Functional Characterization of Ca_Vα2δ Mutations Associated with Sudden Cardiac Death

Background: Missense mutations in Ca_Vα2δ1, an auxiliary subunit of cardiac L-type Ca_V1.2 channels, are associated with arrhythmias.

Results: The reduction in the cell surface density of Ca_Vα2δ1 D550Y/Q917H was sufficient to impair Ca_V1.2 currents.

Conclusion: Defects in the cell surface trafficking of Ca_Vα2δ1 mutants down-regulate L-type currents.

Significance: CACNA2D1 genetic variants may trigger arrhythmias by reducing L-type Ca²⁺ currents.

Abstract

L-type Ca²⁺ channels play a critical role in cardiac rhythmicity. These ion channels are oligomeric complexes formed by the pore-forming Ca_Vα1 with the auxiliary Ca_Vβ and Ca_Vα2δ subunits. Ca_Vα2δ increases the peak current density and improves the voltage-dependent activation gating of Ca_V1.2 channels without increasing the surface expression of the Ca_Vα1 subunit. The functional impact of genetic variants of CACNA2D1 (the gene encoding for Ca_Vα2δ), associated with shorter repolarization QT intervals (the time interval between the Q and the T waves on the cardiac electrocardiogram), was investigated after recombinant expression of the full complement of L-type Ca_V1.2 subunits in human embryonic kidney 293 cells. By performing side-by-side high resolution flow cytometry assays and whole-cell patch clamp recordings, we revealed that the surface density of the Ca_Vα2δ wild-type protein correlates with the peak current density. Furthermore, the cell surface density of Ca_Vα2δ mutants S755T, Q917H, and S956T was not significantly different from the cell surface density of the Ca_Vα2δ wild-type protein expressed under the same conditions. In contrast, the cell surface expression of Ca_Vα2δ D550Y, Ca_Vα2δ S709N, and the double mutant D550Y/Q917H was reduced, respectively, by ≈30–33% for the single mutants and by 60% for the latter. The cell surface density of D550Y/Q917H was more significantly impaired than protein stability, suggesting that surface trafficking of Ca_Vα2δ was disrupted by the double mutation. Co-expression with D550Y/Q917H significantly decreased Ca_V1.2 currents as compared with results obtained with Ca_Vα2δ wild type. It is concluded that D550Y/Q917H reduced inward Ca²⁺ currents through a defect in the cell surface trafficking of Ca_Vα2δ. Altogether, our results provide novel insight in the molecular mechanism underlying the modulation of Ca_V1.2 currents by Ca_Vα2δ.

Introduction

Polymorphic ventricular tachycardia is one of the leading causes of sudden cardiac death in children and young adults (1). These cardiac arrhythmias, which are reported in the absence of structural heart defects, coronary artery disease, or heart failure, are detected in a noninvasive fashion by measuring changes in the QT³ interval on the electrocardiogram. Either excessive prolongation (LQT) or shortening of the QT (SQT) intervals are associated with an increased risk of sudden cardiac death (2). Inherited Mendelian long QT syndrome (LQTS) and short QT syndrome (SQTS) originate from mutations in genes encoding ion channels or channel-interacting proteins (3). Genome-wide linkage studies of families with LQTS have reported 13 LQTS susceptibility genes with 75% of LQTS cases stemming from mutations in KCNQ1 (LQT1), KCNH2 (LQT2), and SCN5A (LQT3) (4). Only recently have mutations of genes coding for subunits forming the L-type Ca²⁺ channel been linked to inherited arrhythmogenic diseases caused by LQTS or SQTS (5). The molecular mechanism underlying the cardiac dysfunction remains unknown in many cases.

The CACNA1C gene encodes the L-type Ca_V1.2 channel that carries the vast majority of the L-type calcium current in the adult heart (6, 7) which in turn initiates the coordinated contraction of the cardiac ventricles (8). Its unique role is substantiated by the observation that homozygous knock-out of the CACNA1C gene is lethal (9, 10). In addition to initiating the coordinated contraction of the cardiac ventricles, Ca_V1.2 channels are critical to the heart's normal rhythmic activity (8). During the cardiac action potential, Ca²⁺ enters the cell through the voltage-dependent L-type Ca_V1.2 generating an inward Ca²⁺ current that contributes to the plateau phase of the action potential (11). The kinetic properties of this channel must be properly timed so that depolarization and contraction are synchronized during the systolic-diastolic cycle of the heart. Even a slight disruption of Ca²⁺ cycling can have a profound impact on action-potential duration and trigger early after depolarizations, ultimately cumulating in lethal arrhythmias (5, 12), including torsade de pointes and ventricular fibrillation (2). Prolonged inward Ca²⁺ current during the plateau phase of the cardiac action potential leads to delays in ventricular myocyte repolarization, a subsequent prolonged QT repolarization interval on the electrocardiogram and a highly arrhythmogenic and potentially lethal substrate. On the other end of the spectrum, any significant decrease in inward currents or significant increase in outward currents may lead to lethal arrhythmias associated with shorter than normal QT repolarization intervals (3).

The L-type Ca_V1.2 channel belongs to the molecular family of high voltage activated Ca_V channels. High voltage-activated Ca_V1.2 channels are hetero-oligomers formed by the main pore-forming Ca_Vα1 subunit in a complex with the cytoplasmic Ca_Vβ auxiliary subunit, the EF-hand protein calmodulin constitutively bound to the C terminus of Ca_Vα1, and the mostly extracellular Ca_Vα2δ subunit (13–18). The full complement of auxiliary subunits is required to produce high voltage-activated Ca_V1.2 channels with the properties of the native channels. Ca_Vβ promotes the cell surface density of Ca_V1.2 channels (19) in part by preventing its degradation by the ubiquitin/proteasome system (20). Co-expression of Ca_Vα2δ subunit with Ca_Vβ-bound Ca_Vα1 increases peak current density and hyperpolarizes the voltage of activation of the L-type Ca_V1.2 (19, 21–23).

Gain-of-function mutations in the pore-forming Ca_Vα1 subunit of the L-type Ca²⁺ channel lead to the highly arrhythmogenic Timothy syndrome. Timothy syndromes 1 and 2 are rare variants of the long QT syndrome (LQT8) (24, 25) characterized by extreme QT-prolongation and gain-of-function mutations in the pore-forming subunit of Ca_V1.2 (24–28). Mutations in the auxiliary subunits forming the L-type Ca_V1.2 channel have also been identified among probands diagnosed with Brugada syndrome, idiopathic ventricular fibrillation, and early repolarization syndrome associated with short QT interval (QTc ≤360 ms) (29–31). According to these studies, the shorter QT interval might result from loss-of-function mutations in the Ca_V genes coding for Ca_Vα1 (CACNA1C), Ca_Vβ (CACNB2), or Ca_Vα2δ (CACNA2D1). Missense mutations D550Y, S709N, S755T, Q917H, and S956T in the Ca_Vα2δ1 protein have been associated with congenital arrhythmias (referred to as short QT syndrome 6 (SQT6) (29–31)), but the molecular mechanism(s) underlying the change in function remains to be established (32).

In this work, we have characterized the Ca_Vα2δ mutations associated with arrhythmogenic activity in regard to their functional impact on the cardiac L-type Ca_V1.2 currents. Using a fluorescence-labeled and extracellularly tagged Ca_Vα2δ subunit, we show here that missense mutations D550Y, S709N and D550Y/Q917H significantly altered the cell surface density of Ca_Vα2δ. Decreasing the cell surface density of Ca_Vα2δ D550Y/Q917H was found to profoundly decrease L-type peak current densities as compared with currents measured with Ca_Vα2δ wild type (WT). Altogether, our data support a model where Ca_Vα2δ modulates channel function without altering the trafficking of Ca_V1.2 and establishes the cell surface density of Ca_Vα2δ relative to the pore-forming Ca_Vα1 as the single most important determinant in the stimulation of Ca_V1.2 currents by Ca_Vα2δ.

EXPERIMENTAL PROCEDURES

Recombinant DNA Techniques

The rabbit Ca_V1.2 (GenBank^TM X15539), the rat Ca_Vβ3 (GenBank^TM M88751) (33), and the rat brain Ca_Vα2δ1 (GenBank^TM NM_012919) (34) were subcloned in commercial vectors under the control of the CMV promoter as described elsewhere (19, 35). The primary sequence of the rat brain Ca_Vα2δ1 clone (1091 residues) is 96% identical to the predicted sequence of the human clone NM_000722 (1091 residues). All residues mutated in the following experiments are conserved between species. The first series of experiments (Figs. 1 and 2) was conducted with a tagged version of Ca_V1.2 whereby the hemagglutinin (HA) epitope (YPYDVPDYA) was inserted in the extracytoplasmic loop of Ca_V1.2 in the S5-S6 linker of domain II between residues 710 and 711, whereas the untagged version of Ca_V1.2 was used in the experiments shown in Figs. 3–11. The rat Ca_Vα2δ1 was subcloned in the pmCherry-N1 vector (Cederlane, Burlington, Ontario, Canada) between the SacI and SalI sites. In this construct, the hemagglutinin (HA) epitope (YPYDVPDYA) was inserted in the extracellular domain of Ca_Vα2. More than 10 different insertion sites were tested in Ca_Vα2, but only one was found to be functional and fluorescently labeled with a FITC-labeled anti-HA (fluorescein isothiocyanate) (results not shown). The insertion site was identified at position Arg-676 in the extracellular domain. Cell surface and total density of Ca_Vα2δ were evaluated by two-color flow cytometry assays using the pmCherry-Ca_Vα2δ1-HA construct. Unless specified otherwise, the flow cytometry experiments and the patch clamp recordings were systematically conducted with this pmCherry-Ca_Vα2δ-1-HA construct.

FIGURE 1.

Ca_Vα2δ1 increases peak current density and hyperpolarizes the activation potential of Ca_V1.2. A, pCMV-Ca_V1.2 (untagged version) was expressed in HEKT cells (Aa), in Ca_Vβ3 stable cells with an empty pCMV plasmid (Ab), in Ca_Vβ3 stable cells with pmCherry-Ca_Vα2δ1 (Ac), and in Ca_Vβ3 stable cells with pmCherry-Ca_Vα2δ1-HA (Ad). In addition, pCMV-Ca_V1.2-HA was expressed with pmCherry-Ca_Vα2δ1-HA in Ca_Vβ3 stable cells in one series of experiments (Ae). Whole-cell current traces were recorded in the presence of 2 mm Ca²⁺ from a holding potential of −100 mV. Time scale is 100 ms throughout. Unless specified otherwise, the current density scale is 10 pA/pF. B, peak current densities increased from −5 ± 1 pA/pF (n = 26) for Ca_V1.2 WT/Ca_Vβ3 with an empty pmCherry vector to −56 ± 3 pA/pF (n = 23) in the presence of Ca_V1.2 WT/Ca_Vβ3 with pmCherry-Ca_Vα2δ1 WT. Peak current densities measured in the presence of the following combinations were not statistically significantly different: Ca_V1.2 WT/Ca_Vβ3 with pmCherry-Ca_Vα2δ1 WT; Ca_V1.2 WT/Ca_Vβ3 with pmCherry-Ca_Vα2δ1-HA; and Ca_V1.2 HA/Ca_Vβ3 pmCherry-Ca_Vα2δ1-HA. Co-expression with Ca_Vα2δ1 WT shifted the voltage dependence of activation of Ca_V1.2 WT/Ca_Vβ3 from E_{0.5, act} = 5 ± 2 mV (n = 26) (no Ca_Vα2δ) to E_{0.5, act} = −10.1 ± 0.5 mV (n = 23) (for Ca_V1.2 WT/Ca_Vβ3 with Ca_Vα2δ1 WT), a significant −15 mV shift in the activation potential. The number in parentheses refers to the number of independent experiments. C, histogram is reporting the distribution of the individual ΔG_act values estimated in kilocalories/mol for each experimental condition (see “Experimental Procedures” for details). The ΔG_act values for Ca_V1.2 WT/Ca_Vβ3 with pmCherry-Ca_Vα2δ1-HA were best fitted by a Gaussian equation with a σ = 0.36 centered around −0.86 kcal mol⁻¹. As seen, the distribution of the ΔG_act values measured in the absence of Ca_Vα2δ1 subunit appeared to be flatter (varying from −0.1 to + 1.5 kcal mol⁻¹) and could not be properly fitted by a Gaussian equation.

FIGURE 2.

Ca_Vα2δ1 stimulates the total protein expression of Ca_V1.2 in HEKT cells. HA-tagged Ca_V1.2 WT was co-expressed transiently either in HEKT cells (A and B, top rows) or in stable Ca_Vβ3 cells (A and B, bottom rows) with or without pCMV-Ca_Vα2δ1. Cell surface expression of Ca_V1.2-HA WT was determined in intact nonpermeabilized cells (A) or after cell permeabilization by flow cytometry using the anti-HA FITC conjugate antibody (B). All histograms display live cells. ΔMFI was used as an index of the cell surface density of Ca_V1.2 in intact nonpermeabilized cells or the total expression (cell surface and intracellular protein density) of Ca_V1.2 in permeabilized cells. As seen, the surface expression of Ca_V1.2 was not significantly improved in the presence of Ca_Vα2δ1 in contrast to the robust stimulation observed with Ca_Vβ3. The total protein expression of Ca_V1.2 was improved after co-expression with Ca_Vα2δ1, but stronger stimulation was observed in the combined presence of Ca_Vβ3 with Ca_Vα2δ1. C, predicted secondary structure of the HA-tagged Ca_V1.2 construct used in the figure. Please note that Ca_Vα2δ1 was not tagged. D, bar graph summarizing the results shown in A and B with the values for the cell surface density in light gray bars and the total protein density in dark gray bars. Each experimental condition was quantified in triplicate. The fluorescence intensity is shown in arbitrary units. The statistical analysis was performed by comparing the ΔMFI values measured in all experimental groups versus ΔMFI values measured for Ca_V1.2-HA alone in intact versus permeabilized cells respectively. **, p < 0.01; ***, p < 0.001.

FIGURE 3.

Increased protein stability of the ternary complex Ca_V1.2-Ca_Vβ-Ca_Vα2δ1 upon arrest of protein synthesis. HEKT cells were transiently transfected simultaneously with the following constructions: A, pCMV-Ca_V1.2 and control plasmid; B, pCMV-Ca_V1.2 and pCMV-Ca_Vβ3; C, pCMV-Ca_V1.2 and pCMV-Ca_Vα2δ1; and D, pCMV-Ca_V1.2, pCMV-Ca_Vβ3, and pCMV-Ca_Vα2δ1. Exactly 24 h after transfection, subconfluent cells were incubated with cycloheximide (100 μg/ml) for up to 48 h to block de novo protein synthesis. At the indicated time points (0 h or no cycloheximide for 6, 12, 24, 36, and 48 h), cell lysates were prepared and fractionated by SDS-PAGE (10%), followed by immunoblotting to visualize Ca_V1.2, Ca_Vβ3, Ca_Vα2δ1, and GAPDH. The protein density of Ca_V1.2 in total lysates was expressed relative to GAPDH and normalized to the protein density measured at time 0. As seen, Ca_Vβ3 alone was more effective than Ca_Vα2δ1 in preventing the degradation of Ca_V1.2, but co-expressing Ca_Vβ3 with Ca_Vα2δ1 further stabilized the protein density of Ca_V1.2. The time courses of the protein degradation for the following conditions, Ca_V1.2 alone and Ca_V1.2/Ca_Vα2δ1, were indistinguishable from one another and from the ΔMFI for FITC in the previous figure suggested that Ca_V1.2 (total density) was more stable in the presence of Ca_Vα2δ1. This difference might arise from the higher sensitivity of the fluorescence assays. Molecular mass of Ca_V1.2 is ∼250 kDa; Ca_Vα2δ1 is 175 kDa; Ca_Vβ3 is 60 kDa; and GAPDH is 40 kDa. E, protein density was estimated with ImageJ (rsbweb.nih.gov). The time course of degradation was measured in three series of experiments. Each symbol represents the mean ± S.E. of the normalized protein density of Ca_V1.2.

FIGURE 4.

Protein expression of Ca_Vα2δ1 is increased in the presence of Ca_V1.2 and Ca_Vβ3. Aa, pmCherry-Ca_Vα2δ1-HA was transiently transfected with a control pCMV plasmid in HEKT cells; Ab, with Ca_V1.2 (untagged) in HEKT cells; Ac, in stable Ca_Vβ3 cells; and Ad with Ca_V1.2 (untagged) in stable Ca_Vβ3 cells. The HA tag of Ca_Vα2δ1-HA was detected using an anti-HA FITC-conjugated antibody. Dead cells were excluded based on the forward scatter/side scatter profiles. Representative two-dimensional plots of mCherry versus FITC fluorescence are shown for each condition as stated (left panels). The vertical line indicates the median fluorescence intensity for FITC to facilitate the visual comparison between the different experimental groups. As seen, the cellular autofluorescence levels increased after permeabilization, which prevents comparison of the absolute fluorescence values between intact nonpermeabilized and permeabilized cells. B, predicted secondary structure of the pmCherry-Ca_Vα2δ1-HA construct used in the figure. Please note that Ca_V1.2 was not tagged either in the N or C termini. C, bar graph shows the ΔMFI measured in the presence of FITC in intact cells for each experimental condition. Experiments were conducted in triplicate, and each bar is the mean ± S.E. of ΔMFI in arbitrary units. Under these conditions, ΔMFI measured for FITC reflects the relative cell surface protein expression of Ca_Vα2δ1. NP, nonpermeabilized; P, permeabilized cells, D, bar graph shows the ΔMFI measured for FITC in permeabilized cells for each experimental condition. Experiments were conducted in triplicate, and each bar is the mean ± S.E. of ΔMFI in arbitrary units. Under these conditions, ΔMFI measured for FITC reflects the total protein expression of Ca_Vα2δ1. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

FIGURE 5.

Cell surface expression of Ca_Vα2δ1 as a function of DNA ratios. A, pmCherry-Ca_Vα2δ1-HA WT (in variable amounts) was transiently transfected with Ca_V1.2 in stable Ca_Vβ3 cells. Protein expression was assessed in flow cytometry assays using the gating strategy explained in Fig. 4. Representative two-dimensional plots of mCherry versus FITC fluorescence are shown for each Ca_Vα2δ1-HA/Ca_V1.2 DNA ratio as follows: ratio 1:8 (0.5 μg for Ca_Vα2δ1-HA); ratio 1:4 (1 μg); ratio 1:2 (2 μg); and ratio 1:1 (4 μg) (left panels). Histogram plots show the distribution of fluorescence intensity of the anti-HA FITC-conjugated antibody staining in intact nonpermeabilized (NP) cells (middle panels) or in permeabilized (P) cells (right panels). B, overlay representation of the data is shown to facilitate visualization of the differences in fluorescence intensity as follows: Ratio 1:8 (filled line); ratio 1:4 (dotted line); ratio 1:2 (dashed line); and ratio 1:1 (full line). C, bar graph shows the normalized ΔMFI measured in the presence of FITC in intact nonpermeabilized (dark gray bars) and permeabilized (hatched bars) cells for Ca_Vα2δ1-HA expressed in the combined presence of Ca_V1.2 and Ca_Vβ3. All ΔMFI values were normalized using the ΔMFI measured in the presence of a Ca_Vα2δ1-HA/Ca_V1.2 DNA ratio of 1:1. As seen, the cell surface and total protein expression of Ca_Vα2δ1 both increase steeply up to 1.5 to 2 μg of cDNA coding for Ca_Vα2δ1-HA WT (ratios 1:3 to 1:2).

FIGURE 6.

Modulation of L-type Ca_V1.2 currents parallels the cell surface expression of Ca_Vα2δ1. A, pCMV-Ca_V1.2 (untagged) was co-expressed in stable Ca_Vβ3 cells with variable pmCherry-Ca_Vα2δ1-HA DNA ratios for Ca_Vα2δ/Ca_V1.2 as follows: Aa, ratio 1:8 (0.5 μg); Ab, ratio 1:4 (1 μg); Ac, ratio 1:2.7 (1.5 μg); Ad, ratio 1:2 (2 μg); and Ae, ratio 1:1 (4 μg). Typical whole-cell current traces were recorded in a 2 mm Ca²⁺ solution from a holding potential of −100 mV. The pmCherry-Ca_Vα2δ1-HA construct used for the patch clamp experiments was identical to the constructs studied in the flow cytometry assays. Time scale is 100 ms throughout. Unless specified otherwise, the current density scale is 10 pA/pF. B, peak current densities increased from −5 ± 1 pA/pF (n = 26) (no insert in the pmCherry vector) to −67 ± 3 pA/pF (n = 163) in the presence of 4 μg of cDNA (ratio 1:1) coding for pmCherry-Ca_Vα2δ1-HA WT. The number in parentheses indicates the number of independent patch clamp recordings. C, histogram reporting the distribution of the individual ΔG_act values (kilocalories/mol) for each concentration of cDNA coding for Ca_Vα2δ1. As seen, the ΔG_act are significantly shifted to the left even when Ca_V1.2 is co-expressed with Ca_Vα2δ1 in a 1:1 ratio. D, inactivation kinetics of the Ca_V1.2 currents expressed with Ca_Vα2δ1 in a 1:8 ratio (gray hatched bars) were in general 20% slower (p < 0.05) than the inactivation kinetics measured after co-expression of Ca_Vα2δ1 in a 1:1 ratio (white striated red bars) (p < 0.05).

FIGURE 7.

Correlation between the cell surface density of Ca_Vα2δ1-HA WT and the changes in ΔG_act and peak current density of Ca_V1.2 currents. Peak current density was normalized to the mean value measured with 4 μg of cDNA for Ca_Vα2δ1-HA (ratio 1:1); ΔΔG_act was calculated relative to the value of ΔG_act = 0.48 ± 0.09 kcal mol⁻¹ for currents measured in the absence of Ca_Vα2δ1; the cell surface density was computed using the normalized ΔMFI measured for FITC in intact nonpermeabilized cells. Despite the large experimental variation, there is a positive correlation between the cell surface density of Ca_Vα2δ1 and the two other parameters (peak current density and ΔΔG_act of macroscopic currents). The variation of ΔΔG_act with the surface density of Ca_Vα2δ1 suggests that Ca_Vα2δ1 influences the fraction of ion channels active in a given gating mode. The correlation could be described by a large number of equations, but it is shown as a linear regression to facilitate visualization.

FIGURE 8.

Cell surface and total protein expression of Ca_Vα2δ1 D550Y/Q917H mutant is significantly decreased. pmCherry-Ca_Vα2δ1-HA WT and mutants were transiently transfected with pCMV-Ca_V1.2 (untagged) in stable Ca_Vβ3 cells. Fluorescence was computed according to the gating strategy described under “Experimental Procedures” and in Fig. 4. Representative two-dimensional plots of mCherry versus FITC fluorescence (left panels) are shown for each Ca_Vα2δ1 mutant as follows: D550Y, S709N, S755T, Q917H, S956T, and D550Y/Q917H (from top to bottom). Histogram plots show the distribution of fluorescence intensity for the anti-HA FITC conjugated antibody staining in intact nonpermeabilized (NP) cells (middle panels) or in permeabilized (P) cells (right panels). The median fluorescence intensity of FITC is indicated by a straight line on the histograms to help comparison between the different mutants. Increasing the DNA amount to 8 μg (Ca_Vα2δ/Ca_V1.2 DNA ratio of 2:1) did not increase further the surface expression of D550Y/Q917H.

FIGURE 9.

A, bar graph shows the normalized ΔMFI measured in the presence of FITC in intact cells for each pmCherry-Ca_Vα2δ1-HA construct (WT and mutants) expressed alone (hatched bars) and in the presence of Ca_V1.2 (untagged) in stable Ca_Vβ3 cells (dark gray bars). As seen, the relative cell surface density measured in intact cells decreased by 29 ± 2, 31 ± 3, and 63 ± 1% for Ca_Vα2δ1 mutants D550Y, S709N, and the double mutant D550Y/Q917H (shown as DY/QH), respectively, when expressed alone (hatched bars) or in stable Ca_Vβ3 cells with Ca_V1.2 (filled dark gray bars). Values were normalized as compared with the ΔMFI for FITC measured for Ca_Vα2δ1-HA WT under the same experimental conditions (p < 0.001). NP, nonpermeabilized. B, bar graph shows the ΔMFI measured for FITC in permeabilized (P) cells for each Ca_Vα2δ1-HA construct (WT and mutants) expressed alone (hatched bars) or in stable Ca_Vβ3 cells with Ca_V1.2 (filled gray bars). All ΔMFI values were normalized as compared with the average ΔMFI measured with Ca_Vα2δ1-HA WT expressed alone in HEKT cells. ΔMFI for FITC measured in permeabilized cells reflects the total protein expression of Ca_Vα2δ1-HA constructs. As seen, the total protein density decreased by 20 ± 1, 21 ± 2, and 24 ± 11% for Ca_Vα2δ1-HA D550Y, S709N, and D550Y/Q917H, respectively (p < 0.05), when compared with the wild-type construct. This small decrease in total protein density of these mutants could account in part for their decreased cell surface density. This decrease was also observed for the constitutive fluorescence of mCherry (data not shown). For all mutants, please note that the total protein density of all Ca_Vα2δ1-HA constructs nearly doubled when co-expressed with Ca_V1.2 in stable Ca_Vβ3 cells. *, p < 0.05; ***, p < 0.001.

FIGURE 10.

Ca_Vα2δ1 short QT mutants functionally modulate the L-type Ca_V1.2 currents. A, pCMV-Ca_V1.2 was co-expressed in stable Ca_Vβ3 stable cells with the following: Ca_Vα2δ1WT (Aa); Ca_Vα2δ1 D550Y (Ab); (Ac) Ca_Vα2δ1 S709N; Ca_Vα2δ1 S755T (Ad); and Ca_Vα2δ1 D550Y/Q917H (Ae). Typical whole-cell current traces were recorded in a 2 mm Ca²⁺ solution from a holding potential of −100 mV. The pmCherry-Ca_Vα2δ1-HA constructs were identical to the constructs used for the flow cytometry assays. Time scale is 100 ms throughout. Unless specified otherwise, the current density scale is 10 pA/pF. B, peak current densities increased from −5 ± 1 pA/pF (n = 26) (no insert in the pmCherry vector) to −67 ± 3 pA/pF (n = 163) in the presence of pmCherry-Ca_Vα2δ1-HA WT (p < 10⁻⁶). Co-expression with Ca_Vα2δ D550Y/Q917H significantly increased peak current densities of Ca_V1.2 as compared with the currents obtained in the absence of Ca_Vα2δ and did not produce currents significantly different from Ca_Vα2δ1 WT (p > 0.1) with a value of 40 ± 10 pA/pF (n = 10). The properties of Ca_Vα2δ1 mutants were estimated from 10 to 15 independent patch clamp recordings. C, histogram reporting the distribution of the individual ΔG_act values (kilocalories/mol) for each Ca_Vα2δ1 mutant. As seen, every single Ca_Vα2δ mutant negatively shifted the activation energy of Ca_V1.2 with a distribution of ΔG_act values that superimposed with the ΔG_act values of the Ca_Vα2δ1 WT. D, bar graph shows the r100 values (the ratio of peak whole-cell currents remaining at the end of a 100-ms depolarization) measured at membrane potentials from −10 to +10 mV. Faster inactivation kinetics yield lower r100 values. Inactivation kinetics of the Ca_V1.2 currents expressed with the Ca_Vα2δ1 mutants S709N and Q917H were not significantly different from the ones measured in the presence of Ca_Vα2δ1 WT. The r100 values of Ca_Vα2δ1 mutants D550Y and S755T were 15% smaller suggesting faster inactivation kinetics for these Ca_Vα2δ1 mutants. *, p < 0.05.

FIGURE 11.

L-type Ca_V1.2 currents are impaired in the presence of Ca_Vα2δ1 D550Y/Q917H mutant. pCMV-Ca_V1.2 (untagged) was co-expressed in stable Ca_Vβ3 stable cells with pmCherry-Ca_Vα2δ-HA WT or pmCherry-Ca_Vα2δ-HA D550Y/Q917H (DY/QH) at the cDNA ratios indicated for Ca_Vα2δ/Ca_V1.2. Typical whole-cell current traces were recorded in a 2 mm Ca²⁺ solution from a holding potential of −100 mV. Time scale is 100 ms throughout. A, DNA ratio for Ca_Vα2δ/Ca_V1.2 of 1:20 with 0.2 μg of DNA for Ca_Vα2δ-HA. Aa, Ca_Vα2δ-HA WT; Ab, Ca_Vα2δ-HA D550Y/Q917H; Ac, average peak current densities increased from −5 ± 1 pA/pF (n = 26) (no insert in the pmCherry vector) to −7 ± 1 pA/pF (n = 9) for mCherry-Ca_Vα2δ1-HA D550Y/Q917H, and −17 ± 2 pA/pF (n = 4) for WT (p < 10⁻⁶). Co-expression with Ca_Vα2δ D550Y/Q917H significantly increased peak current densities of Ca_V1.2 as compared with the currents obtained in the absence of Ca_Vα2δ but were significantly different from Ca_Vα2δ1 WT (p < 0.05). The ΔG_act (in kcal mol⁻¹) values were −0.02 ± 0.01 for D550Y/Q917H and −0.30 ± 0.05 for WT. B, DNA ratio for Ca_Vα2δ/Ca_V1.2 of 1:8 with 0.5 μg of DNA for Ca_Vα2δ-HA. Ba, Ca_Vα2δ-HA WT; Bb, Ca_Vα2δ-HA D550Y/Q917H; Bc, average peak current densities were −18 ± 3 pA/pF (n = 12) for mCherry-Ca_Vα2δ1-HA D550Y/Q917H, and −36 ± 5 pA/pF (n = 12) for WT. Co-expression with Ca_Vα2δ D550Y/Q917H significantly increased peak current densities of Ca_V1.2 as compared with the currents obtained in the absence of Ca_Vα2δ but were significantly different from Ca_Vα2δ1 WT (p < 0.05). The ΔG_act (in kilocalories mol⁻¹) values were −0.29 ± 0.03 for D550Y/Q917H and −0.57 ± 0.09 for WT. C, DNA ratio for Ca_Vα2δ/Ca_V1.2 was 1:1 with 4 μg of DNA for Ca_Vα2δ. Ca, Ca_Vα2δ1WT; Cb, Ca_Vα2δ1 D550Y/Q917H; Cc, average peak current densities are shown under these conditions. QY/QH should read DY/QH for D550Y/Q917H. Numerical values are shown in Table 1. Co-expression with Ca_Vα2δ D550Y/Q917H significantly increased peak current densities of Ca_V1.2 as compared with the currents obtained in the absence of Ca_Vα2δ and were significantly different from Ca_Vα2δ1 WT (p > 0.1).

Site-directed Mutagenesis

All mutants were produced with the Q5 site-directed mutagenesis kit (New England Biolabs, Whitby, Ontario, Canada) according to the manufacturer's instructions. Briefly, amino acid substitutions were performed by nonoverlapping desalted primers. Substitutions are created by incorporating the desired mutation in the center of the forward primer, and the reverse primer is designed so that the 5′ ends of the two primers anneal back-to-back. Following the PCR, a kinase/ligase/DpnI enzyme mixture was added to the amplified DNA for circularization and template removal before transformation into high efficiency DH5α-competent Escherichia coli. Constructs were verified by automated double-stranded sequence analysis (Genomics Platform, IRIC, Université de Montréal, Quebec, Canada). Protein expression of all constructs was confirmed by Western blotting in total cell lysates as described previously (19).

Cell Culture and Transfections

HEK293T or HEKT cells were grown in Dulbecco's high glucose minimum essential medium (DMEM-HG) supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin at 37 °C under 5% CO₂ atmosphere as described elsewhere (19). RT-PCR conducted in these cells failed to highlight the presence of Ca_Vβ and Ca_Vα2δ auxiliary subunits (36). All experiments described herein were conducted in HEKT cells stably transfected with Ca_Vβ3, as this isoform was best at promoting the cell surface density of Ca_V1.2 (19). The use of stable cells limited the problems linked to protein overexpression. Unless specified otherwise, stable Ca_Vβ3 cells (90% confluence) were transiently transfected with similar amounts of DNA (4 μg each for a total of 8 μg total per 10⁶ cells) as follows: Ca_V1.2 WT-HA (Figs. 1 and 2), Ca_V1.2 WT (Figs. 3–11), Ca_Vα2δ1 WT, Ca_Vα2δ1-HA, or empty vector in 10 μl of Lipofectamine 2000 (Invitrogen) using a DNA/lipid ratio of 1:2.5 as described elsewhere (19). In some experiments, the amount of DNA for Ca_Vα2δ was decreased in the 0.2–4.0-μg range to decrease the Ca_Vα2δ/Ca_V1.2 expression ratio, but the total DNA amount was kept constant at 8 μg by using a mock vector.

Antibodies and Reagents

The following primary antibodies were used for Western blots: anti-Ca_V1.2 (Alomone Labs, Jerusalem, Israel) (1:5000); anti-Ca_Vα2δ (Alomone Labs) (1:500); anti-Ca_Vβ3 (Alomone Labs) (1:5000); anti-mCherry (Biovision, Milpitas, CA) (1:20,000); anti-HA (Covance Biotechnology, Quebec, Canada) (1:1000); and anti-GAPDH (Sigma) (1:10,000). Cycloheximide was used at 100 μg/ml for up to 48 h (Sigma). For the flow cytometry assays, the HA tag was detected using the FITC-conjugated mouse monoclonal anti-HA (Sigma) (5 μg/ml). We used the mouse IgG1 isotype control antibody conjugated to FITC (Sigma) to help assess the level of background staining in the flow cytometry experiments (5 μg/ml).

Flow Cytometry Experiments

Flow cytometry experiments were conducted as described elsewhere (19, 35). To determine the cell surface expression level of the tagged proteins, cells were harvested 24 h after transfection, washed in a PBS 1× buffer, and stained with the FITC-labeled anti-HA epitope tag antibody at 5 μg/ml (Clone HA-7, Sigma) at 4 °C for 30 min. To determine the density of both intracellular and extracellular expression of the tagged proteins, cells were fixed and permeabilized using BD Cytofix/Cytoperm^TM fixation/permeabilization solution kit (BD Biosciences, 554714) according to the manufacturer's instructions. Briefly, cells were washed with the 1× BD Perm/Wash buffer containing both FBS and saponin and stained 30 min at 4 °C with the FITC-labeled anti-HA epitope tag antibody. A maximum of 10,000 cells were counted using a FACSAria III® special order research product flow cytometer (BD Biosciences) at the flow cytometry facility located at the Department of Microbiology, Université de Montréal. Relative expression of Ca_Vα2δ was calculated based on Δmean fluorescence intensity (ΔMFI) for each fluorophore (mCherry or FITC) rather than computing the number of fluorescent cells (see below) (19). Three control conditions were always carried out for each series of experiments as follows: (a) untransfected HEKT cells with the murine IgG1-FITC isotype control (5 μg/ml); (b) transfected HEKT cells without the anti-HA FITC antibody but with the murine IgG1-FITC isotype control (5 μg/ml); (c) HEKT cells transfected with the wild-type HA-tagged construct. No fluorescence was detected with the IgG1-FITC isotype control murine (5 μg/ml) nor in untransfected cells with the anti-HA FITC conjugated antibody (5 μg/ml) (data not shown). The fluorescence intensity of each construct transfected in any given condition was measured using six distinct cell dishes for an average of 60,000 cells for each condition. Flow cytometry experiments performed at 36 h to mimic as closely as possible the experimental conditions used for the patch clamp experiments (see below), yielded quantitatively similar results as experiments performed at 24 h (data not shown).

Quantification of Cell Surface Expression

Flow cytometry data were analyzed, and figures were produced using the FlowJo software, version 10 (TreeStar, Ashland, OR). Dead cells were excluded based on forward scatter/side scatter profiles. The FITC-positive cells gate (P2) and the FITC negative cells gate (P3) were set manually. The FITC fluorescence intensity within the region delineated by the P2 and P3 gate was displayed as cell count versus FITC fluorescence intensity (histograms in Figs. 4, 5, and 8). The ΔMFI for FITC was calculated by subtracting the FITC fluorescence density of the FITC-negative cells (P3) from the fluorescence density of the FITC-positive cells (P2). ΔMFI was used as an index of the cell surface density of HA-tagged proteins (either HA-Ca_V1.2 or pmCherry-Ca_Vα2δ1-HA) in intact nonpermeabilized cells or the total expression (cell surface and intracellular protein density) of HA-tagged proteins in permeabilized cells. The two HA-tagged proteins were never expressed together. Because cellular autofluorescence levels are altered by the change in the cell medium, the actual ΔMFI values in arbitrary units cannot be compared between intact nonpermeabilized and permeabilized cells. In Figs. 5 and 9, ΔMFI were pooled and normalized to the average value obtained for the pmCherry-Ca_Vα2δ1-HA WT construct that was expressed under the same conditions and quantified under the same experimental conditions (see above).

Patch Clamp Experiments in HEKT Cells

Whole-cell voltage clamp recordings were performed on isolated cells 30–38 h after transfection using the methods described above in the presence of the peGFP vector (0.2 μg) as a control for transfection. In all cases, the experiments were carried out under optimal transfection conditions after assessment of the mCherry fluorescence of the mCherry-Ca_Vα2δ constructs. Patch clamp experiments were carried out with the Axopatch 200-B amplifier (Molecular Devices, Union City, CA). Electrodes were filled with a solution containing (in mm) 140 CsCl, 0.6 NaGTP, 3 MgATP, 10 EGTA, 10 HEPES, titrated to pH 7.3 with NaOH. Pipette resistance ranged from 2 to 4 megohms. Cells were bathed in a modified Earle's saline solution (in mm) as follows: 135 NaCl, 20 tetraethylammonium chloride, 2 CaCl₂, 1 MgCl₂, 10 HEPES, titrated to pH 7.3 with KOH. PClamp software Clampex 10.2 coupled to a Digidata 1440A acquisition system (Molecular Devices) was used for on-line data acquisition and analysis. Pipette and cell capacitance cancellation and series resistance compensation were applied (up to 80%) using the cancellation feature of the amplifier. Cellular capacitance was estimated by measuring the time constant of current decay evoked by a 10-mV depolarizing pulse applied to the cell from a holding potential of −100 mV. A series of 450-ms voltage pulses were applied from a holding potential of −100 mV at a frequency of 0.2 Hz, from −60 to +70 mV at 5-mV intervals. Unless stated otherwise, data were sampled at 5 kHz and filtered at 1 kHz. Experiments were performed at room temperature (20–22 °C). Activation parameters were estimated from the peak I-V curves obtained for each channel combination and are reported as the mean of individual measurements ± S.E. as described elsewhere (37). Briefly, the I-V relationships were normalized to the maximum amplitude and were fitted to a Boltzmann equation with E_{0.5, act} being the mid-potential of activation. The free energy of activation was calculated using the mid-activation potential shown in Equation 1,

where z is the effective charge displacement during activation, and F is the Faraday constant (38). The r100 ratio is defined as the ratio of peak whole-cell currents remaining after a depolarizing pulse of 100 ms (I_{100 ms}/I_peak) and was used as an indicator of the inactivation kinetics.

Unless specified otherwise (as in Fig. 1), patch clamp experiments were performed with the untagged version of pCMV-Ca_V1.2 WT transfected with pmCherry-Ca_Vα2δ1-HA WT or mutant in stable Ca_Vβ3 cells. Each novel pmCherry-Ca_Vα2δ1-HA mutant was always tested alongside the control condition (pCMV-Ca_V1.2 WT + pmCherry-Ca_Vα2δ1-HA WT in stable Ca_Vβ3 cells) to assess for internal consistency. Experiments performed under the same conditions yielded peak current densities ± 20% between samples and between series of experiments. All experiments were pooled and biophysical properties are reported in Table 1.

TABLE 1.

Biophysical properties of Ca_V1.2/ Ca_Vβ3 channels with and without Ca_Vα2δ1

Ca_V1.2 WT was expressed in stable Ca_Vβ3 cells with pmCherry-no insert or pmCherry-Ca_Vα2δ construct under typical conditions with a 1:1 DNA ratio (4 μg: 4 μg). Biophysical parameters were measured in the presence of 2 mm Ca²⁺ as described elsewhere (19). Activation properties (E_{0.5, act} and ΔG_act) were estimated from the mean I-V relationships and fitted to a Boltzmann equation. The data are shown with the mean ± S.E. of the individual experiments, and the number of experiments is in parentheses. ND means not determined as the signal-to-noise ratio was too small in the absence of Ca_Vα2δ1.

CaV1.2 WT in CaVβ3 stable HEKT cells with 2 mm Ca2+	Electrophysiological properties
	E0.5, act	ΔGact	Peak density	r100 at +5 mV
	mV	kcal mol−1	pA/pF
No CaVα2δ1-HA (pmCherry-no insert)	4 ± 1 (26)	0.5 ± 0.1 (26)	−5.4 ± 0.7 (26)	ND
+ CaVα2δ1 WT no HA (1:1 ratio)	−10.1 ± 0.5 (23)	−0.86 ± 0.05 (23)	−56 ± 3 (23)	0.31 ± 0.02 (23)
+ CaVα2δ1-HA WT (1:1 ratio)	−10.6 ± 0.2 (163)	−0.90 ± 0.03 (163)	−67 ± 3 (163)	0.38 ± 0.02 (163)
+ CaVα2δ1-HA D550Y (1:1 ratio)	−10.9 ± 0.7 (10)	−0.9 ± 0.1 (10)	−61 ± 9 (10)	0.33 ± 0.01 (10)
+ CaVα2δ1-HA S709N (1:1 ratio)	−11.3 ± 0.6 (10)	−0.9 ± 0.1 (10)	−49 ± 10 (10)	0.37 ± 0.02 (10)
+ CaVα2δ1-HA S755T (1:1 ratio)	−11.4 ± 0.4 (20)	−1.0 ± 0.0 (20)	−73 ± 10 (20)	0.29 ± 0.02 (20)
+ CaVα2δ1-HA Q917H (1:1 ratio)	−11.4 ± 0.7 (10)	−1.0 ± 0.1 (10)	−69 ± 9 (10)	0.37 ± 0.02 (10)
+ CaVα2δ1-HA S956T (1:1 ratio)	−10.8 ± 0.6 (10)	−0.9 ± 0.1 (10)	−84 ± 9 (10)	0.33 ± 0.02 (10)
+ CaVα2δ1-HA D550Y/ Q917H (1:1 ratio)	−9 ± 2 (10)	−0.8 ± 0.2 (10)	−40 ± 10 (10)	0.35 ± 0.03 (10)

Statistics

Results were expressed as mean ± S.E. Tests of significance were carried out using the unpaired analysis of variance test embedded in the Origin 7.0 analysis software (OriginLab Corp., Northampton, MA). Data were considered statistically significant at p < 0.05.

RESULTS

HA-tagged Ca_Vα2δ Stimulates Ca_V1.2 Currents

Co-expression of Ca_V1.2 and Ca_V2.2 with the auxiliary Ca_Vα2δ subunit was shown to stimulate whole-cell currents (19, 39). Functional modulation of Ca_V1.2 by the various Ca_Vα2δ constructs was studied after recombinant expression in HEKT cell stably transfected with Ca_Vβ3 (Fig. 1A) as explained earlier (19). The immunoreactivity and the integrity of the constructs were verified by Western blotting (data not shown). As shown in Fig. 1A, whole-cell currents, recorded in the presence of a saline solution containing a physiological concentration of Ca²⁺ (2 mm), were significantly larger when measured in the presence of the Ca_Vα2δ confirming that Ca_Vα2δ stimulates whole-cell currents of Ca_V1.2/Ca_Vβ3 (19). Peak current densities increased from −5 ± 1 pA/pF (n = 26) (no insert in the pmCherry vector) to −56 ± 3 pA/pF (n = 23) in the presence of Ca_Vα2δ WT (Fig. 1B). The increase in peak current densities was associated with a −15-mV leftward shift in the activation potential of Ca_V1.2 from E_{0.5, act} = 5 ± 2 mV (n = 26) (no Ca_Vα2δ) to E_{0.5, act} = −10.1 ± 0.5 mV (n = 23) (with Ca_Vα2δ). Whole-cell currents recorded with untagged Ca_V1.2 WT + mCherry Ca_Vα2δ-HA or with HA-tagged Ca_V1.2 + mCherry Ca_Vα2δ-HA were not significantly different from currents recorded with untagged Ca_V1.2 WT + mCherry Ca_Vα2δ WT. This result validates the constructions and confirms that the HA epitopes inserted in extracellular loops of both Ca_Vα2δ and Ca_V1.2 are not occluding each other. The free energy of activation (ΔG_act) measured in the presence of Ca_Vα2δ was well described by a Gaussian distribution centered around −0.86 kcal mol⁻¹ (Fig. 1C), whereas the ΔG_act measured in the absence of Ca_Vα2δ displayed a broader distribution centered at a value close to ≈0 kcal mol⁻¹. These results are compatible with a model whereby Ca_Vα2δ stimulates peak current density by setting Ca_V1.2 channels in a conformational state very close to the open state (40).

Ca_Vα2δ Improves Total but Not Surface Expression of Ca_V1.2

To evaluate the impact of Ca_Vα2δ on the protein expression of Ca_V1.2, the HA-tagged version of Ca_V1.2 was expressed in HEKT cells and in stable Ca_Vβ3 cells in the presence or absence of Ca_Vα2δ (Fig. 2). Flow cytometry assays were carried out in the presence of the FITC-conjugated anti-HA in intact and in permeabilized cells. Control experiments carried out with a pCMV-Ca_V1.2 control construct that was not HA-tagged, confirmed the specificity of the FITC antibody in these series of experiments (data not shown). Given that the HA epitope is located in the extracellular portion of the protein, the fluorescence intensity for FITC obtained in the presence of intact cells reflects the cell surface density of Ca_V1.2. Fluorescence for FITC was measured after cell permeabilization to confirm the accessibility of the HA epitope. The fluorescence histograms are reported in Fig. 2, A and B, and the averaged mean fluorescence intensities are shown in Fig. 2D. Ca_Vα2δ alone improved total protein expression of Ca_V1.2 without any significant change in the cell surface density of Ca_V1.2. This contrasts with the impact of Cavβ (Fig. 2D) that improves both cell surface and total protein density of Ca_V1.2 (20). In particular, Ca_Vβ3 increased by ≈200% the cell surface density of Ca_V1.2, whereas Ca_Vα2δ barely stimulated cell surface density by 2 ± 3% (n = 3). Co-expression with both Ca_Vβ3 and Ca_Vα2δ did not increase cell surface detection of Ca_V1.2, although total protein stability was improved (Fig. 2, A–D). The former results contrast with confocal imaging data showing that co-expression with both Ca_Vα2δ and Ca_Vβ1b is required to achieve the maximal cell surface staining for Ca_V2.2 (39). We had previously reported similar results with an HA-tagged Ca_V1.2 construct bearing its HA epitope in domain I (19).

The flow cytometry data were validated by cycloheximide chase analysis (Fig. 3, A–C). Ca_Vβ3 increased the protein expression of Ca_V1.2 in total lysates, whereas Ca_Vα2δ alone did not significantly alter protein density of Ca_V1.2. As seen in flow cytometry assays (Fig. 2D), total protein expression of Ca_V1.2 was also increased in the combined presence of Ca_Vβ3 and Ca_Vα2δ auxiliary subunits (Fig. 3, D and E), as also reported by others (20).

Ca_V1.2/Ca_Vβ3 Stabilizes the Expression of Ca_Vα2δ

Ca_Vβ subunits promote the cell surface density of Ca_V1.2, but little is known about their role on the trafficking of Ca_Vα2δ. To investigate the trafficking of Ca_Vα2δ, the pmCherry-Ca_Vα2δ-HA construct was expressed alone in HEKT and in stable Ca_Vβ3 cells in the presence or absence of Ca_V1.2. Two-color flow cytometry experiments were carried out in intact and permeabilized cells to respectively assess cell surface expression and total protein expression, the latter including cell surface and intracellular proteins. In two-dimensional plots, the fluorescence intensity obtained for mCherry was plotted against the fluorescence intensity for FITC. As seen in intact nonpermeabilized cells, the fluorescence intensity was strong for both mCherry and FITC indicating that Ca_Vα2δ is clearly present at the cell membrane (Fig. 4, A and C). Co-expression with Ca_Vβ3 alone did not significantly alter either surface or total expression, whereas co-expression with Ca_V1.2 only slightly improved cell surface and total protein expression of Ca_Vα2δ. However, co-expression with both Ca_V1.2 and Ca_Vβ3 nearly doubled the total protein expression of Ca_Vα2δ but not its cell surface density. It is unlikely that the latter result was due to an unfavorable change in the conformation of the Ca_Vα2δ-HA construct because Ca_V1.2 whole-cell currents were clearly up-regulated by Ca_Vα2δ-HA in patch clamp experiments (see Fig. 1). The relative increase in total protein expression of Ca_Vα2δ was similar whether it was inferred from the constitutive mCherry fluorescence of mCherry-Ca_Vα2δ-HA (data not shown) or from the FITC fluorescence in permeabilized cells (Fig. 4D). These results suggest that Ca_Vα2δ interacts intracellularly with Ca_V1.2 and that the stability of the protein complex is improved by Ca_Vβ3, probably in part driven by the nanomolar interaction between Ca_V1.2 and Ca_Vβ3 (41). Nonetheless, because the increase in total protein density did not translate into a surge in the cell surface density of Ca_Vα2δ, it also suggests that most of the Ca_V1.2 protein complex remains in intracellular compartments.

Channel Modulation Depends upon the Cell Surface Expression of Ca_Vα2δ

The correlation between the protein expression of Ca_Vα2δ and the modulation of Ca_V1.2 channels was quantified by co-expressing different DNA ratios of Ca_Vα2δ and Ca_V1.2. The ΔMFI values for FITC measured in intact and permeabilized cells increased steeply in the range from 1:20 to a 1:2 DNA ratio. Both the cell surface and the total protein densities followed a similar pattern (Fig. 5). Under our experimental conditions, the 1:1 DNA ratio yielded an additional 20% increase in the protein density, but raising further the relative amount of DNA coding for Ca_Vα2δ was found to impair transfection efficiency (data not shown). Patch clamp experiments established that increasing the cell surface expression of Ca_Vα2δ improves peak current density and activation gating of Ca_V1.2 (Fig. 6). Peak current density increased as a function of the cell surface density of Ca_Vα2δ within the range of our experimental conditions (Fig. 7). This suggests that increasing the cell surface density of Ca_Vα2δ improves the macroscopic activation gating of Ca_V1.2 by promoting the channel activation at more negative potentials.

Cell Surface Density Is Impaired in Ca_Vα2δ Genetic Variants

To gain further insight into the molecular mechanism underlying channel modulation by Ca_Vα2δ, we turned to mutations of Ca_Vα2δ associated with cardiac arrhythmias. Burashnikov et al. (29) discovered 23 rare missense variants in three genes encoding subunits forming the Ca_V1.2 L-type calcium channel in 205 patients diagnosed with “J-wave syndromes.” Four genetic variants were identified in the CACNA2D1 gene encoding Ca_Vα2δ1. The mutation S709N was found in two unrelated patients. Two mutations (D550Y and Q917H) were identified in the same individual (29), but their expression profile has yet to been fully characterized (32). Ca_Vα2δ missense mutants D550Y, S709N, Q917H, S956T, the double D550Y/Q917H (29), as well as Ca_Vα2δ S755T mutant associated with SQTS6 (31) were expressed alone or in combination with Ca_V1.2 and Ca_Vβ3, the same subunit composition we used for functional characterization in patch clamp experiments. These six Ca_Vα2δ mutations were co-expressed with Ca_V1.2 WT (no HA) in the maximum 1:1 ratio in stable Ca_Vβ3 cells. The transfection efficiency (as assessed by the number of fluorescent cells) was found to be not significantly different between the HA-tagged Ca_Vα2δ WT and the HA-tagged Ca_Vα2δ mutant constructs. Overall, the total protein density of all mutants estimated from permeabilized cells or from the constitutive mCherry fluorescence (data not shown) significantly increased in the combined presence of Ca_V1.2 and Ca_Vβ3. As shown in Figs. 8 and 9, the ΔMFI for FITC of Ca_Vα2δ S755T, Q917H, and S956T mutants in intact cells under these conditions was similar to the ΔMFI measured for Ca_Vα2δ-HA WT suggesting that the cell surface density of Ca_Vα2δ was not affected by these single mutations. In contrast, missense mutations D550Y and S709N and more significantly the double mutant D550Y/Q917H impaired the cell surface targeting of Ca_Vα2δ-HA with ≈30% reduction for the former mutants and ≈60% decrease for the latter. Doubling the amount of cDNA used for transfection did not improve cell surface expression because the transfection efficiency decreased with larger DNA concentrations. The surface density of D550Y/Q917H proteins remained on average 60% lower than the one measured for the Ca_Vα2δ-HA WT protein at each DNA ratio tested from 1:20 (data not shown) to 1:1. Flow cytometry assays carried out 36 h after transfection yielded similar results suggesting that a 24-h culture time is sufficient to observe the optimal protein expression (data not shown). The fluorescence intensities for FITC in permeabilized cells and for the constitutive mCherry were only decreased by ≈30% when compared with the signals measured for Ca_Vα2δ-HA WT. These data indicate the following: (a) HA tag remained accessible in the double mutant; (b) the cell and the total protein density were both affected by the D550Y/Q917H double mutation albeit to a different extent. Altogether, these data suggest that the double mutation impairs the trafficking of Ca_Vα2δ to a greater extent than protein stability.

Channel Modulation Is Altered in the Presence of Ca_Vα2δ Genetic Variants

The functional impact of the missense Ca_Vα2δ mutations was characterized by electrophysiology after co-expression of Ca_V1.2 and Ca_Vα2δ in stable Ca_Vβ3 cells. In the presence of a 1:1 ratio, the five single point mutations, including D550Y and S709N, increased Ca_V1.2 peak current densities by 10–13-fold with a negative shift in the voltage-dependent gating activation (ΔG_act ≈ −1 kcal mol⁻¹) in a fashion reminiscent of Ca_Vα2δ WT (Table 1). Hence, Ca_Vα2δ D550Y and S709N mutations boosted the peak current densities despite a 30% decrease in the cell surface expression of Ca_Vα2δ. These results are in agreement with Fig. 7 showing that an ≈30% decrease in the cell surface density of Ca_Vα2δ is not sufficient to significantly prevent the up-regulation of macroscopic Ca_V1.2 currents. In contrast, co-expression of Ca_Vα2δ-HA D550Y/Q917H produced currents that were 35% lower than Ca_Vα2δ-HA WT with an average of −40 ± 10 pA/pF (n = 10) compared with −67 ± 3 pA/pF (n = 163) (Fig 10). The experimental variation, however, limits the statistical significance to p < 0.5. Nonetheless, there seems to be a trend toward a decreased function that was also reported by the authors of the original paper (29). We hypothesized that the decrease in the cell surface density of the Ca_Vα2δ D550Y/Q917H mutant could become more significant when expressing Ca_Vα2δ/Ca_V1.2 in a 1:20 ratio. Patch clamp experiments were thus carried out with cDNA ratios of 1:1, 1:8, and 1:20 Ca_Vα2δ/Ca_V1.2 (0.2:4 μg of DNA) (Fig 11). As shown, the behavior of the double mutant departed more significantly from the wild-type protein at lower cDNA concentrations. It is interesting to note that the decrease in peak current density was accompanied by a rightward shift in the activation potentials. Altogether, our results suggest that some Ca_Vα2δ arrhythmogenic mutations may decrease Ca_V1.2 currents in cardiomyocytes as a result of their decreased cell surface density.

DISCUSSION

Ca_Vα2δ Improves the Stability of Ca_V1.2-Ca_Vβ Complexes

It is generally agreed that recombinant expression of Ca_Vα2δ boosts peak current density of Ca_V1.2 in recombinant cells (19) as well as in native mouse cardiomyocytes (21). Whether this modulation is conveyed through an increase in the cell surface density of Ca_V1.2 remains debated (19, 20, 42) in part because the macroscopic peak current density is the product of three parameters as follows: the number of channels in the plasma membrane, the open channel probability, and the single-channel conductance. We have opted to investigate changes in the channel surface density using a sensitive and high throughput flow cytometry fluorescence-based assay. In this assay, the cell samples go through a gating process that excludes dead cells, debris, and aggregates such that the fluorescence intensity reflects protein expression in intact cells of similar morphological properties. This approach required the insertion of a 9-residue HA epitope in the extracellular face of the Ca_V1.2 pore-forming subunit. Such manipulation may interfere with protein expression and/or interaction with other subunits as shown very elegantly in Ref. 39. Indeed, we have tested many constructs before identifying one that preserves the up-regulation of Ca_V1.2 currents by Ca_Vα2δ. As shown in the first series of experiments, whole-cell currents measured with the HA-tagged version of Ca_V1.2 were functionally up-regulated by the tagged and the untagged versions of Ca_Vα2δ. Using the untagged version of Ca_Vα2δ, we evaluated the changes in the surface and the total protein expression of HA-Ca_V1.2 from the relative fluorescence of the FITC-conjugated antibody in intact and permeabilized cells. Total protein density was also measured using Western blotting. Altogether, our results show the following: (a) Ca_Vα2δ alone, unlike Ca_Vβ (20, 35), does not prevent the degradation or improve the cell surface density of Ca_V1.2 proteins; and (b) in the presence of Ca_Vβ3, Ca_Vα2δ improves the stability and/or the total protein expression of Ca_V1.2 proteins. Hence, we show that Ca_Vα2δ up-regulates Ca_V1.2 currents by improving the channel activation gating without significantly increasing the number of Ca_V1.2 proteins at the cell surface.

Recombinant Ca_Vα2δ Is Strongly Expressed at the Membrane

The trafficking of Ca_Vα2δ was studied using a similar fluorescence-based assay, but surface and total protein density were tracked by the external HA epitope and the intracellular mCherry tag, respectively. The constitutive mCherry fluorescence was used as an index of Ca_Vα2δ expression under all conditions for the wild-type protein as well as the Ca_Vα2δ mutants (see below). Patch clamp experiments carried out with the untagged version of Ca_V1.2 but using the HA-tagged mCherry Ca_Vα2δ constructs (WT and mutants) confirmed the latter construct was functional. It is essential to stress that fluorescence cell sorting experiments (protein expression) and whole-cell recordings (channel function) were carried out with cells grown under the same conditions. These data validated that the double modification, insertion of the HA epitope and the mCherry protein, did not prevent functional modulation of the L-type currents by Ca_Vα2δ. Furthermore, the accessibility of the HA epitope was validated in each flow cytometry assay by comparing the relative fluorescence intensities obtained with the FITC-conjugated antibody in permeabilized cells with the constitutive mCherry in intact and permeabilized cells.

Ca_Vα2δ was shown to reach the cell surface when expressed alone (our data in HEKT cells and see Ref. 42). The strong fluorescence intensity suggests that the cell surface density of Ca_Vα2δ is larger than the cell surface density of Ca_V1.2 expressed under similar conditions. Similar observations were made in mouse cardiomyocytes showing that endogenous Ca_Vα2δ is significantly expressed at the plasma membrane (Fig 12). Co-expression with Ca_Vβ3 did not appreciably alter the cell surface density of Ca_Vα2δ in contrast to what we and others have reported for the Ca_Vα1 subunit of Ca_V1.2 (19). Nonetheless, our study highlighted some reciprocal modulation within the proteins forming the L-type Ca_V1.2 channel. Indeed, the total protein expression of Ca_Vα2δ significantly improved in the combined presence of Ca_V1.2 and Ca_Vβ3. The significant increase of total protein expression in the combined presence of Ca_V1.2 and Ca_Vβ3 measured from the relative fluorescence for FITC in permeabilized cells did not result from an improved accessibility of the HA epitope because the constitutive mCherry fluorescence as well as the total protein density measured in Western blots were similarly augmented. Furthermore, our results show that the modulation of L-type currents is a quasi-linear function of the cell surface density of Ca_Vα2δ. It suggests that the interaction between Ca_V1.2 and Ca_Vα2δ is conveyed through one or multiple low affinity binding site(s).

FIGURE 12.

Co-localization of endogenous Ca_V1.2 and Ca_Vα2δ at the plasma membrane in 24-h post-culture myocytes. Ventricles of 1–3-day-old CD1 mice were digested with collagenase type II (Worthington). Cardiomyocytes were collected by centrifugation and cultured at 37 °C in a 5% CO₂ incubator for 24 h in M199 culture medium (Invitrogen) containing 15% fetal bovine serum. A, nuclei were stained with DAPI (1:1000) (Invitrogen). Cultured cells were incubated with wheat germ agglutinin (WGA) 647 (1:200) (Invitrogen) for 10 min and fixed in 2% paraformaldehyde at room temperature for 15 min followed by a final incubation with 0.1% Triton X-100 for 60 min. After overnight incubation at 4 °C with either primary anti-Ca_V1.2 (1:100) (Alomone Labs) and anti-Ca_Vα2δ (1:50) (Santa Cruz Biotechnology) antibodies in 1% donkey serum with 0.05% Triton X-100, cells were incubated with Alexa 555 for Ca_V1.2 (1:800) and Alexa 488 for Ca_Vα2δ (1:800) secondary antibodies (Invitrogen) for 90 min at room temperature. Confocal fluorescent images were captured with a Zeiss LSM 710 confocal microscope system with ×63/1.40 oil objective. Scale bar corresponds to 10 μm. The images were analyzed using FIJI software. B, red channel (either Ca_V1.2 or Ca_Vα2δ) and the green channel (wheat germ agglutinin) are used to produce co-localization pixel maps using co-localization finder plugin. Co-localization is shown in white when the pixels for the green (wheat germ agglutinin) and the red channels (either Ca_V1.2 (top) or Ca_Vα2δ (bottom)) merge. Pearson's correlation coefficient were calculated from six cells using JACoP plugin (45). The Pearson's correlation coefficient values are expected to vary between +1 and −1, with +1 indicating a perfect correlation. The analysis was validated by further establishing that the Costes p value is 100% indicating that co-localization in the regions masked in white is highly probable. In this series of experiments, the Pearson's correlation coefficient values were 0.41 ± 0.03 (n = 6) for Ca_V1.2 and 0.72 ± 0.01 (n = 6) for Ca_Vα2δ indicating that both proteins are found at the plasma membrane in cultured cardiomyocytes.

Cardiac Arrhythmias and Ca_Vα2δ

In the context where the numbers of channelopathy susceptibility genes and mutations identified are increasing rapidly, there is renewed interest in characterizing the molecular mechanisms underlying modulation of L-type Ca_V1.2 channels by its auxiliary subunits, in particular Ca_Vα2δ. In return, structure-function insights gleaned from these disease susceptibility genes have significantly advanced our understanding of the pathophysiology of these syndromes and have paved the way to the development of new treatment strategies. Over the last few years, single missense mutations and/or genetic variants of Ca_Vα2δ have been identified in patients experiencing cardiac arrhythmias associated with repolarizing QT interval (SQT) shorter than normal. Short QT arrhythmias are generally associated with gain-of-function outward currents (potassium currents) or loss-of-function of inward (sodium or calcium) currents. At the time of their identification, the missense mutations in Ca_Vα2δ have been labeled loss-of-function mutations and were proposed to alter L-type Ca_V1.2 peak current density through a change in the membrane expression of Ca_V1.2 (29). Having established a reliable experimental model to define the surface expression and the function of the different components of the L-type Ca_V1.2 channel, we explored the impact of Ca_Vα2δ mutations associated with sudden cardiac death. Recombinant expression in HEKT cells of Ca_Vα2δ single mutants S755T, Q917H, and S956T associated with shorter QT intervals did not cause any significant change in the cell surface expression of Ca_Vα2δ or in the function of Ca_V1.2 currents when expressed using a 1:1 cDNA ratio. This is true also for the mutation Ca_Vα2δ S755T that was reported by others to prevent the up-regulation of Ca_V1.2 currents by Ca_Vα2δ in HEKT cells (31). In contrast, we have observed a, ≈30% decrease in the cell surface density of Ca_Vα2δ D550Y and S709N single mutants accompanied with an ≈15% in the total protein density that remained unchanged up to 36 h after transfection. Despite causing a reduction in the cell surface expression of Ca_Vα2δ, the single mutants D550Y and S709N did not significantly impair the modulation of Ca_V1.2 currents when co-expressed in a 1:1 cDNA ratio. Combining two mutations in Ca_Vα2δ with D550Y/Q917H significantly reduced the cell surface density of Ca_Vα2δ (60%) under the same conditions. Although both the cell surface and the total protein density of the double mutant were reduced, the decrease in the cell surface expression was more important than the reduction in the total protein density. Altogether, these data suggest that the trafficking of Ca_Vα2δ was significantly disrupted by the double mutation. The decrease in the expression of Ca_Vα2δ caused a significant 40% reduction in the peak current density of Ca_V1.2 when co-expressed with Ca_Vα2δ WT in stable Ca_Vβ3 cells. Hence, the combination of two variants that are relatively silent polymorphisms on their own could significantly impair the trafficking of Ca_Vα2δ and consequently reduce L-type Ca_V1.2 currents.

Conclusion

Current genetic studies are placing the homeostasis of Ca²⁺ as a central modulator of cardiac repolarization with genes such as CACNA1C as well as ATP2A2, PLN, PRKCA, SRL, and SLC8A1 (2). Genetic studies have reported a decreased expression of Ca_Vα2δ transcripts in patients suffering from atrial (43) and ventricular fibrillation (44). Our current results confirm that mutations in CACNA2D1 could be a contributing factor in cardiac sudden death associated with a short QT interval. It is important to note that the strongest reduction in the cell surface density of Ca_Vα2δ was observed by combining two genetic variants that had little impact when tested individually. Hence, the role of polymorphisms in CACNA2D1 is not to be neglected. Missense and/or truncation mutations of Ca_Vα2δ could severely impair Ca_V1.2 currents providing that the said mutation significantly decreases the cell surface protein expression of Ca_Vα2δ.