Isolation and Characterization of the 5′-Upstream Region of the Human N-type Calcium Channel α1BSubunit Gene

ω-Conotoxin-sensitive N-type Ca2+ channels, unlike dihydropyridine-sensitive L-type channels, are exclusively expressed in nervous tissues. To understand the molecular basis for neuron-specific expression of the N-type channel, we have isolated genomic clones encoding the human α1B subunit gene, localized to the long arm of chromosome 9 (9q34) by fluorescence in situ hybridization, and characterized its 5′-upstream region. The proximal promoter of the α1B subunit gene lacks a typical TATA box, is highly GC-rich, and contains several sequences for transcription factor binding. Primer extension experiments revealed the presence of two transcription start sites. In vitro transfection study of the α1B subunit-luciferase fusion gene showed that the 4.0-kb 5′-flanking region of the α1B gene functions as an efficient promoter in neuronal cells but not in glioma or nonneuronal cells, consistent with the patterns of the endogenous α1B gene expression in these cells. Deletion analysis of α1B subunit-luciferase fusion gene constructs further revealed the presence of several cis-acting regulatory elements, including a potential repressor located in the distal upstream region (−3992 to −1788) that may be important for the neuron-specific expression of the N-type Ca2+ channel α1B subunit gene.

-Conotoxin-sensitive N-type Ca 2؉ channels, unlike dihydropyridine-sensitive L-type channels, are exclusively expressed in nervous tissues. To understand the molecular basis for neuron-specific expression of the N-type channel, we have isolated genomic clones encoding the human ␣ 1B subunit gene, localized to the long arm of chromosome 9 (9q34) by fluorescence in situ hybridization, and characterized its 5-upstream region. The proximal promoter of the ␣ 1B subunit gene lacks a typical TATA box, is highly GC-rich, and contains several sequences for transcription factor binding. Primer extension experiments revealed the presence of two transcription start sites. In vitro transfection study of the ␣ 1B subunit-luciferase fusion gene showed that the 4.0-kb 5-flanking region of the ␣ 1B gene functions as an efficient promoter in neuronal cells but not in glioma or nonneuronal cells, consistent with the patterns of the endogenous ␣ 1B gene expression in these cells. Deletion analysis of ␣ 1B subunit-luciferase fusion gene constructs further revealed the presence of several cis-acting regulatory elements, including a potential repressor located in the distal upstream region (؊3992 to ؊1788) that may be important for the neuron-specific expression of the N-type Ca 2؉ channel ␣ 1B subunit gene.
Voltage-sensitive Ca 2ϩ channels (VSCC) 1 found in the plasma membranes of many excitable cells regulate calcium entry to mediate a wide variety of physiological functions, encompassing membrane excitability, neurite outgrowth, enzyme regulation, gene expression in cell bodies, and neurotransmitter release from nerve terminals (1,2). A number of Ca 2ϩ channel types have been described based upon biochemical, pharmacological, and electrophysiological studies, including L-, T-, N-, P/Q-, and R-types (3)(4)(5). The skeletal muscle L-type Ca 2ϩ channel, the first one to be defined at the molecular level, is composed of multiple subunits, ␣ 1 , ␣ 2-␦ , ␤, and ␥ (6, 7), and the corresponding cDNAs have been cloned and sequenced (reviewed in Refs. 8 and 9). The cloned skeletal muscle ␣ 1 subunit (␣ 1S ) exhibits structural features common to voltage-gated cation channel gene families and is capable of directing expression of Ca 2ϩ channel activity in heterologous expression systems. Homology screening resulted in the isolation of five additional ␣ 1 subunit cDNAs that encode either dihydropyridine (DHP)-sensitive L-type channels (␣ 1C and ␣ 1D ) or DHP-insensitive high voltage-activated Ca 2ϩ channels (␣ 1A , ␣ 1B , and ␣ 1E ). Distinct isoforms of the ␣ 1C and ␣ 1D subunits are generated by alternative splicing and are responsible for heterogeneity of DHP-sensitive L-type channel subtypes present in a variety of excitable and nonexcitable cells (10 -12). Although multiple isoforms also have been reported, expression of DHP-insensitive Ca 2ϩ channel ␣ 1 subunits seems to be restricted to nervous tissues and the cells of neuronal origin as is the case for the ␣ 1B subunit encoding the -conotoxin GVIAsensitive N-type channel (13).
Little is known about the molecular mechanisms underlying distinct cell-type and tissue-specific expression patterns of VSCC subtypes, despite the progress in our understanding of transcriptional regulation of other ion channel genes. Neuronspecific expression of the rat brain type II Na ϩ channel is regulated by a silencer element located in the 5Ј-flanking region of the gene (14,15). The transcription of a potassium channel gene, Kv1.5, is regulated by glucocorticoids and cAMP in both GH3 cells and cardiac myocytes (16 -18). The expression of the rat DHP-sensitive L-type ␣ 1D subunit gene in NG108-15 cells increases during differentiation in the presence of prostaglandin E 1 or retinoic acid (19). A recent report provided an initial description of the transcriptional regulation of the rat ␣ 1D subunit gene and identified a novel enhancer that consists of an (ATG) 7 trinucleotide repeat sequence (19). Because of their differential expression patterns, it is of interest to examine molecular mechanisms that underlie the regulation of N-type Ca 2ϩ channel ␣ 1B subunit gene expression and compare it with those involved in the L-type ␣ 1D subunit gene expression. In the present study, we report isolation of the genomic clones containing the 5Ј-flanking sequence and its chromosomal location, and we provide an initial characterization of the 4.0-kilobase pairs (kb) upstream promoter region of the human N-type Ca 2ϩ channel ␣ 1B subunit gene.

MATERIALS AND METHODS
Northern Hybridization-A human multiple tissue Northern blot (Clontech Laboratories, Palo Alto, CA) was prehybridized at 42°C for 6 h in a hybridization solution (5 ϫ SSPE, 10 ϫ Denhardt's solution, 50% formamide, 2% SDS, and 100 g/ml salmon sperm DNA) and hybridizied at 42°C for 24 h with the probes labeled with 32 P by random priming. The ␣ 1B probe was a 230-base pair (bp) fragment of the 5Ј region of the ␣ 1B subunit cDNA (corresponding to nucleotide residue number 263-493 as in GenBank TM accession no. M94172) which was generated by a polymerase chain reaction (PCR). The ␣ 1D cDNA probe * 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) U76666.
¶ To whom correspondence should be addressed: Laboratory of Neurochemistry, Bldg. 36 was the 443-bp cDNA insert isolated from p60Z plasmid (12) (corresponding to nucleotide residues 2803-3246 as in GenBank TM accession no. M57682). Following hybridization the blot was rinsed in solution I (2 ϫ SSC and 0.05% SDS) at room temperature for 15 min twice and washed in solution II (0.1 ϫ SSC and 0.1% SDS) at 50°C for 40 min with one change of fresh solution. The blots were exposed to Biomax MR x-ray film (Eastman Kodak Co.) at Ϫ80°C for 3 days.
The genomic inserts isolated from positive plaques were subcloned into pGEM7Zf(ϩ) plasmid (Promega, Madison, WI). Both strands of the genomic inserts were sequenced by the chain termination sequencing method (20). Sequence analysis and data base searches were performed with the GCG software package.
Primer Extension Analysis-A 21-mer antisense oligonucleotide primer (NAPE1, as indicated in Fig. 2), complementary to a portion of the first exon (56 -76; GenBank TM accession no. M94172) of the human N-type calcium channel ␣ 1B gene, was end-labeled with 32 P by T4 polynucleotide kinase. The 32 P-labeled NAPE1 was annealed to 2 g of human neuroblastoma SH-5YSY cell poly(A) ϩ RNA in 40 mM PIPES (pH 6.8), 1.25 mM EDTA (pH 8.0), 125 mM NaCl, and 75% foramide for 1 h at 42°C. Hybrids were ethanol-precipitated and extended by avian myeloblastosis virus (AMV) reverse transcriptase in a mixture containing 0.06 g of actinomycin D. Extension products were analyzed on 8% ployacrylamide-urea sequencing gels.
Fluorescence in Situ Hybridization-The plasmid containing a 9.5-kb human ␣ 1B subunit genomic insert (pNAG Sac2-2) was labeled with biotin-dUTP by nick translation. The labeled probe was combined with human Cot-1 DNA and hybridized to a normal metaphase chromosome from phytohemagglutinin-stimulated peripheral blood lymphocytes in a solution containing 50% formamide, 10% dextran sulfate, and 2 ϫ SSC. Specific hybridization signals were detected by incubating the hybridized slides in fluoresceine conjugated avidin. The slides were counterstained with propidium iodide and analyzed.
Plasmid Construction-The human ␣ 1B gene-luciferase fusion plasmids were constructed by subcloning into the polylinker region of pGL2-Basic vector (Promega) with the following restriction fragments from the human ␣ 1B genomic clone pNAG Sac2-2: a 4.0-kb BamHI/BssHII fragment (Ϫ3992L), a 1.7-kb XhoI/BssHII fragment (Ϫ1788L), and 0.1-kb NotI/BssHII fragment (Ϫ110L). Another deletion constructs were generated by Discrete-Delete ExoIII/mung bean nuclease deletion kit (Epicentre Technologies, Madison, WI). All plasmids were sequenced to determine the deletion end points and to exclude the possibility of recombination in host Escherichia coli. The control plasmids pRSVL (a gift from Dr. Sung O Huh; Sloan-Kettering Institute, New York) and pCMV␤ (Stratagene, La Jolla, CA) contain the Rous sarcoma virus (RSV) promoter fused to the luciferase gene and the cytomegalovirus (CMV) promoter fused to the ␤-galactosidase gene, respectively.
Cell Culture-Human neuroblastoma SH-5YSY and BE(2)-C cells, which were provided by Dr. June Biedler (Sloan-Kettering Institute, New York), were grown in 1:1 Eagle's minimal essential medium and Ham's nutrient medium F12 supplemented with 10% fetal bovine serum (FBS). Two mouse neuroblastoma X rat glioma hybrid cell lines, NG108-15 and 140-3, were grown in Dulbecco's modified Eagle's medium with 10% FBS, 100 M hypoxanthine, 1 M aminopterin, and 16 M thymidine. Murine neuroblastoma NS20Y and human glioma U251 cells were cultured in Dulbecco's modified Eagle's medium with 10% FBS. PC12 cells were maintained in Dulbecco's modified Eagle's medium containing 10% FBS and 5% horse serum. HeLa and HepG2 cells were grown in Eagle's minimal essential medium with 10% FBS. All culture media were supplemented with penicillin G (100 units/ml) and streptomycin (100 g/ml).

Transient Transfection and Luciferase
Assay-Equimolar amounts of the human ␣ 1B gene-luciferase plasmids (3 g were used for the shortest deletion construct Ϫ110L) and pCMV␤ were cotransfected into subconfluent cells in 60-mm culture dishes using LipofectAMINE (Life Technologies, Inc.). Cells were harvested 24 h after transfection and lysed in 1 ϫ cell culture lysis reagent, and activities were assayed using luciferase assay reagent (Promega). The light emitted was integrated over a 15-s interval on a Monolight 2010 luminometer (Analytical Luminescence Laboratory, San Diego, CA) and expressed as light units. ␤-Galactosidase was monitored by an assay kit in the same lysate (Promega). The luciferase activity of fusion constructs was normalized to ␤-galactosidase activity and expressed as a percentage of the RSV promoter activity of pRSVL.
Reverse Transcription (RT)-PCR Analysis-Total RNAs were prepared by a guanidium thiocyanate-phenol extraction method (21). First strand cDNA was synthesized using 1 g of total RNA, which had been treated with RNase-free DNase, by using SuperScript preamplification system (Life Technologies, Inc.). Following the first strand cDNA synthesis, PCRs were done in a 50-l reaction mixture containing 10 mM Tris, pH 8.

RESULTS
Differential Expression of ␣ 1B and ␣ 1D Transcripts-The distribution of ␣ 1B and ␣ 1D Ca 2ϩ channel transcripts in human tissues was examined by Northern blot analysis. A single 10.5-kb transcript of the ␣ 1B gene was present in the brain, but there was no hybridized band detected in other human tissues examined (Fig. 1A). The size of the ␣ 1B mRNA present in human brain is similar to the ϳ10-kb rat brain transcript (13). The ␣ 1D cDNA probe, which is derived from the least conserved intracellular linker of the II and III loop of the ␣ 1D subunit (10), strongly hybridized to the transcripts of two sizes, 9.5 and 8.5 kb, in the brain, placenta, lung, liver, kidney, and pancreas. An additional band of 6.5 kb in size was detected in the kidney and pancreas (Fig. 1B). No bands were detected in the heart and skeletal muscle, indicating that the ␣ 1D probe did not cross hybridize to the L-type ␣ 1C and ␣ 1S subunit transcripts. The 9.5-, 8.5-and 6.5-kb sizes of the human ␣ 1D transcripts are similar to those present in rat brain (10,22), but smaller than the 11-kb transcript found in pancreatic islets (23). Thus, expression of N-type Ca 2ϩ channel ␣ 1B subunit, unlike the DHPsensitive L-type channel which shows a broad distribution pattern in various tissues, is limited to nervous tissues.
Cloning and Nucleotide Sequence of the 5Ј-Upstream Region of Human ␣ 1B Subunit Gene-Screening a human lung fibroblast genomic library yielded two overlapping clones which, taken together, contained 6.5 kb of the 5Ј-upstream region, the first exon, and 2.5 kb of the first intervening sequences. A major portion, approximately 4.7 kb, of the human ␣ 1B genomic clone was sequenced and is shown in Fig. 2. The 4670-nt sequence contains 3991 nt of the 5Ј-flanking region, 430 nt of the first exon (148 nt of 5Ј-untranslated region and 282 nt of coding sequence), and the 249-nt part of the first intron. The exon 1 sequence is identical to that of the human N-type Ca 2ϩ channel cDNA reported by Williams et al. (24).
Identification of the Transcription Initiation Site-Primer extension analysis was performed in order to determine the transcription initiation site of the human ␣ 1B gene. 32 P-Labeled oligonucleotide primer NAPE1 was annealed to poly(A) ϩ RNA of human neuroblastoma SH-5YSY cells, extended by AMV FIG. 2. Nucleotide sequence of 5-upstream region through the first intron of human N-type calcium channel ␣ 1B gene. Two transcription initiation sites are indicated by bent arrows. Nucleotides are numbered with respect to the 3Ј-major transcription start site that is indicated as ϩ1. The coding sequence of the first exon is shown as codon triplets, and the intron sequence is depicted in lowercase letters. The 5Ј-end points of deletion constructs (Ϫ3992, Ϫ1788, Ϫ1559, Ϫ1289, Ϫ1057, Ϫ803, Ϫ347, Ϫ110), and the common 3Ј-end point (ϩ86) are indicated by arrowheads and numbers. Putative transcription factor binding sites are marked by boxes with their names indicated below. A homologous sequence to consensus NRSE is underlined. Another common motif found in several neuron-specific genes is double underlined. The NAPE1 oligonucleotide used for primer extension analysis is shown as a dashed line. reverse transcriptase, and the extended products were separated by polyacrylamide-urea gel electrophoresis. A predominant band of 79 nt and a weaker band of 83 nt were detected on an autoradiogram (Fig. 3, lane 1). No extended products were observed when extension reaction was performed with E. coli tRNA templates which was used as a control for specificity of hybridization (Fig. 3, lane 2). From the size of the extended products and the location of the NAPE1 oligonucleotde primer, we were able to place major and minor initiation sites at 148 and 152 nt upstream of the translation start site, respectively.
Analysis of the sequence immediately upstream of the transcription initiation sites reveals that the ␣ 1B subunit gene promoter contains a CCAAT box (Ϫ59 in antisense orientation) but lacks a typical TATA consensus motif. The sequences surrounding the major (GGTGAGGC) and minor transcription initiation sites (GTCGGGTG) are different from the initiator sequence (CTCANTCT) present in the promoter region of TATA-less genes, including the synapsin I gene (35) (the underlined nucleotide represents the transcription initiation site). In addition, this promoter is highly rich in G ϩ C content and 72 CpG and 85 GpC dinucleotides are located in a region of 500 bp (positions Ϫ400 to ϩ100).
Chromosomal Localization of ␣ 1B Gene-Fluorescent in situ hybridization using the biotin-labeled probe resulted in specific labeling of the distal end of long arm of chromosome 9 (Fig. 4A). A second experiment was carried out in which a chromosome 9 centromere-associated satellite probe was cohybridized with the human ␣ 1B genomic probe to confirm the identity of the specially labeled chromosome. This experiment showed the specific labeling of the centromeric heterochromatin and the distal long arm of chromosome 9 (Fig. 4B). Measurement of 10 specifically hybridized chromosome 9s demonstrated that the human Ca 2ϩ channel ␣ 1B gene is located at a position which is 97% of the distance from the centromere to the telomere of chromosome arm 9q, an area that corresponds to band 9q34. A total of 80 metaphase cells were examined with 65 exhibiting specific signals.
Cell Type-specific Expression by 5Ј-Flanking Region of the ␣ 1B Gene Promoter-To address whether the 5Ј-flanking sequence of the human ␣ 1B subunit gene contains the regulatory sequences utilized in a cell type-specific manner we made a fusion gene construct Ϫ3992L, containing a 4.0-kb 5Ј-flanking sequence of the ␣ 1B gene (Ϫ3992 to ϩ86) linked to the promoterless luciferase reporter vector pGL2-Basic. This plasmid was transfected into a variety of neuronal and nonneuronal cell lines and assayed for luciferase activity. As controls, the plasmids pGL2-Basic and pRSVL were transfected into parallel cultures of each cell line. In all the cell lines tested, the pGL2-Basic plasmid was ineffective in driving expression of luciferase activity, while transfection of the pRSVL resulted in high levels of expression. The results of such an analysis are shown in Fig. 5A. In neuronal cells such as SH-5YSY, BE(2)-C, NS20Y, NG108-15, and PC12 cells, luciferase activities from the ␣ 1B fusion gene construct Ϫ3992L, were approximately 40 -60% of those from the RSV promoter. In contrast, reporter gene expression was very low, maximally 5% of RSV activity, in the glioma cell line U251 as well as in the nonneuronal HeLa and HepG2 cells. Interestingly, the luciferase gene was poorly expressed in one of the mouse neuroblastoma-rat glioma hybrid cell lines, 140-3 cells, consistent with our recent electrophysiological studies showing that 140-3 cells do not expressed any of high voltage-activated currents (36).
RT-PCR, which was carried out to detect the endogenous ␣ 1B mRNA expression in the same cell lines used in transfection studies, yielded the amplified product corresponding to the predicted size of 355 bp in SH-5YSY, BE2(C), NS20Y, NG108-15 and PC12 cells but not in 140-3, U251, HeLa, and HepG2 cells (Fig. 5B). The level of endogenous ␣ 1B gene expression in NS20Y, NG108-15, and PC12 cells, as judged by the intensity of the amplified bands on agarose gels, seemed to be higher than in SH-5YSY and BE(2)-C cells, suggesting that there is a good correlation between reporter gene expression from the ␣ 1B -luciferase fusion gene construct and endogenous ␣ 1B gene expression. Taken together these results, we conclude that the 4.0-kb 5Ј-flanking sequence contains the cis-regulatory elements important for directing expression of the ␣ 1B gene in The Distal Upstream Region of the ␣ 1B Gene Promoter for Neuron-specific Expression-To locate a cis-acting regulatory element for neuron-specific expression of the ␣ 1B gene, a series of ␣ 1B -luciferase fusion plasmids were constructed and transfected into NS20Y cells and HeLa cells. Progressive 5Ј deletions between nucleotides Ϫ3992 and Ϫ110 were made from the Ϫ3992L construct using either specific restriction enzymes or exonuclease III digestion protocols. As shown in Fig. 6A, a deletion from Ϫ3992 to Ϫ1788 resulted in an approximately 10-fold increase in luciferase activity in HeLa cells, but no change in NS20Y cells, indicating the presence of a repressor element between Ϫ3992 and Ϫ1788 that inhibits the reporter gene expression in HeLa cells. Further deletions of the region between Ϫ1788 and Ϫ1289 had little effect on luciferase activity in both cell lines (Fig. 6A). However, removal of the region from Ϫ1289 to Ϫ1057 resulted in a small but significant 2.0fold increase only in HeLa cells, suggesting that this region may contain another weak repressor element. Extension of 5Ј deletions to nucleotide Ϫ110 gradually reduced the luciferase activity in NS20Y and HeLa cells (Fig. 6, A and B). On the basis of these results, it is likely that at least two negative regulatory elements with different strengths, distal (Ϫ3992 and Ϫ1788) and proximal (Ϫ1289 and Ϫ1057) region, may play critical roles in the neuron-specific regulation of the N-type Ca 2ϩ channel ␣ 1B subunit. DISCUSSION In this study we report the cloning, chromosomal localization, and molecular analysis of the 5Ј-flanking region of the human N-type Ca 2ϩ channel ␣ 1B subunit gene. A single 10.5-kb ␣ 1B mRNA transcript was detected only in the brain among the human tissues examined, whereas L-type ␣ 1D mRNAs were detected in a variety of tissues (Fig. 1). The ␣ 1B transcripts FIG. 5. Cell type-specific expression of the human ␣ 1B subunit gene. A, cell type-specific expression directed by the 5Ј-flanking sequence of the human ␣ 1B gene. The ␣ 1B -luciferase fusion gene Ϫ3992L was cotransfected with internal control plasmid pCMV␤. The luciferase activity was normalized to ␤-galactosidase activity and expressed as a percentage of the RSV promoter activity in that cell. The histograms show the mean Ϯ S.E. of three independent experiments, each of which was performed in triplicate. B, RT-PCR analysis of endogenous ␣ 1B gene expression. The first-strand cDNA was reverse-transcribed using total RNA from each cell and amplified with a specific ␣ 1B gene primer and a GAPDH primer for 30 cycles. The PCR products were electrophoresed on an 1.5% agarose gel.
FIG. 6. Deletion analysis of the human ␣ 1B gene promoter. A series of ␣ 1B gene fragments containing progressive 5Ј deletions and a common 3Ј end at ϩ86 nucleotide were inserted upstream of promoterless luciferase gene pGL2-Basic. The promoter activity of each ␣ 1Bluciferase fusion gene in NS20Y and HeLa cells was determined and expressed as a percentage of the RSV promoter activity (A) or as a fold of repression (B). Results are the mean Ϯ S.E. from at least three separate transfection experiments. A repression fold was defined as a ratio of luciferase activity in NS20Y cells to that of HeLa cells expressed as a percentage of luciferase activity of the reference construct pRSVL. Statistical significance was calculated using Student's t test. *p Ͻ 0.01 were generated from the single ␣ 1B gene, which was mapped to the distal end of the long arm of human chromosome 9 (Fig. 4), utilizing the major transcription start site located at 148 nt and the minor start site located at 152 nt 5Ј-upstream from the ATG translation start site (Fig. 3). The 4.0-kb 5Ј-flanking sequence of the ␣ 1B gene contained a promoter which was capable of directing expression of the ␣ 1B transcript in neuronal cells and repressing its expression in nonneuronal cells (Fig. 5). Deletion analysis of ␣ 1B subunit-luciferase fusion gene constructs indicated the presence of cis-acting regulatory elements located in the distal upstream region (Ϫ3992 to Ϫ1788) that may be critical for the neuron-specific expression of the ␣ 1B subunit gene (Fig. 6).
Over the past decade, considerable progress has been made in elucidating molecular mechanisms for the transcriptional activation of tissue-and cell type-specific expression of genes in nonneuronal cell types, such as erythrocytes, lymphocytes, and hepatocytes. More recently, the molecular bases for neuronspecific gene expression has also been examined (reviewed in Refs. 37 and 38). The transcription factor, termed neural-restrictive silencer element (NRSE)-binding factor (NRSF) (39 -41) binds a 21-bp NRSE sequence present in the 5Ј-upstream region of neural-specific genes to selectively repress the transcription of these genes in nonneuronal cells.
The restricted expression of the N-type ␣ 1B gene in the central nervous system and wide distribution of L-type ␣ 1D transcripts provide us with an excellent opportunity to examine and compare molecular bases governing Ca 2ϩ channel gene expression. Consistent with the broad mRNA expression within and outside of the central nervous system, we did not find any sequences with similarity to the NRSE sequence in the 5Јupstream region of the rat ␣ 1D gene. Furthermore, the transcription of the rat ␣ 1D gene is regulated by both cis-acting positive and negative elements in the 5Ј promoter region and by an enhancer that consists of (ATG) 7 trinucleotide repeats (19). Inspection of 5Ј-flanking sequence of the human ␣ 1B gene, however, revealed the nucleotide sequence (NRSE-␣ 1B ) homologous to the NRSE (nucleotide Ϫ810 to Ϫ789 as shown in Fig.  2). Although overall sequence identity to the 21-bp NRSE consensus sequence is 57%, the 5Ј half of 10-bp NRSE-␣ 1B fragment showed a 80% sequence identity to that of the NRSE. We have subcloned NRSE-␣ 1B into the 5Ј upstream of SV40 promoter linked to luciferase reporter gene to test whether or not an NRSE-␣ 1B could function as a repressor element in nonneuronal cells. Luciferase activity assay showed that one or two copies of this putative motif did not affect the SV40 promoter activity, whereas one copy of the NRSE from SCG 10 gene was sufficient to repress its activity to 30% of the control in HeLa cells (data not shown). Since the promoter activity of 4.0-kb 5Ј flanking region of the human ␣ 1B gene in various cell lines was in excellent agreement with RT-PCR analysis of the endogenous ␣ 1B mRNA expression (Fig. 5), we used the two of these lines, NS20Y and HeLa, to search for the cis-acting regulatory elements further 5Ј upstream of the ␣ 1B gene. In vitro transient transfection of truncated ␣ 1B -luciferase fusion gene indicated that the region between Ϫ3992 and Ϫ1788 contains a repressor element(s) responsible for the neuron-specific expression of the N-type ␣ 1B subunit gene (Fig. 6). Since sequence analysis did not indicate the presence of any sequence with similarity to the NRSE, a repressor functional in the N-type ␣ 1B subunit gene may be distinct from the ones already identified. Further studies are required to establish whether this region contains a unique neuron-specific element that is capable of binding a NRSF.
Our results in the present study suggest that selective repression by negative cis-regulatory elements is responsible for neuron-specific expression of the human Ca 2ϩ channel ␣ 1B subunit gene as is the case for the type II Na ϩ channel and other genes exclusively expressed in the nervous system. In addition to the NRSE that is the primary determinant for the tissue specificity, the core promoters are also important for conferring substantial neuronal specificity to several genes such as synapsin I, II, and myelin basic protein (34,42,43). In contrast, the activity of the human ␣ 1B subunit gene core promoter (Ϫ110L plasmid) was apparently similar in NS20Y and HeLa cells (Fig. 6A), indicating that the core promoter itself does not confer the neuron specificity to the ␣ 1B gene. However, we cannot rule out the possibility of concerted interactions between cell type-specific distal upstream repressor element(s) and the general minimal promoter.
In summary, we have presented an initial characterization of the human N-type Ca 2ϩ channel ␣ 1B subunit gene and identified a region in the 5Ј-upstream of the gene (Ϫ3992 and Ϫ1788) that contains negatively acting cis-regulatory elements responsible for neuron-specific expression of the ␣ 1B gene. Further deletion analyses of the region between Ϫ3992 and Ϫ1788 and the studies of the DNA-protein interactions between transcription factors and the putative repressor elements should help to elucidate the molecular mechanisms of transcriptional regulation underlying spatiotemporal expression of VSCC ␣ 1 subunit genes in the nervous systems.