INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

In EC¹ coupling an action potential activates DHP receptors in the transverse tubular membrane, which somehow open ryanodine receptors in the sarcoplasmic reticulum to release Ca²⁺ and initiate muscle contraction. The skeletal muscle DHPr is a slowly activating L-type Ca²⁺ channel (2) that also controls Ca²⁺ release, working as a voltage sensor (3, 4). How DHPrs interact with ryanodine receptors during EC coupling is still not clear. Most experiments on EC coupling have been done in frog skeletal muscle fibers, although the cDNAs of skeletal muscle DHPr have been cloned from rabbit (5), fish (carp) (6), and human (7). Functional differences exist between frog and mammalian skeletal muscle DHPrs. Seeking additional insights into the molecular basis of skeletal muscle EC coupling, we cloned and expressed a skeletal muscle DHPr from a frog.

The frog DHPr (Fg1S) was cloned from mRNA of skeletal muscle of Rana catesbeiana by combining RT-PCR with construction and screening of a cDNA library. The molecule has high homology with other DHPrs at transmembrane domains, voltage sensors (5), pore regions (8), and DHP binding sites (9). Despite those conserved features, Fg1S has also several differences with known DHPrs.

The cloned rabbit skeletal muscle L-type Ca channel has been expressed in dysgenic myotubes (4) and in Xenopus oocytes (10). However, functional expression of this channel in mammalian cell lines has been difficult (11-14). In our hands, the functional expression of Fg1S also occurred infrequently. In contrast, a chimera (named 1C-FSDI) of domain I of Fg1S in a rabbit cardiac background was expressed easily and at high levels.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

RNA Preparation-- Total RNA samples were isolated from the liver, cardiac, and skeletal muscle of a frog by the acid guanidinium-phenol-chloroform method (15). Poly(A)⁺ RNA was purified from the total RNA by using the Mini-oligo(dT) cellulose spin kit (5 Prime  3 Prime, Inc., Boulder, CO).

cDNA Cloning by RT-PCR-- Total RNA harvested from frog skeletal muscle was reverse transcribed using random primers. Based on the conserved amino acid sequences of known DHP receptors, two pairs of degenerate primers (C5/C6 and C12/J3) were synthesized for amplifying a cDNA segment (segment 7) between IIIS6 and IVS6 and a segment (segment 6) between IS1 and IIS3 of the channel molecule. Primer C5 (5'-CGCGGAATTC TT(T,C) ATG ATG AA(C,T) AT(T,C,A) TT(T,C) GT-3', with base degeneracy represented in parentheses) had an EcoRI restriction site at the 5' end (bold) fused to the DNA sequence of FMMNIFV in IIIS6. Primer C6 (5'-CGG CGG ATCC A NAG CAT(A,G)TA (A,G)AA (A,G)CT (A,G,T)AT (A,G)AA-3') had a BamHI site at the beginning fused to the cDNA sequence of FISFYML in IVS6. Primers C12 (5'-AT(A,C,T) GTN GA(A,G) TGG AA(G,A) CCN TT-3') and J3 (5'-AA (A,G)CA (A,G)TC (A,G)AA NC(G,T) (A,G)TT (A,G)AA-3') were derived from the conserved amino acid sequences IVEWKPF at the beginning of domain IS1 and FNRFDCF in domain IIS3, respectively. Based on the cDNA sequences of cloned frog cDNA segments 6 and 7, nondegenerate primers (J4 and J5) were designed to amplify the cDNA (segment 9) between domains II and IV. Primer J4 (5'-GCC CTT GGT TTC CAG TCA TAT TTC-3') corresponds to the amino acid sequence ALGFQSYF in IIS2, and J5 (5'-CAT ACC AAG GCT GAT GGT GTT CAG-3') presents the cDNA sequence of LNTISLGM in IVS1. Control reactions were performed in the absence of reverse transcriptase to rule out possible contamination by genomic DNA. First-strand cDNA synthesis kits were from either Boehringer Mannheim or Pharmacia Biotech, Inc. PCR products were subcloned into pGEM-T vectors (Promega, Madison, WI).

cDNA Cloning by Constructing and Screening a Frog Skeletal Muscle cDNA Library-- Because there are no conserved sequences at both C and N termini of known Ca channels, no gene-specific primers could be designed for RT-PCR. A frog skeletal muscle cDNA library was constructed instead, and the probes derived from RT-PCR clones (segments 6 and 7) were used to screen it and to obtain both 5' and 3' cDNAs. Double-stranded cDNA was synthesized using the Superscript Choice system for cDNA synthesis (Life Technologies, Inc.). The size-fractionated cDNA (>1 kb) was ligated into the ZAP Express vector (Stratagene, La Jolla, CA). The recombinant phage vectors packaged with a protein coat (Gigapack II Gold, Stratagene) were used to infect XL1-Blue-MRF' bacteria (Stratagene) to construct the library. Positive clones pBK/5-13 (covering the 5' untranslated region), pBK/5-8, pBK/5-10 (covering the 3' untranslated region), and pBK/5-5 were obtained by screening the library.

DNA Sequencing and Analysis-- All cDNA clones were sequenced using the Sequenase 2.0 DNA sequencing kit (Amersham Corp.) and the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin-Elmer). Sequence similarity searches of GenBank^TM were performed with PCGENE (IntelliGenetics, Mountain View, CA) and DNASIS (Hitachi-Software, South San Francisco, CA).

Full-length Channel Construction-- Based on suitable restriction enzyme sites (Fig. 1), RT-PCR clones (segments 6 and 9) and the library clones (5-13, 5-5, and 5-8) were ligated to obtain the full-length molecule (Fg1S) in the mammalian expression vector pCR3.1 (Invitrogen, San Diego, CA).

Northern Blot Analysis-- Total RNA (20 µg) isolated from various frog tissues was separated on a 1% agarose-formaldehyde gel, transferred to Nytron BA-S-supported nitrocellulose membranes (Schleicher & Schuell), and cross-linked by baking for 2 h at 80 °C. The cDNA probe, including Fg1S nucleotides 4859-5600, was cut out from clone pBK/5-8 by NotI and XmnI digestion. The probe was radiolabeled with [-³²P]dCTP using random primers (Ready to Go DNA labeling beads, Pharmacia). The membrane was hybridized with the probe for 24 h at 42 °C in 6 × SSC, 50% formamide, 0.5% SDS, 100 µg/ml sheared salmon sperm DNA, and 5 × Denhardt's solution. The high-stringency wash was performed in 0.5 × SSC and 0.1% SDS at 60 °C. X-ray film (X-Omat AR, Eastman Kodak Co.) was exposed to the membrane at 80 °C. Glyceraldehyde-3-P dehydrogenase probes were also used for hybridization to estimate relative total RNA for each tissue.

Chimeric Channel Construction-- cDNA coding for domain I of rabbit cardiac channel (1C) at amino acid positions 165-422 was replaced by the corresponding cDNA segment from either Fg1S or rabbit skeletal muscle 1S to obtain chimeras.² The cDNA of 1C has unique restriction sites, MfeI at nucleotide 689 of IS1 and BamHI at nucleotide 1456 of IS6. These unique sites also appear at the corresponding positions of Fg1S (491 and 1258, respectively). The cDNA (491-1258) was cut out from Fg1S and then ligated into MfeI and BamHI predigested pCR3/1C to produce the frog chimeric channel 1C-FSDI. MfeI and BamHI sites are not present at the corresponding positions of domain I of rabbit 1S, and the sites were introduced by PCR. The PCR product, which covers at 415-1185 of rabbit 1S, was cut by MfeI and BamHI and then ligated into pCR3/1C to produce the rabbit chimeric channel 1C-RSDI.

Transfection Procedure-- Large T-antigen-transformed human embryonic kidney cells (tsA201, a gift from Dr. M. M. Hosey, Northwestern University, Chicago, IL) were maintained in DMEM (Sigma) containing 10% fetal bovine serum (Biowhittaker, Walkersville, MD), 100 units/ml penicillin/streptomycin (Sigma) at 37 °C in 5% CO₂. Rabbit skeletal muscle 2/ (in pMT2) was a gift from Dr. T. Tanabe (Tokyo Medical/Dental University, Tokyo, Japan). Rabbit cardiac 1C was carried in pCR3 and rat brain 2a in pCMV. Seven micrograms of each cDNA (total <25 µg) were used for transfection by calcium phosphate precipitation (16). Expression efficiency was defined as the fraction of cells that had Ca²⁺ currents among those selected and patched.

Electrophysiological Studies-- Ionic currents were recorded by whole-cell patch clamp. Voltage clamping, pulse generation, and data acquisition were carried out with an Axopatch 200 A (Axon Instruments, Foster City, CA) and a 16-bit digital conversion card in a PC-compatible computer. The analog-to-digital and digital-to-analog routines were written by Ivan Stavrovsky in our laboratory. Pipettes were from borosilicate glass (Corning 7052; Garner Glass, Claremont, CA) heat polished to a tip outside diameter of 2-4 µm for a resistance of 2-4 M. Ion currents were recorded with a pipette solution containing (in mM) 150 Cs⁺, 125 Asp, 15 Cl, 5 MgATP, 10 HEPES, and 10 EGTA, pH 7.6, adjusted by CsOH, and an external solution containing (in mM) 130 NaCl, 10 HEPES, and 10 Ca²⁺ or Ba²⁺, pH 7.3. Experiments were performed at 20-23 °C. Single, nonclustered cells were chosen for recording. Cells were pulsed from a holding potential of 90 mV. Asymmetric currents were obtained by subtraction of scaled control currents elicited with pulses from 130 to 100 mV.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Sequence Alignments of RT-PCR and Library Clones-- The composite cDNA sequence of Fg1S was obtained by sequencing multiple overlapping clones. The sequence map of all clones is shown in Fig. 1. The 5'-end sequence, including an untranslated region, is present in the library clone 5-13. The 3'-end, including an untranslated region and a portion of the poly(A) tail, is covered by library clones 5-8 and 5-10. The cDNA sequences of the library clones (5-13, 5-5, 5-8, and 5-10) are equal to those of RT-PCR clones (segments 6, 7, and 9), with several single-base differences: 1) clone 5-13 has CGT⁷⁸⁰, whereas segment 6 and the full-length molecule have CGG⁷⁸⁰; both CGT and CGG encode arginine; 2) clones 5-13 and 5-5 have G⁸⁷⁶TA encoding valine, whereas segment 6 and the full-length cDNA have A⁸⁷⁶TA, encoding isoleucine; 3) clone 5-5 has T²¹⁵¹CT (serine), whereas segment 9 and the full-length cDNA have G²¹⁵¹CT (alanine) at this position; and 4) segment 7, clone 5-8, and the full-length cDNA have CTC⁴²⁷⁴, whereas clone 5-10 has CTT⁴²⁷⁴. Both CTC and CTT code for leucine. These single-base differences could be explained by polymorphism, also observed in 1 cDNA clones from rabbit skeletal muscle (5).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Cloned cDNA segments. cDNA segments 6, 7, and #9 were obtained from RT-PCR and subcloned into pGEM-T. Clones 5-13, 5-5, 5-8, and 5-10 were obtained from the cDNA library in pBK(CMV) plasmid vectors. Sequences in the overlapping regions are identical, except for several base differences described under "Results." The restriction sites indicated on each clone were used for construction of the full-length Fg1S.

cDNA and Deduced Amino Acid Sequence of Fg1S-- The 5600-nucleotide cDNA sequence of Fg1S and its amino acid sequence are available (GenBank^TM accession number AF037625). The polyadenylation signal sequence, AATAAA (17), is found between nucleotides 5554 and 5559, 16 nucleotides upstream of the poly(A) tail. Translation of the deduced channel protein (1688 amino acids) starts at nucleotide 303 with the first methionine and ends at nucleotide 5367, which is followed by the stop codon TGA. Transmembrane domains and pore regions were deduced from an alignment of Fg1S with rabbit cardiac (Rb1C) (18), skeletal (Rb1S) (5), and carp skeletal muscle (Cp1S) DHP receptors (6). Like other L-type Ca²⁺ channels, Fg1S has four putative transmembrane domains, but its C terminus is shorter than that of Rb1S and Cp1S. Table I lists the homology between different portions of Fg1S and rabbit and carp L-type Ca²⁺ channels. The homology is high with all L-type channels, but in every domain the highest homology is with Rb1S.

View this table: [in this window] [in a new window]   Table I Degrees of homology of Fg1S with other L-type Ca channels Rb1C, rabbit cardiac channel; Rb1S, rabbit skeletal muscle channel; Cp1S, Carp skeletal muscle channel. Note the greater homology with the skeletal muscle channels and the high degree of homology with Rb1S in the critical II-III linker.

Conserved Structure of a Voltage Sensor-- The structural basis of the voltage sensitivity has been located at transmembrane segments S4 for voltage-gated K⁺, Na⁺, and Ca²⁺ channels (19-21). An alignment of the S4 segments of Fg1S with rabbit and carp channels (Fig. 2A) shows identical charges and almost perfect homology, consistent with the requirements for the voltage sensor of EC coupling and a voltage-operated ion channel.

View larger version (49K): [in this window] [in a new window]   Fig. 2.   Sequence alignment of Fg1S with other DHP receptors. Rb1C, 1 subunit of rabbit cardiac channel; Rb1S, 1 subunit of rabbit skeletal muscle channel, Cp1S, 1 subunit of carp skeletal muscle channel. Dashes mark conserved amino acid residues in Rb1C. A, alignment of S4 in the four domains. B, putative pore regions. C, regions involved in DHP binding.

Conserved Structure for High-affinity Ca²⁺ Selection in Pore Regions-- The pore regions of ion channels have been localized to the linkers between segments S5 and S6 (22). Four glutamic acid residues at equivalent positions in pore regions have been demonstrated to determine the high-affinity Ca²⁺ binding in L-type channels (23), and other cloned high-voltage-activated Ca²⁺ channels also have these conserved glutamic acid residues at equivalent positions. The alignment of pore regions of Fg1S, rabbit, and carp channels is shown in Fig. 2B. Fg1S has conserved sequences in those regions, as well as the critical glutamic acid residues.

Conserved Structure for DHP Binding-- DHP binding affinity is an established criterion to distinguish L- and non-L-type Ca channels (24). The DHP binding site of L-type Ca channels appears to involve regions between transmembrane segments IIIS6 and IVS6, where residues Tyr¹⁰⁴⁸, Ile¹⁰⁴⁹, and Ile¹⁰⁵² in IIIS6 and Tyr¹³⁶⁵, Met¹³⁶⁶, Ile¹³⁷², and Ile¹³⁷³ in IVS6 are critical (25). Fg1S has a sequence almost identical to other DHP receptors in the DHP binding regions, including all required residues at corresponding sites (Tyr¹⁰⁴⁶, Ile¹⁰⁴⁷, Ile¹⁰⁵⁰, Tyr¹³⁵², Met¹³⁵³, Ile¹³⁵⁹, and Ile¹³⁶⁰). The alignment of DHP binding regions of Fg1S with rabbit and carp DHP receptors is shown in Fig. 2C.

Conserved Structure for Skeletal Muscle-Type EC Coupling-- Unlike cardiac cells, in skeletal muscle Ca²⁺ release is triggered by DHPrs without requirement for entry of extracellular Ca²⁺ (26, 27). A crucial molecular determinant of skeletal muscle type EC coupling is at the intracellular loop between domains II and III of rabbit 1S (28). The II-III loop of Fg1S has high (86%) homology with Rb1S (Table I), suggesting that Fg1S is a skeletal muscle-type voltage sensor.

PKA Phosphorylation Sites-- Skeletal muscle L-type Ca²⁺ channels are regulated by phosphorylation. In rat skeletal muscle fibers, both Ca²⁺ current and intramembranous charge movement are increased by PKA (29), whereas in frog fibers, only Ca²⁺ currents are increased by PKA (29) or cAMP (30). Serine 687, in the intracellular II-III loop of rabbit 1S, is a major PKA phosphorylation site in vitro (31). Fg1S has no Ser or Thr at the corresponding position. In intact rabbit skeletal muscle myotubes other putative PKA sites have been located at Ser¹⁷⁵⁷ and Ser¹⁸⁵⁴ at the C terminus of the full-length channel (32). Fg1S, with a shorter C terminus, has no corresponding sites. By searching the derived amino acid sequence, we located a PKA phosphorylation site motif, RRXS (33), at the intracellular I-II loop of Fg1S (Ser⁴²⁴).

mRNA Distribution in Native Tissues-- To determine the tissue specificity of the cloned molecule, we carried out Northern blot analysis. Total RNA from frog heart, liver, and skeletal muscle tissues was hybridized with a cDNA probe derived from Fg1S, covering the 3'-untranslated region and the diverse sequence of the C terminus. Despite the differences in glyceraldehyde-3-P dehydrogenase signal between lanes, the ethidium bromide staining gel (data not shown) indicated that approximately the same amount of total RNA was loaded on each lane. The blot revealed a single RNA band in skeletal muscle (Fig. 3), consistent with the result found in mammalian skeletal muscle (34). There were no detectable transcripts in any of the other tissues, even using longer exposure times (data not shown).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Northern blot analysis of Fg1S. Total RNA from various tissues (20 µg each) of Rana catesbeiana was electrophoresed on an agarose-formaldehyde gel, blotted to a nitrocellulose membrane, and probed with Fg1S cDNA segment 4859-5600. A glyceraldehyde-3-P dehydrogenase (GAPDH) probe was used to evaluate relative RNA on the membrane for each tissue.

Expression of Full-length Fg1S-- tsA201 cells were chosen as a functional expression system for the cloned frog channel because HEK cells have no endogenous L-type Ca²⁺ currents (35), and large T antigen-transformed HEK cells (tsA201) express rabbit cardiac calcium channels efficiently (16).

View this table: [in this window] [in a new window]   Table II Expression efficiency and current properties in cells transfected with Fg1S One hundred forty cells were stable after whole-cell patching. The currents recorded in 15 of those were studied and included, either because they were >0.3 pA/pF (the native nontransfected cells) or because they had L-type characteristics (all others). IP is the normalized peak current, and 1/2 is the time to half-activation at the maximum of the current-voltage dependence.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Ionic currents in cells transfected with Fg1S. A, Ba2+ currents in a cell transfected with 2a alone, with or without 5 µM Bay K in the bath. B1 and B2, Ba2+ currents in a Fg1S- plus 2a-transfected cell in reference (B1) or with Bay K (B2), elicited by pulses to voltages between 40 and +60 mV at 20 mV intervals. C1 and C2, Ca2+ (C1) and Ba2+ (C2) currents in a cell transfected with Fg1S plus 2a plus 2. Cells were pulsed from a holding potential of 90 mV.

Studies of Chimeric Channels-- Because functional expression of Fg1S was so infrequent, we carried out additional expression tests with a chimera of Fg1S and rabbit 1C. In rabbit channels domain I determines the slow activation kinetics (37). Because the currents induced by the full Fg1S activated rapidly, neither domain I nor any other in the frog molecule was expected to make the activation of the chimeric construct slower. Still, we chose to splice domain I of Fg1S to rabbit 1C (to make 1C-FSDI), because a similar chimera with DI from Rb1S (named 1C-RSDI) should have slower kinetics, thus providing overall control of the procedures. In all cases, 2a cDNA was cotransfected with the 1 constructs.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Functional expression of 1C and its chimeras. ICa and IBa in a cell transfected with rabbit wild-type 1C (A), the rabbit chimera 1C-RSDI (B), or the frog skeletal-rabbit cardiac 1C-FSDI (C). All cells were co-transfected with 2a. Cells were held at 90 mV. Pulses were at 10-mV intervals in the range of 20 to +40 mV for ICa and 30 to +30 mV for IBa.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Kinetic properties of currents in cells transfected with 1C or its chimeras. (A) activation kinetics. 1/2, determined in records such as those in Fig. 5 as the time to one-half of peak amplitude, averaged in seven cells for each type of molecule. Bars, ±S.E. , FSDI; , a1C; , RSDI. B, inactivation kinetics. Ratio of ICa (open symbols) or IBa (filled symbols) at 100 ms, over peak current in the same record, calculated in records such as those in Figs. 4 (for Fg1S, triangles) and 5 (for 1C, circles, or its frog chimera FSDI, squares) and averaged (numbers of cells listed in Table II).

Current-dependent Inactivation of Expressed Channels-- L-type Ca²⁺ currents of cardiac muscle inactivate by a slow voltage-dependent mechanism and a faster, current-dependent mechanism. Initiation of the latter has been ascribed to Ca²⁺ binding to a stretch of the 1C subunit, with similarities to an EF-hand (38), a view contested by Zhou et al. (39). Despite the discrepancy, there is agreement that the C terminus is largely responsible for ion-dependent inactivation. Moreover, this is a characteristic that distinguishes cardiac L-type from skeletal muscle channels, in which ion-dependent inactivation appears not to exist (1, 40). Two signature characteristics of this mechanism are a U-shaped voltage dependence and a much more marked effect when the current is carried by Ca²⁺ rather than Ba²⁺. We found both characteristics in the expressed current with both the full Fg1S and its chimera 1C-FSDI. Fig. 6B plots the ratio between ionic current amplitude at 100 ms during a pulse and peak amplitude (reached usually at 20-30 ms during the pulse), as a function of pulse voltage. The filled symbols plot the results for I_Ba, a monotonically decreasing dependence, indicative of a lack of significant ion-dependent inactivation at 100 ms. The open symbols, representing the I_Ca results, show by contrast the characteristic U shape, indicative of current-dependent inactivation, equally marked when the 1 subunit is cardiac (circles), frog skeletal (triangles), or the FSDI chimera (squares).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

We have obtained a cDNA from frog skeletal muscle and named it Fg1S. It has 5600 nucleotides and codes for a protein with 1688 amino acids, highly homologous with the DHP receptors, particularly rabbit skeletal muscle DHPr. As for all known L-type Ca²⁺ channels, the deduced protein has four putative transmembrane domains, a positively charged S4 in each domain, conserved Ca²⁺ channel pore regions, and putative DHP binding sites. The intracellular II-III loop, a molecular determinant of skeletal muscle-type EC coupling, is conserved very well in Fg1S. Additionally, Northern blot analysis showed that Fg1S is transcribed in skeletal muscle but not in heart or liver. For these reasons, the molecule qualifies as a skeletal muscle-specific DHPr.

Functional expression of this cDNA was possible, but only ~10% of the transfected cells that were studied had large L-type inward currents compared with 80% of the cells transfected with the frog and rabbit cardiac chimera. The reason for this low efficiency of expression, consistent with previous attempts to express the 1S in mammalian cell lines (11-14), is unknown. Nevertheless, the good functional expression of the chimeric channel is further evidence that the cloned Fg1S encodes an L-type Ca²⁺ channel.

Surprisingly, the large L-type currents induced by transfection with Fg1S had fast activation, unlike that of the "slow" Ca²⁺ current of skeletal muscle, and a decay during 100 ms pulses with the characteristics of ion-dependent inactivation usually associated with cardiac I_Ca. Based on the existence of a fast and a slow phase in the activation of Ca currents (41), it has been proposed that there are two types of voltage-dependent Ca²⁺ channels in frog muscle. Fg1S might encode the "fast" Ca²⁺ channel. On the other hand, slowly activating Ca²⁺ currents of frog muscle can be switched to a fast activation mode by prior depolarization (42, 43). Therefore, the fast phase seen in the normal unconditioned current (and the currents induced here by Fg1S) may not correspond to a separate channel population but to a fraction of the slow channels showing fast gating kinetics in the steady state.

Because expression of the full Fg1S was seldom successful, we used part of its sequence in a chimera with a cDNA that expresses efficiently in tsA201 cells. We did achieve a much higher expression efficiency with the chimera, and the expressed currents were still cardiac like in their activation rates, as expected given the fast activation of the currents obtained with the full clone. Therefore, slow I_Ca activation kinetics, which for rabbit skeletal muscle is determined by domain I (37) and for carp skeletal muscle appears to be linked to domains III and IV (44), may for frog require other interactions that are lost when Fg1S is expressed in mammalian cells. Other protein molecules such as Ca²⁺ release channels, which appear to determine high current density (13), may be required for establishing native kinetics.

A related kinetic property of the expressed currents is the presence of ion-dependent inactivation, as revealed by voltage and ion dependence of the reduction of current after the peak. In skeletal muscle, I_Ca is not believed to undergo Ca²⁺-dependent inactivation (1, 40, 45). The currents obtained by expression of the Fg1S had a rapid phase of decay, to ~75% of their peak amplitude after 100 ms, that was clearly greater in Ca²⁺ than Ba²⁺. This decay therefore qualifies as Ca²⁺-dependent inactivation. Its presence is not inconsistent with the skeletal muscle origin of the channels because, as shown by Adams and Tanabe (1) on cardiac-skeletal chimeras expressed in HEK293 cells, the mechanism only requires the absence of a skeletal muscle-specific stretch in the primary sequence, rather than the presence of a cardiac-specific portion. These authors showed that deletion of the distal 211 residues in the (492-residue-long) C terminus of rabbit skeletal Rb1S restored Ca²⁺-dependent inactivation to the expressed currents. Fg1S, with the C-terminal 172 residues missing, has a structure close to that of the rabbit deletion mutant. The fact that it displays Ca²⁺-dependent inactivation supports the interpretation that the distal C-terminal segment is what prevents skeletal muscle channels from displaying this type of inactivation.

Despite its high homology with other DHP receptors, Fg1S has unique sequence features. The major PKA phosphorylation sites of rabbit 1S are not conserved at corresponding positions of Fg1S. These structural differences may explain differences in regulation between mammalian and amphibian skeletal muscle DHPrs.

The C terminus of Fg1S (after IVS6) is 172 amino acid residues shorter than that of the corresponding full-length rabbit 1S. Rabbit and rat skeletal 1 subunits purified from T-tubule membrane and intact muscle cells in culture (46, 47) have a predominant form with an apparent molecular mass of 175 kDa (1₁₇₅), which is truncated at the C terminus near residue 1690, and a minor form of 212 kDa, containing the complete amino acid sequence (48). Fg1S, with its 1688 amino acid residues, has nearly the same size as 1₁₇₅. Expression of a truncated rabbit cDNA with an even shorter C terminus (1662 amino acid residues, the same deletion mutant used in the work by Adams and Tanabe (1) mentioned earlier) fully restored both EC coupling and Ca current in dysgenic myotubes (2). This suggests that Fg1S may be sufficient to function in vivo as both a voltage sensor for EC coupling and a Ca channel.

In summary, a skeletal muscle-specific cDNA has been cloned in the frog. Structurally and functionally, this molecule codes for the main subunit of an L-type Ca²⁺ channel. This channel activates rapidly, resembling the minor fast component of native I_Ca, and inactivates in a voltage- and Ca²⁺-dependent manner, as expected from its truncated C terminus. Its sequence is adequate for skeletal muscle-type EC coupling, although its functionality in this regard remains to be shown. Its similarities and differences, both functional and structural, with other L-type Ca²⁺ channels, make this molecule an interesting new tool for studies of Ca²⁺ current and Ca²⁺ release gating, as well as modulation and gene expression of L-type Ca²⁺ channels in skeletal muscle.