Adrenergic Ca_V1.2 Activation via Rad Phosphorylation Converges at α_1C I-II Loop

Abstract

Rationale:

Changing activity of cardiac Ca_V1.2 channels under basal conditions, during sympathetic activation, and in heart failure is a major determinant of cardiac physiology and pathophysiology. Although cardiac Ca_V1.2 channels are prominently up-regulated via activation of protein kinase A, essential molecular details remained stubbornly enigmatic.

Objective:

The primary goal of this study was to determine how various factors converging at the Ca_V1.2 I-II loop interact to regulate channel activity under basal conditions, during β-adrenergic stimulation, and in heart failure.

Methods and Results:

We generated transgenic mice with expression of Ca_V1.2 α_1C subunits with: 1) mutations ablating interaction between α_1C and β subunits; 2) flexibility-inducing polyglycine substitutions in the I-II loop (GGG-α_1C); or 3) introduction of the alternatively spliced 25-amino acid exon 9* mimicking a splice variant of α_1C up-regulated in the hypertrophied heart. Introducing three glycine residues that disrupt a rigid IS6-AID helix markedly reduced basal open probability despite intact binding of Ca_Vβ to α_1C I-II loop, and eliminated β-adrenergic agonist stimulation of Ca_V1.2 current. In contrast, introduction of the exon 9* splice variant in α_1C I-II loop, which is increased in ventricles of patients with end-stage heart failure, increased basal open probability but did not attenuate stimulatory response to β-adrenergic agonists when reconstituted heterologously with β_2B and Rad or transgenically expressed in cardiomyocytes.

Conclusions:

Ca²⁺ channel activity is dynamically modulated under basal conditions, during β-adrenergic stimulation, and in heart failure by mechanisms converging at the α_1C I-II loop. Ca_Vβ binding to α_1C stabilizes an increased channel open probability gating mode by a mechanism that requires an intact rigid linker between the β subunit binding site in the I-II loop and the channel pore. Release of Rad-mediated inhibition of Ca²⁺ channel activity by β-adrenergic agonists/PKA also requires this rigid linker and β binding to α_1C.

Graphical Abstract

INTRODUCTION

In cardiomyocytes, Ca²⁺ influx through L-type Ca_V1.2 channels commences the process of excitation-contraction coupling via the triggering of Ca²⁺ release from ryanodine receptors. In the failing heart, dysfunctional regulation of Ca_V1.2 channels can trigger electrical abnormalities leading to early Ca²⁺-mediated after-depolarizations, arrhythmias, and sudden death.

Voltage-gated Ca²⁺ channels are comprised of a pore-forming α₁ subunit ¹ and a cytosolic β subunit that interacts with the α-interaction domain (AID) in the intracellular linker between domains I and II (I-II loop) of the α₁ subunit ^2–4 (Figure 1A). When expressed heterologously, binding to the β subunit is obligatory for α_1C trafficking to the plasma membrane, and for normalizing channel activation and inactivation gating properties ^5–8 via a mechanism that requires a rigid IS6-AID helix linker ^{9, 10}. β-adrenergic agonists, via activation of protein kinase A (PKA), increase Ca²⁺ influx through Ca_V1.2 ^{11, 12}, an important component of the physiological ‘fight-or-flight’ response that contributes to the increased contractility of the heart during exercise. The overall features of this regulation − PKA-dependent enhanced whole-cell current amplitude, a hyperpolarizing shift in voltage-dependence of channel activation, and increased open probability (P_o) − are well-established ^{11, 12}, yet essential molecular details remained stubbornly enigmatic for decades. Recently, we showed that binding of Ca_Vβ subunits to α_1C is required for β-adrenergic stimulation of Ca_V1.2 channels and positive inotropy in the heart ¹³, and that the Ca²⁺ channel inhibitor Rad ¹⁴, which binds to the β subunit, is the functionally relevant PKA target in the Ca_V1.2 complex ¹⁵.

Figure 1. Electrophysiological properties of pseudo-wild-type and AID-mutant Ca²⁺ channels.

(A) Schematic of rabbit cardiac α_1C subunit topology showing β subunit binding to α-interaction domain (AID) motif, and the position of the 9* exon in I-II loop. Rad interaction with both β subunit and the plasma membrane are shown. WT, mutant GGG, and mutant AID motif in the I-II loop of α_1C. (B) Diagrams showing the binary transgene system that allow expression of FLAG-DHP-resistant (DHP*) α_1C only when both reverse tetracycline-controlled transactivator (rtTA) (top diagram) and doxycycline are present (Tet-ON). The lower diagram shows cDNA for FLAG-DHP-resistant (DHP*) α_1C ligated behind seven tandem tetO sequences. (C) Anti-FLAG (upper) and anti-β immunoblots (lower) of anti-FLAG antibody immunoprecipitation of cardiac homogenates of non-transgenic (NTG), pseudo-wild-type α_1C, GGG-α_1C and AID-mutant α_1C mice. Representative of 3 experiments. (D-F) Single channel Ba²⁺ currents are shown. Channel closures are labeled “c” and openings are downward deflections to the open level (slanted gray curves, labeled “o”) in the absence of nisoldipine (top 2 rows) and presence of nisoldipine (bottom 2 rows). Bottom: P_o versus voltage relationship, averaged over multiple patches. N= 5, 7, 6, from left to right. (G) Graph of P_o for Ca²⁺ channels recorded from non-transgenic (NTG), pseudo-wild-type (pWT) α_1C and AID-mutant α_1C cardiomyocytes. Kruskal-Wallis P= 4 × 10⁻⁴; Dunn’s multiple comparison test P-values in panel. (H-I) Histograms show distribution of single-trial average P_o obtained from DHP-resistant one-channel patches from pseudo-wild-type α_1C and AID-mutant cardiomyocytes.

Beyond these, the I-II loop is also subject to alternative splicing that tunes channel function and interacting proteins in a cell-type specific manner. The inclusion of alternatively-spliced exon 9* is observed at high levels in the smooth muscle and at lower but variable expression in adult heart that increases in animal models of hypertrophy and in the peri-infarct zone after myocardial infarction in mice ^16–18. This variant results in the insertion of 25 amino acid residues C-terminal to the AID ^18–20 and tunes channel activation. In all, the modulatory landscape supported by the Ca_V1.2 domain I-II linker appears rich and multifaceted, involving the β subunits, RGK proteins, phosphoregulation by PKA, and alternative splicing, all poised to precisely tune Ca²⁺ influx into cardiomyocytes.

Several important mechanistic unknowns persist impeding in-depth pathophysiological understanding. First, although Ca_Vβ – α₁ interaction is obligatory for Ca_V1.2 trafficking in heterologous cells, we found it to be dispensable for trafficking to the dyad, for basal function and for initiating excitation-contraction coupling in adult cardiomyocyte ¹³, thus raising fundamental questions about the functional role of Ca_Vβ subunits in cardiomyocytes. Second, it is unknown how distal conformational changes involving Rad interaction with the Ca_Vβ subunit and phosphorylation-dependent signaling are ultimately conveyed to the channel pore-domain. Third, how alternative splicing of the domain I-II linker contributes to this overall regulatory scheme including downstream effects of PKA activation remains to be fully-elucidated. Importantly, the pathogenesis of heart failure has been long-suspected to reshape this regulatory framework, although the precise changes remain largely undefined.

To dissect these possibilities, we measured baseline channel gating properties and the strength of adrenergic modulation of Ca_V1.2 in three transgenic mouse models: (1) Our previously-established AID mutant where Ca_V1.2 is incapable of binding to the β subunit ¹³, (2) Triple-glycine substitution (GGG-α_1C) of the rigid I-II linker, which connects the pore-domain with the AID. These mutations have been previously shown to disrupt coupling between these two domains ^{9, 10, 21}, and (3) The insertion of 25 amino acid residues C-terminal to the AID corresponding to the introduction of exon 9* variant (9*-α_1C) (Figure 1A). Our findings identify the IS6-AID linker as a vital molecular element for transducing the PKA-induced activation of Ca_V1.2, and as a molecular rheostat for Ca_V1.2 activity whereby distinct structural changes elicited by β subunits, RGK proteins, and alternative splicing bidirectionally tune Ca²⁺ influx into the heart in both normal physiology and during heart failure.

METHODS

Data Availability.

All supporting data are available within the article and its online data supplement.

For details on the experimental procedures, see the materials and methods section in the online data supplement.

RESULTS

Generation of inducible, cardiac-specific α_1C transgenic mice.

We created several mice lines with inducible, cardiac-specific expression of dihydropyridine (DHP)-resistant, FLAG-epitope-tagged α_1C (Figure 1B). We have previously shown that control transgenic FLAG-tagged DHP-resistant α_1C subunits, termed pseudo-wild-type α_1C have similar properties as native cardiac Ca²⁺ channels ²². To study properties of Ca_V1.2 devoid of the β subunit, we used transgenic mice¹³ expressing rabbit α_1C with a disrupted AID via alanine substitutions of 3 conserved residues, Y467, W470 and I471, that are essential for binding of β subunit ²¹. We further generated transgenic mice (Figure 1B) expressing FLAG-tagged DHP-resistant α_1C with a substitution of three glycine residues (GGG) for the AKA motif in the IS6-AID linker ¹⁰, designated GGG-α_1C (Figure 1A). All three lines, pseudo-WT-, AID- and GGG-α_1C were crossed with transgenic mice expressing reverse transcriptional transactivator in the heart (αMHC-rtTA) ²³ (Figure 1B), yielding mice with doxycycline-inducible α_1C expression. As expected, channels with a disrupted AID motif do not bind β subunits, assessed by anti-FLAG antibody immunoprecipitation of cleared homogenates (Figure 1C). By comparison, GGG-α_1C channels exhibit robust β subunit binding similar to pseudo-wild-type α_1C (Figure 1C).

Binding of β to α_1C enhances channel openings.

Previously, we showed that in cardiomyocytes, Ca_V1.2 channels without β subunits can be transported to the dyad and can generate currents that mediate normal excitation-contraction coupling ¹³. To determine whether β binding to α_1C alters basal channel gating in cardiomyocytes, we utilized low-noise single-channel recordings of acutely isolated cardiomyocytes, employing Ba²⁺ as a charge carrier. Stochastic channel openings, which reflect near-steady-state open probability (P_o) at each voltage, were elicted by a slow voltage ramp ^{24, 25}. Ca²⁺ channels in cardiomyocytes from non-transgenic mice were inhibited by 300 nM nisoldipine (Figure 1D, middle). In the transgenic mice, however, there is a mixture of transgenic nisodipine-resistant channels and endogenous nisoldipine-sensitive channels (Figure 1E–F). As dihydropyridines are known to allosterically modify channel gating ²⁶, partial blockade of nisoldipine-resistant channels may be a confounding factor. To obviate this possibility, we obtain ~80–120 stochastic records from each patch and subsequently apply nisoldipine to identify resistant transgenic channels. This process allows us to unambiguously establish baseline function of mutant Ca_V1.2. Indeed, exemplar traces from DHP-resistant pseudo-wild-type α_1C channels confirm channel openings in the presence of nisoldipine. Measurements of steady-state P_o as a function of voltage were obtained by averaging many records and by normalizing the unitary current level. Reassuringly, steady-state P_o-V relationships of pseudo-wild-type α_1C channels are similar to that of non-transgenic channels. The DHP-resistant AID-mutant α_1C channels, however, displayed a striking 3.5-fold reduction in maximal P_o compared to DHP-sensitive endogenous Ca²⁺ channels from non-transgenic Ca²⁺ mice and DHP-resistant channels from pseudo-wild-type α_1C mice (Figure 1G). To further elucidate changes in elementary channel gating mechanisms, we scrutinized single-trial average open probabilities (${\overline{P}}_{\text{o}}$) from one-channel patches. A dash-line discriminator with ${\overline{P}}_{\text{o}}=0.1$ was used to identify low activity versus high-activity traces. Thus analyzed, pseudo-wild-type α_1C channels switched between epochs of no openings or blanks (40.6% of traces), low activity (33.0%), and high activity (26.4%) (Figure 1H) consistent with previous studies. By comparison, AID-mutant α_1C channels exhibited a distinct pattern (p<0.0001 by χ² test of independence) with rare sojourns to the high-activity mode (7.1%) and a higher propensity for blank (45.7%) and low activity sweeps (47.2%) (Figure 1I). Thus β subunit binding appears to stabilize the high-activity gating mode. We conclude that the interaction between the β subunit and the α_1C subunit essentially modulates Ca_V1.2 channel activity in cardiomyocytes by enhancing channel openings.

Increased flexibility of IS6 linker reduces basal Ca²⁺ channel activity in cardiomyocytes.

Having established the importance of β subunits in upregulating Ca_V1.2 channel activity, we considered whether the rigid I-II linker between the IS6 pore helix and the AID is essential for tuning channel function. Introduction of the three glycine residues in the IS6-AID linker had no effect on the subcellular localization and functional expression of Ca_V1.2 in cardiomyocytes. Anti-FLAG antibody immunofluorescence studies on fixed cardiomyocytes showed that GGG-α_1C Ca_V1.2 channels demonstrated a striated z-disk pattern consistent with localization to the surface membrane and transverse-tubules (t-tubules) (Figure 2A). Similar to cardiomyocytes expressing pseudo-wild-type α_1C or AID-mutant α_1C transgenic channels, field-stimulated contraction of cardiomyocytes isolated from GGG-α_1C transgenic mice persisted in the presence of 300 nM nisoldipine, which is sufficient to block excitation-contraction coupling induced by endogenous Ca_V1.2 channels in non-transgenic mice (Figure 2B), indicating that the GGG-α_1C Ca_V1.2 channels are localized correctly and flux sufficient Ca²⁺ to evoke Ca²⁺-induced Ca²⁺ release in cardiomyocytes.

Figure 2. Electrophysiological properties of GGG Ca²⁺ channels.

(A) Immunostaining of pseudo-wild-type (pWT) α_1C and GGG-α_1C cardiomyocytes. Anti-FLAG and Alexa 594-conjugated secondary antibodies, and nuclear labeling with Hoechst stain. Negative control omitted anti-FLAG antibody. Images obtained with confocal microscopy. Scale bar = 20 μm. (B) Percent contraction of sarcomere length in the presence of nisoldipine for cardiomyocytes isolated from NTG and GGG-α_1C mice. Cardiomyocytes were field-stimulated at 1-Hz. Unpaired t-test. (C) Graph of τ₁ and τ₂ inactivation for pseudo-wild-type α_1C (n=14 from 4 mice) and GGG-α_1C (n=26 from 3 mice) cardiomyocytes. Mann-Whitney test. Inset: Exemplar tracing of normalized Ca²⁺ current in response to a step depolarization to +10 mV. Fits were obtained using two exponentials by Clampfit. (D) Graph of conductance density-voltage relationship for nisoldipine-resistant Ca²⁺ channels recorded from pseudo-wild-type α_1C (n=22 from 4 mice) and GGG-α_1C (n=20 from 3 mice) cardiomyocytes. Mean ± SEM. P < 0.0001 by one-way ANOVA; Sidak’s multiple comparison test P-values in panel. (E) Single-channel Ba²⁺ currents in absence and presence of nisoldipine. Bottom: P_o versus voltage relationship averaged over multiple patches. N=6. (F) Graph of P_o for Ca²⁺ channels recorded from non-transgenic (NTG), pseudo-wild-type (pWT) α_1C, and GGG α_1C cardiomyocytes. NTG and pWT α_1C data are the same as in Figure 1. P= 4 × 10⁻⁴ by Kruskal-Wallis. Dunn’s multiple comparison test P-values in panel. (G) Histogram shows distribution of single-trial average P_o obtained from DHP-resistant one-channel patches from GGG mutant cardiomyocytes.

When GGG-α_1C is co-expressed in Xenopus oocytes with β_2B, one of the major β₂ isoforms in the heart, channels displayed accelerated voltage-dependent inactivation but slowed Ca²⁺-dependent inactivation ¹⁰. Here, we assessed aggregate inactivation of Ca²⁺ channels in the heart with Ca²⁺ as a charge carrier. The inactivation kinetics of nisoldipine-resistant GGG-α_1C Ca²⁺ currents at +10 mV test potential was significantly faster compared to pseudo-wild-type α_1C controls (Figure 2C). Therefore, in adult cardiomyocytes Ca_V1.2 channels comprised of transgenic GGG-α_1C have faster overall inactivation kinetics as compared to transgenic pseudo-wild-type Ca_V1.2 channels likely reflecting accelerated kinetics of voltage-dependent inactivation.

Given that β subunits upregulate Ca_V1.2 channel P_o, we considered whether disruption of the rigid IS6-AID linker might reverse this effect. Consistent with this possibility, the conductance-voltage (G-V) relationships, normalized to cell capacitance, of nisoldipine-resistant transgenic mutant GGG-α_1C channels was reduced compared to pseudo-wild-type α_1C (Figure 2D). To directly assess changes in P_o, we used low-noise single-channel recordings of acutely isolated cardiomyocytes from the GGG-α_1C transgenic mice. Exemplar records show DHP-resistant GGG-α_1C channels exhibit sparse channel openings (Figure 2E, top), a distinct gating pattern compared to pseudo-wild-type α_1C and non-transgenic channels which undergo high-activity flickery openings (Figure 1). Ensemble average P_o-V relationship (Figure 2E, bottom) and bar-graph summary of maximal P_o (Figure 2F) from individual patches show a striking 3.5-fold reduction in maximal P_o compared to both pseudo-wild-type α_1C and non-transgenic channels (Figure 2F). Interestingly, this reduced basal activity of GGG-α_1C is reminiscent of β-less AID-mutant channels suggesting that disruption of the rigid IS6-AID linker may be akin to the uncoupling of the β subunit from the channel pore ¹⁰. To further scrutinize this possibility, we assessed average P_o from individual trials for one channel patches of GGG-α_1C. Unlike pseudo-wild-type α_1C channels, the single-trial ${\overline{P}}_{\text{o}}$ distribution of the GGG-α_1C channels was restricted to either blank (73.4%) or low activity sweeps (26.6%) with no evidence of high activity traces (Figure 2G). As GGG-α_1C are fully capable of β subunit binding (Figure 1), these results suggest that the rigidity of the linker between the pore-domain and I-II loop may be a structural requirement for the high-activity gating configuration. As AID-mutant channels exhibit some propensity for high-activity traces, one attractive possibility is that the IS6-AID linker may switch between rigid and flexible conformations, with β subunit binding to the AID serving to stabilize the rigid helical linker conformation, an outcome also supported by X-ray crystallographic and circular dichroism experiments ^2–4.

9*-α_1C splice variant increases basal open probability.

Having established the I-II loop as a vital regulator of channel openings, we considered whether alternative splicing in this domain might tune channel gating. The 9* exon, which encodes a 75-nucleotide sequence within the I-II loop (Figure 3A), is expressed at a high level in aortic smooth muscle and has lower expression in non-diseased adult human and rat heart. However, the 9* exon-containing channels are increased in rodent models of hypertrophy and in the perinfarct zone ^16–18. This altered pattern of α_1C splicing in rodent models raises the possibility of pathological inclusion of exon 9* in human cardiac disease, an outcome yet to be observed clinically. As such, we sought to determine whether the frequency of exon 9* splice variant is changed in humans with end-stage heart failure. Samples from patients undergoing LVAD implantation at Columbia-NY Presbyterian Hospital were acquired in the operating room, and compared to samples obtained from donor hearts without heart failure (Online Tables I–III). Exon 9* transcript expression was increased in humans with heart failure compared to control samples (Figure 3B).

Figure 3. Expression and electrophysiological properties of 9* splice variant-α_1C Ca²⁺ channels in heart.

(A) Schematic of rabbit cardiac α_1C subunit topology showing β subunit binding to AID motif in I-II loop, and insertion of 9* exon. WT and 9* splice variant sequence in the I-II loop of α_1C. (B) Graph of normalized 9* exon mRNA expression from patients with advanced heart failure (HF) undergoing implantation of a left ventricular assist device and control, no-HF patients. Mean ± SEM. N=19 for HF, 10 for no-HF. Mann-Whitney test for P-value in panel. (C) Anti-FLAG (upper) and anti-β immunoblots (lower) of anti-FLAG antibody immunoprecipitation of cardiac homogenates of non-transgenic (NTG), 9*-α_1C and pseudo-wild-type (pWT) α_1C mice. Representative of 3 experiments. Input is 6% of immunoprecipitation. (D) Immunostaining of 9*-α_1C cardiomyocyte. Anti-FLAG and Alexa 594-conjugated secondary antibodies, and nuclear labeling with Hoechst stain. Images obtained with confocal microscopy. Scale bar = 20 μm. (E) Graph of conductance density-voltage relationship for nisoldipine-resistant Ca²⁺ channels recorded from pseudo-wild-type α_1C and 9*-α_1C cardiomyocytes. pWT α_1C curve is the same as in Figure 2D. Mean ± SEM. N= 20 cells from 3 mice. P < 1.0 ×10⁻⁴ by one-way ANOVA; Sidak’s multiple comparison test P-values in panel. (F) Single-channel Ba²⁺ currents in the absence of nisoldipine (left) and presence of nisoldipine (right). (G) P_o versus voltage relationship, averaged over multiple patches. N=10. (H) Histogram shows distribution of single-trial average P_o obtained from DHP-resistant one-channel patches from 9*-expressing cardiomyocytes. These channels largely adopt high P_o gating mode with a marked reduction in the fraction of blank sweeps.

To determine the functional consequence of exon 9* splice inclusion in cardiomyocytes, we created transgenic mice with cardiac-specific expression of DHP-resistant Ca²⁺ channels containing exon 9*, but with all other mutually exclusive exons typical of cardiac variants. The 9*-α_1C channels still bound β subunits (Figure 3C) similar to pseudo-wild-type α_1C channels, and trafficked to the surface membrane and t-tubules, as demonstrated by the striated z-disk pattern of immunofluorescence (Figure 3D). Whole-cell electrophysiological analysis of nisoldipine-resistant transgenic mutant 9*-α_1C channels revealed a signicant increase in the G-V relationship normalized to cell capacitance in comparison to pseudo-wild-type α_1C channels (Figure 3E). Changes in whole cell current may stem from alterations in channel trafficking, unitary conductance, or baseline open probability. To dissect these mechanistic possibilities, we undertook low-noise single-channel recordings to determine whether the increased normalized conductance of 9*-α_1C reflected a genuine increase in channel P_o. In the presence of nisoldipine, the DHP-resistant 9*-α_1C channels exhibited robust channel openings (Figure 3F) with the ensemble average demonstrating a marked increase in maximal P_o (0.26 ± 0.03, mean ± s.e.m) (Figure 3G) in comparison to pseudo-wild-type α_1C (0.15 ± 0.015, mean ± s.e.m). Furthermore, examination of single-trial ${\overline{P}}_{\text{o}}$ distribution revealed a virtual elimination of blank traces (~0% for 9* versus 40.5% for pseudo-wild-type α_1C -see Figure 1H), and increased propensity for the high activity gating mode (56.7%) (Figure 3H). Thus, the insertion of the 9* exon into the I-II loop upregulates basal voltage-dependent opening of Ca²⁺ channels by enhancing channel availability and stabilizing the high P_o gating configuration. Interestingly, this functional signature is reminiscent of Ca_V1.2 behavior in myocytes from failing human hearts, which also show increased availability and P_o ²⁷.

β-adrenergic upregulation of Ca_V1.2 also requires a rigid IS6-AID linker.

Recently, we determined that the mechanism of adrenergic stimulation of Ca_V1.2 requires constitutive pre-inhibition of Ca_V1.2 mediated by Rad interaction with the Ca_V channel β subunit ¹⁵. PKA phosphorylation of Rad at conserved sites in its C-terminus alters its interaction with the Ca_V channel β subunit and relieves constitutive inhibition ¹⁵. As increased flexibility of the IS6-AID linker effectively decouples β subunit mediated regulation, we hypothesized that the rigid IS6-AID linker may be also essential for adrenergic upregulation of Ca_V1.2 currents. Consistent with this possibility, at the single channel level, Rad inhibited Ca_V1.2 channels have decreased availability and increased propensity for low activity gating mode, akin to GGG-α_1C channels ^{9, 10}.

In cardiomyocytes isolated from mice expressing transgenic pseudo-wild-type α_1C, forskolin (an adenylyl cyclase activator) induced an increase in the nisoldipine-insensitive maximal conductance (G_max) (Figure 4A–B, G), and shifted the V₅₀ for activation (Figure 4H), consistent with our prior studies ^{13, 22, 28, 29}. Ca²⁺ currents through transgenic GGG- α_1C channels, however, were not stimulated by forskolin (Fig 4C–D, G–H). By comparison, forskolin increased the G_max of 9* exon-containing Ca_V1.2 channels by a mean of 1.4-fold (Fig 4E–F, G), and shifted the V₅₀ for activation (Figure 4H) suggesting that the 9* exon still preserves responsiveness to PKA modulation. For both pseudo-wild-type α_1C and 9* Ca_V1.2 channels, the forskolin-induced enhancement of Ca²⁺ current was greatest at hyperpolarized potentials and decreased as the test potential approached the reversal potential of Ca²⁺ (Figure 4I), consistent with prior observations ¹². The GGG Ca_V1.2 channels failed to respond to forskolin at any test potential (Figure 4I).

Figure 4. β-adrenergic regulation of GGG and 9* Ca²⁺ channels in cardiomyocytes.

(A, C, E) Current-voltage relationships before and after 10 μM forskolin in the presence of 300 nM nisoldipine. Representative of pseudo-wild-type α_1C: n=20, GGG mutant: n= 22 and 9* n=22 cardiomyocytes. Insets: Exemplar whole-cell Ca_V1.2 currents recorded from freshly dissociated cardiomyocytes of pseudo-wild-type, GGG and 9* α_1C transgenic mice. Pulses from −70 mV to +10 mV before (black traces) and 3 minutes after (red traces) 10 μM forskolin in presence of nisoldipine. (B, D, F) Graphs of conductance density-voltage relationship for nisoldipine-resistant Ca²⁺ channels recorded from pseudo-wild-type α_1C, GGG-α_1C, and 9*-α_1C before (black trace) and after (red trace) forskolin. Mean ± SEM. P<1.0 ×10⁻⁴ by repeated measures ANOVA; Sidak’s multiple comparison test P-values are in panels. (G) Fold-change in G_max. Shown are means ± SEM; P<1.0 ×10⁻⁴ by Kruskal-Wallis test; Dunn’s multiple comparison test P-values in panel. (H) Boltzmann function parameter, V₅₀. Shown are means + SEM; Paired two-tailed t-test. (I) Ratio of Ca²⁺ current after forskolin treatment to Ca²⁺ current before treatment of cardiomyocytes with forskolin for pseudo-wild-type, 9* and GGG-α_1C cardiomyocytes. Mean ± SEM. Pseudo-wild-type n=20, GGG mutant: n= 22 and 9* n=22 cardiomyocytes from at least 3 mice for each group. P<1.0 ×10⁻⁴ by Kruskal-Wallis test; Dunn’s multiple comparison test P-values in panel for pseudo-WT vs. GGG. * P<1.0 ×10^-4.

We used reconstitution studies in HEK293T cells ¹⁵ to further gain insights into the mechanisms by which the β subunit and I-II loop modulate adrenergic regulation of Ca_V1.2. In cells transfected with only α_1C + β_2B, superfusion of forskolin did not affect the Ba²⁺ currents (Figure 5A). In contrast, applying forskolin to cells expressing α_1C + β_2B + Rad increased the current as we have previously described ¹⁵. Consistent with our findings in cardiomyocytes, in cells transfected with GGG-α_1C forskolin failed to increase G_max or shift the voltage-dependence of activation in a hyperpolarizing direction (Figure 5C–D, G–H). In contrast, applying forskolin to cells expressing 9*-α_1C, β_2B, and Rad increased the G_max by a mean of 1.6-fold, and shifted the V₅₀ for activation (Figure 5E–H).

Figure 5. Rad-mediated β−adrenergic regulation of Ca_V1.2 requires rigid I-II linker.

(A–F) Ba²⁺ current elicited by voltage ramp every 10 s, with black traces obtained before and blue traces obtained after forskolin. The pseudo-WT (pWT) α_1C, β_2B and Rad were heterologously expressed in HEK293T cells. Representative of top row: 15, 17, 12 cells, left to right; bottom row: 16, 21, 12 cells, left to right. (G) Fold-change in maximum conductance (G_max) induced by forskolin. Mean ± SEM; Two-tailed unpaired t-test. (H) Boltzmann function parameter V₅₀. Mean ± SEM; Two-tailed paired t-test.

Flow-cytometry Förster resonance energy transfer (FRET) 2-hybrid assay ³⁰ was utilized to determine whether the presence of GGG substitutions or 9* exon altered β subunit binding to the I-II loop, which is required for β-adrenergic regulation of Ca_V1.2 in heart ¹³ or the PKA-induced reduction in Rad binding to the β_2B subunit. Robust interaction is detected between Cerulean-tagged β_2B subunit and Venus-tagged I-II loop (Online Figure I A, E). As would be expected, the mutation of the AID of the I-II loop markedly reduced binding (Online Figure I B), whereas the GGG substitutions or insertion of the 9* exon did not affect binding (Online Figure I C–E). As we previously reported ¹⁵, there is a strong interaction between the Cerulean-tagged β_2B subunit and Venus-tagged WT Rad (Online Figure II A–B,K). This interaction was markedly weakened, however, by co-expression of PKA catalytic subunit. The basal and PKA-dependent reduction of binding between Rad and β_2B was unaffected by co-expression of WT α_1C (Online Figure II C–D,K). Similarly, expression of the AID-mutant α_1C (Online Figure II E–F), GGG-α_1C (Online Figure II G–H) or 9*-α_1C (Online Figure II I–J) had no effect on the basal or the PKA-dependent reduction in binding between Rad and β_2B (Online Figure II K). These results suggest that the PKA-dependent dissociation of Rad and β_2B is neither dependent upon the binding of β to α_1C nor is it perturbed by the alterations in the I-II loop induced either by the GGG substitution or insertion of the 9* exon.

DISCUSSION

This work has examined mechanisms of regulation of cardiac Ca_V1.2 channel gating by three essential factors with important physiological and pathophysiological consequences that converge at the α_1C subunit I-II loop— auxiliary Ca_Vβ subunits, sympathetic activation, and alternative splicing. Overall, we find that PKA-modulation of Ca_V channels in heart and HEK cells is dependent on both Rad phosphorylation and a rigid IS6-AID linker. We discuss our findings on these three inter-related regulatory mechanisms in the context of previously published reports.

Many reconstitution studies in heterologous mammalian cells established the idea that auxiliary Ca_Vβ subunits were necessary for trafficking of Ca_V1.2 channels to the plasma membrane, and that this depended on high-affinity Ca_Vβ binding to a discrete α₁-interaction domain (AID) in the α_1C I-II loop ^{5, 7, 31–37}. Beyond trafficking, Ca_Vβ binding also boosted Ca_V1.2 P_o and produced a hyperpolarizing shift in the voltage-dependence of channel activation ^{36, 38}. These gating effects were deduced to require formation of a rigid helix spanning IS6 and AID because they were selectively eliminated by a triple glycine substitution that disrupts the continuous helix ¹⁰. In adult cardiomyocytes, cardiac-specific excision of the dominant cardiac Ca_Vβ₂ isoform reduced β₂ protein levels by 96%, yet resulted in only a 26% reduction in whole-cell Ca_V1.2 current, providing a first hint that, by contrast to heterologous cells, Ca_Vβ binding to α_1C may not be obligatory for forming functional Ca_V1.2 channels at the cell surface ³⁹. We explicitly confirmed this by showing that transgenic mice expressing a DHP-resistant α_1C mutant that does not bind Ca_Vβ, nevertheless, yielded robust nisoldipine-resistant whole-cell Ca²⁺ currents indicating that the channels made it to the surface sarcolemma ^29,13. While Ca_Vβ does not appear necessary for surface trafficking of Ca_V1.2 in adult cardiomyocytes, it remained unclear whether this also extended to the impact on channel P_o. Here, we unambiguously show using single-channel recordings that Ca_Vβ binding to α_1C in cardiomyocytes enhances Ca_V1.2 channel P_o by 3.5-fold. The single-channel gating signature of AID-mutant channels was dominated by null (46%) and low-activity (47%) sweeps, with rare sojourns into a high-activity (7%) gating mode. By contrast, Ca_Vβ-bound pseudo-WT channels displayed more high-activity sweeps (26%) and correspondingly lower null (41%) and low-activity (33%) gating modes. GGG-α_1C channels displayed only blanks and low-activity gating. These results are consistent with the interpretation that in adult cardiomyocytes Ca_Vβ induces high-P_o gating by stabilizing a continuous helix linking IS6 to AID. Absence of Ca_Vβ binding (as occurs with AID-mutation) reduces the propensity for stabilizing a continuous helix and accordingly decreases fractional occupancy of the high-P_o gating mode. GGG-α_1C completely dispels the continuous helix and, therefore, these channels do not sojourn into the high-P_o mode. In the cryo-electron microscopy structure of the homologous Ca_V1.1 channel, comparison of the two conformations, class Ia and II, revealed signficant shifts between the C-terminal end of IS6 and I-II helix of the α₁ subunit, and the β subunit ⁴⁰. Substitution of the three glycine residues in either one of the two conformations (Figure 6A) likely alters the conformation and increases the flexibility of the I-II loop and the position of the β subunit.

Figure 6. Schematics of role of GGG and 9* in modulating β-adrenergic regulation of Ca_V1.2.

(A) Ribbon representation of the published two Ca_V1.1 conformations determined by cryo-EM ⁴⁰ with GGG, AID and exon 9* mutations/insertions in the I-II loop. Left: 5GJV, Right: 5GJW. (B) Proposed models of β−adrenergic regulation of WT, GGG-α_1C and 9*-α_1C channels. Basal state (left) and after β-adrenergic agonist (stimulated, right). β-agonist-induced activation of adenylyl cyclase (AC) leads to activation of PKA. PKA phosphorylates several residues on Rad, causing dissociationof Rad from the Ca_V1.2 complex and therefore increased Ca²⁺ influx. Under basal conditions, the GGG-α_1C channels have reduced basal open probability and no response to β-adrenergic stimulation. In contrast to the polyglycine substitution, modifying the I-II loop by introduction of the 9* splice variant increased the basal open probability, yet the stimulatory response to β−adrenergic agonists was preserved.

We recently reported that β-adrenergic regulation of cardiac Ca_V1.2 channel requires Ca_Vβ binding binding to α_1C I-II loop, ¹³ and PKA phosphorylation of Rad, a small G-protein that inhibits Ca²⁺ channels via binding to Ca_Vβ subunit ¹⁵. Here, we report that GGG-α_1C channels expressed in cardiomyocytes, or reconstituted with Rad in heterologous cells, do not display PKA-mediated up-regulation of whole-cell current density, even though they bind Ca_Vβ. This result suggests the rigid IS6-AID helical linker is another essential requirement for transduction of sympathetic regulation of Ca_V1.2. The exclusively low-P_o gating mode of GGG-α_1C channels is reminiscent of the behavior of Rad-inhibited channels ¹⁵. It is intriguing to speculate that Rad interaction with Ca_Vβ may structurally alter the IS6-AID linker as a mechanism for channel inhibition.

Finally, we examined the impact of the α_1C 9*-splice variant that is elevated in the failing heart. Interestingly, the 9*-α_1C channels had increased basal P_o and were exclusively recorded in high activity gating modes, with no sojourns into mode 0 gating. Given the proximity of the 9* exon to the AID, and the role of the continuous IS6-AID helix in promoting high activity Cav1.2 gating, one explanation is that the 9* exon may increase basal P_o by further stabilizing the rigid IS6-AID helical linker. Another explanation is that insertion of the 9* exon may affect the function of the voltage-sensor domain of domain II. Previous single-channel experiments have indicated that Ca_V1.2 channels in failing hearts have an elevated basal P_o which was putatively attributed to enhanced phosphorylation of the channel ⁴¹. Our results suggest that the increased Ca_V1.2 P_o observed in heart failure may be due, in part, to the emergence of the alternatively spliced 9*-α_1C variant in this condition.

While powerful tools for electrophysiological assessment of α_1C mutants in vivo, certain limitations of our transgenic inducible over-expression models prevent examination of whether and how these α_1C mutant contribute to both basal and sympathetic regulation of cardiac contractility, and arrhythmogenesis in vivo. Chief among these is that same strategy that allows us to isolate the DHP-resistant transgenic channels for electrophysiological assessment in vivo—application of nisoldipine—blocks endogenous Ca_V1.2 channels in both heart and vasculature, thus leading to hypotension and confounding the experiments. Furthermore, modest over-expression of WT α_1C can initiate hypertrophy and heart failure ^{42, 43}. A direct assessment of whether exon 9* is detrimental could theoretically be achieved with an inducible 9*-α_1C knock-in or knock-out strategy, yet due to the genomic structure of exons 9 and 10 in α_1C, creating inducible 9*-α_1C knock-in or knock-out mice lines would be quite challenging, likely requiring the insertion of a large minigene.

Notwithstanding these limitations, these results provide key insight into the mechanism underlying the β-adrenergic stimulation of Ca²⁺ current and contractility in the heart. β subunit binding to the α_1C subunit promotes transitions to a high P_o state. Upon β-adrenergic stimulation, PKA phosphorylation of Rad releases the Rad-induced inhibition of the Ca²⁺ channels (Figure 6B). The Rad-inhibited channels are the heart’s functional reserve of Ca²⁺ channels, likely having minimal effects on excitation-contraction coupling at rest, but primed to respond to β-adrenergic agonists upon release of Rad-induced inhibition. Thus, therapeutic release of Rad-mediated inhibition of Ca²⁺ channels could be inotropic.

Major Resources Table

Animals (in vivo studies)
Species	Vendor or Source	Background Strain	Sex	Persistent ID / URL
NA	NA	NA	NA	NA

Genetically Modified Animals
	Species	Vendor or Source	Background Strain	Other Information	Persistent ID / URL
Pseudo-WT-α1C	Mouse	Transgenic Mouse Facility Columbia University	B6CBAF2	Generated at Columbia University	NA
GGG-α1C	Mouse	Transgenic Mouse Facility Columbia University	B6CBAF2	Generated at Columbia University	NA
AID-α1C	Mouse	Transgenic Mouse Facility Columbia University	B6CBAF2	Generated at Columbia University	NA
9*-α1C	Mouse	Transgenic Mouse Facility Columbia University	B6CBAF2	Generated at Columbia University	NA
FVB/N-Tg(Myh6-rtTA)8585Jam/Mmmh	Mouse	MMRC	FVB	NA	RRID: MMRRC_ 010478-MU

Antibodies
Target antigen	Vendor or Source	Catalog #	Working concentration	Lot # (preferred but not required)	Persistent ID / URL
FLAG (rabbit)	Sigma	F7425	1:200 (ICC)		https://www.sigmaaldrich.com/catalog/product/sigma/f7425
Monoclonal Anti-FLAG M2-Peroxidase (HRP) antibody	Sigma	A8592	1:1000 (WB)		https://www.sigmaaldrich.com/catalog/product/sigma/a8592?lang=en&region=US
Beta common (custom)	Yen-Zym		1:1000		epitope: mouse residues 120–138:DSYTSRPS-DSDVSLEEDRE
AlexaFluor 594-goat anti rabbit	Thermo	A11037	1:200	2079421	https://www.thermofisher.com/antibody/product/Goat-anti-Rabbit-IgG-H-L-Highly-Cross-Adsorbed-Secondary-Antibody-Polyclonal/A-11037

DNA/cDNA Clones
Clone Name	Sequence	Source / Repository	Persistent ID / URL
mouse Rrad mRNA		NCBI/GenBank	XM_006531206
human Xacnb2 mRNA		NCBI/GenBank	NM_201590.3
rabbit cardiac Cacna1c mRNA		NCBI/GenBank	X15539
rabbit smooth muscle Cacna1c mRNA		NCBI/GenBank	X55763

Cultured Cells
Name	Vendor or Source	Sex (F, M, or unknown)	Persistent ID / URL
HEK293T cells	ATCC		CRL-3216
HEK293 cells	ATCC		CRL-1573

Data & Code Availability
Description	Source / Repository	Persistent ID / URL

Other
Description	Source / Repository	Persistent ID / URL
polyethylenimine (PEI)	Polysciences	23966-2; https://www.polysciences.com/default/polyethylenimine-linear-mw25000
cycloheximide	Sigma	01810-1G; https://www.sigmaaldrich.com/catalog/product/sigma/c4859?lang=en&region=US
Trypsin 0.05%	Thermo Fisher	25300-054; https://www.thermofisher.com/order/catalog/product/25300054?SID=srch-hj-25300-054#/25300054?SID=srch-hj-25300-054
FBS	Thermo Fisher	10438-026; https://www.thermofisher.com/order/catalog/product/10438026?SID=srch-hj-10438-026#/10438026?SID=srch-hj-10438-026
DMEM	Thermo Fisher	11995-065; https://www.thermofisher.com/order/catalog/product/11995065?SID=srch-hj-11995-065#/11995065?SID=srch-hj-11995-065
Lipofectamine 2000	Thermo Fisher	11668-019; https://www.thermofisher.com/order/catalog/product/11668019?SID=srch-hj-11668-019#/11668019?SID=srch-hj-11668-019
Pen/Strep Glutamine	Thermo Fisher	10378-016; https://www.thermofisher.com/order/catalog/product/10378016?SID=srch-hj-10378-016#/10378016?SID=srch-hj-10378-016
optiMEM	Thermo Fisher	31985-062; https://www.thermofisher.com/order/catalog/product/31985062?SID=srch-hj-31985-062#/31985062?SID=srch-hj-31985-062
Attachment Factor	Thermo Fisher	S-006-100; https://www.thermofisher.com/order/catalog/product/S006100?SID=srch-hj-S-006-100#/S006100?SID=srch-hj-S-006-100
Doxycycline Rodent Diet (200 mg/kg)	Bio Serv	S3888; https://us.vwr.com/store/product/15984109/doxycycline-rodent-diet-sterile-bio-serv
Triton X-100	IBI Scientific	IB07100; https://www.ibisci.com/products/triton-x-100
Calpain inhibitor I	Sigma	A6185; https://www.sigmaaldrich.com/catalog/product/sigma/a6185
Calpain inhibitor II	Sigma	A6060; https://www.sigmaaldrich.com/catalog/product/sigma/a6060
Complete Mini tablets	Roche	04693159001; https://www.sigmaaldrich.com/catalog/product/roche/04693159001?lang=en&region=US
Protein A beads	GE Healthcare	17-0780-01; https://www.sigmaaldrich.com/catalog/product/sigma/ge17078001?lang=en&region=US
Forskolin	Santa Cruz Biotechnology	SC-7562; https://www.scbt.com/p/forskolin-66575-29-9
Isoproterenol	Sigma	I5627; https://www.sigmaaldrich.com/catalog/product/sigma/i5627
Nisoldipine	Santa Cruz Biotechnology	SC-212396; https://www.scbt.com/p/nisoldipine-d7
Amphotericin B	Sigma	A9528; https://www.sigmaaldrich.com/catalog/product/sigma/a9528?lang=en&region=US
SYBR Green qPCR Master Mix	Thermo	https://www.thermofisher.com/order/catalog/product/4309155#/4309155
RNeasy Mini kit	Qiagen	https://www.qiagen.com/us/products/discovery-and-translational-research/dna-rna-purification/rna-purification/total-rna/rneasy-mini-kit/#orderinginformation
High-Capacity cDNA RT kit	Thermo Scientific	https://www.thermofisher.com/order/catalog/product/4368814#/4368814
Red ANTI-FLAG M2 Affinity Gel	Sigma	F2426; https://www.sigmaaldrich.com/catalog/search?term=F2426&interface=All&N=0&mode=match%20partialmax&lang=en&region=US&focus=product
Liberase	Sigma	5401119001; https://www.sigmaaldrich.com/catalog/product/roche/libtmro?lang=en&region=US

NOVELTY AND SIGNIFICANCE.

What Is Known?

• Changing activity of cardiac CaV1.2 channels under basal conditions, during sympathetic activation, and in heart failure is a major determinant of cardiac physiology and pathophysiology.

• Activation of β-adrenergic receptors results in a multi-fold upregulation of Ca_V1.2 currents, which is dependent upon PKA phosphorylation of Rad.

• It is unknown how distal conformational changes involving Rad interaction with the Ca_Vβ subunit and phosphorylation-dependent signaling are ultimately conveyed to the channel pore-domain.

• How alternative splicing of the I-II linker of the pore-forming a_1C subunit contributes to this regulatory scheme remains to be fully-elucidated.

What New Information Does This Article Contribute?

• Introducing flexibility into the rigid IS6-AID helix markedly reduced basal open probability despite intact binding of Ca_Vβ to α_1C I-II loop, and eliminated β-adrenergic agonist stimulation of Ca_V1.2 current.

• The α_1C 9*-splice variant, which is elevated in the failing human heart, causes an increased basal open probability but did not attenuate stimulatory response to β-adrenergic agonists.

• PKA-modulation of Ca_V channels in heart and HEK cells is dependent on both Rad phosphorylation and a rigid IS6-AID linker

β-adrenergic stimulation of Ca_V1.2 current is vital for sympathetic nervous system regulation of cardiac contractility. Recently, we identified Rad, not Ca_V1.2, as the functionally-relevant target of PKA. At baseline, Rad inhibits Ca_V1.2 by binding to Ca_V β subunit. Upon β-adrenergic activation, PKA phosphorylation of Rad releases this interaction and inhibition. How is the signal from Rad and the β subunit transmitted to the channel’s gate? To answer this question, we created transgenic mice featuring the introduction of three glycine residues that disrupt the rigid helix between pore and the α-interaction-domain. Ca²⁺ channels from these mice displayed markedly reduced basal open probability and were insensitive to β-adrenergic agonist stimulation. In contrast, introduction in the α_1C I-II loop of the exon 9* splice variant, which is increased in ventricles of patients with end-stage heart failure, increased basal open probability but did not attenuate stimulatory response to β-adrenergic agonists in cardiomyocytes. We speculate that the increased Ca_V1.2 open probability observed in heart failure may be due, in part, to the emergence of the alternatively spliced 9*-α_1C variant. Taken together, adrenergic stimulation of Ca_V1.2 requires an intact rigid linker between the binding site of β subunit in the I-II loop and the channel pore.

Nonstandard Abbreviations and Acronyms:

AID

α-interaction domain

αMHC

α-myosin heavy chain

CaM

calmodulin

DHP

Dihydropyridine

FRET

Förster resonance energy transfer

G-V

conductance-voltage

G_max

maximal conductance

HF

heart failure

P_o

open probability

PKA

protein kinase A

pWT

pseudo-wild-type

RGK

Rad, Gem, Kir proteins

rtTA

reverse transcriptional transactivator