Ribozyme Ablation Demonstrates That the Cardiac Subtype of the Voltage-sensitive Calcium Channel Is the Molecular Transducer of 1,25-Dihydroxyvitamin D 3 -stimulated Calcium Influx in Osteoblastic Cells*

1,25-Dihydroxyvitamin D 3 (1,25(OH) 2 D 3 ) stimulates transmembrane influx of Ca 2 1 through L-type voltage-sensitive Ca 2 1 channels (VSCCs) in ROS 17/2.8 osteoblastic cells. Ca 2 1 influx modulates osteoblastic activities including matrix deposition, hormone responsiveness, and Ca 2 1 -dependent signaling. 1,25(OH) 2 D 3 also regu-lates transcript levels encoding VSCCs. L-type VSCCs are multisubunit complexes composed of a central pore-forming a 1 subunit and four additional subunits. The a 1 subunit is encoded by one gene in a multimember fam-ily, defining tissue-specific subtypes. Osteoblasts syn-thesize two splice variants of the a 1C cardiac VSCC sub- type; however, the molecular identity of the 1,25(OH) 2 D 3 -regulated VSCC remained unknown. We created a ribozyme specifically cleaving a 1C mRNA. To increase target ablation efficiency, the ribozyme was inserted into U1 small nuclear RNA (snRNA) by engi-neering the U1 snRNA expression cassette, conferring the ribozyme transcript with stabilizing stem-loops at both sides and the Sm binding site that facilitates local-ization into nucleoplasm. After transfection of ROS 17/ 2.8 cells with U1 ribozyme-encoding

high threshold for voltage-dependent activation and sensitivity to dihydropyridine type organic channel blockers characterize L-type VSCCs that consist of five distinct subunits (␣ 1 , ␣ 2 , ␤, ␥, and ␦) encoded by four genes (3). The ion translocating function of the L-type channel is conferred by its pore forming ␣ 1 subunit that contains four homologous repeats, each having six putative transmembrane segments (S1-S6) (1)(2)(3)(4). At least four subtypes have been identified and ascribed to coding gene differences; ␣ 1C , ␣ 1D , ␣ 1F , and ␣ 1S correspond to cardiac muscle, neuroendocrine tissue, retina, and skeletal muscle, respectively. The diversity of the channel is further generated by alternative splicing into various isoforms, as is the case for the ␣ 1C subunit (5).
L-type VSCCs have been identified in osteoblasts of various origins (6 -9), where they are thought to play important roles in regulating osteoblast responses to external stimuli. In particular, L-type VSCCs transduce the rapid effects of 1,25(OH) 2 D 3 and depolarization, leading to Ca 2ϩ influx in various osteoblastic cells (6,7,10). In the longer term, 1,25(OH) 2 D 3 treatment alters gene expression in osteoblastic cells, resulting in increased expression of matrix proteins including osteopontin and osteocalcin and reduced expression of the VSCC ␣ 1C subunit (11,12). However, the molecular identity of the VSCC involved in Ca 2ϩ flux induced by 1,25(OH) 2 D 3 and depolarization has never been demonstrated conclusively.
Recent developments in ribozyme technology provide powerful means to selectively inhibit expression of specific target genes. Ribozymes are a class of small catalytic RNA molecules that recognize specific substrate RNA molecules by their complementary nucleotide sequence, cleaving the substrate RNA as an endoribonuclease at enzymatic rates (13). Ribozyme expression in transfected target cells thus can inhibit the functional expression of a specific gene. Factors that contribute to ribozyme efficiency in transfected cells are expression levels, stability against rapid degradation, correct folding for exposure to the target, and co-localization and processing at sites where target transcripts are produced and accumulated. By all of these criteria, the U1 small nuclear RNA (snRNA) expression cassette provides an excellent vehicle for ribozyme delivery and expression (14 -16). This cassette has been successfully used to introduce a ribozyme to inhibit fibrillin 1 expression in a human osteosarcoma cell line (16).
U1 snRNA is a highly expressed, stable small RNA (164 nucleotides) involved in both spliceosome and catalytic processing during pre-mRNA splicing (14). The U1 snRNA expression cassette has potent, constitutive upstream elements comparable in strength with the SV40 early promoter (15). At the 5Ј-end, there is a trimethylguanosine 5Ј cap. At both ends, there are stable stem-loop structures, with the 3Ј loop having a high GC content. These structures confer resistance to exonucleases (14). U1 snRNA has a conserved binding site, referred to as the Sm binding site, for the small nuclear ribonucleoproteins. The consensus sequence is RAU 3-4 NUGR. The hypermethylation of the 5Ј cap structure and the binding to the small nuclear ribonucleoprotein enable U1 snRNA to accumulate in the nucleoplasm, where the newly transcribed ␣ 1C mRNA molecules are most abundant (14). In consideration of these advantages of the U1 snRNA expression cassette, the ribozyme vehicle used in these studies was derived from an engineered U1 snRNA expression vector (16), maintaining the conserved Sm binding sequence AAUUUGUGG.
This study was designed with the aim of producing a hammerhead ribozyme specifically targeting the ␣ 1C subtype of the L-type VSCC, transfecting it into ROS 17/2.8 cells in the snRNA expression cassette, and determining whether efficient ablation of the ␣ 1C transcript would reduce or eliminate Ca 2ϩ influx stimulated by 1,25(OH) 2 D 3 or depolarization. Our results provide conclusive evidence that the cardiac subtype of the L-type VSCC present in osteoblastic cells, including ROS 17/2.8 cells and primary cultures, can serve as the major plasma membrane transducer of external stimuli including both 1,25(OH) 2 D 3 -stimulated and depolarization-induced Ca 2ϩ influx.

EXPERIMENTAL PROCEDURES
Generation of Partial cDNA Encoding ␣ 1C mRNA Target-For selecting and testing ribozyme activity in vitro, a cDNA copy of partial ␣ 1C mRNA was ligated into pGEM ® -3Z vector (Promega, Madison, WI). Total RNA was extracted from cultured ROS 17/2.8 cells with RNA STAT-60 TM (Tel-Test, Inc., Friendswood, TX). The upstream primer CCC1 (5Ј-GGTAATCACCCGAAGGAGCAAG-3Ј) in the untranslated region and the downstream primer CCC3 (5Ј-ATGAATTCGCGTT-GGGGTGGAAGAGTAGTC-3Ј; ATGAATTC represents an extra modified sequence) in the open reading frame were selected to amplify a 734-base pair fragment corresponding to bases Ϫ160 to 566 of the rat ␣ 1C subtype. cDNA was synthesized using reverse transcriptase (RT) and random hexamer priming (Advantage TM RT for PCR Kit, CLON-TECH, Palo Alto, CA).
Polymerase chain reactions (PCRs) were performed as follows: 30 cycles of 1 min at 95°C, 1 min at 55 or 65°C, and 2 min at 72°C. The PCR products were subjected to electrophoresis in 1.2% (w/v) agarose gels and visualized with ethidium bromide. The 734-base pair amplimer was excised, extracted with the QIAEX II gel extraction kit (Qiagen, Valencia, CA), and ligated into a pGEM ® -3Z vector. Subsequently, the ligated product was transformed into JM109 competent cells (Promega) for amplification. At least 10 colonies were selected, and the cloned DNA and ligation junctions were sequenced to verify the identity, sequence, and orientation of the insert. A ligated plasmid, termed pGEM␣ 1C 7, was utilized to perform in vitro transcription. pGEM␣ 1C 7 contains an SP6 promoter upstream from the cloned DNA corresponding to ␣ 1C mRNA of the L-type VSCC.
Identification of a Specific Ribozyme Target Site in the ␣ 1C mRNA-The DNA sequence encoding ␣ 1C transcript between primers CCC1 and CCC3 in rat ROS 17/2.8 cells was aligned and compared with other rat subtypes of VSCC ␣ 1 subunits, including ␣ 1S (GenBank TM accession no. X05921), ␣ 1D (GenBank TM accession no. M57682), and reported sequence for ␣ 1C (GenBank TM accession no. M67516). The computer program Align Plus (Scientific and Educational Software, State Line, PA) was used for sequence analysis and identification of subtype-specific sequences (see Fig. 1). A search for the nucleotide triplet NUX was made (where N is any nucleotide, and X is A, U, or C). NUX represents the cleavage triplet of hammerhead ribozymes (17,18). The secondary RNA structure was then analyzed using the RNA MFold program (by Mike Zuker; available on the World Wide Web). Sites buried within complex predicted secondary structures were eliminated from consideration (19), and a site at nt 6 was selected for this study. The target sequence (25 nt) of ribozyme 6 was compared with available sequences in GenBank TM and found to be unique to ␣ 1C .
Ribozyme Design-A hammerhead ribozyme targeting ␣ 1C mRNA was designed according to published procedures (13). The ribozyme contained a 24-nt hammerhead domain flanked by 10 -12 nt complementary to the targeted region of the ␣ 1C mRNA. To construct the ribozymes in vitro, two strands of DNA template containing a SP6 RNA polymerase promoter sequence and ribozyme sequence were synthesized. The ribozyme was targeted to a GUC at nt 6 of ␣ 1C mRNA, relative to the translational site, and was thus named ribozyme 6.
In Vitro Testing of Ribozyme Activity-The ribozyme-encoding oligonucleotides were incubated in 10 mM Tris (pH 8.0) and 50 mM NaCl at 94°C for 5 min and then cooled gradually to room temperature to allow annealing into double-stranded DNA. Substrate RNA was synthesized from pGEM␣ 1C 7 that contains a SP6 promoter upstream from the cloned DNA corresponding to partial ␣ 1C mRNA. The plasmid pGEM␣ 1C 7 was linearized with SacI and then isolated from an agarose gel and extracted using the QIAEX II gel extraction kit. Transcription of substrate and ribozyme RNA was performed with the RiboMAX TM large scale RNA production system (Promega). The transcription reaction mixture contained 2.5 g of linearized plasmid or 1.5 g of ribozyme DNA templates and 25 mM each ATP, CTP, GTP, and UTP. In addition, a prescribed amount of SP6 RNA polymerase and buffers was added as suggested by the manufacturer's instructions in a total volume of 100 l. Following transcription, the DNA templates were removed from the reaction mixture with RQ1-RNase-Free DNase. Unincorporated nucleotides were removed from the RNA transcript by size exclusion chromatography through a small prepacked MicroSpin TM G-50 column (Amersham Pharmacia Biotech).
The target RNA and ribozyme 6 were mixed, and the in vitro cleavage reaction was performed in 50 mM Tris (pH 8.0), 20 mM MgCl 2 at 37°C for various times. Prior to the addition of MgCl 2 , the mixture was heated at 95°C for 2 min and then quick chilled on ice. The cleavage reaction was terminated by the addition of a 3-fold (v/v) excess of Ambion (Austin, TX) loading buffer containing formaldehyde, formamide, and 6 mM EDTA. Cleavage products were resolved by electrophoresis in 1.9% (w/v) agarose/formaldehyde gels and then stained with ethidium bromide and visualized after destaining with RNase-free water.
Construction of the Eukaryotic Expression Vector Encoding Ribozyme-The ␣ 1C ribozyme used for transfection was designed and modified from the ribozyme tested in vitro. The vector for expressing this ribozyme in transfected cells was constructed from the plasmid pZeoU1EcoSpe (16). pZeoU1EcoSpe contains the pZeoSV plasmid DNA modified by excising the SV40 promoter, SV40 polyadenylation site, and polylinker at the BamHI sites. In constructing the pZeoU1EcoSpe, a U1 snRNA expression cassette in pUC13 (20,21) was excised with BamHI digestion and ligated into the BamHI sites of the modified pZeoSV. Two rounds of site-directed mutagenesis were then performed to change 4 nt flanking the Sm protein binding site of U1 snRNA, creating unique EcoRI and SpeI restriction sites. The 5Ј-flanking region of the inserted U1 snRNA expression cassette possesses a promoter/ enhancer comparable in strength with the SV40 early promoter (15).
The transfection vector expressing ribozyme 6 was termed pU1RB␣ 1C 6, and the vector expressing the control RNA was designated pU1control. In construction of pU1RB␣ 1C 6 and pU1control, the Sm binding sequence AATTTGTGG (AAUUUGUGG in mRNA) was inserted prior to the template DNA sequence. The DNA sequence of pU1control differs from that of pU1RB␣ 1C 6 only in that the nucleotides corresponding to those in ribozyme 6 were randomly selected. Oligonucleotides were synthesized and chemically phosphorylated at the 5Ј-end (Integrated DNA Technologies, Inc., Coralville, IA). The singlestranded oligonucleotides were annealed and ligated into the EcoRI and SpeI sites of pZeoU1EcoSpe to create pU1RB␣ 1C 6 and pU1control. The sequence and orientation of the inserts as well as the U1 snRNA expression cassette were confirmed by DNA sequencing.
Cell Culture-ROS 17/2.8 rat osteoblastic sarcoma cells were cultured in Ham's F-12 medium/Dulbecco's modified Eagle's medium-high glucose (1:1) containing 10% fetal calf serum at 37°C in a humidified incubator containing 5% CO 2 as described previously (11). Zeocin-resistant clones, into which an anti-␣ 1C ribozyme construct or control plasmid had been transfected (see below) were maintained in the same medium supplemented with 50 -150 g/ml Zeocin (Invitrogen Inc., Carlsbad, CA). The cell lines were weaned from an initial selection concentration of 225 g/ml (see below).
Transfection of ROS 17/2.8 Cells-pU1RB␣ 1C 6 and pU1control were linearized with ApaI and cleaned with Wizard DNA Clean-Up Resin (Promega). ROS 17/2.8 cells were grown to 50 -80% confluency and transfected with the linearized plasmids. Transfections were performed with LipofectAMINE TM (Life Technologies, Inc.) according to the manufacturer's protocol. At 72 h post-transfection, the cells were passaged to less than 25% confluency in medium containing 225 g/ml Zeocin. After 14 days, widely spaced clonal colonies were scraped and aspirated into a micropipettor as described (22). The cell clumps were added to wells containing 0.15 ml of trypsin EDTA (22). After 15 min, the cell clumps were dispersed by pipetting, and the trypsin reaction was terminated with 1 ml of medium containing 10% (v/v) serum and 225 g/ml Zeocin. The medium was replaced with fresh medium after the cells had attached firmly. To ensure plasmid retention, the Zeocin concentration was maintained at 50 -150 g/ml throughout all remaining phases of experimentation.
RT-PCR Analysis of Ribozyme 6 and ␣ 1C Expression-Total RNA extraction and reverse transcription with random hexamer were performed as described above. For detection of ribozyme expression, PCR was performed with primers PR␣ 1C 6 (5Ј-CATTCTGATGAGTCCGT-GAGG-3Ј) and PRU1 (5Ј-GGAAAGCGCGAACGCAGT-3Ј). For detection of control plasmid expression, a nucleotide sequence termed PRcntr (5Ј-GCAGTTGACCGAAGAGTTGGA-3Ј) was used with PRU1. The amplification conditions were as follows: 45 s at 95°C, 45 s at 58°C, and 45 s at 72°C for 30 -35 cycles. PCR products were electrophoresed in 1.4% agarose gels and visualized with ethidium bromide.
For detection of the mRNA encoding ␣ 1C , PCR was performed with primers CCC1 and CCC3 as described above. PCR also was performed with a downstream primer pair near the 3Ј-end used previously in our laboratory (23) and termed PR11 (5Ј-AGAACCAGAGACAATGTGTG-GAATA-3Ј) and PR12 (5Ј-TACAGGCTAGCATGATATCTTGCCA-3Ј).
Confocal, Immunofluorescent Analysis of ␣ 1C Expression-Cells were grown overnight, rinsed with Hanks' balanced salt solution three times, and fixed with 2.5% paraformaldehyde in phosphate-buffered saline for 10 min. Fixed cells were blocked for 20 min with blocking solution (1% gelatin, 0.05% casein, and 0.2% donkey serum) and then rinsed three times with phosphate-buffered saline. Commercial rabbit anti-␣ 1C primary antibodies that do not react with either ␣ 1S or ␣ 1D subtypes (Alomone Laboratories, Jerusalem, Israel) were applied at a dilution of 1:500 in blocking solution for 1 h. A second antibody solution containing donkey anti-rabbit Ig-Texas Red (Amersham Pharmacia Biotech) at a dilution of 1:10 in blocking solution was applied for 1 h. Prior to visualization, the cells were rinsed three times in buffer containing 0.1% (w/v) gelatin and 0.05% (w/v) casein for 10 min each. Finally, the cells were prepared with Fluoromount G (Fisher) and observed with a Zeiss 510 LSM confocal fluorescence microscope (Carl Zeiss, Inc., Thornwood, NY) as described previously (24). It should be noted that Western blot analyses using a variety of osteoblastic cells confirm the specificity of the ␣ 1C antibody. 2 Calcium Uptake Assays-Cultured cells were assayed for Ca 2ϩ uptake using previously published procedures (25). Cells were seeded at a density of 50,000 cells/ml onto 3.5-cm dishes and grown to approximately 50% confluency. Culture medium was aspirated, and the cells were washed with resting buffer (132 mM NaCl, 5 mM KCl, 1.3 mM MgCl 2 , 1.2 mM CaCl 2 , 10 mM D-glucose, and 25 mM Tris-HCl, pH 7.4). Then the cells were incubated for 2 min with either 37°C or room temperature resting buffer, resting buffer plus 1 nM 1,25(OH) 2 D 3 , or depolarizing buffer (5 mM NaCl, 132 mM KCl, 1.3 mM MgCl 2 , 1.2 mM CaCl 2 , 10 mM D-glucose, and 25 mM Tris-HCl, pH 7.4). 1,25(OH) 2 D 3 from stock solution (1 M) in absolute ethanol was added into the resting buffer, and the same amount of ethanol was added into all other test tubes. All uptake solutions contained 12.5 Ci/ml 45 Ca 2ϩ (NEN Life Science Products). Uptake was terminated by aspiration of the labeling solution followed by three washes with ice-cold resting buffer. Cellassociated 45 Ca 2ϩ was extracted by a 2-h incubation with 0.5 N NaOH and measured by liquid scintillation counting. Two-tailed t tests were performed to assess the significance of the difference between each stimulated group compared with its control in resting buffer.

Identification of Subtype-specific Ribozyme Cleavage Sites in
the Osteoblastic ␣ 1C Transcript-The first objective was to verify the sequence of ␣ 1C transcript produced in ROS 17/2.8 cells for identification of ribozyme cleavage sites. The cDNA corresponding to the 5Ј region of ␣ 1C in ROS 17/2.8 cells that encompasses the initiating AUG codon was cloned and sequenced. This region was chosen for ribozyme targeting because the mRNA near the translational start site is more likely to be exposed on the outer surface of the three-dimensional structure of the folded transcript. Only a few nucleotide discrepancies upstream of the translational start site were found in a comparison of the ␣ 1C mRNA from ROS 17/2.8 osteoblastic sarcoma cells with the ␣ 1C mRNA sequence in GenBank TM (26); the coding sequences were identical. For selection of subtype-specific ribozyme cleavage sites in the ␣ 1C mRNA in ROS 17/2.8 cells, the mRNA sequence of ␣ 1C mRNA was compared with those of ␣ 1D from rat brain mRNA (27) and ␣ 1S from rabbit skeletal muscle mRNA (28) obtained from GenBank TM . No corresponding rat sequence of ␣ 1S was found in the GenBank TM data base. Both K cat and K cat /K m for GUC and CUC cleavage sites are comparatively higher than for other potential sites. Because ␣ 1C mRNA is not an abundant transcript in ROS 17/2.8 cells, both K cat and K m were considered (29). GUC at nt 6, relative to the translational site of ␣ 1C , was selected as the best potential ribozyme target site. Fig. 1 shows the design and specificity of the ribozyme that was termed ribozyme 6. MFold program analysis of the ␣ 1C transcript further showed that the ribozyme 6 cleavage site was likely to be in a region of the predicted folded structure accessible for efficient ribozyme cleavage ( Fig. 2A).
In Vitro Testing of Ribozyme Cleavage of the Substrate ␣ 1C mRNA-To assess the ability of ribozyme 6 to cleave the ␣ 1C transcript, reactions were performed in vitro with cloned products. To create a synthetic ␣ 1C mRNA substrate, a region of ␣ 1C mRNA that encompasses the selected subtype-specific target sites was inserted into the pGEM3 vector. This vector contains a SP6 promoter directly upstream of the cloned ␣ 1C cDNA that corresponds to nt Ϫ160 to 566 relative to the translational start site. Transcription of the sense RNA construct by SP6 RNA 2 K. Brubaker and R. Duncan, personal communication.
FIG. 1. Ribozyme 6 is designed to specifically cleave the ␣ 1C subtype of the VSCC. Sequences encoding ␣ 1C , ␣ 1D , and ␣ 1S subtypes of the L-type VSCC were obtained from GenBank TM (accession numbers M67516, M57682, and X05921, respectively). The rat osteoblastic ␣ 1C sequence present in ROS 17/2.8 cells was deduced from a cloned cDNA produced using primers CCC1 and CCC3 (see "Experimental Procedures") and found to match the published sequence for ␣ 1C in the complementary region to ribozyme 6. The GUC site chosen for ribozyme targeting is located at nt 6 (relative to start codon AUG designated nt 1-3) and contains a sequence unique to the ␣ 1C VSCC subtype. Because this sequence is not present in either ␣ 1S or ␣ 1D subtypes, the ribozyme will not hybridize with or cleave these transcripts if present.
polymerase yielded a 783-nt RNA substrate that contains 726 nt from the ␣ 1C mRNA, plus 57 nt at the ends (46 plus 11 nts) contributed by the pGEM3 vector. Ribozyme 6 was synthesized from double-stranded synthetic DNA oligonucleotide as described (30).
The synthetic ␣ 1C mRNA substrate and ribozyme were mixed as described in the legend to Fig. 2B. Cleavage was assessed after various incubation times in the cell-free system. As shown in Fig. 2B, cleavage of the ␣ 1C target by ribozyme 6 occurred efficiently within the first 1.5 h (lane 3). The 783-base substrate was cleaved by ribozyme 6 into limit 571-base and 212-base fragments that appeared to be fairly stable for at least 6 h. Two other ribozymes targeting ␣ 1C sites were also designed and cloned (data not shown). Ribozyme 6 was the most effi-cient, consistent with its target being located in the proximity of the start site.
Expression of Ribozyme 6 mRNA in Transfected Cells-Ribozyme 6 and control double-stranded DNA oligonucleotides were inserted separately into a modified U1 snRNA expression vector as described under "Experimental Procedures." The predicted secondary structures of both transcripts of U1-ribozyme and U1-control chimeric RNA were compared with the parent U1 snRNA using the RNA MFold program. As shown in Fig. 3, the U1-ribozyme 6 chimeric transcript was predicted to preserve the 5Ј and 3Ј stem-loop structures of U1 snRNA. The Sm binding site (AAUUUGUGG) of U1-ribozyme also is predicted to maintain its single-stranded character seen in U1 snRNA. While these models do not necessarily reflect the actual structure of the RNA molecules under study, they provide useful models for the design of active ribozymes.
The ribozyme-encoding and control plasmids were transfected into ROS17/2.8 cells using LipofectAMINE TM and subcloned as described under "Experimental Procedures." Multiple, stable, and Zeocin-resistant clones transfected with ribozyme 6 were screened for ␣ 1C mRNA expression by RT-PCR analysis. The clone that demonstrated the most efficient ablation of ␣ 1C mRNA as revealed by PCR was selected for further study.
For detecting the transcripts of U1-ribozyme and U1-control chimeric genes, specific primers were designed as described under "Experimental Procedures." The RT-PCR analysis shown in Fig. 4 demonstrates that ribozyme 6 was expressed only in the selected clonal cells transfected with ribozyme 6-encoding plasmids, while control ribozyme transcripts were found only in clonal cells transfected with control plasmids. As expected, both ribozyme and control transcripts were absent in the parental ROS 17/2.8 cells. The cDNA bands migrate at a size corresponding to the expected 66 base pairs, demonstrating that the ribozyme and control RNA constructs were expressed in the respective clonal cells.
Inhibition of ␣ 1C Expression in ROS 17/2.8 Cells-Selected clonal cells were analyzed by RT-PCR and immunofluorescence microscopy for ␣ 1C transcript and protein expression. As shown in Fig. 5, the ␣ 1C mRNA levels were strikingly reduced in cell lines transfected with ribozyme-encoding plasmids, while the ␣ 1C mRNA levels in cell lines transfected with control plasmids were comparable with those found in the original ROS 17/2.8 cells. The primers (CCC1 and CCC3) employed in Fig. 5, A and B, amplified ␣ 1C transcripts that spanned the predicted ribozyme cleavage site shown in Fig. 1. It was considered that short versions of ␣ 1C transcripts containing 3Ј sequences might be present in cells transfected with active ribozyme. To address this, a downstream primer pair (PR11 and PR12) that amplifies region IV of ␣ 1C transcripts corresponding to bases 3551-4276 relative to the translational start site of rat ␣ 1C was used in RT-PCRs. No products were detected, suggesting that 5Јtargeted, ribozyme ablation of ␣ 1C transcript leads to complete degradation. The amplimer size of the PR11/PR12 product has been revised from earlier estimates (11) of 740 nt to a more precise determination of 726 nt, shown in Fig. 5C, lanes 1 and  3. Similar amounts of glyceraldehyde-3Ј-phosphate dehydrogenase were amplified in reactions from all three cell lines (Fig.  5C, lanes 1Ј-3Ј), indicating that ␣ 1C targets were specifically affected. In Fig. 5B, additional nonspecific DNA bands were observed when the annealing temperature was decreased from 65 to 55°C during the PCR process, but ␣ 1C transcripts remained undetectable. The inability to detect ␣ 1C transcript in ribozyme 6-transfected cells, even under lower stringency PCRs, points to the efficiency of the pU1RB␣ 1C 6 construct.
The selected clonal cells were further characterized by im- munofluorescence using specific antibodies to ␣ 1C protein and confocal microscopy. The antibodies are specific to the ␣ 1C subtype of the VSCC as determined by the manufacturer. As shown in Fig. 6, A-C, the number of detectable immunoreactive cell surface ␣ 1C VSCCs was substantially reduced in clonal cells harboring ribozyme-coding pU1RB␣ 1C 6 (Fig. 6B). In contrast, clonal cells transfected with pU1control (Fig. 6C) showed a pattern and density of ␣ 1C protein expression similar to that  1 and 1Ј), and subclones were transfected with ribozyme encoding pU1RB␣ 1C 6 (lanes 2 and 2Ј) or with pU1control (lanes 3 and 3Ј). Preparation of the ribozyme and control constructs and transfection and subcloning were as described under "Experimental Procedures." Reverse transcription reactions were performed with random hexamer primers. PCR was performed using ribozyme-specific primers (lanes 1, 2, and 3) as shown in Fig. 3. The control chimeric RNA was detected using control chimeric RNA-specific 5Ј primer (lanes 1Ј, 2Ј, and 3Ј) and the same 3Ј primer as that used to detect the ribozyme.  1 and 1Ј) or subclones transfected with ribozyme encoding pU1RB␣ 1C 6 (lanes 2 and 2Ј) or with control pU1control (lanes 3 and 3Ј). Reverse transcriptase reactions were performed with random hexamer primers. PCR reactions shown in lanes 1, 2, and 3 were performed using primers CCC1 and CCC3 (see "Experimental Procedures") located near the 5Ј-end or subtype-nonspecific primers PR11 and PR12 (23) located closer to the 3Ј-end of the ␣ 1C transcript. The annealing temperature of the PCR shown in A was 10°C higher than that used in the reaction shown in B and C. Only ␣ 1C subtype products were detected in all cases. The PCRs shown in lanes 1Ј, 2Ј, and 3Ј were performed using glyceraldehyde-3Ј-phosphate dehydrogenase (G3PDH)-specific primers and provided a loading assessment.
FIG. 3. Sequence and predicted secondary structure of chimeric ribozyme transcript of pU1RB␣ 1C 6. Ribozyme 6 DNA was inserted into a modified U1 snRNA expression vector to produce a modified anti-␣ 1C ribozyme-producing vector termed pU1RB␣ 1C 6. The hammerhead ribozyme is located at the center of the transcript flanked by stem structures at both ends. The 5Ј cap structure was methylated by 1 or 3 methyl (m) groups. A Sm binding site sequence (AAUUUGUGG) was included. Primers PR␣ 1C 6 (5Ј-CATTCTGATGAGTCCGTGAGG-3) (sense) and PRU1 (5Ј-GGAAAGCGCGAACGCAGT-3) (antisense) were designed to amplify the ribozyme transcripts. The number at ␣ 1C mRNA indicates the position relative to the translational start site. An arrow indicates the cleavage site.
found in the parental ROS 17/2.8 cells (Fig. 6A). This finding demonstrated that ribozyme 6 successfully inhibited cell surface expression of ␣ 1C protein as well as transcript in transfected cells.
ROS 17/2.8 Cells Ablated of ␣ 1C Transcript Do Not Respond to 1,25(OH) 2 D 3 and Depolarization-45 Ca 2ϩ uptake assays were performed using subconfluent cell monolayers as described under "Experimental Procedures." As shown in Fig. 7, cells in which ␣ 1C of VSCC was ablated by active ribozyme 6 lost the ability to respond to either depolarization or 1,25(OH) 2 D 3 . In contrast, cells transfected with control plasmids continued to express ␣ 1C and maintained their sensitivity to 1,25(OH) 2 D 3 and high K ϩ -stimulated depolarization reflected by a statistically significant 2-3-fold increase in 45 Ca 2ϩ influx. As seen previously (6), slightly higher levels of stimulated 45 Ca 2ϩ influx were seen when cells were depolarized. Overall, treatment with 1,25(OH) 2 D 3 increased 45 Ca 2ϩ influx 2.2-fold, compared with 2.8-fold after depolarization. Together, these results revealed that the ␣ 1C subtype of VSCC is responsible for 1,25(OH) 2 D 3 -stimulated influx in ROS 17/2.8 cells. The VSCC represents a major pathway for Ca 2ϩ entry into bone cells (6,32). Several laboratories, including our own, have demonstrated the presence and activation of VSCCs in osteoblasts treated with 1,25(OH) 2 D 3 (6), parathyroid hormone (8), and epidermal growth factor (32) and during depolarization (6). The development of 1,25(OH) 2 D 3 analogs with differential actions on nuclear receptor-mediated or membrane-initiated pathways provided a powerful approach to dissect the dual effects of 1,25(OH) 2 D 3 , demonstrating conclusively that the rapid activation of cell surface VSCCs is independent of the classical nuclear receptor (12,33). Vitamin D 3 analogs that bind to the nuclear receptor but fail to stimulate Ca 2ϩ influx have clinical utility because they stimulate bone formation with a lesser tendency to produce hypercalcemia compared with 1,25(OH) 2 D 3 (34). This implies that activation of VSCCs may be of pathophysiological importance in the development of hypercalcemia.
Early studies established that depolarization and hormonestimulated Ca 2ϩ influx into osteoblasts could be inhibited by dihydropyridines (6), indicating that the VSCCs involved in this influx are of the high threshold, or L-type. Since the first cDNA encoding the ␣ 1S skeletal muscle Ca 2ϩ channel subunit was cloned (28), two major additional subtypes of L-type VSCCs were discovered: the cardiac muscle subtype, ␣ 1C , and the neuroendocrine subtype, ␣ 1D . Recently, a new L-type subtype (␣ 1F ) from retina also was reported (35). Major subtypes of VSCCs are frequently found outside the excitable tissue where they were first discovered; for example, the expression of the ␣ 1C subtype was detected in rat brain as well as cardiac tissue (26).
Since the initial discovery that Ca 2ϩ influx by 1,25(OH) 2 D 3 was transduced by L-type VSCCs in ROS 17/2.8 osteoblastic cells (6), many efforts have been made to discern the exact molecular nature of the Ca 2ϩ channel(s) present in osteoblastic cells. L-type VSCCs are the best studied Ca 2ϩ channel type and appear to be the major VSCC in osteoblasts, although T-type channels may be expressed transiently during development (36,37). Three types of cDNA sequences corresponding to ␣ 1S , ␣ 1C , and ␣ 1D were detected by Barry et al. (8) in the rat UMR-106 osteosarcoma cell line. Two alternate splice isoforms of the ␣ 1C subtype were identified in rat ROS 17/2.8 osteoblastic sarcoma cells and primary osteoblasts in our laboratory, but the other two subtypes were undetected (11). The molecular identification of the cardiac subtype was confirmed recently in primary cultures of rat calvarial osteoblasts, where it was proposed that the ␣ 1C is involved in development of epidermal growth factor responsiveness of these cells (32). This finding was consistent with our previous report (11) that the ␣ 1C subtype was present in primary cultures of cells grown from neonatal rat calvaria. Taken together, these studies indicate that the ␣ 1C subtype of the VSCC is the major subtype expressed both in primary cultures and cell lines of osteoblastic phenotype. Nonetheless, a collective limitation of all prior studies is that they rely upon circumstantial evidence to predict relationships between structure and function. No evidence has been presented to definitively link the presence of ␣ 1C VSCC subtype to a demonstrated function.
The approach taken in this study involved ribozyme-mediated ablation of the ␣ 1C transcript in transfected osteoblastic cells, followed by functional assays using stable transfectants. For the study to be definitive, it was essential that ␣ 1C transcript levels be reduced to negligible levels. Antisense approaches have been used previously in this system (38), but we believed more efficient ablation would be achieved using a ribozyme approach with catalytic efficiency. Considering that targeting the initiating AUG codon or adjacent sequences may increase the antisense effect of ribozymes (39,40), ribozyme-6 was chosen for transfection and expression in ROS 17/2.8 cells.
Because the alternative splice that occurs in the osteoblast ␣ 1C mRNA is downstream in the coding region at transmembrane segments IVS3-IVS4 (11), ribozyme-6 was expected to cleave transcripts encoding both of these isoforms. We also engineered a ribozyme construct in which ribozyme 6 was inserted into a U1 snRNA expression cassette (16). This cassette offered the advantage of producing a ribozyme that should be highly expressed, resistant to exonucleases, and capable of concentrating in the nucleoplasm, where newly transcribed ␣ 1C transcripts should be most abundant.
In this study, we prepared stable sublines of ROS 17/2.8 cells in which ␣ 1C expression is highly inhibited by expression of active ribozyme 6. Control cell lines expressing inactive ribozyme constructs also were prepared. The selected clonal cells showing most efficient ribozyme-mediated ablation of ␣ 1C transcripts, as well as clonal control transfectants, were employed to investigate the relationship between ␣ 1C and Ca 2ϩ influx modulated by 1,25(OH) 2 D 3 and depolarization. 45 Ca 2ϩ uptake assays were performed in subconfluent cell monolayers of sublines expressing active and inactive ribozymes, as well as in the parental line. As presented under "Results," cells expressing active ribozyme 6 lost their rapid Ca 2ϩ response to both 1,25(OH) 2 D 3 and depolarizing high K ϩ buffer. In contrast, cells transfected with control plasmids maintained their sensitivity to 1,25(OH) 2 D 3 and high K ϩ buffer, demonstrated by rapid 45 Ca 2ϩ influx at levels comparable with nontransfected cells. Taken together, these data establish that the ␣ 1C subtype of the L-type channel is the depolarization and 1,25(OH) 2 D 3 -responsive VSCC in osteoblasts.
The results presented here integrate and extend previous findings regarding the nature of the osteoblast VSCC. For the first time, we definitively demonstrated that the rapid membrane-initiated effects of 1,25(OH) 2 D 3 on Ca 2ϩ influx in osteoblasts are transduced by the ␣ 1C subtype of the VSCC. By defining the cardiac subtype to be the molecular form of VSCCs responsive to 1,25(OH) 2 D 3 and depolarization, we are now in a position to further explore the physiological importance of the VSCC in osteoblast function and to begin to define the signaling pathways activated directly by the VSCC.