A novel long N-terminal isoform of human L-type Ca2+ channel is up-regulated by protein kinase C.

Human L-type voltage-dependent Ca(2+) channels (alpha(1C), or Ca(v)1.2) are up-regulated by protein kinase C (PKC) in native tissues, but in heterologous systems this modulation is absent. In rat and rabbit, alpha(1C) has two N-terminal (NT) isoforms, long and short, with variable initial segments of 46 and 16 amino acids, respectively. The initial 46 amino acids of the long-NT alpha(1C) are crucial for PKC regulation. However, only a short-NT human alpha(1C) is known. We assumed that a long-NT isoform of human alpha(1C) may exist. By homology screening of human genomic DNA, we identified a stretch (termed exon 1a) highly homologous to rabbit long-NT, separated from the next known exon of alpha(1C) (exon 1b, which encodes the alternative, short-NT) by an approximately 80 kb-long intron. The predicted 46-amino acid protein sequence is highly homologous to rabbit long-NT. Reverse transcriptase PCR showed the presence of exon 1a transcript in human cardiac RNA. Expression of human long-NT alpha(1C) in Xenopus oocytes produced Ca(2+) channel enhanced by a PKC activator, whereas the short-NT alpha(1C) was inhibited. The long-NT isoform may be the Ca(2+) channel enhanced by PKC-activating transmitters in human tissues.

only inhibited by PKC; the enhancement could not be reconstituted (15). The reason for the inability to reproduce the PKC modulation of human L-type channels remained unknown.
The main, pore-forming subunit of cardiac/smooth muscle L-type channel (␣ 1C or Ca v 1.2), also present in the brain, is the product of the ␣ 1C gene, CACNA1C (16). Several splice variants of CACNA1C are known (17,18). The resulting isoforms of human ␣ 1C protein show differential distribution in human tissues, and in failing versus normal myocardium. They play important roles in Ca 2ϩ -dependent inactivation, oxygen sensing, and drug sensitivity (18 -22). However, the genomic structure of the beginning of N-terminal region of human ␣ 1C is not entirely clear. In the two best studied mammalian species, rat and rabbit, two N-terminal isoforms of ␣ 1C cDNA are known, which most probably represent variable splicing products of the same gene (23). These splice variants encode long-and short-NT ␣ 1C proteins, with variable initial segments of 46 and 16 aa, respectively (23)(24)(25)(26) (the total length of the cytosolic part of the NT region of ␣ 1C is ϳ154 aa in the long-NT ␣ 1C ). Traditionally, the short-NT isoform is called "neuronal" in rat and "smooth muscle" in rabbit, whereas the long-NT isoform is considered "cardiac" in rabbit. However, in rat, ␣ 1C protein containing the long-NT is found in both heart and brain (27). The known human ␣ 1C is highly homologous to short-NT isoforms of rat and rabbit; the long-NT isoforms of rat and rabbit are also highly homologous to each other. Because the human L-type channel is up-regulated by PKC, and because the initial 20 aa of the long-NT isoform are crucial for this regulation in rabbit ␣ 1C (27,28), we reasoned that a long-NT isoform of human ␣ 1C should also exist.
Here we demonstrate the presence of an exon encoding the initial 46-aa long-NT segment in the human genomic DNA. The existence of ␣ 1C RNA containing this segment was confirmed by RT-PCR gel analysis and DNA sequencing of the RT-PCR product. Using the RT-PCR products, we have demonstrated that cDNA coding for a full-length long-NT ␣ 1C isoform expressed in Xenopus oocytes produces a Ca 2ϩ channel that is enhanced by a PKC activator. The identification of the long-NT isoform of human ␣ 1C will enable the study of the molecular mechanisms of PKC modulation, which has previously been hampered by the inability to reconstitute this modulation in expression systems.

EXPERIMENTAL PROCEDURES
Tissues and Oocyte Culture-All experiments with human and animal tissues were approved by the Tel Aviv University Helsinki Committee and the Sackler School of Medicine Animal Use Committee, respectively. Rat atria and ventricles were obtained from 19-day-old Wistar rats after decapitation performed under ether anesthesia. Xenopus frogs were maintained and operated on, and oocytes were prepared as described (29). Each oocyte was injected with 2.5 ng of RNAs of ␣ 1C , ␣ 2 /␦, and ␤2A subunits and incubated for 3-4 days at * This work was supported by Israel Basic Research Grant 47/00-16.0. 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.
RT-PCR Analysis-5 g of total human cardiac RNA purchased from Ambion, Inc. (catalog no. 7966, lot 110P43B) was reverse-transcribed with SuperScript II reverse transcriptase (Invitrogen) with primer 3 (see text below and Fig. 2A). Each PCR reaction (50 l) contained 2.4 l of the product of RT reaction, 1 l of 10 mM dNTPs, 20 -50 pmol of primers, 5 l of 10ϫ PCR buffer, 2 l of Mg 2ϩ (2 mM), and 0.7 l of Taq DNA polymerase (Promega). PCR was performed under the following conditions: 95°C for 1 min, 49°C for 1 min, and 72°C for 2 min, repeated 35 times. The final elongation was performed at 72°C for 5 min. The PCR products were analyzed on a 1% agarose gel.
The primers used for RT and PCR were: No. 1, CTTCGAGCCTTT-GTTCAG (nt 4 -21 from A in ATG of exon 1a); No. 2, TCAATGAGAAT-ACGAGGA (nt 5-22 in exon 1b; numbering from ATG of the short-NT isoform ␣ 1C,77 (17) DNA Constructs-cDNAs of ␣ 2 /␦ and ␤2A were as described (30). cDNA of human short-NT isoform ␣ 1C,77 (Ref. 31; GenBank TM accession No. Z34815) was obtained from Dr. R. Zuhlke and subcloned into pGEM-HJ vector (which contains 5Ј-and 3Ј-UTRs of Xenopus ␣-globin flanking the polylinker). In the resulting construct, termed ␣ 1C,77S , the 5Ј-UTR of ␣-globin is followed by a BamHI restriction site and then immediately by the initial ATG. The coding sequence of ␣ 1C,77 is followed by the 3Ј-UTR of ␣-globin and then by a SalI site. The DNA of the long-NT, termed ␣ 1C,77L , was constructed as follows. The human heart cDNA obtained in the RT reaction described above was subjected to PCR with the reverse primer No. 3 (see above) and a forward primer ATTCGCGGATCCATGCTTCGAGCCTTTGTT that overlaps the first 18 nt of the coding part of exon 1a (starting with ATG) and also creates a BamHI restriction site preceding ATG. The PCR product was digested with BamHI and MfeI (a unique MfeI site is present in exon 3 upstream of primer No. 3) and inserted between these restriction sites into BamHI/MfeI-digested ␣ 1C,77S in place of the original DNA segment flanked by these sites. The 5Ј portion of the resulting DNA was sequenced. The sequence of the BamHI-MfeI segment obtained by this RT-PCR subcloning procedure was 100% identical to that predicted for the long-NT splice variant (based on the DNA sequence of chromosome 12), in which exon 1a is followed by exons 2 and 3 (see Fig. 1C). RNA was synthesized as described (29).
Electrophysiology-Whole cell currents were recorded using the Gene Clamp 500 amplifier (Axon Instruments, Foster City, CA) using the two-electrode voltage clamp technique, as described (30), in a solution containing 40 mM Ba(OH) 2 , 50 mM NaOH, 2 mM KOH, and 5 mM HEPES, titrated to pH 7.5 with methanesulfonic acid. PMA and bisindolylmaleimide (BIS) were purchased from Sigma. Oocytes were treated with BIS essentially as described (32). In brief, the oocyte was injected with 30 nl of a water solution of BIS at 150 M and, in addition, incubated in 5 M BIS 2-4 h before recording. Stimulation, data acquisition, and analysis were performed using pCLAMP software (Axon Instruments).

RESULTS AND DISCUSSION
The initial exon of the previously described (17) human ␣ 1C clone, which we now designate exon 1b, contains a 5Ј-UTR and encodes the first 16 aa of the short-NT ␣ 1C protein. To find out whether DNA sequences that encode an alternative, long-NT isoform of human ␣ 1C exist, we performed a standard BLAST search of human genomic DNA, 2 using as our query the cDNA sequence encoding the first 46 aa of the rabbit cardiac long-NT clone (24). The initial search showed the presence of a highly homologous sequence in the contig 3810573 3 from human chromosome 12, locus p13.3 ( Fig. 1A; the region of the initial search is highlighted by bold letters). This is within the region where the gene of ␣ 1C , CACNA1C, is located. Significantly, the sequence upstream from this segment showed significant homology to the 5Ј-UTR of the rabbit ␣ 1C (Fig. 1A), supporting the possibility that it may be part of an initial exon of a human L-type channel. We have designated this tentative exon as 1a.
Within the presumably protein-coding part of exon 1a, a 45base-long DNA segment, starting at base 16 (from the initial ATG) also shares 40% homology with the protein-coding part of exon 1b (Fig. 1A).
The availability of the draft sequence of the human chromosome 12 on the NCBI Human Genome site allowed us to map the location of the putative exon 1a relative to the known exons of the human ␣ 1C gene (Fig. 1B). Exon 1a precedes exon 1b and is separated from the latter by ϳ80 kb. We assume that this is a large intron. Two other large introns are found, according to the draft map of the chromosome, between exon 1b and exon 2 (ϳ60 kb) and between exons 3 and 4 (ϳ330 kb). These large introns have not been fully sequenced previously; introns of Ͼ5 and 2.4 kb have been reported (17). The location of exon 1a supports the possibility that it constitutes the first alternative exon of the CACNA1C gene. The screening procedure also confirmed the presence and the correct spatial location on chromosome 12 of all the of constant and alternative exons described by Soldatov in 1994 (17) (Fig. 1B show 10 of the 50 exons, excluding exon 1a).
Our working hypothesis was that exons 1a and 1b are alternative initial exons of the human CACNA1C gene (Fig. 1C). Exons 2 and 3 are probably constant (17), as supported by the conservation of the corresponding protein sequences in known long-and short-NT ␣ 1C isoforms in human, rat, rabbit, and mouse ( Fig. 1D and Refs. 23, 27, and 28). The protein sequences of N termini of the proposed human ␣ 1C short-and long-NT isoforms are shown in Fig. 1D and are compared with the rabbit long-NT ␣ 1C . The segment corresponding to exon 2 starts after aa 46 in the long-NT isoform and after aa 16 in the short-NT isoform. The protein segment encoded by exon 3 starts at aa 155, exactly at the proposed junction between the cytosolic NT and the first transmembrane segment of the channel, IS1 (24).
To further substantiate our working hypothesis, we performed RT-PCR on total human cardiac RNA. cDNA was obtained by reverse transcription using primer No. 3 ( Fig. 2A), which corresponds to the 3Ј end of exon 3. The analytical PCR was done with primers corresponding to segments of chromosome 12 DNA, as published at the NCBI Human Genome site, within the putative exons and introns of the 5Ј region of CACNA1C. The location of primers is shown in Fig. 2A. Clear bands were detected for all DNAs that contained exon 1a (  6 and 7) and ϳ1.4 -2.4 kb upstream from the beginning of exon 1a (lane 8) did not yield signals. Another cDNA segment that included protein-coding sequences of exon 1a, 2, and (part of) 3 was obtained for subcloning purposes (see "Experimental Procedures") and sequenced. The size and the DNA sequence of this RT-PCR product were exactly as expected for [exon 1a ϩ exon 2 ϩ exon 3 up to MfeI site] without exon 1b and encoded the first 46 aa of the proposed long-NT isoform of ␣ 1C followed by protein sequences highly conserved in all known isoforms of ␣ 1C . These results strongly support the presence in human cardiac tissue of an RNA species encoding the proposed long-NT isoform shown in Fig. 1, C and D. A band corresponding to the DNA of exons [1b ϩ 2 ϩ 3] was also obtained; the size was exactly as expected for a short-NT isoform, without 1a (Fig. 2B, lane 2). Under standard condi-tions the band corresponding to [exon 1b ϩexon 2] was not detected (Fig. 2B, lane 4), but varying the conditions of PCR revealed this band, although it still remained relatively weak (Fig. 2C). Thus, RNA of the previously described short-NT isoform is also present in human cardiac tissue.
Previously, the existence of a long-NT ␣ 1C protein in rat has been demonstrated by Western blot with a polyclonal antibody directed against the unique initial 46 aa of the rabbit long-NT isoform (27). The same antibody detected a Ͼ220 kDa protein (presumably the L-type Ca 2ϩ channel) in a human colon cancer cell line (33). These results corroborate the presence of the long-NT ␣ 1C protein in human tissues. Taken together, our data and those of Ref. 33 support the notion that human cardiac (and probably other) tissues contain two N-terminal isoforms of ␣ 1C protein, a long-NT and a short-NT one, which are products of alternatively spliced RNA transcripts of the 5Јterminal region of the CACNA1C gene (as shown in Fig. 1, C  and D).
According to our initial hypothesis, the long-NT ␣ 1C should be up-regulated by the activation of PKC. To test this prediction, we constructed a cDNA encoding a long-NT ␣ 1C on the basis of a human short-NT ␣ 1C DNA, ␣ 1C,77 (19). Both cDNAs were subcloned into the pGEM-HJ vector and verified by DNA sequencing (see "Experimental Procedures"). The long-NT clone was designated ␣ 1C,77L and the short-NT clone ␣ 1C,77S . The corresponding RNAs were synthesized in vitro and injected into Xenopus oocytes with RNA of the ␣ 2 /␦ subunit with or without RNA of the ␤2A subunit (30). The voltage-dependent Ba 2ϩ currents (I Ba ) via the expressed ␣ 1C,77L and ␣ 1C,77S Ca 2ϩ FIG. 1. The newly identified exon 1a and the proposed DNA and protein structure of two NT isoforms of human ␣ 1C . A, nucleotide homology between rabbit heart (RH) long-NT cDNA, the homologous part of human chromosome 12 corresponding to the proposed exon 1a (ex1a), and the coding sequence of the known exon 1b (ex1b). Full homology is indicated by asterisks. Bold letters show the protein-coding sequence of long-NT isoforms. B, the location of exons and introns according to the draft map of chromosome 12 available at the NCBI site. C, the DNA sequences corresponding to proposed alternative splice variants encoding long-and short-NT isoforms of ␣ 1C . The middle section shows the 5Ј part of the CACNA1C gene, with boxes representing exons and angled lines representing introns. D, comparison of the protein sequences of rabbit cardiac ␣ 1C and the two proposed isoforms of the human ␣ 1C . Asterisks show fully conserved aa; bold letters highlight the PKC phosphorylation sites responsible for the inhibitory effect of PMA in rabbit ␣ 1C (35); the underlined residues are those that differ in rabbit and human long-NT ␣ 1C . channels ranged from 100 to 800 nA without ␤2A and from 1500 to 7000 nA with ␤2A. The kinetics of I Ba of ␣ 1C,77L and ␣ 1C,77S (e.g. Fig. 3C) were very similar to each other and to those directed by rabbit cardiac ␣ 1C . The other parameters also resembled those previously reported for L-type Ca 2ϩ channels of various species. For instance, as described previously for the rabbit ␣ 1C (30,34), coexpression of the ␤2A subunit increased the current amplitude 9.5 Ϯ 0.4-fold in ␣ 1C,77L and 9.3 Ϯ 0.6-fold in ␣ 1C,77S , and shifted the voltage dependence of activation to more negative potentials, as shown by normalized current-voltage curves (Fig. 3A). Peak I Ba was similar in the long-NT and short-NT isoforms; the small difference was statistically significant (Fig. 3B), but at present we do not know whether this reflects differences in protein expression or in the gating properties of the two isoforms.
The prediction that the long-NT human ␣ 1C should be enhanced by PKC has been fully confirmed. The main effect of the phorbol ester ␤-PMA (a PKC activator) was to increase the Ba 2ϩ current of ␣ 1C,77L coexpressed with ␣ 2 /␦ with or without ␤2A (Fig. 3, C-E). As in the rabbit ␣ 1C channel (13), the increase reached a maximum within 5-8 min and was followed by a decrease, sometimes below the basal current level (Fig.  3D). In contrast, the short-NT ␣ 1C,77S channels were inhibited by PMA (Fig. 3D), in line with the previous report (15). On average, PMA increased the current via the long-NT channels by 22.7 Ϯ 4.4% in the absence of ␤2A subunit and by 21.2 Ϯ 4.2% in its presence (Fig. 3E). The PMA-induced increase in I Ba was fully blocked by the specific PKC inhibitor, bisindolylmaleimide ( Fig. 3E; bar marked BIS), supporting the notion that the enhancement of I Ba by PMA was indeed mediated by PKC. Thus, in general, the modulation of the human long-NT isoform by PKC is similar to that of the rabbit cardiac channel. With rabbit cardiac L-type channel, coexpression of ␤2A substantially weakens the enhancement of the current by PMA (13), whereas in the human channel this action of ␤2A is less pronounced. It would be of interest to see whether this is due to one of the few differences in the primary amino acid composition of the 46-aa-long initial NT segment in human and rabbit (Fig. 1D, underlined amino acids).
In a different expression system (human embryonic kidney cell line tsA-201), only inhibition of the rabbit long-NT ␣ 1C by PMA has been observed (35). The inhibition crucially depended on the presence of each of the two threonines, Thr 27 and Thr 31 (shown in bold in Fig. 1D), in which phosphorylation by PKC was proposed to underlie this modulation. It has been proposed that the lack of PKC-induced enhancement in tsA-201 cells may result from the absence of a specific PKC isozyme (35). At present, we do not know whether the phosphorylation of Thr 31 (Thr 27 is absent in the long-NT human ␣ 1C ; Fig. 1D) plays a role in any of the PKC effects observed in Xenopus oocytes. We suspect that the mechanism of PMA-induced inhibition of ␣ 1C in oocytes is different from that observed in tsA-201 cells, because a short-NT isoform of rat ␣ 1C was not sensitive to PMA in this system (35), whereas the homologous short-NT human ␣ 1C is inhibited by PKC in Xenopus oocytes (Ref. 15 and Fig. 3). Furthermore, the decrease of I Ba is still observed in the presence of BIS (Fig. 3E), whereas in tsA-201 cells PKC inhibitors blocked the effect of PMA (35). The PMA-induced inhibition of human L-type Ca 2ϩ channels observed in Xenopus oocytes requires further study.
Because the most widely observed effect of PKC activators on human L-type Ca 2ϩ channels is an enhancement of the current (sometimes accompanied by an inhibition (5-10)), and because this regulation is reproduced with the long-NT ␣ 1C in Xenopus oocytes, we propose that the long-NT ␣ 1C is the isoform that underlies the PKC-induced enhancement. The homologous long-NT rabbit ␣ 1C isoform behaves in the same way when expressed in oocytes (13,28), probably by means of an identical molecular mechanism as in human ␣ 1C . The four aa following the initial methionine in rabbit long-NT ␣ 1C , LRAL, are necessary for PKC-induced enhancement. Because this protein segment contains no putative phosphorylation sites, we have proposed that PKC phosphorylation occurs elsewhere in ␣ 1C or in an auxiliary protein, whereas the first five aa play a role in PKC anchoring or in channel gating (27). The short-NT isoform lacks these four amino acids; its initial 15-aa segment (following the first methionine) carries a partial homology to amino acids 6 -20 of the long-NT isoform (Fig. 1D). In rabbit long-NT ␣ 1C , the first 20 aa play the role of the inhibitory gating element, which reduces the open probability of the L-type channel 4 (27,28). It would be of great interest to see whether this is also the case in human long-NT ␣ 1C . We hope that the discovery of the long-NT isoform of human ␣ 1C and the demonstration of its modulation by PKC, as described in this report, will provide the basis and the incentive for future studies of PKC targets (␣ 1C or an auxiliary protein?) and of the interaction between the phosphorylated and gating parts of ␣ 1C in PKC modulation.