduces PTH mRNA. Similar regulation was observed for PHF in rat parathyroid cells (9). We show here that HCaRG expression is regulated in a manner similar to PTH and PHF in PTC isolated from the rat.
To date, very few extracellular calcium-negative responsive genes have been cloned. Parathormone was the first gene described to possess a negative calcium-responsive element (nCARE) in its 5Ј-flanking region (37). Several types of nCARE have been reported; type 2 is a regulatory element consisting of a palindromic core sequence and several upstream T nucleotides originally described in the PTH gene. Its transcriptional inhibitory activity is orientation-specific. The nCARE core is present in an Alu repeat in 111 copies in the human genome, suggesting the possibility that other genes may possess functional nCARE (38). With the properties described in the present study, HCaRG may be one of them.
HCaRG is not only expressed in the parathyroid gland but also in most organs tested, although at highly variable levels. Elevated HCaRG levels have been noted consistently in the tissues of genetically hypertensive animals, suggesting abnormalities of HCaRG regulation in several organs of SHR that could be due to either 1) decreased extracellular calcium levels, 2) an abnormal response to extracellular calcium, 3) abnormal transcription/stability of HCaRG mRNA in hyper-tensive rats, or 4) a combination of these. A state of negative calcium balance has been described in SHR that could support the first possibility. On the other hand, 2-fold higher HCaRG mRNA levels were observed in PTC from SHR than from WKY at normal calcium concentration (Fig. 2C). Thus, the modest reduction of calcemia in hypertension will not be the sole explanation of increased levels, suggesting increased expression or decreased degradation of this gene product in hypertension.
No homologous protein sequence to the HCaRG open reading frame was found in the SWISSPROTEIN data base. The HCaRG coding sequence contains one consensus motif known as the EF-hand or helix-loop-helix calcium motif (Fig. 4, dashed  box). This motif generally consists of a 12-residue, calciumbinding loop flanked by two ␣-helices. Eight of the 10 most conserved amino acids are present in HCaRG protein. Usually, the basic structural/functional unit consists of a pair of calcium-binding sites rather than a single helix-loop-helix motif. The HCaRG coding sequence contains only 1 EF-like motif, but it is possible that its high ␣-helix content favors coiled-coil interactions and dimerization of the protein. Pairing of the two EF-hand motifs may enhance its calcium function. Hodges and collaborators (39,40) have demonstrated that domain III of troponin C (a synthetic 34-residue calcium-binding domain) can form a symmetric two-site homodimer in a head-to-tail arrangement in the presence of calcium (41). Similarly, a 39residue proteolytic fragment containing calcium-binding site IV of troponin C was shown to form a dimer (42). These studies and others (43)(44)(45) have demonstrated that dimerization of single helix-loop-helix structures controls calcium affinity and that even homodimers can bind two calcium molecules with positive cooperativity (40). Hydrophobic interactions at the interface between calcium-binding sites appear to stabilize the calcium domains. Our in vitro translation studies showed the appearance of a protein band of about 43 kDa under nonreducing conditions. HCaRG protein might form reductant-sensitive, noncovalent homodimers compatible with its putative high ␣-helix content, but the existence of a functional calcium domain in HCaRG protein remains to be established. Several characteristics of HCaRG are similar to those of S100A2 protein, a calcium-binding protein of the EF-hand type that is preferentially expressed in the nucleus of normal cells but down-regulated in tumors (44). As with HCaRG, S100A2 expression is down-regulated by calcium (46,47).
We also cloned the human homolog of HCaRG from a VSMC cDNA library, using a 437-bp fragment of rat HCaRG as a probe. The coding sequence was found to be 80% homologous to the rat sequence and to contain the putative EF-hand domain. A restriction fragment length polymorphism permitted us to localize the HCaRG locus on chromosome 7 of rats. 3 The gene was assigned within a 4.4-centimorgan region on the long arm of chromosome 7 between Mit 3 and Mit 4 markers. By analogy, we suggested the assignment of HCaRG on human chromosome 8q21-24. In a recent search of HCaRG homologous se-quences in GenBank TM , homologies were found with three chromosome 8 clones containing zinc finger protein 7. It was therefore possible to localize HCaRG on chromosome 8q24.3, confirming our initial assignment. This region contains loci involved in several bone diseases, including osteopetrosis and multiple exostosis, and several human neoplasms (48,49).
Many DNA-binding proteins utilize zinc-containing motifs to bind DNA. Other classes of DNA-binding proteins have a DNA recognition domain at their N terminus that dimerizes to form a two-chain coiled-coil of ␣-helices, also known as a leucine zipper. We identified four overlapping leucine zipper regions conserved in the rat and human sequence, and the high ␣-helix content of HCaRG makes it a possible DNA-binding protein.
We are currently investigating this possibility. It has been shown that nuclear receptors require the ligand-dependent recruitment of co-activator proteins to effectively stimulate gene transcription (50). The nuclear receptor interaction domain of these factors is highly conserved and contains the consensus sequence LXXLL. This motif is sufficient for liganddependent interaction with nuclear receptors (51). We have identified one of these motifs in HCaRG. Nuclear localization of HCaRG protein makes this gene a potential transcription regulator.
Recently, a new transcription factor from the rat kidney (Kid-1) was identified (52)(53)(54)(55). It was reported that Kid-1 mRNA levels declined after renal injury secondary to ischemia (55). Similarly, decreased HCaRG mRNA levels are seen when epithelial cells are dedifferentiated and proliferate (following ischemia and reperfusion). In the model of unilateral ischemic injury, it was shown that contralateral uninephrectomy attenuates apoptotic cell death and stimulates tubular cell regeneration (28 -31). We demonstrate here that HCaRG mRNA levels decreased 3 and 6 h after ischemia in contrast to c-myc expression which is correlated with hyperplastic responses (31). We also observed that its levels are lower in all fetal organs tested when compared with adult organs and lower in tumors and the cancerous cell lines tested. It is possible that the gene product may exert a negative effect on growth. This was confirmed by the stable expression of HCaRG in HEK293 cells. We found that HCaRG overexpression had a profound inhibiting effect on HEK293 cell proliferation. This was shown not only by lower cell number but also by lower DNA synthesis, suggesting that the effect seen was not due to a death-promoting effect of HCaRG.
In conclusion, we have cloned the cDNA of a novel gene that is regulated negatively by extracellular calcium and presents greater expression in several organs of the genetically hypertensive rat model, which is known to demonstrate negative calcium balance. HCaRG mRNA levels are rapidly regulated by calcium, perhaps via the action of calcium receptor signaling. Comparison of HCaRG mRNA levels in fetal organs with those in adult organs and normal and tumor cells showed that HCaRG is more expressed in all adult normal tissues tested. We also report that HCaRG mRNA levels are modulated during ischemia-reperfusion injury, which mimics kidney ontogeny. Furthermore, its nuclear localization, identified motifs, and patterns of expression make this gene a potential regulator of cellular proliferation.