Skeletal Muscle and Brain Isoforms of a &Subunit of Human Voltage-dependent Calcium Channels Are Encoded by a Single Gene*

Clones of the B1-subunit of the voltage-dependent calcium channel (VDCC) from human skeletal muscle and hippocampus cDNA libraries, and from human genomic libraries, were isolated using a human skeletal muscle B1 cDNA probe generated by polymerase chain reaction. The skeletal muscle p1 cDNA (&M) encodes a protein of 523 amino acids that is 97% identical to the rabbit skeletal muscle &subunit. Two different cDNAs, @,Bl and &B2, were obtained from the human hippocampus library. The BIBl transcript encodes a protein of 478 amino acids that is identical to the skeletal muscle &subunit (BIM), except for an internal region of 52 amino acids. The B1B2 transcript encodes a protein of 596 amino acids. The O1B2 polypeptide is identical to the &Bl polypeptide at amino acids 1-444; however, it has a unique 152 amino acid carboxyl terminus. Like BIBl, it differs from BIM at the internal 52 amino acids. Analysis of the B1 gene structure dem- onstrates that these three cDNAs represent transcripts encoded by a single Dl gene. Transcripts from the B1 gene were detected in RNA from skeletal muscle, and but not in RNA from liver, stomach, or

*This work was supported by a grant from the Department of Veteran Affairs (to K. H.) and by a grant from the Waisman Center on Mental Retardation and Human Development (to R. G. G.). 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 thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) M92300 (genomic sequence), M92301 (@,M), M92302 (&Bl), and
Recipient of a B. B. Sankey Anesthesia Advancement Award and a Foundation for Anesthesia Education and Research Young Investigator Award.

M92303 (PlB2).
The skeletal muscle DHP-sensitive VDCC consists of five distinct subunits, al, a2, p, y, and 6 (Caterall, 1988). a1, P, and y are encoded by separate genes, whereas a2 and 6 are produced by proteolytic cleavage of a larger precursor encoded by a single gene (De Jongh et al., 1991). When introduced into Xenopus oocytes (Mikami et al., 1989) or into fibroblasts (Perez-Reyes et al., 1989;Tanabe et al., 1990) in the absence of the other subunits, the al-subunit forms a channel that is DHP-sensitive and exhibits many of the characteristics of the native calcium channel. The subunit composition of cardiac, smooth muscle, and neuronal DHP-sensitive VDCCs has not been completely determined. Analysis of purified cardiac and brain VDCC subunits with antibodies to the skeletal muscle VDCC subunits suggests that there is a high degree of homology between the al-and a2/6-subunits expressed in different tissues (Cooper et al., 1987;Ahlijanian et al., 1990). Northern blot analysis (Ruth et al., 1989) and monoclonal antibody binding  demonstrate that the &subunit is expressed in skeletal muscle and brain. The ysubunit appears to be expressed exclusively in skeletal muscle (Jay et al., 1990;Bosse et al., 1990).
In order to determine the specific contribution of each of the subunits to channel function, al-subunits from different tissues have been coexpressed with a2/6-, 8-, and y-subunits from skeletal muscle in Xenopus oocytes and L-cells (Singer et al., 1991;Mori et al., 1991;Varadi et al., 1991;Lacerda et al., 1991). Although the resulting channel properties depend both on the a1 subtype and the recipient cell type, the general conclusion is that the production of calcium channels having normal physiological properties is greatly enhanced by the coexpression of the a&, p-, and y-subunits (for review, see Catterall (1991) and Miller (1992)). The skeletal muscle psubunit has the largest effect on functional expression of the skeletal and cardiac muscle and several forms of the brain a]subunits. Coexpression of these al-subunits and the skeletal muscle &subunit increases the calcium current and accelerates activation and inactivation more than 10-fold (Singer et al., 1991;Mori et al., 1991;Varadi et al., 1991;Lacerda et al., 1991).
Skeletal muscle P-subunits have a molecular mass of approximately 58 kDa. They are phosphorylated by CAMPdependent protein kinases, protein kinase C, and cGMPdependent protein kinase (Curtis and Catterall, 1985;Takahashi et al., 1987;Jahn et al., 1988) but are not glycosylated nor do they copurify with membranes . Features of the primary structure of the rabbit skeletal f i l polypeptide (Ruth et al., 1989) and biochemical data suggest that it is a peripheral membrane protein that interacts with an intracellular domain of the a,-subunit and that it may modulate a1 function through a phosphorylation pathway.
The skeletal muscle PI cDNA detects transcripts of 1.6 and 22967 1.9 kb in skeletal muscle RNA and a 3-kb transcript in rabbit brain RNA (Ruth et al., 1989). It is not known whether these transcripts are products of the same gene or of two or more closely related genes. In order to determine the relationship between the skeletal muscle and brain p1 transcripts, we compared the nucleotide sequence of Dl cDNAs from human skeletal muscle and hippocampus cDNA libraries to portions of the gene(s). Our results demonstrate that at least one skeletal muscle and two brain transcripts are the product of the same p1 gene.

EXPERIMENTAL PROCEDURES
Methodology-Except where noted, recombinant DNA methods were adapted from protocols described by Sambrook et al. (1989) and Ausubel et al. (1987).
Libraries-The human fetal skeletal muscle cDNA library was kindly provided by Dr. L. Kunkel (Koenig et al., 1987). The human hippocampus cDNA library and the human placental cosmid library were purchased from Stratagene Inc.

5'-CGCGCGCGAATTCACNCCNCCNCCNCAYGG-3', correspond-
Screening the Libraries-Degenerate oligonucleotide primers, B1, ing to amino acids 248-253, and B3, 5"GCGCGAATTCGGNG-GRTGNGTNGCYTTCCA-3', corresponding to amino acids 454-460 of the rabbit skeletal muscle P1-subunit (Ruth et al., 1989), were synthesized by the University of Wisconsin Biotechnology Center. Each primer contained an EcoRI restriction site in addition to the 17-20 nucleotides of sequence corresponding to the PI peptide sequence. These primers were used in a polymerase chain reaction (PCR) using human skeletal muscle cDNA as a template (see below for conditions). The PCR product was cloned into the EcoRI site of pBluescript KS' (Stratagene, Inc.). DNA sequence analysis confirmed the identity of the insert in one clone, PCRD3, which was subsequently used as a probe to screen the human cDNA and genomic libraries.
Sequencing-DNA sequence was determined by the method of Sanger et al. (1977) using the Sequenase kit (United States Biochemical Corp.) and double-stranded plasmid as a template. Nucleotide and polypeptide sequence analyses were performed using GCG software (Devereaux et al., 1984) and a VAX computer.
PCRs"A11 PCRs contained a DNA template (see below), 200 p~ of each deoxynucleotide triphosphate, 1 p~ of each primer, 50 mM KCI, 1.5 mM MgC12, 10 mM Tris-HC1, pH 8.3, 0.001% gelatin, and 1.25 units of Taq DNA polymerase (Perkin-Elmer Cetus) in a total volume of 50 pl. The DNA was denatured for 2 min at 94 "C followed by 25-35 cycles of amplification with each cycle consisting of 45 s at 94 "C, 60 s at 47-62 "C, and 2 min at 72 "C. After the final cycle, an additional extension for 10 min at 72 "C was performed. The template DNA used was either 250-500 ng of genomic DNA, first strand cDNA made from 0.5 pg of polyadenylated RNA, or 1 ng of cloned DNA. The number of cycles and the annealing temperature were empirically determined for each primer pair. A 10-pl aliquot of each PCR product was electrophoresed on a 1.5% agarose gel in 0.045 M Tris borate, pH 8.3, 1 mM EDTA, stained with ethidium bromide, and visualized under UV light. The primers used in this study were B5,5' CGGAG-TACTTGGAAGCCTAT 3'; B7,5' TCCAGTCTGGGAGATGTGGT CGGAAGCAAGGT 3'; and B14, 5'CATGGCATGTTCCTGCTC-CT3'.
RNA Isolntion and cDNA Synthesis-Heart, liver, kidney, spleen, stomach, brain, and skeletal muscle tissues were collected from ICR mice. RNA was isolated using the protocol of Chirgwin et al. (1979), and polyadenylated RNA was purified using oligo(dT)-cellulose columns or messenger affinity paper (Amersham Corp.). Polyadenylated RNAs from human brain and heart were purchased from Clontech Laboratories Inc. First strand cDNA was made using polyadenylated RNA as template (Kawasaki and Wang, 1989).
Cloning PCR Products-PCR products were cloned into plasmid vectors by treating the product with the Klenow fragment of Escherichia coli DNA polymerase I, kinasing the blunt-ended PCR product with T4 polynucleotide kinase, and ligating the fragments into the S m I site of pBluescript KS' (Stratagene Inc.). 3'; B8,5' ACCAGGATGATGGGCCTCAT 3'; B12,5'CTGGTCCC-

RESULTS
Human cDNAs-cDNAs from the fetal skeletal muscle library and the hippocampus cDNA library were obtained by screening with PCRD3.
DNA Sequence Comparison-The DNA sequences of a skeletal muscle cDNA and portions of the hippocampus cDNAs were determined. Some of the hippocampus cDNAs contained a 450-bp insert with no homology to the rabbit skeletal muscle cDNA (Ruth et al., 1989). The sequence of this insert contained stop codons in all three forward frames and was flanked by consensus splice sites, so it most likely represents an unspliced intron present in about half of the hippocampus cDNAs.
The correctly spliced hippocampus cDNAs fall into two classes, plBl and &B2, representing two different transcripts. Fig. 1 shows a comparison of the hippocampus PlBl and &B2 cDNA and the skeletal muscle PIM cDNA. The sequences of the P1M and the BIBl cDNAs are identical with the exception of a region of 154 bp encoding 52 amino acids in BIM that has been substituted by 19 bp encoding a different 7 amino acids in plBl and /&Be. Both the 5'-and 3"noncoding sequences of DIM and &Bl are identical. The poly(A) tail is located 108 bp from the end of the coding sequence in both cDNAs. The sequence of P1B2 is identical to PlBl from nucleotide 1 to 1482. However P1B2 contains a novel 3' end encoding 152 amino acids and a 1.8-kb 3'-untranslated region.
Gene Structure-The two observed differences between the human skeletal muscle (&M) and brain isoforms (&Bl and &B2) and the difference between &Bl and &B2 could arise as a consequence of the use of alternative exons present in a single gene. Alternatively, these isoforms could be the products of several closely related genes. We examined the & gene structure in order to distinguish between these two possibilities. Several cosmids that contained portions of the PI gene(s) were isolated and used to determine the gene structure in the regions where the three cDNAs differ.
To investigate the difference between BIM and &Bl at amino acids 210-261 in PIM and 210-216 in &Bl, primers B7 and B8 that flank this region in the cDNAs were used in a PCR assay. A single 2-kb PCR product was obtained when either cosmid D3-3 or genomic DNA was used as template. This genomic fragment was cloned and sequenced. Exons were identified by comparing the genomic sequence with the DIM and &Bl cDNA sequences. The genomic sequence (shown in Fig. 2) contains two exons, one that is present in PIBl (and &B2) and the other that is present in PIM. Consensus splice sites flank each of the putative exons. Thus the difference between PIM and PlBl arises from the use of tissuespecific exons contained within a single gene. The structure of this region of the gene is represented in Fig. 4A.
To investigate the difference between the 3' ends of plBl and &B2, three primers; B5 (present in both &Bl and B1B2) B14 (present only in &Bl), and B12 (present only in P1B2) were used in PCR assays using cDNAs, cosmids, and genomic DNA as templates. Fig. 3A summarizes the results of PCRs using primers B5 and B14. A 228-bp fragment was produced when &Bl cDNA, cosmid D3-2, and genomic DNA were used as templates. As expected, no product was seen when P1B2 cDNA was used as template. Fig. 3B summarizes the results of PCRs using primer B5 and B12. A 154-bp product was produced when p1B2 was used as template. No product was observed when PIBl was used as template. A 1.85-kb fragment was produced when genomic DNA or cosmid D3-2 DNA was used as template.
The data shown in Fig. 3 allow us to deduce the P1 gene structure in this region which is summarized in Fig. 4. First, plBl cDNA and genomic DNA are colinear between primers B5 and B14, indicating that there are no introns in this region. DNA sequence analysis of a portion of cosmid D3-2 confirms that this portion of PIBl is colinear with genomic FIG. 1. Nucleotide sequence and conceptual translation of the human skeletal muscle DIM and hippocampus 6 , B l and @,B2 cDNAs. The sequence of &B2 (82) is shown in its entirety, with its conceptual translation shown directly below the sequence. Positions where the DIM ( M ) and/or PlBl (BI ) nucleotide sequence and amino acid sequence differ are shown above the P1B2 sequence. Lowercase letters represent untranslated sequences, and uppercase letters represent translated sequences. Gaps, represented by dots, were introduced to maximize alignment. These gaps are included in the coordinates of the nucleotide sequences.

CCAACAGCTTTGTCCGCCACGGCTCAGCGGAGTCCTACACCAGCCGTCCATCAGACTCTGATGTATCTCTGGAGGAGGACCGGGAAGCCTTAAGGAAGGA N S F V R O C S A E S Y T S R P S D S D V S L E E D R E A L R K E AGCAGAGCCCCAGGCATTAGCGCAGCTCGAGAACGCCAAGACCAAGCCAGTGGCATTTCCTGlGCGCAC~TGTTGCCTACAATCCCTCTCCAGGGGAT
.

TCGACCCMTCTlCCAGCTGGCCCGGACCClTCAGTTGGlCGCTCTGGATGClGACACCATCAATCACCCAGCCCAGCTGTCCAAGACCTCGClGGCCCC E R I F E L A R T L O L V A L D A D T I N H P A O L S K T S L A P
. ~ ~~

G O P P G L Y P S S H P P C R A C T L R A L S R O D T F O A D T P G
GCACCCGAAACTCTGCCTACACGGAGCTGGGAGACTCATGTGTGGACATGGAGACTGACCCCTCA~GGGCCCAGGGCTTGGAGACCCTGCAGGGGGCGC ttattttgtaaaaaaataagatgagcggcaaaaaaaaaaaaaaaaa l W . 6 DNA from nucleotide 1354 to its poly(A) tail (data not shown). Second, when primers specific to the 3' region of &Bl (primers B5 and B14) or &B2 (primers B5 and B12) are used in PCRs with cosmid D3-2 as the template, products of 228 bp and 1.85 kb are produced, respectively. Thus the B12 primer sequence (from the B2 exon) is located 1.6 kb downstream of the B14 primer sequence. The 1.85-kb B5-Bl2 PCR product was cloned into the pCRlOOO vector (Invitrogen, Inc.) and portions of it were sequenced. Fig. 4B summarizes the structure of the gene in this region and shows the sequence of the exon-intron junctions used in the P1B2 transcript.

E G G G P V L C R N K N E L E G U G R C V Y I R *
These data show that the PI primary transcript undergoes a novel splice in &B2, skipping over both the termination codon and the polyadenylation signal that are used in PIM and &Bl. The completely processed &B2 transcript is 1.98 kb longer than the plBl transcript and includes one or more 3' exons.
As shown in Fig. 4, all of the differences between the DIM, /IIBl, and P1B2 cDNAs are the result of alternative splicing of a single o1 gene.
Tissue Specificity-First strand cDNA was synthesized from polyadenylated RNA isolated from mouse brain, heart, liver, kidney, spleen, and stomach. These cDNAs were used in PCRs with primers B7 and B8 which span amino acids 191-280 in PIM and amino acids 191-235 in PlBl and P1B2.
No product was detected in reactions using cDNA made from liver, kidney, or stomach RNA. A 265-bp product was amplified in PCRs using cDNA made from skeletal muscle RNA or the plM cDNA clone. A 130-bp product was amplified in PCRs using cDNA made from brain, heart, or spleen RNA or the PIBl or D1B2 cDNA clones (Fig. 5).
Peptide Sequence Motifs-A comparison of peptide sequence motifs was undertaken in an attempt to detect potentially significant differences between the brain and the skeletal muscle polypeptides. The peptide sequence of human &B1, human PlB2, and human PIM (this study) and the Tissue-specific use of the skeletal muscle and brain exons. A photograph of PCR products obtained using the B7 and B8 oligonucleotide primers and first strand cDNA derived from RNA from skeletal muscle, brain, heart, spleen, kidney, liver, and stomach. rabbit skeletal muscle DIM (Ruth et al., 1989) were scanned for over 400 protein sequence motifs using the GCG software (Devereaux et al., 1984). Four different motifs were identified protein kinase C phosphorylation sites, CAMP-dependent protein kinase phosphorylation sites, casein kinase phosphorylation sites, and N-glycosylation sites. The location of these sites is shown in Fig. 6. The majority of these sites were conserved between the rabbit and human sequences and are found in the BlB1, &B2, and both the human and rabbit DIM isoforms. However, a potential protein kinase C phosphorylation site at serine 238 is located in a region exclusive to the FIG. 6. Summary of peptide sequence motifs identified in the &subunit. The conceptual translation of the human skeletal muscle PIM cDNA ( h M ) is shown in its entirety. Amino acids that differ in the rabbit skeletal muscle PIM cDNA ( r M ) from Ruth et al. (1989) are shown above hM, and amino acids that differ in the human hippocampus PIBl cDNA (hB1) and the human hippocampus P1B2 cDNA (hB2) are shown beneath hM. Dots represent identical amino acids, and asterisks represent gaps that were introduced to maximize alignment. Predicted phosphorylation sites for protein kinase C (filled circles), casein kinase I1 (open squares), and CAMP-dependent protein kinase (*) are shown above the sequence. Filled squares indicate potential N-glycosylation sites. The 6-bp conserved regions within the a-helical repeat identified by Ruth et al. (1989) are shown as shaded boxes. Boxed amino acids are specifically discussed in the text. skeletal muscle isoforms. &Bl and &B2 contain a potential protein kinase C phosphorylation site at serine 209 that is not present in PIM. Instead the human and rabbit PIM isoforms have a potential casein kinase phosphorylation site at serine 209. P,B2 lacks the potential protein kinase C phosphorylation sites at serine 495 and serine 509 present in PIM and &Bl but has potential casein kinase phosphorylation sites at serine 509, serine 540, serine 563, and serine 590 and a potential protein kinase C site at threonine 535. A CAMPdependent protein kinase phosphorylation site, threonine 205, is present in all four sequences and can be phosphorylated in vitro (De Jongh et al., 1989). Serine 182 has also been phosphorylated in vitro by CAMP-dependent protein kinase (Ruth et al., 1989); however, the replacement of this residue with glycine in the human polypeptides suggests that this phosphorylation may not be physiologically relevant. Ruth et al. (1989) identified a block of 8 similar amino acids (marked as shaded boxes in Fig. 6) repeated within four ahelical domains. The second of these a-helical domains is located in a region that is present in DIM, but not in plBl and &B2. Whether the absence of one of the repeats contributes to a functional difference between the isoforms remains to be determined.

DISCUSSION
A single gene encodes three different P1-subunits, including the &-subunit of the skeletal muscle dihydropyridine-sensitive calcium channel and at least two other &-subunits that are expressed in the brain, heart, and spleen. A polypeptide of 523 amino acids (&M) is produced in skeletal muscle, whereas two polypeptides of 478 (plBl) and 596 amino acids (B1B2) are produced in the brain. The sequences of PIM and plBl are identical with the exception of a region of 52 amino acids in &M that is replaced by 7 different amino acids in &Bl. DNA sequence analysis of the corresponding portion of the D l gene demonstrates that this difference is the result of use of one exon in skeletal muscle transcripts and another exon in brain transcripts. The two brain polypeptides, &Bl and &B2, are identical from amino acids 1-444, but differ at their carboxyl termini. plBl shares the same 34-amino acid carboxyl terminus as PIM. P,B2 has a unique 152-amino acid carboxyl terminus. Analysis of the gene structure in this region demonstrated that the primary transcript undergoes a novel splice in P1B2, skipping over both the termination codon and the polyadenylation signal used in DIM and plBl. The &B2 transcript is 1.87 kb longer than the &Bl and P1M transcripts and differs by the use of one or more 3' exons which we have not yet characterized.
The P1B2 cDNA probably corresponds to the 3.0-kb transcript seen using rabbit brain RNA on Northern blots (Ruth et al., 1989). The &Bl cDNA is 1.9 kb in length and has the same 5' end and uses the same polyadenylation site as the skeletal muscle DIM cDNA. Since the PIBl cDNA contains a complete open reading frame and a poly(A) tail, it probably represents a minor class of transcripts not detected on the Northern blots. Alternatively, it is possible that the &Bl transcript is actually 3.0 kb in length but utilizes a large 5'untranslated region not included in the PlBl cDNAs described here.
While the manuscript was in preparation, Pragnell et al. (1991) reported the sequence of a rat brain p1 cDNA, and Williams et al. (1992) reported the sequence of a human brain B cDNA that they called p2. These cDNAs correspond to our B1B2 and plBl cDNAs, respectively. In addition, cDNAs from two other p genes have recently been identified. Perez-Reyes et al. (1992) reported a rat brain /?-subunit cDNA that represents the transcript of the p2 gene that is expressed in heart, lung, and brain. Hullin et al. (1992) described a collection of cDNAs from heart, aorta, and brain. Several of these cDNAs represent transcripts from the p2 gene, whereas others represent transcripts from a third gene, p3.
Polymerase chain reactions that distinguish the skeletal muscle (DIM) and brain (plBl and P1B2) transcripts from the P1 gene in RNA isolated from various tissues suggest that the brain contains only the brain isoforms (plBl and P1B2), and skeletal muscle contains only the skeletal muscle isoform Brain and Muscle VDCC @-Subunits Encoded by a Single Gene (&M). Brain isoforms (plBl and/or P1B2 or some closely related gene) are also expressed in the heart and spleen. No P1-subunit RNA was detected in liver, kidney, or stomach.
Detection of p1 transcripts in the heart and spleen is in contrast to the results of Ruth et al. (1989) who observed 6subunit transcripts of 3.0 kb in rabbit brain RNA and 1.6 and 1.9 kb in rabbit skeletal muscle RNA and no hybridization with polyadenylated RNA from smooth muscle or heart. This discrepancy may be due to the increased sensitivity of PCR as compared with Northern analysis, or it could result from the detection of a closely related gene that is expressed in heart and spleen. The observed PCR products in heart and spleen RNA are not likely to be derived from either the p2 or p3 gene, since within the region containing the B7 and B8 primers, both genes are less than 70% identical to the p1 gene.
Therefore, the cardiac and spleen PCR products are produced from transcripts from the p1 gene or another more closely related gene or genes.
The identification of two p1 isoforms expressed in the brain raises questions about their association with specific channels.
A recent study showed that a monoclonal antibody to PIM immunoprecipitates both a DHP receptor and a w-conotoxin GVIA receptor in rabbit brain membrane preparations . Two polypeptides of 58 and 78 kDa were detected with this monoclonal antibody in partially purified w-conotoxin receptors from rabbit brain membranes. It seems likely that both the 478-amino acid &Bl-subunit (approximately 52 kDa) and the 596-amino acid P1B2-subunit (approximately 64 kDa) were isolated in this preparation. Several different peptide sequence motifs were identified in an analysis of the conceptual translation products of the rabbit PIM cDNA (Ruth et al., 1989), the human DIM cDNA, and the human plBl and p1B2 cDNAs. The @-subunit from the skeletal muscle DHP-sensitive calcium channel has been phosphorylated by a protein kinase intrinsic to isolated triads (Imagawa et al., 1987) and by CAMP-dependent protein kinases, protein kinase C, and a cGMP-dependent protein kinase (Curtis and Catterall, 1985;Takahashi et al., 1987;Jahn et al., 1988;De Jongh et al., 1989). Our analysis shows that the majority of the phosphorylation sites are conserved between the rabbit and human sequences and are present in all three (PIM, PIBl, and plB2) isoforms. However, some potential phosphorylation sites are found exclusively in the skeletal muscle (&M) isoforms or the brain (plBl and P1B2) isoforms or only in the P1B2 isoform. The observation that a subset of the phosphorylation sites are conserved in all four peptide sequences while others are exclusive to either the brain or skeletal muscle isoforms implies specific functional correlates. However, the significance of the differences between the brain and skeletal muscle isoforms cannot be assessed without first characterizing the functional differences between the brain and skeletal muscle p1 polypeptides.
The tissue-and cell-specific functions of different calcium channels may depend not only on the al-subunit expressed but also on the specific a2/6-, p-, and y-subunits expressed (see Catterall, 1991). In order to dissect the contribution of each of these subunits, it is essential to understand the genetic relationship and pattern of expression of tissue-specific isoforms of each subunit. Our identification of three isoforms of the human P1-subunit encoded by a single gene provides a key tool in the dissection of the role of the pl-subunit in the tissueand cell-specific regulation of voltage-gated calcium channels.