Introduction

Top Abstract Introduction Genomic Organization Promoters and Expression... Ca2+ Sensors of the... Modulation of Sensitivity to... Truncated Forms and Mutants... References

The L-type Ca²⁺ channel is a ubiquitously expressed voltage-gated ion channel supporting inward current of Ca²⁺ ions that play a central role as an intracellular second messenger in many processes ranging from gene expression to cardiac and smooth muscle contraction. A number of L-type Ca²⁺ channel blockers have been developed for the treatment of cardiovascular disorders. 1,4-Dihydropyridines, which have proven to be the most potent inhibitors of the channel, were used for its biochemical identification in rabbit skeletal muscle T-tubules as a protein complex composed of the pore-forming _1S subunit and auxiliary ₂, , and  subunits. Two dihydropyridine-sensitive genetic variants of _1S have been identified and cloned from heart (_1C) and pancreatic beta cells (_1D), with the cardiac _1C isoform being expressed in the vast majority of eukaryotic cells. Activity of the _1C channel is highly regulated by membrane potential, other calcium channel auxiliary subunits (₂, ), as well as by feedback dependence on permeating Ca²⁺, which is mediated by calmodulin. The genetic regulation of the _1C Ca²⁺ channel subunit occurs through species-specific variability and largely tissue-specific alternative splicing. In this review, we briefly discuss genomic organization of the _1C subunit gene and then focus attention on alternative splicing that generates functionally distinct Ca²⁺ channel isoforms as potentially new pharmacological targets.

	Genomic Organization

Top Abstract Introduction Genomic Organization Promoters and Expression... Ca2+ Sensors of the... Modulation of Sensitivity to... Truncated Forms and Mutants... References

The _1C L-type Ca²⁺ channel is subject to complex genetic regulation that gives rise to a variety of channel subtypes. Both genomic variability and alternative splicing of the primary transcript appear to contribute to this complexity. The human _1C subunit gene (CACNL1A1) is composed of 53 identified exons (Fig. 1) (Soldatov, 1994) and is located in the distal region of chromosome 12p13 (Sun et al., 1992; Powers et al., 1992). There are two other ₁ subunits (_1D and _1S) of the L-type Ca²⁺ channel subfamily that are encoded by different genes located in different chromosomes.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Hypothetical transmembrane topology of the polypeptide chain of the channel-forming 1C subunit of the human L-type Ca2+ channel. Shown are four repeats (I, II, III, IV) each consisting of six putative transmembrane segments. Both NH2 and COOH termini are in the cytoplasm. Transmembrane segments S4 contain from 3 to 5 positively charged amino acid residues and presumably form the voltage sensor of the channel. Bold bars mark segments encoded by numbered exons. Arrows point to transmembrane segments encoded by alternative exons 1/1A, 8/8A, 21/22, and 31/32. Open lines represent sequences encoded by exons 7, 33, and 45, which are subject to constitutive splicing (Soldatov, 1992, 1994). Boxed are the Ca2+-sensing domains L (1572-1598) and K (1599-1651) coded by alternative exons 40 to 42 that are subject to alternative splicing leading to 1C,86 (Soldatov et al., 1997). Motifs involved in Ca2+-induced inactivation are shown in bold letters: M1 (1572-1576), M8 (1580-1587), M3 (1600-1604), and M4 (1630-1634). The combined M1 + M3 mutant was found to be deprived of Ca2+-induced inactivation similar to 1C,77L, 1C,77K (Soldatov et al., 1998), and 1C,86 (Soldatov et al., 1997). Amino acids that did not contribute to Ca2+-induced inactivation are shown in italics. The CaM-binding IQ motif (1624-1635) (Zühlke and Reuter, 1998) is underlined. The full-length domain L (1572-1599) and the sequence encompassing M3 and M8 (1582-1604) both show high-affinity CaM-binding in low (50 nM) and high (2 µM) [Ca2+]free. The high-affinity Ca2+ sensor has been identified as sequence 1572-1587 (Romanin et al., 2000). A larger panel of sequences from similar regions of other channel types has been investigated by Pate et al. (2000).

The exact size of the _1C subunit gene remains to be determined. Based on the assumption that 10 kb is a rough size approximation of unknown introns, the gene was estimated to span more than 150 kb of the human genome (Soldatov, 1994). However, the high-resolution visual mapping of stretched DNA by fluorescent in situ hybridization (termed fiber-FISH) has shown that the distance from exon 10 to the 3'-end is in fact 170.4 kb (Liu et al., 1998). Thus the total length of the human CACNL1A1 gene is at least 70 kb greater than earlier estimates. In addition, we have found a repetitive element of three paired exon 45/46-related sequences in the _1C subunit gene (Soldatov et al., 1998b). Two of them, clones g6-20 and g12-5, were found by the fiber-FISH technique to be located within 59.1 kb downstream of the polyadenylation site, i.e., 230 kb away from exon 10. Thus, the total size of the human CACNL1A1 gene may be 300 kb.

The idea that the _1C gene contains several repetitive 3'-terminal sequences is indirectly supported by the fact that cDNA coding for human and nonhuman Ca²⁺ channel _1C subunits significantly diverge downstream from exon 44. This divergence may originate from variability in the 3'-terminal part of the gene. In the absence of data indicating that the 3'-end of CACNL1A1 may be subject to alternative splicing, we hypothesize that sequences representing g12-5 and g6-20 could become silent during evolution of this very complex gene.

A new exon 45/46-related sequence g8-19 has been identified in the human genome and mapped by FISH to the 12p11.2 and 12p13.2-p13.1 bands (Soldatov et al., 1998b). These positions were not recognized by DNA probes generated from the 5'- and 3'-terminal regions of the _1C gene. It is possible that hybridization of g8-19 to two loci occurs because 1) this DNA may in fact be a chimeric clone that contains sequences from 12p11.2 and 12p13.2-p13.1 that are not normally contiguous or 2) g8-19 may contain repetitive elements other than the exon 45/46-related sequence, which are recognized by both loci. However, given the unique similarity of its exon 45/46-related sequences to the _1C gene, an alternative hypothesis is that g8-19 belongs to a new gene or pseudogene of the same ₁ family.

	Promoters and Expression Regulation

Top Abstract Introduction Genomic Organization Promoters and Expression... Ca2+ Sensors of the... Modulation of Sensitivity to... Truncated Forms and Mutants... References

Human _1C transcripts are highly homologous to those from mice, rat, and rabbit except exons 17 and 44-50. Exon 1 comprising an initiation codon and the 5'-untranslated region also appears to be subject to alternative splicing. Indeed, the _1C transcripts identified in rabbit heart (Mikami et al., 1989) and rat aorta (Koch et al., 1990) have exon 1 (subsequently referred to as exon 1A) different from _1C transcripts in rabbit lung (Biel et al., 1990), rat heart (Snutch et al., 1991), human fibroblasts (Soldatov, 1992), and hippocampus (N. M. Soldatov, unpublished observation) thus suggesting that both exon 1 isoforms of _1C are not species- but rather tissue-specific. Unlike these two isoforms, the human heart _1C cloned by Schultz et al. (1993) has exon 1 and an upstream part of exon 2 deleted. Since the putative splice acceptor site that would be employed in this case does not conform to the consensus sequence, the proposed shortened N terminus of this _1C isoform must be validated by genomic DNA sequences. In fact, exon 1A with the respective 5'-untranslated region appears to be present in human chromosome 12 at the genomic DNA region upstream of the identified exon 1 (N. Dascal, personal information). This exon contributes to the _1C transcripts in human cardiac tissue.

The promoter region for the exon 1A isoform of rat _1C (Mikami et al., 1989; Koch et al., 1990) lacks a canonical TATA sequence but has a consensus Inr element corresponding to the major transcription start site (Liu et al., 2000). Seven possible additional transcription initiation sites were identified within a 100-nt distance around Inr, as well as a number of potential regulatory elements in a 2-kb sequence upstream of the major 5'-cap site. Those elements included five consensus sequences for transcription factor Nkx2.5, which is specific for developing heart and exhibits modest transcriptional activation. Other elements included the potentially important hormone-responsive elements associated with response to steroid hormones, the cAMP-responsive element, and AP-1 that is regulated by mitogen-activated protein kinase signal transduction pathways. Other elements identified are two muscle determination factors MyoD and MEF2 and several other putative elements including C/EBPb, Oct-1, GRE, NF-E3/C-Ets, and STATX.

The effect of some of these factors on _1C transcripts and protein levels has been experimentally documented. As determined by dihydropyridine (DHP) binding assay, serum deprivation increased the amount of _1C channels in human fibroblasts that returned to baseline when serum was reintroduced (Dudkin et al., 1988). The individual mitogens including epidermal growth factor, basic fibroblast growth factor, and insulin reduced the amount of DHP receptors in cells (Soldatov et al., 1988). In cardiac myocytes, incubation in high Ca²⁺ increased both transcription of _1C channels and DHP binding (Davidoff et al., 1997). In contrast, exposure to phenylephrine, an ₁ adrenergic agonist with signal transduced as cytosolic Ca²⁺ elevation via receptor-operated calcium channels, decreased _1C transcripts and L-type calcium currents (Maki et al., 1996). Isoproterenol and 8-bromo-cAMP increased Ca²⁺ channel mRNA and peak calcium current (Maki et al., 1996). These findings all support a role for the _1C Ca²⁺ channel in mitogenic responses and cellular proliferation. Whether the opposite effects of norepinephrine-mediated increase of [Ca²⁺]_i via receptor-operated Ca²⁺ channels compared with direct increase of [Ca²⁺]_i in the medium are due to specific involvement of a different set of regulatory elements remains to be studied.

Fifteen exons of the human _1C gene have been established to be subject to alternative splicing. The regulation of Ca²⁺ channel expression through alternative splicing (Perez-Reyes et al., 1990) is particularly interesting in terms of regional structural diversity. Alternative splicing affects regions encoding transmembrane segments IIIS2, IVS3, as well as the intracellular C-terminal tail and linkers between repeats I, II, and III. Systematic study of the expression of _1C splice variants has not yet been undertaken but may soon become possible with the development of new proteomic methods such as matrix-assisted laser desorption/ionization mass spectrometry. Limited data available at this time indicate that alternative splicing of _1C may occur in a species- and tissue-specific fashion and can generate channel isoforms with altered functional properties. One of these splicing events at transmembrane IVS3 is related to developmental regulation of Ca²⁺ channel isoform expression (Diebold et al., 1992). Our preliminary data (Soldatov et al., 2001) show that switch of the exon 22 to 21 isoform of _1C occurs in response to suppression of proliferative stimuli in human aortic smooth muscle in an age-specific manner with cells from donors older than about 50 years expressing exon 21 and cells from younger donors not expressing exon 21. However, only few alternative isoforms of _1C show properties of functionally distinct Ca²⁺ channel subtypes.

	Ca2+ Sensors of the Ca2+ Channel

Top Abstract Introduction Genomic Organization Promoters and Expression... Ca2+ Sensors of the... Modulation of Sensitivity to... Truncated Forms and Mutants... References

Useful information to help understand the molecular bases of Ca²⁺-dependent regulation of the L-type Ca²⁺ channel activity has been obtained from studies of alternative exons 40-43 that were identified in a partial _1C transcript of human hippocampus (Soldatov, 1994). Reconstruction of the alternative exons into the _1C-coding sequence of the conventional (_1C,77) channel has resulted in a Ca²⁺-insensitive isoform (_1C,86) of the channel (Soldatov et al., 1997) with 80 amino acids in the second quarter of the cytoplasmic tail of the channel replaced by 81 nonidentical amino acids (Fig. 1). _1C,86 conducted Ba²⁺ and Ca²⁺ currents with almost identical fast kinetics of inactivation and did not show a U-shape dependence of the time course of inactivation on membrane voltage, which is the characteristic feature of Ca²⁺-induced inactivation. By replacement of large segments of this 81-amino acid domain of _1C,86 into the _1C,77 channel, it was found (Soldatov et al., 1998a) that Ca²⁺-induced inactivation is independently determined by two sequences that we have called domains L and K (Fig. 1). Ca²⁺-induced inactivation was found to partially depend on several shorter sequences identified in these domains (shown in Fig. 1 in bold letters). Only their combined mutation within or between domains L and K removed the Ca²⁺ sensitivity of the channel inactivation. Some of the identified important partial motifs have been later confirmed by deletion analysis (Zühlke and Reuter, 1998).

Both domains L (Pate et al., 2000; Romanin et al., 2000) and K (Peterson et al., 1999; Qin et al., 1999; Zühlke et al., 1999) were found to bind calmodulin (CaM) involved in Ca²⁺-induced inactivation. In addition, domain L contains a highly specific Ca²⁺ sensor (K_d  100 nM) composed of motifs M1 and M8, both of which are important for Ca²⁺-dependent inactivation of the channel. Ca²⁺ loading of this Ca²⁺ sensor was shown to modulate the CaM affinity of CaM-binding site in domain L. The IQ motif in domain K (Fig. 1) appears to have a role in the binding of Ca²⁺-loaded CaM. Translocation or sliding of CaM with possible involvement of motif M3 or independent binding of CaM to the two sites appears to regulate different stages of the channel activity.

It is known that Ca²⁺-saturated CaM can adopt different conformations upon interaction with CaM-binding domains, from an extended dumbbell shape to a globular shape wrapped around the target peptide (Elshorst et al., 1999). There are examples when either one or both Ca²⁺-binding halves of the CaM molecule are needed for activation, or their binding occurs with no effect. Whether CaM is able to form a triple complex with two discrete determinants in L and K (Mouton et al., 2001), or they represent distinct or alternative binding sites (Pate et al., 2000) remains to be verified by direct structural studies. The constitutive nature of CaM binding, however, explains why CaM inhibitors did not affect Ca²⁺-induced inactivation in earlier studies (Zühlke and Reuter, 1998).

The human _1C channel did not show current facilitation by strong positive voltage prepulses (N. M. Soldatov and H. Reuter, unpublished observation). Interestingly, the I1624A mutation in the IQ motif of _1C,77 revealed CaM-dependent facilitation of Ca²⁺ but not Ba²⁺ current in response to an applied train of high-frequency depolarizations (Zühlke et al., 1999). This property was unmasked in cardiac myocytes by intracellular application of a domain L-derived peptide (1579-1604) (Pate et al., 2000). Collectively, these data suggest a new principle of modulatory control over Ca²⁺ signaling mediated by the _1C channel particularly in cardiac and vascular cells.

The cross-talk between two CaM-binding sites is transduced into Ca²⁺-driven inactivation by a yet unknown mechanism. Replacements in the cytoplasmic 80-amino acid segment leading to _1C,86, _1C,77L, and _1C,77K variants affected voltage sensors for activation and inactivation (Soldatov et al., 1997, 1998a), and reduced open probability and single channel conductance (Kepplinger et al., 2000b) thus suggesting that both Ca²⁺-sensing domains interact with a pore region and affect the voltage-gating mechanism of the channel. Combined disruption of motifs L and K in _1C,86 completely eliminated the characteristic run-down property of the channel (Kepplinger et al., 2000a). This region may involve determinants for the rescue effect of calpastatin (Romanin et al., 1991).

Increasing evidence demonstrates clustering of recombinant channels in the plasma membrane of expressing cells. By the photobleaching of _1C,77 N-terminally labeled by enhanced yellow fluorescent protein, it was found that a mean cluster contains approximately 40 channels (Harms et al., 2001). Application of label-fracture and cryothin-sectioning techniques with high-resolution immunogold-labeling to guinea pig ventricular myocytes allowed direct demonstration of _1C channels clustering, which occurred predominantly in plasma membrane domains overlying junctional sarcoplasmic reticulum (Gathercole et al., 2000). Therefore, membrane targeting and subsequent clustering appear to be important features of _1C for its function as a trigger of intracellular Ca²⁺ release involved in excitation-contraction coupling in cardiac muscle. We found that determinants for both features are located in the same 80-amino acid segment of the tail (1572-1651). Indeed the _1C,86 channel labeled N-terminally by green fluorescent protein was predominantly distributed within the cytoplasm. Only a minor fraction of _1C,86 was able to incorporate into the plasma membrane, and this may correspond to low current density that is generally observed upon _1C,86 expression (Kepplinger et al., 2000b). The membrane-bound fraction of the labeled _1C,86 did not form characteristic clusters.

	Modulation of Sensitivity to DHP

Top Abstract Introduction Genomic Organization Promoters and Expression... Ca2+ Sensors of the... Modulation of Sensitivity to... Truncated Forms and Mutants... References

Free Ca²⁺ is known to be important for the high-affinity interaction of Ca²⁺ channels with DHP calcium channel blockers. However, disruption of Ca²⁺ sensors increased DHP sensitivity of _1C,86 3.5-fold compared with _1C,77 over a wide range of potentials (Soldatov et al., 1997; Zühlke et al., 1998). Thus, the modulation of DHP sensitivity observed in _1C,86 occurs in a Ca²⁺-independent manner.

An opposite but also voltage-independent modulation of the sensitivity to isradipine was found to be caused by a replacement of exon 8 for 8A (Fig. 1) leading to the _1C,105 isoform with a modified transmembrane segment IS6 (Zühlke et al., 1998). Unlike the _1C,86, _1C,72, and _1C,105 channels, _1C,70, produced by substitution of exon 22 for 21, showed an altered voltage-dependent inhibition of Ba²⁺ current by isradipine at very negative potentials (Soldatov et al., 1995). The slope of the IC₅₀ curve at 90 mV was significantly less steep in _1C,70 than in _1C,77, causing a 2.5-fold difference in the inhibitory potency of the drug between the two channels. These results suggest that the external portion of the putative transmembrane segment IIIS2, encoded by exons 21 and 22, experiences voltage-dependent conformational changes that alter DHP binding. The voltage dependence of isradipine action is more pronounced in the exon 21 than in the exon 22 isoform of the channel.

The identified sites of modulation for isradipine inhibition are located in different regions of _1C but outside of the high-affinity binding site for DHPs. These altered pharmacological properties of the _1C channel isoforms imply that alternative splicing may contribute to the tissue specificity and to age-related changes in the clinical effects of DHP calcium antagonists, at least in vascular smooth muscle (Abernethy and Schwartz, 1999).

	Truncated Forms and Mutants of 1C

Top Abstract Introduction Genomic Organization Promoters and Expression... Ca2+ Sensors of the... Modulation of Sensitivity to... Truncated Forms and Mutants... References

Screening of brain and fibroblast transcripts has revealed a dominant truncated form of the human _1C calcium channel subunit, which originates from the utilization of an alternative splice acceptor site at the 5'-end of exon 15, conforming better for splice acceptor requirements than the functional one (Soldatov, 1994). This leads to a 73-base pair deletion and interruption of the reading frame in the region of transmembrane segment IIS6 in as many as 75% of _1C transcripts. The role of this and the other much less abundant truncated form produced by a 12-nt insertion at the 3'-end of exon 16 remains to be clarified. Recently, two proteins produced by truncations of the _1C gene in the region of exons 17-19 have been identified in rabbit sarcolemma and sarcoplasmic reticulum (Wielowieyski et al., 2001). These forms may have a role in excitation-contraction coupling or in sequestering auxiliary  subunits by the _1C- interaction site that is retained in the repeat I-II linker.

Structure-functional studies of naturally occurring mutations that affect Ca²⁺ channel properties are of particular interest as they may allow identification of new therapeutic targets. One such mutation was originally identified as a single nucleotide conversion g²²⁵⁴  a in two independent human fibroblast _1C subunit transcripts (Soldatov, 1992). This produced substitution by Thr of the invariant Ala752 residue located at the cytoplasmic end of the highly conserved transmembrane segment IIS6, which significantly impaired voltage-gated inactivation (Soldatov et al., 2000). Such a "leaky" mutant may cause Ca²⁺ overload of the cell and cytotoxicity; however, this remains to be proven.

The remarkable structure-functional diversity of the _1C calcium channel requires further systematic investigation. An interesting approach for exploration of functional links in calcium channel regulation was gained with coexpression of Ca²⁺ channel isoforms with receptors that have in vivo functional interaction with the Ca²⁺ channel. For example, coexpression of the _1C,77 channel with the angiotensin type IA receptor in Xenopus oocytes allowed study of regulation of the L-type Ca²⁺ channel by angiotensin. This regulation was mediated via IP₃-induced intracellular Ca²⁺ release and occurred at the molecular motif responsible for the Ca²⁺-induced inactivation of the channel (Oz et al., 1998). Use of diverse functional isoforms of Ca²⁺ channel as biosensors and the measurement of voltage-gated Ca²⁺ and Ba²⁺ currents in such systems offers new opportunities to investigate pharmacological properties of coexpressed receptors and to study the mechanism of in vivo drug effects. In this particular example, these findings were helpful in understanding a vascular interaction between angiotensin II and calcium antagonist drugs seen in clinical study (Andrawis et al., 1992).

Recent discoveries of the molecular bases of Ca²⁺ channel inactivation mechanisms, particularly of its Ca²⁺ dependence, and their evolving role in excitation-contraction coupling in cardiac and vascular cells point to a necessity of detailed structural investigation of the involved regions, particularly bearing Ca²⁺ sensors, using diffraction and NMR methods. Careful investigation of intramolecular protein-protein interactions critical for activation and termination of Ca²⁺ current will obviously help develop new therapeutic targets based on new principles.