Diversity and Developmental Expression of L-type Calcium Channel β2 Proteins and Their Influence on Calcium Current in Murine Heart*

By now, little is known on L-type calcium channel (LTCC) subunits expressed in mouse heart. We show that CaVβ2 proteins are the major CaVβ components of the LTCC in embryonic and adult mouse heart, but that in embryonic heart CaVβ3 proteins are also detectable. At least two CaVβ2 variants of ∼68 and ∼72 kDa are expressed. To identify the underlying CaVβ2 variants, cDNA libraries were constructed from poly(A)+ RNA isolated from hearts of 7-day-old and adult mice. Screening identified 60 independent CaVβ2 cDNA clones coding for four types of CaVβ2 proteins only differing in their 5′ sequences. CaVβ2-N1, -N4, and -N5 but not -N3 were identified in isolated cardiomyocytes by RT-PCR and were sufficient to reconstitute the CaVβ2 protein pattern in vitro. Significant L-type Ca2+ currents (ICa) were recorded in HEK293 cells after co-expression of CaV1.2 and CaVβ2. Current kinetics were determined by the type of CaVβ2 protein, with the ∼72-kDa CaVβ2a-N1 shifting the activation of ICa significantly to depolarizing potentials compared with the other CaVβ2 variants. Inactivation of ICa was accelerated by CaVβ2a-N1 and -N4, which also lead to slower activation compared with CaVβ2a-N3 and -N5. In summary, this study reveals the molecular LTCC composition in mouse heart and indicates that expression of various CaVβ2 proteins may be used to adapt the properties of LTCCs to changing myocardial requirements during development and that CaVβ2a-N1-induced changes of ICa kinetics might be essential in embryonic heart.

Cardiac contractions require Ca 2ϩ influx in cardiomyocytes from the extracellular fluid, which leads to Ca 2ϩ release from the sarcoplasmic reticulum via ryanodine receptors (1).
This Ca 2ϩ -induced Ca 2ϩ release (CICR) 4 causes a marked increase in intracellular Ca 2ϩ concentration for short periods of time and underlies cardiac contraction (2,3). The Ca 2ϩ influx into cardiac myocytes is mediated by high voltage-activated L-type Ca 2ϩ channels (LTCCs), which are heteromultimeric complexes comprised predominantly of the pore-forming CaV␣ 1 subunit and the auxiliary CaV␤ subunit (4). In heart, the principal CaV␣ 1 subunit, CaV␣ 1 c (CaV1.2), is encoded by the Cacna1C gene (5). Four genes (Cacnb1-4) encoding CaV␤ subunits have been identified that are expressed in the heart of different species including human, rabbit, and rat (6,7,8).
CaV␤ proteins are ϳ500 amino acid cytoplasmic proteins that bind to the CaV␣ 1 I-II intracellular loop (9) and affect channel gating properties (4), trafficking (10,11), regulation by neurotransmitter receptors through G-protein ␤␥ subunit activation (12), and sensitivity to drugs (13). The CaV␤ primary sequence encodes five domains, arranged V1-C1-V2-C2-V3. V1, V2, and V3 are variable domains, whereas C1 and C2 are conserved (14). Structural studies reveal that C1 and C2 form a SH3 domain (Src homology 3 domain) and a NK domain (nucleotide kinase domain), respectively (15). Although C1-V2-C2 makes the CaV␤ core, in heart the V1 region appears critical for the kinetics of I Ca and heart function. Accordingly a mutation in the V1 region of the Cacnb2 gene was recently identified as an underlying cause of Brugada syndrome (16).
In mice-targeted deletion of the Cacnb2 gene (17) but not of Cacnb1 (18), Cacnb3 (19,20), or Cacnb4 (21) leads to a morphologically and functionally compromised heart, which causes severe defective remodeling of intra-and extra-embryonic blood vessels and death at early embryonic stages both when the Cacnb2 gene was targeted globally or in a cardiac myocytespecific way (17). Although these results point to an essential role of CaV␤2 for I Ca and cardiac function, the existence of various CaV␤2 splice variants and heterogeneity of the expressed CaV␤2 proteins require further studies on the subunit composition of LTCCs in the mouse heart. In addition and in view of the growing number of preclinical studies using mouse models carrying definite Ca 2ϩ channel subunits as transgenes in heart tissue, the identification of the relevant gene products underlying the endogenous mouse cardiac L-type channel is essential. Recent mouse models (e.g. 22,23,24) carrying a rat CaV␤2 splice variant ("rat CaV␤2a") (25) expressed in rat and rabbit brain (26), but not in rabbit heart (26), have only escalated this requirement, because it has never been shown that the mouse orthologue of this variant is endogenously expressed in the mouse heart.
So far, five CaV␤2 variants varying only in the V1 domain have been identified from different species (25,27,28) and in human heart these variants have been obtained mainly by RT-PCR approaches (29,30). In contrast, there is little information on the CaV␤ proteins present in mouse heart, their respective splice variants, and expression ratios. We therefore started to study CaV␤ expression in the murine heart using Western blots and cDNA cloning and to reveal their functional impact on LTCCs formed by the murine CaV1.2 protein.
Construction and Screening of cDNA Libraries-Murine total RNA was isolated from 695 mg (postnatal day 7 (P7)) or 750 mg (adult, Ͼ8 weeks) heart tissue using RNAGold (peqlab). Reverse transcription was performed with 5 g of poly(A) ϩ RNA from adult or P7 mouse heart using a CaV␤2 gene specific primer (GSP, 5Ј-GGA GTG TGC TCT GTC) or hexameric random primers and the Superscript TM Plasmid System with Gateway Technology for cDNA Synthesis and Cloning (Invitrogen). cDNA libraries were constructed in pcDNAII (Invitrogen) and transformed into ElectroMAX TM DH10B TM cells. Bacteria were grown on Duralon-UV nylon membranes (Stratagene). 5 ϫ 10 5 recombinant clones were screened with radioactive cDNA probes covering the conserved C1 and C2 regions of CaV␤2 (C1 431 bp, exons 3-6; C2 591 bp, exons 8 -13). After 3 h of prehybridization, hybridization was carried out overnight at 55°C for random-primed and at 60°C for gene-specific primed libraries, respectively. Prehybridization and hybridization solution contained 450 mM sodium chloride, 45 mM sodium citrate, 0.2% Ficoll, 0.2% polyvinylpyrrolidone, 0.2% bovine serum albumin, and 150 g/ml denatured salmon sperm DNA, with the hybridization solution containing random primed [ 32 P]dCTP-labeled probe (10 5 cpm/l). After washing steps with solutions containing decreasing concentrations of sodium chloride (450 to 45 mM) and sodium citrate (45 to 4.5 mM), 0.1% SDS, membranes were exposed to x-ray films for 8 and 16 h, respectively. Positive clones were isolated and sequenced.
Isolation of Murine Cardiomyocytes-Embryonic and adult cardiomyocytes were isolated as described before (17,31).
RT-PCR-For RT-PCR, we used the Superscript TM One Step RT-PCR with Platinum Taq System (Invitrogen). 10 -12 cardiomyocytes were selected by patch pipette, pooled in Eppendorf tubes, and directly used for reverse transcription and PCR. CaV␤2-N-specific forward primers (N1, 5Ј-ATG GTC CAA AGC GAC ACG TC; N3, 5Ј-ATG CAG TGC TGC GGG CTG; N4, 5Ј-ATG CTT GAC AGG CAG TTG GTG; N5, 5Ј-ATG AAG GCC ACC TGG ATC AG) and the common reverse primer (5Ј-CTC TCT GTT CGT GCT GTA GC) were used. Positive controls were done in the presence of 5 ng of the respective CaV␤2a-N plasmid used as template, negative controls in the presence of H 2 O instead of template; reactions were performed in parallel. PCR (39 cycles) conditions were: reverse transcription at 50°C for 25 min, one denaturation step at 94°C for 2 min to inactivate the reverse transcriptase and to activate PCR polymerase, denaturation at 94°C for 30 s, annealing at 57°C for 30 s and extension at 72°C for 35 s.
Cell Culture and Transfection-HEK293 cells and COS cells were grown in MEM and Dulbecco's modified Eagle's medium (Invitrogen), respectively, supplemented with 10% fetal calf serum (Invitrogen) and maintained under standard cell culture conditions (37°C, 5% CO 2 ).
COS cells were transiently co-transfected with 2 g of CaV␤2 cDNA-encoding plasmids. HEK293 cells were transiently co-transfected with 0.7 g full-length CaV1.2 subunit together with 0.7 g of the different murine cardiac CaV␤2 cDNA encoding plasmids and 0.7 g eGFP. Transfection was carried out with Fugene6 (Roche Applied Sciences). Electrophysiological recordings in GFP-positive cells were obtained 48 -72 h after transfection.
Electrophysiological Recordings-For whole-cell Ca 2ϩ current recordings, HEK293 cells expressing CaV1.2 and CaV␤2 were bathed in a solution containing (in mM): tetraethylammonium chloride 140, MgCl 2 1, CaCl 2 1.8, HEPES 10, pH 7.4 (TEA-OH). Borosilicate patch pipettes (BioMedical Instruments) were filled with a solution containing (in mM): CsCl 120, MgCl 2 3, Mg-ATP 5, EGTA 10, HEPES 5, pH 7.4 (CsOH), and had resistances ranging from 2.0 to 4.0 M⍀. Currents were filtered at 1.67 kHz and digitized at a 5-kHz interval. I Ca was normalized to cell size. Currents were activated from the holding potential of K90 mV every 5 s by step depolarizations from Ϫ70 to ϩ70 mV in 10 mV increments for 400 ms to obtain current-voltage (I-V) relationships. In some HEK293 cells just transfected with CaV1.2 and GFP lacking CaV␤2, a small but clear I Ca could be discerned (maximal I Ca at 0 mV Ϫ2.23 pA/pF, n ϭ 24). Therefore only cells transfected with CaV1.2 and one of the CaV␤2 isoforms, with a current density larger than Ϫ3.5 pA/pF at 0 mV were included in the analysis to ensure that both transfected proteins CaV1.2 and CaV␤2 were present. Currentvoltage (I-V) relationships from individual cells were fitted with I ϭ G max (V-E rev )/(1 ϩ exp(-(V-V 1/2,act )/k)) (32), where G max represents the whole-cell conductance, E rev the apparent reversal potential, V 1/2,act the voltage for half-maximal activation of the current and k the slope factor. The time to peak was measured 1.2 ms after the beginning of the depolarization to avoid measuring capacitance artifacts. The steady-state activation curve was constructed by normalizing the current density measured between Ϫ70 mV and 0 mV to that at 0 mV and fitted with a Boltzman equation I/I 0mV ϭ 1 ϩ (A-1)/(1 ϩ exp((V-V 1/2,ss-act )/k)), with V 1/2,ss-act as the voltage of half-activation, k the slope factor. Steady-state inactivation was measured by a double-pulse protocol; cells were depolarized from a holding potential of Ϫ90 mV to potentials from Ϫ100 to ϩ40 mV for 5 s in 10-mV increments; subsequently the test pulse to 0 mV for 400 ms was applied. Steady-state inactivation curves were normalized to the current after the prepulse to Ϫ100 mV, averaged, and fitted with a Boltzman equation I/I max ϭ 1/(1 ϩ exp((V-V 1/2,inact )/k) with V 1/2,ss-inact as the voltage of half-inactivation and k the slope factor. Voltage protocols were applied, and currents recorded with an EPC9 patch-clamp amplifier (HEKA, Germany) controlled by the software PulseFit (HEKA). A P/4 protocol was used in all experiments to subtract linear leak and capacitance. Data were compared in GraphPad Prism using a one-way analysis of variance test.

CaV␤ Protein Expression in Adult Heart and Brain from Mice
We made use of specific antibodies for CaV␤1, CaV␤2, CaV␤3, and CaV␤4 to detect the CaV␤ subunits in protein fractions from adult mouse heart and brain by Western blot. The CaV␤1, CaV␤3, and CaV␤4 proteins are readily detectable in adult brain (Fig. 1A, 75 g of protein per lane) but not in the adult heart although 200 g of cardiac proteins were applied per lane. In contrast, the CaV␤2 protein is detected in brain (75 g of protein per lane) and in heart (15 g of cardiac protein was applied per lane), demonstrating that CaV␤2 is the predominant CaV␤ protein in the adult mouse heart. Accordingly transcripts of the Cacnb2 gene of ϳ2, ϳ3.5, ϳ4, and ϳ5 kb are readily detectable in poly(A) ϩ RNA isolated from adult mouse heart (Fig. 1B).

CaV␤2 Protein Expression in Developing Mouse Hearts
The CaV␤2 protein is detectable in the developing heart tube approximately at embryonic day (E)8.5. To study its expression during cardiac development, immunoblots with protein fractions prepared from hearts from E10.5 to E17.5, neonatal mice (P1), P7, and adult mouse hearts were analyzed for CaV␤2, CaV␤3, and the CaV␣ 1 c subunit expression. In early embryonic heart, a ϳ72-kDa CaV␤2 protein is expressed (Fig. 1C). Interestingly, the CaV␤2 protein detected in brain lysate shares molecular weight with this embryonic ϳ72-kDa CaV␤2 protein (Fig. 1C). In later stages of embryonic development a second CaV␤2 protein of ϳ68 kDa is co-expressed, whereas in the adult heart only the ϳ68-kDa CaV␤2 protein (Fig. 1, A and C) but not the ϳ72-kDa protein is detectable. The ϳ55-kDa CaV␤3 protein, which is expressed in brain (Fig. 1, A and D) is detectable to a very minor degree in the protein fractions from embryonic heart but not in adult heart (Fig. 1, A and D) whereas the expression of the CaV␣ 1 c proteins parallels CaV␤2 protein expression during cardiac development (Fig. 1C). Apparently, different CaV␤2 proteins are expressed during cardiac development leading to changes in LTCC protein composition. The CaV␤3 protein is primarily expressed in brain (Fig. 1, A and D) but also in heart at early embryonic stages (Fig. 1D). Immunoblots clearly show expression of CaV␤3 in heart lysates from wild type mice at E12.5 and E14.5 but not in lysates of hearts from CaV␤3-deficient mice (Fig. 1D, CaV␤3 Ϫ/Ϫ ). However, CaV␣ 1 c and CaV␤2 subunits are the main constituents of embryonic and adult murine cardiac LTCC channels.
The molecular basis of the different CaV␤2 proteins detected by Western blot analysis is not known but could be caused by the expression of splice variants of the Cacnb2 gene, which has been described in human, rabbit, and rat heart (7, 29, 33). In the . C, expression of CaV␤2 (␤2) and Cav1.2 (␣ 1 c) during mouse heart development. Immunoblot of microsomal proteins from adult heart (h adult, 25 g) and brain (b adult, 50 g) and lysates of hearts (h) taken at embryonic day (E) 10.5, 11.5, 12.5, 14.5, 17.5, 1 day (P1), and 7 days (P7) post partum and antibodies for CaV␤2 and CaV1.2 as indicated. * indicates myosin, which is present in protein lysates and recognized by the CaV␤2 antibodies; it is not detectable in microsomal protein fractions, which were obtained after differential centrifugation. D, expression of CaV␤3 in embryonic heart. Immunoblot of heart protein lysates (150 g per lane) taken at E12.5 and E14.5 from wild type and CaV␤3-deficient mice. Brain microsomal membrane proteins (20 g) from wild type and CaV␤3-deficient mice were used as controls. The blot was stripped thereafter and incubated in the presence of the antibody for CaV␤2 to control loading. OCTOBER 30, 2009 • VOLUME 284 • NUMBER 44

CaV␤2 Diversity and Developmental Expression in Murine Heart
following we identified and characterized CaV␤2 variants in mouse heart by the unbiased approach of constructing and screening of cDNA libraries.

Structure of the Cacnb2 Gene, Strategy to Construct cDNA Libraries, and Isolation of CaV␤2 Variants
The murine Cacnb2 gene is localized on chromosome 2, comprises 20 protein coding exons and extends over a region of ϳ383 kbp. The six 5Ј exons 1A, 1B, 2A, 2B, 2C, and 2D (nomenclature according to the human Cacnb2 gene by Foell et al. (29)) encode for alternate V1 regions and are scattered among the 5Ј ϳ270 kbp of the gene (Fig. 2A). The alternate V2 regions are encoded by the mutually exclusive exons 7A, 7B, and 7C. No splicing events of CaV␤2 V3 region have been described so far. Splicing of V1 and V2 regions could be responsible for the CaV␤2 protein pattern detected in Fig. 1C.
To get hold of all CaV␤2 V1 and V2 splice variants expressed in murine heart we used the following strategies. Poly(A) ϩ RNA was isolated from hearts taken from adult animals and from P7 animals; in the latter, the ϳ68-kDaand ϳ72-kDa CaV␤2 proteins are co-expressed (Fig. 1C). We used these RNAs and two types of primers for construction of cDNA libraries: To obtain the nucleotide sequences of the V1 and V2 regions, for which splicing events were most probably expected, the cDNA first strand was primed by oligonucleotides complementary to the 5Ј-end of the C2 domain, and cDNA library screening was done by a probe covering the C1 domain. Second, we constructed a random-primed cDNA library using the poly(A) ϩ RNA from P7 hearts, which was screened with probes covering the nucleotides encoding the C1 and the C2 domains. The latter approach should identify CaV␤2 variants with differing C1, C2, or V3 domains.
Altogether 60 independent cDNA clones were isolated from the three libraries and sequenced. The majority of clones encoded the sequences of the V1 and V2 regions of CaV␤2 ( Fig. 2 and supplemental Table S1). The random-primed clones also contained the C2 and V3 domains. The V2 domain was encoded in all clones by exon 7A (Fig. 2, A and B). According to the nomenclature of Foell et al. (29), these clones are of the CaV␤2a type (a for the A-type exon 7). Differences were only observed within the V1 domain which gave rise to N termini of type N1 (exons 1A plus 2A), N3 (exon 2B), N4 (exon 2C), and N5 (exon 2D) (Fig. 2). 38 out of the 60 sequenced clones encoded the CaV␤2a-N4 variant, demonstrating that this variant is the predominant CaV␤2a variant in hearts from adult and 7-day-old mice. The CaV␤2a-N5 (5 clones out of 60) and CaV␤2a-N1 variants (16 clones out of 60) are expressed to an intermediate extent. The CaV␤2a-N3 splice variant was only detected once among the 60 CaV␤2 cDNA clones indicating that this variant is only present in a very minor fraction of cardiac LTCCs if at all. The N3-type N terminus is encoded by exon 2B and the corresponding mouse, human, and rat amino acid sequences of this exon are identical and start with 1 Met-Gln-Cys-Cys (Fig. 2C). The cysteine residues at position 3 and 4 in the N3 terminus of the rat protein have been identified as sites of palmitoylation (34). No N2-type N terminus-encoding , and -N5 (2D), and STOP the common termination. Arrow indicates GSP, gene-specific primer, which is complementary to C2 and used to specifically prime the cDNA libraries. B, scheme of the four CaV␤2 variants identified in murine heart; they differ in V1, but not in C1 (starting with amino acids GSA), V2 (exon 7A), C2, and V3. Colors of V1 correspond to colors of coding exons. 34 of 60 independent clones coded for CaV␤2a-N4, 16 for CaV␤2a-N1, five for CaV␤2a-N5, and one for CaV␤2a-N3. C, amino acid alignment of the V1 domains of CaV␤2 from mouse (this study), rat and human; from the mouse sequences differing amino acids are highlighted in green; ϳ, no amino acid residue; * predicted sequence. GenBank TM accession numbers: N1 mouse BC109156, this study FM872408; N1 human AF423191; N2 mouse predicted from genomic sequence; N2 rat AY190119; N2 human AF423190; N3 mouse this study FM872406; N3 rat NM_053851; N3 human AF423189; N4 mouse this study AM259383; N4 rat AF_423193; N4 human NM_201590; N5 mouse L20343, this study FM872407; N5 rat AY190120; N5 human NM_201570.

CaV␤2 Diversity and Developmental Expression in Murine Heart
cDNA was identified although the N2-coding exon 1B is present within the mouse Cacnb2 gene. Supplemental Fig. S1 shows an alignment of the amino acid sequences derived from the four types of clones identified. They only differ in the N termini N1, N3, N4, and N5 whereas the remaining sequences starting with the amino acid residues GSA (exon 3, Fig. 2) are identical within the four CaV␤2 splice variants. The sequences following V1 comprise three serine residues, which in the rat CaV␤2-N3 orthologue have been suggested to be phosphorylated in the heart in vivo and in vitro (Ser-459, Ser-478, and Ser-479 in CaV␤2a-N3 (35)).

CaV␤2 Expression in Isolated Cardiac Myocytes
So far the results demonstrate expression of predominantly CaV␤2a-N4, CaV␤2a-N1, and CaV␤2a-N5 in murine heart. For construction of the cDNA libraries, poly(A) ϩ RNA had been isolated from the entire heart; thus, RNA from fibroblasts, endothelial cells, and neurons was included. To refine the expression analysis, 10 -12 cardiomyocytes from adult and embryonic hearts were subjected to combined cDNA synthesis and PCR reactions using primer pairs specific for N1-, N3-, N4-, and N5-type N termini. Analysis of the PCR products revealed that the CaV␤2-N1, -N4, and -N5 splice variants are expressed in adult and embryonic cardiomyocytes (Fig. 3). No PCR product was obtained for CaV␤2-N3 demonstrating that CaV␤2-N3 is not expressed in cardiomyocytes confirming the result obtained by cDNA library screening.

In Vitro Reconstitution of the in Vivo CaV␤2 Protein Expression Pattern
Considering that CaV␤2a-N1, CaV␤2a-N4, and CaV␤2a-N5 are expressed in cardiac myocytes, we wondered if these proteins are sufficient to explain the CaV␤2 protein pattern observed in Western blot experiments (Fig. 1C). Therefore, the three proteins were separately expressed in COS cells and lysates of these cells were used for Western blots (Fig. 4). As controls we used protein fractions from hearts of adult mice (Fig. 4, h adult) and of P7 mice (Fig. 4, h P7). The electrophoretic mobilities of CaV␤2a-N4 and CaV␤2a-N5 proteins expressed in COS cells (Fig. 4) resemble the mobility of the ϳ68-kDa CaV␤2 protein endogenously expressed in the adult heart (Fig. 4, h adult) and in the P7 heart (Fig. 4, h P7). In contrast, the CaV␤2a-N1 protein runs slightly slower (Fig. 4), very much like the ϳ72-kDa CaV␤2 protein endogenously coexpressed with the ϳ68-kDa CaV␤2 protein in the P7 heart. Apparently the N4-and N5-variants are the predominant CaV␤2 proteins expressed in the adult mouse heart, whereas in P7 hearts the CaV␤2a-N1 protein is additionally co-expressed. This assumption is supported by the finding that the protein patterns of mixtures of the respective COS cell lysates (Fig. 4) resemble the pattern observed in P7 hearts.
In summary, three CaV␤2 protein variants are expressed in mouse heart with CaV␤2a-N1 being expressed predominantly in embryonic stages, followed later by the additional expression of CaV␤2a-N4 and -N5, which become more and more the prevailing CaV␤2 proteins in the maturing heart.
Next we wanted to study the influence of these CaV␤2 proteins on LTCC currents (I Ca ). To keep the heterologous expression system as close to the murine Ca 2ϩ channel as possible, we wanted to co-express only the cDNAs of the murine CaV1.2 and the CaV␤2 variants in HEK293 cells. No full-length murine cardiac CaV1.2 cDNA was available or obtainable, why we amplified murine full-length CaV1.2 cDNA from mouse heart (supplemental "Experimental Procedures," Fig. S2, and Table  S6). The exons 1A and 1B are supposed to encode the N terminus of the cardiac (1A) and smooth muscle CaV1.2 (1B) proteins in rabbit (5,36), rat, and human. Depending on the presence of either exon the CaV1.2 clones were referred to as CaV1.2a and CaV1.2b.

Different Modulation of L-type Ca 2؉ Channel Currents (I Ca ) by the CaV␤2 Variants
Current Density and Steady-state Activation and Inactivation-First, we analyzed current densities at different test potentials in HEK293 cells co-expressing CaV1.2a and each of the CaV␤2 variants. As controls we used non-transfected HEK293 cells and HEK293 cells just expressing CaV1.2a. Ca 2ϩ currents were recorded in response to voltage steps of 400-ms duration to Ϫ70 mV up to ϩ70 mV in 10-mV increments from a holding  Immunoblot of COS cell lysates expressing the mouse CaV␤2 variants CaV␤2a-N5, -N4, -N1, -N1 plus -N5, -N1 plus -N4 and -N1 plus -N4 plus -N5, microsomal membrane proteins from adult heart (h adult, 50 g), and protein lysate from heart of P7 mice (h P7, 100 g) using the antibody 425 for CaV␤2. *, compare with legend in Fig. 1C. OCTOBER 30, 2009 • VOLUME 284 • NUMBER 44 potential of Ϫ90 mV. Fig. 5A shows a family of current traces in CaV1.2a/CaV␤2a-N4-expressing cells; the average currentvoltage (I-V) relationships for the different co-expressions are shown in Fig. 5B. Recordings in HEK293 cells co-expressing CaV1.2a and a CaV␤2 protein reveal characteristic I Ca , which were not observed in control HEK293 cells.

DISCUSSION
In a first series of experiments we show that CaV␤2 is the major CaV␤ protein expressed in the adult mouse heart and that a switch of CaV␤2 protein isoform expression occurs during development.
Whereas numerous studies described cloning of Ca 2ϩ channel subunits and Ca 2ϩ channel composition in the human, guinea pig, and rat heart (5,7,25,29,39,40,41) the molecular "make-up" of the mouse cardiac LTCC is much less defined, although several mouse lines have been created with Ca 2ϩ channel subunit transgenes as preclinical models for heart diseases (22,23,24,42,43).
CaV␤2 Protein Expression in Heart during Development-In the mouse heart, LTCCs are already functional at E9.5 (17). Analysis showed expression of the CaV␤2 protein throughout embryonic development whereas CaV␤1 and CaV␤4 protein expression was not detectable. The CaV␤3 protein is very weakly expressed in the embryonic hearts but not in hearts from perinatal and adult mice. We observed a switch in CaV␤2 protein expression from a ϳ72-kDa protein, mainly expressed at early embryonic stages to a ϳ68-kDa protein in hearts from adult mice.
Screening of cDNA Libraries for CaV␤2 Subunit Variants and Reconstitution of the CaV␤2 Protein Expression Pattern-To identify possible CaV␤2 variants we constructed cDNA libraries which were screened by probes covering the conserved domains C1 and C2 which together with V2 comprise the CaV␤2 core. Most groups (29,30,37,44) used strategies aimed at the specific amplification of a DNA fragment to obtain CaV␤2 variants, which is critically determined by the primer pair used for PCR. In contrast, the screening of cDNA libraries represents a rather unbiased approach to identify which mRNA is expressed. In addition, this approach makes it possible to study the frequencies of given cDNAs, provided a sufficient number of clones are available. In the case of cardiac CaV␤2, sixty independent cDNA clones were identified, subcloned, and sequenced allowing an estimate of the types of CaV␤2 splice variants expressed in mouse heart and their naturally occurring frequencies. By this method, only potential CaV␤2 variants lacking the C1-V2-C2 core region (30) may have been missed. We confirmed the results by PCR amplification of the CaV␤2 variants identified by cDNA library screening using isolated cardiomyocytes as template. The combined approaches identified three CaV␤2 variants expressed in murine cardiomyocytes, CaV␤2a-N1, -N4, and -N5, which only differed in their V1 regions encoded by alternate exons. Reconstitution of the protein pattern of ϳ72and ϳ68-kDa proteins in vitro by co-expression of CaV␤2a-N1, -N4, and -N5 can explain the protein pattern obtained in protein fractions from mouse heart at different developmental stages. CaV␤2a-N1 is predominantly expressed in hearts from embryonic and neonatal mice, whereas CaV␤2a-N4 and -N5 are isoforms expressed in hearts from neonatal and adult mice. The CaV␤2a-N1 protein is not detectable in protein fractions from adult heart. Results from cDNA library screening indicate a low number of CaV␤2 N1-type protein in the adult heart, which might have escaped detection with available antibodies.
The N2-type variant comprising the second protein-coding exon (exon 1B) was not identified at all. It encodes a twelve amino acid residues sequence (MDQASGLDRLKI), which is predicted to be alternatively spliced to the third protein coding exon (exon 2A); the orthologue exons have been identified in rat (rat CaV␤2c, GenBank TM Acc. No. AY190119) rabbit (rab-