Absence of the gamma subunit of the skeletal muscle dihydropyridine receptor increases L-type Ca2+ currents and alters channel inactivation properties.

In skeletal muscle the oligomeric alpha(1S), alpha(2)/delta-1 or alpha(2)/delta-2, beta1, and gamma1 L-type Ca(2+) channel or dihydropyridine receptor functions as a voltage sensor for excitation contraction coupling and is responsible for the L-type Ca(2+) current. The gamma1 subunit, which is tightly associated with this Ca(2+) channel, is a membrane-spanning protein exclusively expressed in skeletal muscle. Previously, heterologous expression studies revealed that gamma1 might modulate Ca(2+) currents expressed by the pore subunit found in heart, alpha(1C), shifting steady state inactivation, and increasing current amplitude. To determine the role of gamma1 assembled with the skeletal subunit composition in vivo, we used gene targeting to establish a mouse model, in which gamma1 expression is eliminated. Comparing litter-matched mice with control mice, we found that, in contrast to heterologous expression studies, the loss of gamma1 significantly increased the amplitude of peak dihydropyridine-sensitive I(Ca) in isolated myotubes. Whereas the activation kinetics of the current remained unchanged, inactivation of the current was slowed in gamma1-deficient myotubes and, correspondingly, steady state inactivation of I(Ca) was shifted to more positive membrane potentials. These results indicate that gamma1 decreases the amount of Ca(2+) entry during stimulation of skeletal muscle.

High voltage activated Ca 2ϩ channels are oligomeric protein complexes composed of the ion conducting ␣ 1 protein that is tightly associated with the auxiliary subunits ␣ 2 /␦, ␤, and ␥. Each of the subunits is encoded by a separate gene, selected from seven ␣ 1 , three ␣ 2 /␦, four ␤, and two ␥ genes, some of which exist as splice variants. Various ␣ 1 proteins define different types of Ca 2ϩ channels, which differ in current properties, pharmacology, and G-protein-dependent modulation. In skeletal muscle, the high voltage activated L-type Ca 2ϩ channel or dihydropyridine receptor functions as voltage sensor for excitation contraction coupling and is responsible for the Ltype Ca 2ϩ current present in this tissue (1,2). The purified skeletal muscle L-type Ca 2ϩ channel is a heterooligomeric complex of the ␣ 1S , ␤1, ␣ 2 /␦-1, or ␣ 2 /␦-2 and ␥1 subunits (3)(4)(5). The ␥1 subunit is a membrane-spanning protein (6,7) encoded by a single-copy gene consisting of four translated exons (8,9). A similar organization has been described for the gene of a novel ␥ subunit, ␥2 or stargazin (10), which is the target of the stargazer mutation in mice. Whereas ␥2 mRNA is expressed in adult mouse brain but not in other mouse tissues like heart, kidney, or skeletal muscle (10), the ␥1 subunit is exclusively expressed in skeletal muscle (6,7,9,11).
Up to now, functional data on the ␥1 subunit were only obtained in expression systems. Coexpression of ␥1 with ␣ 1S , ␤1, and ␣ 2 /␦-1 in Xenopus oocytes (12) and L-cells (13) did not reveal significant effects on Ca 2ϩ currents. Due to low expression of the ␣ 1S skeletal muscle Ca 2ϩ channel subunit (12,14,15), the role of ␥1 has been studied by heterologous coexpression with the cardiac ␣ 1C subunit in Xenopus oocytes (16,17) and human embryonic kidney (293) cells (11,18). In these studies, coexpression of the ␥1 subunit shifted the steady state inactivation of ␣ 1C -induced I Ba to more negative membrane potentials, accelerated current activation and inactivation, and, in one study (17), increased peak currents. Considering the discrepancies obtained through coexpression of ␥1 with either ␣ 1C or ␣ 1S and the considerable structural and functional diversity of ␣ 1S and ␣ 1C (1,19), it does not appear feasible to simply transfer the results obtained with ␣ 1C to the "correct" subunit ␣ 1S .
To determine directly the role of the ␥1 subunit in L-type channel activity in skeletal muscle in vivo, we therefore used gene targeting to inactivate the ␥1 gene. In the absence of the ␥1 subunit, peak Ca 2ϩ current amplitudes in skeletal muscle myotubes were increased and the steady state inactivation was shifted to more positive potentials. These results demonstrate that a specific modulatory function of the skeletal muscle Ltype Ca 2ϩ channel can be assigned to the ␥1 subunit, which contributes to tissue-specific differences between high voltage activated L-type Ca 2ϩ channels.

EXPERIMENTAL PROCEDURES
Gene Targeting and Generation of ␥1 Mutant Mice-␥1 DNA was isolated from a 129 SvJ murine genomic library (Genome Systems, St. Louis, MO), and the structure of the murine ␥1 gene has been characterized (9). The replacement type targeting vector was constructed as follows. The multiple cloning site of the pBluescript-SK vector (Stratagene) was replaced by the multiple cloning sites 5Ј-CGC GCA ATC  GAT TTG GAT CCA ACT CGA GAT CCA TGG CCG TCG ACG TGC  GGC CGC AAC CGC GGA GCG CG-3Ј (plasmid pCON1) and 5Ј-CGC  GCA AGC GGC CGC AAG AGC TCC CGT CGA CCT GGT ACC TGA  AGC TTA TCC GCG GAG CGC G-3Ј (plasmid pCON2). To obtain the vector pKO3, the 5.1-kb 1 BamHI/NcoI fragment, representing the 5Ј non-translated region of the ␥1 gene and nucleotides 1-73 encoding the * This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. 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. initial 24 amino acid residues of ␥1, and the 1.8-kb SalI/NotI fragment from pGEM7 (KJ1) (20) comprising the neomycin resistance gene, were subcloned in pCON1. The 3.75-kb SalI/Asp718 fragment containing intron sequence 3Ј of exon 1 of the ␥1 gene and the herpes simplex thymidine kinase cassette from pNTK (21) were subcloned in pCON2 to yield pKO5. The targeting vector pKO␥ was obtained by subcloning the 5.9-kb NotI/SstII fragment from pKO5 into the pKO3 vector 3Ј of the neomycin resistance gene. About 1.1 ϫ 10 7 R1 embryonic stem (ES) cells (22) were electroporated with ClaI-linearized pKO␥ (25 g; setting of 800 V and 3 F, Bio-Rad Gene Pulser) and plated on irradiated G418-resistant embryonic feeder cells (23). Recombinant clones were selected with G418 (0.5 mg/ml) and 1-(2Ј-deoxy-2Ј-fluoro-␤-D-arabinofuranosyl-)-5-iodouracil (0.2 M), and 11 out of 334 double-resistant colonies showed correct homologous recombination with the targeting vector as confirmed by screening with both the 5Ј-and 3Ј-probes. Germline chimeras were obtained by injecting mutant ES cells of the clone 212 into C57Bl/6 blastocysts and transferring the blastocysts into the uteri of pseudopregnant recipients. Chimeric mice were mated with C57Bl/6 females to test for germline transmission or with 129/SvJ females to obtain inbred lines carrying the mutated ␥1 gene.
Genotyping of ES Cells and Mice-For Southern blot analysis genomic DNA from ES cells or mouse tails (24) were digested with EcoRV or SstI, respectively, electrophoresed on a 0,6% agarose gel, and transferred to a nylon membrane (Hybond-N). Hybridizations were carried out with 32 P-labeled 5Ј-and 3Ј-probes ( Fig. 1A) for 12 h at 62°C as described (25). After hybridizations, the blots were washed in 2ϫ SSC, 1% (w/v) SDS for 15 min at 65°C, followed by a second wash in 0.4ϫ SSC, 1% (w/v) SDS at 65-70°C for 30 min, and exposed to x-ray films at Ϫ80°C for 2 days. A polymerase chain reaction-based assay was developed for rapid offspring genotyping using primer pairs to produce 370-and 570-bp fragments, representing the wild-type and the mutant allele, respectively.
Northern Blot and Immunoblot Analysis-Northern blot analysis was carried out as described (26) using a polymerase chain reactionamplified 437-bp fragment consisting of the ␥1 cDNA (9) 3Ј of the deleted sequence of exon 1. Preparation of skeletal muscle microsomes and immunoblot analysis were performed as described (9).
Cell Culture-Primary cultures of mouse skeletal myotubes were prepared essentially as described before (27) with minor modifications. Briefly, after decapitation of neonatal mice and removal of skin from fore-and hindlimbs, muscles were minced (ϳ350 m) using a McIlwain tissue chopper. The tissue was then digested at 37°C for 30 -40 min in Rodent Ringer solution (in mM): NaCl 146, KCl 5, CaCl 2 2, MgCl 2 1, glucose 11, HEPES 10, pH 7.4) with added 2 mg/ml collagenase I (Sigma). The tissue was pressed through a 177-m mesh to remove clumps of tissue. After addition of an equal amount of Rodent Ringer solution, cells were pelleted by centrifugation at 200 ϫ g for 10 min. Cells were resuspended in DMEM (Life Technologies, Inc.) with 10% fetal calf serum (Life Technologies, Inc.), 10% horse serum (Life Technologies, Inc.), and 100 IU/ml penicillin and 100 g/ml streptomycin (Sigma). To reduce the number of fibroblasts in the culture, cells were preplated for 1-1.5 h before finally plating them in 35-mm plastic Petri dishes (Falcon). The next day, the medium was replaced, and the following day, replaced by DMEM with 10% horse serum, 100 IU/ml penicillin, and 100 g/ml streptomycin. Myotubes were used between 4 and 9 days in vitro.
Electrophysiology-Ca 2ϩ currents were recorded with the whole-cell patch-clamp configuration of the patch-clamp technique (28) at room temperature. Patch pipettes were fabricated from Kimax-51 borosilicate glass (inner diameter, 1.5 mm; outer diameter, 1.8 mm). After heat-polishing, pipette resistances varied between 1.5 and 3 megohms with the solution used. For recording, cells were bathed in (in mM): tetra methylammonium-chloride 146, CaCl 2 10, MgCl 2 1, glucose 10, HEPES 10, adjusted to pH 7.4 with CsOH. The pipette solution contained (in mM): cesium aspartate 145, MgCl 2 5, EGTA 20, Mg-ATP 5, HEPES 10, adjusted to pH 7.2 with CsOH. Currents were recorded with a EPC9 patch-clamp amplifier (HEKA, Lambrecht, Germany), filtered at 1.5 kHz, and digitized at 5 kHz. A 10-mV liquid junction potential correction was applied to all voltages. Cell capacitance and series resistance were cancelled electronically prior to each pulse. To determine currentvoltage (IV) relationships, Ca 2ϩ currents were activated every 5 s by 200-ms step depolarization from a holding potential of Ϫ90 mV to potentials between Ϫ70 and ϩ50 mV in 10-mV increments. L-type Ca 2ϩ currents were measured at the end of each depolarizing pulse. For measuring steady state inactivation, the cell was depolarized for 5 s to potentials from Ϫ100 mV to 0 mV (20-mV increments) and subsequently to potentials from Ϫ40 mV and ϩ15 mV (5-mV increments). Ca 2ϩ current amplitude was then measured at the end of a 200-ms test pulse to ϩ20 mV. For linear leak and capacitance subtraction, a P/4 protocol (leak holding potential Ϫ120 mV) was used in all measurements preceding the test pulse. Individual IV curves were fitted with a modified Boltzman equation of the form (29) with g max being the maximal whole-cell conductance, E rev the reversal potential, V 1 ⁄2 the voltage for half-maximal activation, and k the slope factor. For comparison of the activation kinetics of the L-type Ca 2ϩ channel current in ␥1ϩ/ϩ cells and ␥1Ϫ/Ϫ cells, current traces at a depolarization of ϩ20 mV were averaged and fitted with a double exponential (normalized to the beginning of the test pulse). To compare steady state inactivation of ␥1ϩ/ϩ cells and ␥1Ϫ/Ϫ cells, current amplitudes at the end of the 20-mV test pulse were normalized to the current at the end of the 20-mV test pulse following the prepulse to Ϫ100 mV. The data were fitted with a Boltzman equation of the form I/I max ϭ 1/(1 ϩ exp((V Ϫ V 1 ⁄2)/k), with V 1 ⁄2 being the voltage of half-inactivation and k a measure for the steepness. In some experiments, cells were depolarized to ϩ20 mV for 350 ms every 5 s. After a control recording of approximately 60 -80 s when the current had reached a steady amplitude, (Ϫ)-BayK 8644 (1, 3, 5, or 5.6 M) (a gift from Bayer, Germany, or purchased from RBI) was added into the bath to stimulate the L-type Ca 2ϩ channels. Currents were analyzed at the end of the pulse. Cell-attached single-channel recordings were carried out to compare the single-channel conductance of L-type Ca 2ϩ channels in wild-type and ␥1-deficient cells. To obtain the single-channel conductance, Ca 2ϩ channels were activated by voltage ramps from Ϫ100 mV to ϩ100 mV and returned to Ϫ100 mV at a rate of 1 mV/ms before returning to the holding potential (Ϫ90 mV or Ϫ40 mV). Patches were depolarized every 5 s. Current traces were filtered off-line at 1 kHz before leak subtraction in IGOR Pro, WaveMetrics, Inc. Leak subtraction was carried out by fitting a polynomial to a blank sweep or averaging several blank sweeps, which were then subtracted from the sweep containing channel activity. A linear fit for the prolonged channel openings was used to obtain the single-channel conductance.
Isradipine Binding Assays-Hindlimb muscles were dissected from wild-type and mutant mice, immediately homogenized in 20 volumes of 50 mM Tris-HCl, 150 mM NaCl, pH 7.4, containing 0.1 mM phenylmethylsulfonyl fluoride, 1 mM phenanthroline, 1 mM iodoacetamide, 1 mM benzamidine, 1 M pepstatin A, 1 g/ml antipain, and 1 g/ml leupeptin (protease inhibitor mix) and centrifuged at 6000 ϫ g for 15 min at 4°C. The pellet was rehomogenized in the same buffer and centrifuged twice. The combined supernatants were then centrifuged at 42,000 ϫ g for 40 min at 4°C. The pellet containing microsomal membranes was rehomogenized in 50 mM Tris-HCl, 150 mM NaCl, pH 7.4, containing the protease inhibitor mix and used for equilibrium binding of (ϩ)[ 3 H]isradipine ((ϩ)-[methyl-3 H]PN 200-110, 82 Ci/mmol; Amersham Pharmacia Biotech). Assays were performed in duplicate using microsomal membranes obtained from two independent preparations for both muscle from wild-type mice and and ␥1Ϫ/Ϫ mice, respectively. Microsomes (15-23 g of microsomal protein) were incubated in 250 l of 50 mM Tris-HCl, pH 7.4, 1 mM CaCl 2 , 0.5 mM MgCl 2 in the presence of 0.035-9.81 nM [ 3 H]isradipine and in the presence or absence of 1 M (Ϯ)isradipine. Incubation was stopped after 90 min at 4°C by the addition of 3 ml of ice-cold solution A (8.5% (mass/volume) polyethylene glycol 6000, 100 mM HEPES, pH 7.4). The suspension was immediately poured on GF/C filters. After washing twice with solution A, the filter was subjected to liquid scintillation counting. Protein concentrations were determined using the BCA method (Pierce).

Targeted Disruption of the Skeletal Muscle Calcium Channel
␥1 Subunit Gene-One allele of the ␥1 gene was mutated by replacing the 3Ј-part of its first translated exon and the adjacent splice donor site (Fig. 1A) with the neomycin resistance gene (neo r ) in murine embryonic stem cells. Chimeric males were obtained that transmitted the mutant allele to their progeny. These heterozygous F1 offspring were intercrossed to produce F2 homozygous ␥1Ϫ/Ϫ mice, as confirmed by Southern blot analysis of tail DNA (Fig. 1B). The mutation was trans-mitted at Mendelian ratio, suggesting normal fetal and embryonic development of homozygous mutant mice. ␥1Ϫ/Ϫ mice grew and reproduced normally and were indistinguishable from their wild-type littermates. The deletion of the ␥1 gene and the lack of expression was confirmed by Northern and Western blot analysis (Fig. 1, C and D). No transcripts could be detected in poly(A) ϩ RNA prepared from ␥1Ϫ/Ϫ skeletal muscle using as probes either the mouse ␥1 cDNA covering the translated exons 2, 3, and 4 ( Fig. 1C) or the 160-bp fragment of exon 1 which has been replaced by neo r (data not shown). Using polymerase chain reactions, ␥1 transcripts encoding the complete ␥1 protein could be amplified from ␥1ϩ/ϩ and ␥1ϩ/Ϫ poly(A) ϩ RNA but not from poly(A) ϩ RNA isolated from ␥1Ϫ/Ϫ skeletal muscle. In addition, the DNA fragments encoding exon 2 or exons 3 and 4 could be amplified from ␥1ϩ/ϩ and ␥1ϩ/Ϫ poly(A) ϩ RNA but not from ␥1Ϫ/Ϫ poly(A) ϩ RNA (data not shown), indicating that the replacement of part of exon 1 and of the adjacent intron completely prevents expression of ␥1-specific transcripts. Accordingly, the ␥1 protein (ϳ32 kDa) is recognized by a ␥1 subunit-specific polyclonal antibody (9) in skeletal muscle microsomes prepared from ␥1ϩ/ϩ mice and ␥1ϩ/Ϫ mice but not in microsomes prepared from ␥1Ϫ/Ϫ mice (Fig.  1D). No transcripts of the neuronal ␥2 subunit (10) could be detected in skeletal muscle RNA from wild-type or ␥1-deficient mice.
Functional Characterization of Calcium Channels in ␥1-deficient Myotubes-To study the function of the ␥1 subunit in skeletal muscle, primary cultures were obtained from fore-and hindlimb muscles of 1-3-day-old wild-type and ␥1Ϫ/Ϫ littermatched mice. Both ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes exhibited spontaneous contractions after day 4 in culture and were used for electrophysiological recordings up to 9 days. Fig. 2 shows representative recordings of Ca 2ϩ channel activity at different test potentials from ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes of littermatched mice. At positive test potentials, a large sustained L-type inward Ca 2ϩ current was observed in ␥1ϩ/ϩ and ␥1Ϫ/Ϫ cells ( Fig. 2A). In all cells, the activation threshold was around Ϫ20 mV and the peak amplitude of the current was at ϩ20 mV (Fig. 2B). In addition to L-type currents, additional transient Ca 2ϩ currents at test potential ϾϪ60 mV (present at Ϫ20 mV in Fig. 2A, right panel) were observed representing T-type Ca 2ϩ channels (27). This channel type was not consistently observed in all cells but, when present, was not obviously altered by the ␥1 subunit deletion (data not shown). The peak whole-cell current through L-type Ca 2ϩ channels was found to be significantly larger in ␥1Ϫ/Ϫ cells compared with ␥1ϩ/ϩ cells, as evident from the representative examples and the averaged EcoRV fragments obtained from wild-type (10.0 kb) and targeted alleles (8.2 kb) were detected by the 5Ј probe. Using the 3Ј probe SstI fragments of 6.5 and 5.2 kb representing wild-type and mutated alleles, respectively, were detected. Absence of additional random integrations of the targeting construct was checked by hybridization with a cDNA probe representing a 634-bp PstI fragment of the neo r cassette (data not shown). C, Northern blot analysis of poly(A) ϩ RNA (9 g), extracted from skeletal muscle of ␥1ϩ/ϩ, ␥1ϩ/Ϫ, and ␥1Ϫ/Ϫ mice. The blot was hybridized with the nucleotide 236 -672 cDNA fragment encoding the translated exons 2, 3 and 4 of the murine ␥1; lower panel, hybridization of the same filter with a human glyceraldehyde-3-phosphate dehydrogenase cDNA probe. D, immunoblot analysis of ␥1 expression in skeletal muscle microsomes (70 g) from ␥1ϩ/ϩ, ␥1ϩ/Ϫ, and ␥1Ϫ/Ϫ mice. current-voltage (IV) relationship (Fig. 2, A and B). Individual IV curves were fitted with a Boltzman equation, resulting in a significantly different whole-cell conductance g of 0.45 Ϯ 0.07 nS/pF for ␥1Ϫ/Ϫ cells compared with 0.37 Ϯ 0.05 nS/pF for ␥1ϩ/ϩ cells (n ϭ 11, p Ͻ 0.05). This effect is not due to a difference of cell cultivation length, as ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes were prepared on the same day and measured in a paired fashion. Furthermore, expression of L-type Ca 2ϩ currents did not vary between days 5 and 9 in ␥1-deficient or wild-type cells (data not shown). To look for differences between ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes concerning selectivity, gating properties, and activation kinetics of L-type Ca 2ϩ channels, we analyzed the experiments in more detail. The apparent reversal potential E rev (65.7 Ϯ 2.2 mV for ␥1ϩ/ϩ and 66.3 Ϯ 1.8 mV for ␥1Ϫ/Ϫ) as well as the half-maximal voltage for activation V 1 ⁄2 (8 Ϯ 1 mV for ␥1ϩ/ϩ and 7 Ϯ 1 mV for ␥1Ϫ/Ϫ) and the slope factor for activation k (5.8 Ϯ 0.4 mV for ␥1ϩ/ϩ and 6.2 Ϯ 0.3 mV for ␥1Ϫ/Ϫ) remained unchanged. Activation kinetics of the L-type Ca 2ϩ current were also not altered by the elimination of the ␥1 subunit as shown in Fig. 2C. Depicted are averaged (n ϭ 11 for ␥1ϩ/ϩ and n ϭ 13 for ␥1Ϫ/Ϫ), normalized current traces (test pulse to ϩ20 mV) from ␥1-deficient and wild-type mice fitted with a double exponential function (time constants of 67.9 and 4.2 ms for ␥1ϩ/ϩ and 58.1 and 5.7 ms for ␥1Ϫ/Ϫ). Together, these measurements indicate that Ca 2ϩ selectivity, gating, and activation kinetics of L-type Ca 2ϩ channels in skeletal muscle are not significantly influenced by the ␥1 subunit.
The increased L-type Ca 2ϩ current amplitude can in principle be explained in one of three ways. 1) The single-channel conductance is increased.
2) The open probability is increased.
3) The number of functional Ca 2ϩ channels in the plasma membrane is increased. Because the open probability of skeletal muscle L-type Ca 2ϩ channels is rather low and open times can be extremely short, we used the calcium channel agonist (Ϫ)-BayK 8644 to increase open times of the channels.
(Ϫ)-BayK 8644 (at 1 M) did increase the peak current amplitude of L-type currents at ϩ20 mV by 64% in both ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes (␥1ϩ/ϩ: from Ϫ13.4 Ϯ 1.6 to Ϫ22.9 Ϯ 2.3 pA/pF, n ϭ 9; ␥1Ϫ/Ϫ: from Ϫ22.1 Ϯ 2.1 to Ϫ39.6 Ϯ 4.7 pA/pF, n ϭ 9), indicating a comparable sensitivity of native and ␥1deficient skeletal calcium channels toward dihydropyridines. The effect of (Ϫ)-BayK 8644 was already saturated at this concentration, as higher concentrations (between 3 and 5.6 M) did not increase the peak current any further in either ␥1ϩ/ϩ or ␥1Ϫ/Ϫ myotubes (n ϭ 22). Thus, the relative difference of current amplitude of L-type channels between ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes was not changed by application of (Ϫ)-BayK 8644. Therefore, we were able to compare single-channel properties in the presence of (Ϫ)-BayK 8644. To measure singlechannel conductances, a voltage ramp protocol as shown in Fig.  3A was used. Typical examples of current responses from ␥1ϩ/ϩ and ␥1Ϫ/Ϫ myotubes with no single-channel activity and with channel activity are overlaid in Fig. 3B. Subtracting blank traces from a trace with channel activity, we could fit a line to the current traces shown in Fig. 3C to get the singlechannel conductance of each experiment. On average, there was no significant difference between recordings form ␥1ϩ/ϩ and ␥1Ϫ/Ϫ cells (␥1ϩ/ϩ: 13.0 Ϯ 0.8 pS, n ϭ 4; ␥1Ϫ/Ϫ: 12.3 Ϯ 0.3 pS, n ϭ 3). These values are very similar to the ones (14.5 pS) from a previous study measured under slightly different conditions (30). Next, we tested whether a difference in the amount of ␣ 1S protein in ␥1ϩ/ϩ and ␥1Ϫ/Ϫ skeletal muscle might be responsible for the difference in current densities by binding studies using the dihydropyridine isradipine as a ligand. However, the densities of isradipine binding sites were not significantly different in ␥1ϩ/ϩ and ␥1Ϫ/Ϫ muscles (␥1ϩ/ϩ: B max , 1150 Ϯ 192 fmol/mg protein; K D , 0.53 Ϯ 0.04 nM; n ϭ 3; ␥1Ϫ/Ϫ: B max , 1236 Ϯ 175 fmol/mg protein; K D , 0.51 Ϯ 0.02 nM; n ϭ 3), implicating that no significant differences in the amount of functional ␣ 1S protein exist. Because both singlechannel conductance and amount of ␣ 1S protein do not account for the difference in current densities, we conclude that loss of the ␥1 subunit increases the open probability of L-type Ca 2ϩ channels in skeletal muscle.
To compare inactivation kinetics of L-type Ca 2ϩ channels, long depolarizations to test potentials between Ϫ20 and ϩ20 mV were applied to wild-type and ␥1-deficient cells. In the absence of the ␥1 subunit, L-type Ca 2ϩ currents inactivated more slowly than in wild-type myotubes from litter-matched control mice, as illustrated by the representative current traces in Fig. 4A. At the end of the 5-s depolarization to Ϫ5 mV, 60.0 Ϯ 5.8% of the current was inactivated in wild-type cells (n ϭ 11) compared with only 48.4 Ϯ 4.7% in ␥1Ϫ/Ϫ cells (n ϭ 17). The effect on time-dependent inactivation was most noticeable at negative test potentials (Fig. 4A). At depolarizations more positive to ϩ10 mV, the difference in inactivation following 5-slong depolarizations was no longer observed. Because L-type Ca 2ϩ currents at negative test potentials are very small due to the low open probability, we used a prepulse protocol followed by a test pulse to investigate the inactivation properties in more detail. Fig. 4B shows representative current traces from wild-type (left) and ␥1Ϫ/Ϫ cells (right) depolarized to ϩ20 mV following a 5-s-long prepulse to varying potentials (of which only the last few milliseconds are seen). As shown in Fig. 2, current amplitudes in ␥1Ϫ/Ϫ cells are larger than the ones in ␥1ϩ/ϩ cells. When compared with the test potential of Ϫ100 mV, it is obvious that steady state inactivation of L-type channels in ␥1Ϫ/Ϫ cells occurs at more depolarized potentials compared with ␥1ϩ/ϩ cells (Fig. 4B). This finding is summarized in Fig.  4C, in which normalized steady state inactivation is plotted as a function of the prepulse potential. The inactivation curve of Ca 2ϩ currents from ␥1Ϫ/Ϫ cells is shifted to more positive potentials and is also steeper than that from ␥1ϩ/ϩ cells. This behavior is also reflected in the slope factor k and half maximal-inactivation potential V 1 ⁄2, which were found to be significantly different for ␥1-deficient and wild-type cells (k ϭ 5.7 mV, V 1 ⁄2 ϭ 0.9 mV for ␥1Ϫ/Ϫ cells (n ϭ 12); k ϭ 10.4 mV, V 1 ⁄2 ϭ Ϫ7.7 mV for ␥1ϩ/ϩ cells (n ϭ 10); p Ͻ 0.05). DISCUSSION We used gene targeting to generate mice that lack the ␥1 subunit of the multisubunit skeletal muscle L-type Ca 2ϩ channel. In ␥1-deficient myotubes, (i) the peak amplitudes of L-type Ca 2ϩ currents are increased, (ii) the time-dependent inactivation of I Ca is decelerated, and (iii) the steady state inactivation curve is shifted to more positive potentials. As a result, Ca 2ϩ influx through L-type Ca 2ϩ channels is increased in ␥1-deficient skeletal myotubes.
The changes of inactivation properties of I Ca due to the inactivation of the ␥1 gene corresponds to previous results obtained after coexpression of ␥1 with the cardiac L-type Ca 2ϩ channel complex ␣ 1C ␤2(or ␤1)␣ 2 ␦-1 (11,16,17,18). In these studies, the major effects of ␥1 were to accelerate current inactivation and to shift the steady state inactivation curve to more negative potentials. The latter effect may be due to the addition of negative charges into the vicinity of the external mouth of the channel. The ␥1 subunit is membrane-spanning protein (6,7), and the extracellular loop between the predicted membrane-spanning segments 1 and 2 is especially rich in negative charges. This loop may be close to the voltage sensing parts of the native ␣ 1S in skeletal muscle but also of the recombinant cardiac ␣ 1C . Interestingly, this negatively charged extracellular loop region is almost completely conserved among the ␥1 proteins of various species (9) and corresponds to a similarly charged region of the neuronal ␥2 protein (10). Accordingly, in neurons of mutant mice that exhibit a mutation in the ␥2 gene, enhanced Ca 2ϩ entry through voltage-activated Ca 2ϩ channels due to reduced steady state inactivation of Ca 2ϩ currents has been implicated (10,31).
In contrast to the modulatory role of ␥1 on current inactivation, the observed increase of peak Ca 2ϩ current amplitude in ␥1-deficient myotubes has not been anticipated from the results of previous expression studies. In those studies, ␥1 did not have a consistent effect on the amplitude of I Ba through recombinant cardiac L-type Ca 2ϩ channels (11,16,18) or even increased the measured currents (17). Obviously, on native skeletal muscle Ca 2ϩ channel activity, ␥1 subunits act as a brake and Ca 2ϩ channel activity increases in their absence and leads to enhanced Ca 2ϩ influx in response to depolarization. This increased Ca 2ϩ entry could increase the store of Ca 2ϩ for release from the sarcoplasmic reticulum or slow the Ca 2ϩ release process by changing the sensitivity of the ryanodine receptor toward activation (32,33).
In skeletal muscle, single depolarizations simultaneously activate only a small fraction (Ͻ5%) of the L-type Ca 2ϩ channels (34), and during single twitches, contraction is not dependent on extracellular Ca 2ϩ (35). In contrast, repetitive depolarizations at high frequency substantially increase Ca 2ϩ influx through skeletal muscle L-type channels (36 -38) and, subsequently, contractile force of muscle fibers may be increased. This Ca 2ϩ channel potentiation has been shown to be due to phosphorylation by cAMP-dependent protein kinase at positive membrane potentials and results in a negative shift in the voltage dependence of channel activation and slowing of channel deactivation (39). The ␥1 subunit, which dominates the inactivation process of I Ca , making it faster and more sensitive to voltage, might fine-tune this process.
In summary, deletion of the ␥1 subunit of the skeletal muscle L-type Ca 2ϩ channel specifically alters channel activity. ␥1Ϫ/Ϫ mice are viable, and their phenotype differ from the phenotype seen in mice with mutations in the ␣ 1S (40,41) or ␤1 subunit (42) of the same multisubunit channel, muscular dysgenic (mdg) and ␤1Ϫ/Ϫ. The slow voltage activated L-type Ca 2ϩ current is absent or 10 -20-fold decreased in mdg mice (43) and ␤1Ϫ/Ϫ mice (44), respectively, and homozygous mutants lack excitation-contraction coupling (40,42). The availability, now, of a viable mutant that lacks the ␥1 subunit allows to study the functional impact of ␥1 on Ca 2ϩ channel activity and excitation-contraction coupling in parallel in embryonic skeletal muscle cells, in isolated muscle fibers and in the adult animal. FIG. 4. Steady state inactivation is shifted to more depolarized potentials in ␥1؊/؊ cells. A, representative current traces of cells depolarized for 5 s to Ϫ5 mV are shown, indicating that time-dependent inactivation is faster in wild-type cells than in ␥-deficient cells. The right panel shows the percentage decrease of current from the peak to the end of the pulse after 5 s (n ϭ 11 for ␥1ϩ/ϩ cells, n ϭ 17 for ␥1Ϫ/Ϫ cells). B, representative current traces during the 200-ms step depolarization to ϩ20 mV following the indicated prepulse potential for a ␥1ϩ/ϩ cell (left) and a ␥1Ϫ/Ϫ cell (right). C, steady state inactivation (normalized to Ϫ100 mV prepulse potential) is plotted as a function of the respective prepulse potential for ␥1ϩ/ϩ cells (n ϭ 10) and ␥1Ϫ/Ϫ cells (n ϭ 12) prepared from litter-matched mice. Data are fitted with a Boltzman function.