Targeted disruption of the voltage-dependent calcium channel α₂/δ-1-subunit

Abstract

Cardiac L-type voltage-dependent Ca²⁺ channels are heteromultimeric polypeptide complexes of α₁-, α₂/δ-, and β-subunits. The α₂/δ-1-subunit possesses a stereoselective, high-affinity binding site for gabapentin, widely used to treat epilepsy and postherpetic neuralgic pain as well as sleep disorders. Mutations in α₂/δ-subunits of voltage-dependent Ca²⁺ channels have been associated with different diseases, including epilepsy. Multiple heterologous coexpression systems have been used to study the effects of the deletion of the α₂/δ-1-subunit, but attempts at a conventional knockout animal model have been ineffective. We report the development of a viable conventional knockout mouse using a construct targeting exon 2 of α₂/δ-1. While the deletion of the subunit is not lethal, these animals lack high-affinity gabapentin binding sites and demonstrate a significantly decreased basal myocardial contractility and relaxation and a decreased L-type Ca²⁺ current peak current amplitude. This is a novel model for studying the function of the α₂/δ-1-subunit and will be of importance in the development of new pharmacological therapies.

cardiac L-type voltage-dependent Ca²⁺ channels (L-VDCCs) are heteromultimeric polypeptide complexes of α₁-, α₂/δ-, and β-subunits. The α₁-subunit is autoregulatory and harbors the channel pore, gating machinery, and modulatory drug binding sites (30). The accessory subunits (α₂/δ and β) affect channel kinetics and are involved in the trafficking and insertion of the α₁-subunit into the membrane. The α₂-subunit is closely associated with an extracellular loop of the α₁-subunit (15) and linked to a small protein called δ (2, 9). Both the α₂ and δ are encoded by the same gene, separated by proteolytic cleavage, and extracellularly linked through a disulfide bridge (9). Currently, four α₂/δ-subunits, each encoded by separate genes, have been identified (4). The α₂/δ-1, originally cloned from skeletal muscle (10), is ubiquitously distributed (18), with high levels of protein expression in brain, heart, skeletal, and smooth muscle (13). It possesses a stereoselective, high-affinity binding site for gabapentin (GBP) (1-aminomethylcyclohexane acetic acid; Neurontin), widely used to treat epilepsy, postherpetic neuralgic pain, and sleep disorders (8, 13). Of the remaining α₂/δ-subunits, only α₂/δ-2 has also shown a propensity for GBP binding, but the affinity is much lower than α₂/δ-1 (18, 24).

It is clear that the α₂/δ-subunits of VDCCs are associated with some disease states. A spontaneous mutation in the α₂/δ-2 murine gene resulted in the Ducky mouse, demonstrating clinical signs of epilepsy (1, 3). Similar results to those obtained for the Ducky mouse were found in an α₂/δ-2 knockout mouse (17). These results included ataxic gait, seizures, signs of cerebellar degeneration, and premature death. There were some differences noted in the pathology of the cerebellar region for these two mice. There were no pathological abnormalities noted in the hearts of the α₂/δ-2 null mice. In conscious mice, there was a trend toward bradycardia observed during ECG and transthoracic echo. However, while under the influence of isoflurane, the null mice showed a significantly less-pronounced decrease in heart rate. A mutation in the α₂/δ-4 gene introduced a premature stop codon that truncated one third of the protein and led to a cone-rod dysfunction in the visual system of mice (35). In addition, a mutation of the α₂/δ gene (unc-36) in Caenorhabditis elegans has been identified, manifesting in a lack of adaptation (desensitization) to dopamine and serotonin, a phenotype indistinguishable from the unc-2 phenotype that is caused by lack of the α₁-subunit (25). Localization analysis of a gene encoding the polymorphous α₂/δ-1-subunit to chromosome 7q provided some evidence of the segregation of flanking markers in families susceptible to malignant hyperthermia (16). The gene was localized to chromosome 5 in the mouse in an earlier study (7). In addition, the α₂/δ-1-subunit has been shown to be upregulated in response to spinal injury and neuropathic pain (20, 36, 37). However, the effect of a knockout of the α₂/δ-1-subunit in an in vivo model has not been accomplished.

We have developed a viable conventional α₂/δ-1 knockout mouse using a construct targeting exon 2 of α₂/δ-1. As we expected, the α₂/δ-1-subunit knockout mice display a lack of the high-affinity GBP binding site and altered L-type Ca²⁺ current in cardiomyocytes, without causing a significant change in the expression of other Ca²⁺ channel subunits.

METHODS

The investigation conforms to the Guide for the Care and Use of Laboratory Animals, published by The National Institutes of Health. The animals in the present study were handled according to approved protocols and animal care regulations at the University of Cincinnati. With the exception of the mice used for genotyping, the animals used for experimentation, including binding, Western blot analysis, histopathology, ex vivo analysis, and electrophysiology were all between 2 and 8 mo of age.

Generation of the α₂/δ-1 knockout mouse.

A 12-Kb genomic fragment containing exon 2 of the α₂/δ-1 gene encoding residues I33-D59 [82 base pairs (bp)] was isolated and sequenced from a lambda mouse genomic DNA library (129 SVJ, Stratagene; La Jolla, CA). Mutations were introduced into exon 2 to include XhoI and NotI sites. The XhoI site was inserted using the nucleotides from L44 and V45 and the NotI site used L47, A48, and K49. Similarly, a NcoI site was introduced into the long arm (5′; 4.0 Kb) and an NdeI site into the short arm (3′; 2.3 Kb). The fragment between sites NcoI and NdeI was placed in the vector thyamine kinase gene elegans (pMCTK), and the neomysin resistance gene was introduced into the XhoI-NotI of exon 2. Figure 1A represents a schematic of the gene-targeting construct. This construct was electroporated into murine embryonic stem cells. After positive (G-418) and negative selection (gancyclovir), genomic DNA was isolated from 288 embryonic stem cell lines. These cell lines were screened by PCR to identify the cell lines with an integration of the full-length targeting region of the knockout construct. Following the testing of 10 of the positive cell lines, one cell line was found in which site-specific targeting occurred. This positive line was microinjected into C57Bl/6 blastocysts and implanted into F1 (C57Bl/6X129J) pseudopregnant females. Two separate injections into blastocysts yielded 12 pups. From these two litters, seven chimeric animals were identified (5 of them at the level of 100% chimerism and 2 at 80% chimerism). The 100% chimeras were bred with Black Swiss females. The resulting offspring were genotyped via PCR to determine which male chimeras had gametes capable of passing the mutant allele to progeny. Screening of pups using PCR genotyping confirmed heterozygous (+/−) and null (−/−) animals.

Fig. 1.

Conventional knockout (KO) of the α₂/δ-1 gene. A: schematic of the α₂/δ-1 ablation construct, the location of the α₂/δ-1 gene (chromosome 5), and the resultant homologous recombination. B: diagram of the PCR for confirmation of the α₂/δ-1 gene-targeting event. Approximate primer locations are depicted by small black lines and either P1, P2, or P3. The 15 bp (V45–K49) of the wild-type (WT) that was subsequently deleted in the targeted allele is located at the line indicating P2. C: agarose gel electrophoresis of the PCR products. The WT allele yields an 83-bp product, and the targeted allele yields a 459-bp product. The PCR control consisted of a 1:1 mixture of WT and (−/−) tail DNA. D: diagram for the tail-prep PCR genotyping. The PCR primers are shown in their approximate location and are listed as A, B, and C. E: representative DNA agarose gel of the PCR genotyping results. The PCR products were 346 [WT and heterozygous (Het)] and 635 (Het and KO) bp. pMCTK, thymidine kinase gene cassette; Ctr, control; Ex2, exon 2; Neo, neomysin; M, marker.

PCR genotyping for wild-type, +/−, and −/− α₂/δ-1 murine pups.

DNA was isolated from tail preps of 3-wk-old murine pups. The DNA was then added to a cocktail containing Taq Polymerase (NEB), Taq buffer, dNTPs and three primers: primer A (5′-CATGGGTGGACAAGATGCAAG-3′), primer B (5′-CTGCACGAGACTAGTGAGACG-3′), and primer C (5′-CATTCTCAAGACTGTAGG-3′). Using the first nucleotide of exon 2 as the initial bp, the primers were located as follows: primer A begins at 9, primer B begins at 1,448, and primer C begins at 2,065 (Fig. 1D). The PCR conditions were 35 cycles of 62°C annealing for 1 min, 72°C extension for 1 min, and 94°C denaturing for 1 min. PCR products were run on a 2.5% agarose gel. The expected bands were 346 bp for the wild-type (WT) and 635 bp for the −/− (Fig. 1E). The heterozygous mice had both bands.

Determination of the targeting construct integration and knockout of the α₂/δ-1-subunit.

To confirm that the targeting construct integrated at the homologous genomic site (i.e., exon 2 of the α₂/δ-1 VDCC subunit-encoding gene), we attempted Southern blot analysis using probes from flanking regions of the construct. However, four independent attempts at this failed, due to the highly repetitive nature of the DNA in the intronic sequences flanking exon 2. We then took advantage of the fact that 15 bp of sequence were deleted from the second exon (Fig. 1B, small line indicating P2) during construction of the gene-targeting construct. This deletion consists of the bp corresponding to amino acids V45–K49. We designed a PCR primer that is homologous to this region (Fig. 1B, P2, 5′-GTTTTTGCCAGTGTGACAAGG-3′). An additional 3′ primer was designed within the neomycin resistance cassette of the construct (P3, 5′-CGCCTTCTATCGCCTTCTTGAC-3′), and both were designed to be compatible with a common 5′ primer (P1, 5′-TGGAGCTTTCTTTCTTCTGATTCC-3′). These primers are located at −28 (P1), 34 (P2), and 403 (P3) with respect to the first nucleotide of exon 2. The PCR was designed such that the WT allele resulted in a product of 83 bp in length, whereas the targeted allele produced a 459-bp band (Fig. 1B). Importantly, P2 will not bind to the targeted allele, since 15 of its 21 bp are missing (3 bp 5′ and 3 bp 3′ of the deletion). However, if the targeting construct is integrated elsewhere in the murine genome, both the 453- and the 82-bp products would be generated. Thus this experimental procedure clearly demonstrates that the targeting construct successfully replaced exon 2 of the α₂/δ-1 gene.

GBP binding.

The [³H]GBP (American Radiolabeled Chemicals, St. Louis, MO) binding assay was performed as described by Gee et al. (12). Membranes (20 μg/ml) of mouse tissues were incubated with [³H]GBP at room temperature for 60 min. A separation of bound from free ligand was effected by filtration through 0.3% polyethyleneimine-soaked GF/B filters. The filters were washed with 3 × 4 ml of 10 mM HEPES (pH 7.4). Radioactivity retained by the filters was determined by scintillation counting. Nonspecific binding was defined in the presence of 100 μM unlabeled GBP.

Western blots for the Ca²⁺ channel subunits.

Total protein, isolated as described previously (14), was separated on SDS-PAGE gel, transferred to nitrocellulose, and analyzed with the following primary antibodies: voltage-dependent L-type Ca²⁺ channel (Ca_v1.2) (ACC-003, 1:1,000, Alomone), calsequestrin (PAI-913, 1:2,500, ABR), β_2a [1:2,500; generous gift from M. M. Hosey (6)], α₂/δ-1 (D219, 1:1,000; Sigma), α₂/δ-2 (Ac-REIYKDNRNLFEVQENEPC-amide, 1:500; custom-made antibody by Johnson & Johnson, not published), α₂/δ-3 [1:200; custom-made antibody by Johnson & Johnson (24)], α₂/δ-4 [1:100; custom-made antibody by Johnson & Johnson (24)], Ca_v2.1 (C1353, 1:200; Sigma), Ca_v2.2 (AB5154, 1:100; Millipore), β₃ (C1978, 1:500; Sigma), and the protein loading normalized to GAPDH (4300, 1:4,000; Ambion).

Histopathology.

Following anesthesia with ketamine (100 mg/kg)-xylazine (10 mg/kg), body weight (BW) measurements were taken. Hearts were removed via bilateral thoracotomy, weighed, flushed with PBS (pH 7.4), fixed in 10% buffered formalin, and processed for histopathology following routine methods.

Ex vivo analyses, working heart.

Control and α₂/δ-1 (−/−) age-matched mice of either sex were anesthetized with ketamine (100 mg/kg)-xylazine (10 mg/kg) and injected with 1.5 IU of heparin intraperitoneally to prevent intracardiac blood coagulation. Isolated hearts were placed in the working heart mode as previously reported (27). Positive and negative first derivatives of intraventricular pressure (dP/dt), duration of contraction (time to peak pressure), and relaxation (time to half relaxation) were calculated. Data are presented as means ± SD. Statistical analysis was carried out using Student's t-test for unpaired observations. Values of P < 0.05 were regarded as statistically significant.

Isolation of cardiomyocytes.

Single ventricular myocytes were enzymatically dissociated from isolated hearts of WT and α₂/δ-1 (−/−) age-matched mice of either sex, as previously reported (40), with modifications. In brief, the heart was rapidly excised and placed in Tyrode solution containing (in mM) 120 NaCl, 5.4 KCl, 1.2 NaH₂PO₄, 5.6 glucose, 20 NaHCO₃, 1.6 MgCl₂, 10 2,3-butanedione monoxime, and 5 taurine, infused with 95% O₂-5% CO₂. All solutions were filtered and equilibrated with 95% O₂-5% CO₂ for at least 20 min before use. The heart was perfused in retrograde mode for 4 to 5 min, followed by perfusion with buffer containing 1 mg/ml collagenase Type II (Worthington) at 37°C. After 2 min of enzyme perfusion, 50 μM CaCl₂ was added to the solution. After ∼5 min, when the heart was demonstrating characteristics of global digestion, the enzyme was recirculated. The heart was perfused for an additional 8–12 min or until flow rate surpassed preenzyme flow rate. After perfusion, the ventricles were separated from the atria and minced. Only Ca²⁺-tolerant cells with clear cross striations and without spontaneous contractions or significant granulation were selected for experimental studies.

Electrophysiological measurements.

The electrophysiological studies were performed as previously reported (22). The following protocol was used to obtain steady-state inactivation parameters of L-type Ca²⁺ current (I_Ca). From a holding potential of −80 mV, a series of 1-s prepulses (from −80 to +50 mV in 10-mV increments) were applied, followed by a 6-ms repolarization to the holding potential and a subsequent test depolarization to +10 (WT) or +20 mV [α₂/δ-1 (−/−)].

Peak currents at the test pulse were normalized to the peak current amplitude elicited after the −80-mV prepulse and plotted as a function of the prepulse voltage. All recordings were made after 3 min of membrane rupture at room temperature (22–24°C). These experiments were carried out in 1.8 mM Ca²⁺, using Ca²⁺ as the charge carrier.

Data analysis.

The individual activation data [current-voltage (I-V) curves] were fitted to standard Boltzmann equation in the form I_Ca = G_max(V_m − V_rev)/{1 + exp[−(V_m − V_0.5)/k]}, where G_max is the maximal conductance, V_m is the membrane voltage, V_rev is the I_Ca reversal potential, V_0.5 is the half-activation potential, and k is the slope factor. The obtained parameters of G_max and V_rev were then used to calculate fractional conductance at each V_m, G/G_max, using the equation: G/G_max = I_Ca[G_max(V_m − V_rev)], where G is the total macroscopic conductance at V_m. The G-V curves were plotted with the values obtained from the fit of the I-V curves, using the following form of Boltzmann equation: G/G_max = 1/{1 + exp[−(V_m − V_0.5)/k]}.

The time course of decay of I_Ca was fitted to a two-exponential equation by Chebyshev method on Clampfit version 6.03 (Axon Instruments). The fit interval started at 60 ms and ended at 440 ms.

For the steady-state inactivation curves, the normalized amplitude of the inward Ca²⁺ current evoked by the test pulse (+10 or +20 mV) is plotted as a function of prepulse voltage.

RESULTS

Confirmation of the integration of the targeting construct and knockout of the subunit.

Tail DNA isolated from homozygous mice (Fig. 1C, lanes 1–4) produced only the 459-bp PCR product, whereas the WT DNA produced only the 83-bp product (Fig. 1C, lanes 9–12). PCR from heterozygous mouse DNA (Fig. 1C, lanes 5–8) produced both bands. We performed a PCR control that consisted of a 1:1 mixture of homozygous and WT tail DNA. PCR of this DNA resulted in both PCR products, confirming that the homozygous DNA does not inhibit the production of the 83-bp product. These results demonstrate that the α₂/δ-1 knockout gene-targeting event occurred precisely as predicted and indeed does replace the normal endogenous allele. The genomic-targeting PCR confirmed the results of the PCR genotyping for WT, +/−, and −/− mice.

Loss of the high-affinity GBP binding site.

The VDCC α₂/δ-1-subunit possesses a high-affinity binding site for GBP. To determine the differences in protein expression in α₂/δ-1 (−/−) mice, a GBP radioligand binding experiment was performed with skeletal muscle, brain, and heart tissue. As shown in Fig. 2 and Table 1, the α₂/δ-1 (−/−) animals displayed no measurable or very low levels of ³H[GBP] binding in skeletal muscle and brain. The density of GBP binding in the α₂/δ-1 (+/−) animals was also reduced compared with WT in either brain or skeletal muscle. Though the α₂/δ-1 binding site in WT heart tissue was very low (713 fmol/mg), approximately fivefold lower than that in WT skeletal muscle and brain tissues, the deletion of the α₂/δ-1-subunit further reduced the GBP binding sites in heart.

Fig. 2.

Comparison of [³H]gabapentin (GBP) binding sites in different tissues from α₂/δ-1 (−/−) mice. Membranes (20 mg/ml) of brain (left), skeletal muscle (middle), and heart (right) from WT, +/−, and −/− mice were incubated with varying concentrations of [³H]GBP at room temperature for 60 min. The results are representative of 3 experiments, with each point assayed in duplicate. The data were analyzed using Graphpad PRISM. The computer-drawn curves represent the best fit to the data by a 1-site binding model. cpm, Counts per minute.

Table 1.

Comparison of [³H]GBP binding sites in different tissues from α₂/-δ1 WT (+/+), heterozygous (+/−), and knockout (−/−) mice

	Bmax, fmol/mg	Kd, nM
Brain
WT	3,359±29	25±0.5
+/−	2,757±390	36±8.9
−/−	570±81	25±5
Skeletal muscle
WT	3,343±961	21±3
+/−	2,268±370	20±4
−/−	145±18*	11±4
Heart
WT	713±227*	27±12
+/−	423±202*	27±18
−/−	201±193*	20±7

Values are means ± SD. GBP, gabapentin; B_max, binding maximum; K_d, dissociation equilibrium constant; WT, wild-type.

^*Value is out of detectable range.

Western blot analyses of the Ca²⁺ channel subunits.

The murine cardiac L-type VDCC complex is composed of Ca_v1.2 (α_1C), α₂/δ-1-, and β₂-subunits. To determine whether an ablation of α₂/δ-1 expression altered the expression of other subunits in the complex, we first performed Western blot analyses of Ca_v1.2 and β_2a-subunit protein levels on total protein from heart tissue from WT and −/− mice. No significant changes were observed (Fig. 3A). Since four α₂/δ-isoforms have been identified to date, we wanted to evaluate whether the deletion of the α₂/δ-1-isoform may result in a compensatory increase in expression of the other isoforms. In addition, since it is well known that the α₂/δ-1-subunit also associates with other Ca²⁺ channel pore-forming α₁-subunits, such as Ca_v2.1, Ca_v2.2, and Ca_v2.3, to form P/Q-, N-, and R-type Ca²⁺ channels, respectively, the expression of these subunits was also investigated. To quantify the impact on expression of these proteins, Western blot analyses of total protein, obtained from heart, brain, and skeletal muscle from WT, +/−, and −/− α₂/δ-1 mice, were performed for each Ca²⁺ channel isoform. Consistent with the GBP binding data, α₂/δ-1 protein expression was not detected at all in the α₂/δ-1 (−/−) mouse heart, brain, or skeletal muscle. In contrast, no significant changes in protein expression were observed for any of the other Ca²⁺ channel subunit isoforms tested (Fig. 3B).

Fig. 3.

Expression of Ca²⁺ channels in α₂/δ-1-deficient mice. A: Western blot analyses of Ca²⁺ channel isoform voltage-dependent L-type Ca²⁺ channel (Ca_v1.2) and β_2a-subunits in heart tissue normalized to calsequestrin. B: Western blot analyses of α₂/δ-isoforms (1, 2, 3, and 4), Ca_v2.1, Ca_v2.2, and β₃-subunit normalized to GAPDH in brain, heart, and skeletal muscle (SM) from α₂/δ-1 WT, +/−, and −/− mice. The numbers reported as molecular mass are in kilodaltons.

Clinical characteristics and morphology.

The α₂/δ-1 (−/−) animals clinically appear normal, although males display a tendency for bladder dilatation. Grossly, the α₂/δ-1 (−/−) hearts appear normal with no obvious cardiac hypertrophy or dilatation (Fig. 4). At 8 mo, the average heart weight (HW) and BW were not significantly different between the two groups; α₂/δ-1 (−/−) HW was 172.5 ± 18.9 mg, and BW was 29.3 ± 2.7 g (n = 4) compared with the WT HW of 154.3 ± 12.9 mg and BW of 28.8 ± 1.8 g (n = 7), respectively. HW-to-BW ratios also were not significantly different between α₂/δ-1 (−/−) and WT (5.9 ± 0.42 and 5.37 ± 0.28 mg/g, respectively). The majority of the animals were female. When analyzed by sex, HW [163.3 ± 20.2 (n = 3) vs. 140 ± 9.1 mg (n = 5)], BW (28.1 ± 3.02 vs. 25.5 ± 1.4 g), and HW-to-BW ratios (5.83 ± 0.51 vs. 5.49 ± 0.011) for α₂/δ-1 (−/−) and WT, respectively, were not significantly different.

Fig. 4.

Hematoxylin and eosin slides of WT and −/− hearts.

Diminished cardiac contractility.

We performed ex vivo working heart studies in WT (n = 6) mice and α₂/δ-1 (−/−) (n = 7) littermates (Table 2), hypothesizing that the deletion of the α₂/δ-1-subunit would result in a diminished contractility that was not dependent on altered vascular resistance or loading conditions. Indeed, when compared with WT, basal contractility and relaxation (+dP/dt and −dP/dt) were significantly lower in α₂/δ-1 (−/−) hearts.

Table 2.

Contractile parameters in WT (+/+) and α₂/δ-1 (−/−) hearts from isolated working heart preparations

Parameters	WT	α2/δ-1 (−/−)
n	6	7
LVSP, mmHg	119.7±31.4	95.7±8.1
LVDP, mmHG	−17.5±6.8	−7.3±9.2
LVEDP, mmHg	2.4±4.5	13.1±10.1
+dP/dt, mmHg/s	4,159±893	3,134±618*
−dP/dt, mmHg/s	3,727±515	2,934±695*
Heart rate, beats/min	352±16.5	348±66.7
TPP, ms/mmHg	0.47±0.07	0.52±0.14
TR1/2, ms/mmHg	0.70±0.15	0.70±0.17

Values are means ± SD; n, number of hearts. LVSP, left ventricular systolic pressure; LVDP, left ventricular diastolic pressure; LVEDP, left ventricular end-diastolic pressure; +dP/dt, maximal rate of pressure development; −dP/dt, maximal rate of pressure decline; TPP, time to peak pressure, normalized to peak pressure; TR_1/2, time to half relaxation pressure, normalized to half relaxation pressure.

^*P < 0.05.

Ca²⁺ channel properties.

In whole cell patch-clamp measurements (I-V curves) on cardiomyocytes derived from WT and α₂/δ-1 (−/−) mice, the L-type VDCC I_Ca peak current density was significantly decreased by ∼45%, compared with WT (I_Ca density; Fig. 5A, and Table 3). There was a significant increase in the half-activation potential (V_0.5, more than 4-fold) and the slope factor (k, 28%) between the WT and α₂/δ-1 (−/−) cardiomyocytes. Depolarizing pulses of different amplitudes were applied from a holding membrane potential of −50 mV. The inward currents reached their peak values at around +20 mV in the α₂/δ-1 (−/−) and +10 mV for WT (steady state activation) mice. There was a hypopolarizing shift in the voltage for half-maximal activation (V_0.5, right shift ∼8.1 mV; Fig. 5B, and Table 3). The slope parameter (k_activation) value was not statistically different between the two groups. For the steady-state inactivation curves, the normalized amplitude of the inward Ca²⁺ current evoked by the test pulse (+10 or +20 mV) was plotted as a function of prepulse voltage. The voltage of V_0.5 was shifted to a hypopolarizing direction by ∼13 mV (Fig. 5C, and Table 3). The steepness of the inactivation curve (k_inactivation) did not change significantly between groups. The time course of the decay of I_Ca was analyzed by fitting a double exponential equation by Chebyshev. The fit interval started at 60 ms and ended at 440 ms. The time constant of the rapid component (τ_fast) was significantly different depending on the pulse (+10 or +20 mV), suggesting the interference of a Ca²⁺-dependent inactivation process, since Ca²⁺ was used as a charge carrier. The slow time constant (τ_slow), representing voltage-dependent inactivation independent of current amplitude, was slowed in α₂/δ-1-deficient cardiomyocytes (Table 3). Cell capacitance was significantly decreased in α₂/δ-1 (−/−) cardiomyocytes (Table 3).

Fig. 5.

Voltage dependence and kinetics of L-type Ca²⁺ current (I_Ca) activation. A: voltage dependence of peak inward current densities [current-voltage (I-V) curve]. B: steady-state activation curves. Means and means ± SE are from 27–44 experiments for each group [WT and α₂/δ-1 (−/−), respectively]. C: steady-state inactivation curves. Inactivation fraction was determined by using a 2-pulse protocol. Data are means ± SE from 8 experiments for each group. Lines represent Boltzmann fits. D: representative Ca²⁺ current traces. Currents were elicited by 380-ms depolarizing pulses between −40 and +50 mV or between −30 and +60 mV from a holding potential −60 mV at 0.2 Hz. G, conductance.

Table 3.

Comparison of WT and α₂/δ-1 (−/−) mice on I_Ca

	WT	n	α2/δ-1 (−/−)	n
ICa density, pA/pF	5.76±0.30 (+10 mV)	35/5	3.17±0.19* (+20 mV)	47/7
V0.5, mV (I-V curve)	−2.32±1.18		8.56±1.12†
k, mV (I-V curve)	5.29±0.17		6.79±0.18†
V0.5, mV (steady-state activation)	−3.50±1.15	27/5	4.66±1.14†	38/7
k, mV (steady-state activation)	4.87±0.14		5.88±0.12
V0.5, mV (steady-state inactivation)	−15.15±1.71	8/3	−2.18±2.59†	8/6
k, mV (steady-state inactivation)	5.54±0.42		6.14±0.43
τfast, ms	19.90±1.25	23/5	44.74±3.67†	28/7
τslow, ms	83.73±2.40		158.92±13.39†
Cell capacitance, pF	222.25±11.84		171.78±6.77†

Values are means ± SD; n, number of experiments performed (number of animals used to generate the myocytes/number of myocytes measured). I_Ca density, peak current density; I-V, current-voltage; V_0.5 (I-V curve), half-activation potential; V_0.5 (steady-state activation), half-maximal activation; V_0.5 (steady-state inactivation), half-maximal inactivation; k, slope factor; τ_fast, time constant of the rapid component; τ_slow, slow time constant.

^*P < 0.05;

^†P < 0.001.

DISCUSSION

The role of the accessory subunits on VDCC kinetics has been closely studied using heterologous coexpression systems. From these studies, it has been shown that both α₂/δ- and β-subunits increase membrane expression of the α₁-subunit. The accessory subunits enhance current amplitude and density, as well as regulating the kinetics of activation and inactivation (5, 34, 38, 39). The coexpression of both accessory subunits appears synergistic, rather than additive, with respect to membrane expression of the α_1C-subunit and additive for the voltage-dependence of activation of the channel. In addition, whereas the α₂/δ-subunit increases Ca²⁺ channel density, the β-subunit both increases channel density and facilitates an opening of the channel (38). However, since uninjected oocytes express endogenous barium currents (28), as well as endogenous α₁-, β-, and α₂/δ-subunits (4, 29, 32), the studies on regulatory functions of α₂/δ- and β-subunits are inevitably “contaminated” by these endogenous channels.

This is the first successful attempt at knocking out the α₂/δ-1-subunit in vivo. The genomic DNA PCR verified that the integration of the targeting construct was at exon 2 of the gene. This confirmed the results of the genotyping PCR. Furthermore, the Western blot analysis and the binding data clearly demonstrate the lack of α₂/δ-1 protein in these animals. The ablation of the α₂/δ-1-subunit did not have any effect on the expression levels of the other Ca²⁺ channel subunits. This makes an ideal system for studying the α₂/δ-1-subunit since there are no compensatory mechanisms that could possibly confuse the relevance of the data.

A comparison of the electrophysiological properties in isolated cardiomyocytes from the α₂/δ-1 (−/−) and WT mice clearly showed that the absence of the α₂/δ-1 gene in the heart results in an attenuated I_Ca amplitude, a decrease in I_Ca density, an increase in the time constants (fast and slow), and a hypopolarizing shift in activation and inactivation. Thus, recapitulating studies on heterologous systems (5, 34, 38, 39), the knockout of the α₂/δ-1 gene either reduces the number of channels targeted to the membrane and/or changes the kinetics of the existing channels, as shown by the increased membrane potential necessary to achieve half-maximal activation and inactivation, changing the open/closed status of the channels. In other words, the results demonstrate that the channels are slower to activate as well as inactivate, thus decreasing Ca²⁺ current. Single-channel analyses are planned to decipher this mechanism. Basal contractility and relaxation were also significantly decreased in ex vivo working heart experiments, further indicating decreased Ca²⁺ load.

Several studies have been performed using small interfering RNAs to disrupt the α₂/δ-1 in muscle cell lines (11, 21, 23, 33), although most have been aimed at skeletal muscle, measuring the effects of α₂/δ-1 on the α_1S-subunit (11, 21, 23). One of these systems involves a dysgenic muscle cell line, GLT, which lacks the α₁-subunit (31) and has the α₂/δ-1-subunit targeted to the cytoplasm and plasma cell membrane (23). Investigating this system revealed a major role for the α₂/δ-1 subunit in excitation-contraction coupling in both skeletal (21) and cardiac tissue (33) but did not demonstrate a function in targeting of the α₁-pore subunit to the membrane. Reconstitution with the α_1C-pore subunit, mimicking cardiac muscle, and silencing of the α₂/δ-1-subunit resulted in a decreased current amplitude, with a depolarizing shift in voltage dependence of activation as well as decreased activation and inactivation kinetics (33). Interestingly, in a recent study examining skeletal myotubes in C2C12 cells, in which the α₂/δ-1-subunit had been silenced, Ca²⁺ current was similar to that of WT cells. However, a potential role for the α₂/δ-1-subunit in extracellular signaling involving migration and spreading of skeletal myocytes was reported (11).

Homozygous knockout of the α_1C-subunit is embryologically lethal before 14.5 days of development (26). Conditional smooth muscle-specific knockout of the Ca_v1.2 (SMACKO mouse) results in a complete loss of L-type current in single smooth muscle electrophysiology (19). The life span of these animals ranges 21–28 days following tamoxifen-induced inactivation of the Ca_v1.2 gene. Protein expression of Ca_v1.2 was <10% in these mice (19). In contrast to the α₂/δ-2 knockout mouse, the α₂/δ-1 knockout does not show obvious signs of seizures, which is most likely due to the different localizations of the individual isoforms. The α₂/δ-2 isoform is predominantly located in the central nervous system, explaining the manifestation of ataxia and seizures (17).

We have developed a knockout model that, while obviously having a profound affect on the regulation of the α_1C-subunit, is not lethal. It represents a novel model for studying the function of the α₂/δ-1-subunit and characterizing its effects, particularly with respect to cardiovascular function, epileptogenesis, and nociception, potentially leading to new pharmacological therapies.

GRANTS

This work was supported by US National Heart, Lung, and Blood Institute Grants NIH-T32-HL-07382 (to A. Schwartz), R01-HL-079599 (to A. Schwartz), and P01-HL-22619 (to A. Schwartz).