Molecular Mechanism of Calcium Channel Regulation in the Fight-or-Flight Response

Abstract

During the fight-or-flight response, the sympathetic nervous system stimulates L-type calcium ion (Ca²⁺) currents conducted by Ca_V1 channels through activation of β-adrenergic receptors, adenylyl cyclase, and phosphorylation by adenosine 3′,5′-monophosphate–dependent protein kinase [also known as protein kinase A (PKA)], increasing contractility of skeletal and cardiac muscles. We reconstituted this regulation of cardiac Ca_V1.2 channels in non-muscle cells by forming an autoinhibitory signaling complex composed of Ca_V1.2Δ1800 (a form of the channel truncated at the in vivo site of proteolytic processing), its noncovalently associated distal carboxyl-terminal domain, the auxiliary α₂δ₁ and β_2b subunits, and A-kinase anchoring protein 15 (AKAP15). A factor of 3.6 range of Ca_V1.2 channel activity was observed from a minimum in the presence of protein kinase inhibitors to a maximum upon activation of adenylyl cyclase. Basal Ca_V1.2 channel activity in unstimulated cells was regulated by phosphorylation of serine-1700 and threonine-1704, two residues located at the interface between the distal and the proximal carboxyl-terminal regulatory domains, whereas further stimulation of channel activity through the PKA signaling pathway only required phosphorylation of serine-1700. Our results define a conceptual framework for Ca_V1.2 channel regulation and identify sites of phosphorylation that regulate channel activity.

INTRODUCTION

The “fight-or-flight” response is a conserved behavior of vertebrates experiencing fear, stress, or intense exercise. Release of epinephrine from the adrenal medulla and norepinephrine from sympathetic nerve endings activates β-adrenergic receptors and leads to an increase in L-type Ca²⁺ currents in cardiac and skeletal muscle, which in turn increase the force of contraction of skeletal muscle and the beating rate and contractility of the heart (1–3). This regulation of L-type calcium currents occurs through activation of protein kinase A (PKA)–mediated phosphorylation (4–8), but the molecular mechanism is unknown.

Ca_V1.1 and Ca_V1.2 channels, which conduct L-type Ca²⁺ currents in skeletal and cardiac muscle, respectively, are composed of pore-forming α₁ subunits, plus auxiliary α₂δ and β subunits (9–17). The α₁ subunits of Ca_V1.1 and Ca_V1.2 channels each exist as two size forms produced by in vivo proteolytic processing of the C terminus in skeletal and cardiac muscle (18–22) and in brain (23, 24). Expression of complementary DNA (cDNA) encoding full-length Ca_V1.2 channels in non-muscle cells yields only full-length α₁ subunits with no evidence of proteolytic processing (25). Deletion of the distal C terminus increases Ca_V1.2 channel activity (26, 27), and coexpression of the distal C terminus results in formation of a noncovalent autoinhibitory complex with the proximal C terminus of the truncated channel (28). The distal C-terminal domain effectively inhibits Ba²⁺ currents and reduces coupling efficiency between voltage sensor movement and opening of the pore (21, 28).

Ca_V1 channels in skeletal and cardiac muscle bind A-kinase anchoring protein 15 (AKAP15) (29–31), and β-adrenergic regulation of these channels requires PKA anchored by AKAP15 (32, 33). The AKAP15-PKA complex interacts with a site located in the proteolytically cleaved distal C terminus, suggesting that the distal C terminus remains associated with the remainder of the channel and mediates PKA-dependent regulation in cardiac myocytes (28, 33, 34).

Physiological regulation of L-type currents by PKA has not been reconstituted in non-muscle cells, impeding detailed analysis of the regulatory properties of calcium channel mutants. Previous studies of PKA-dependent modulation of cloned Ca_V1.2 channels expressed in non-muscle cells have examined full-length channels [for instance, (35, 36)]. Here we show that PKA-dependent modulation of Ca_V1.2 channel activity with a dynamic range similar to that seen in cardiac myocytes can be reconstituted in non-muscle cells after optimized expression of the different components of the Ca_V1.2-AKAP15 autoinhibitory signaling complex. Using this reconstitution system, we have identified sites of protein phosphorylation that are both necessary and sufficient for physiological regulation of Ca_V1.2 channels in non-muscle cells. These sites are located at the interface between the proximal (PCRD) and the distal (DCRD) C-terminal regulatory domains, well positioned to mediate disinhibition of the auto-inhibitory signaling complex and increase Ca_V1.2 channel activity in the fight-or-flight response.

RESULTS

Reconstitution of an autoinhibitory Ca_V1.2 channel signaling complex

The distal C-terminal domain inhibits channel activity when cDNA encoding distal_1822–2171 is coexpressed at a 5:1 molar ratio with that encoding Ca_V1.2 channels truncated at position 1821 (28). For the experiments described here, we cotransfected cDNA encoding Ca_V1.2 channels truncated at A¹⁸⁰⁰, the exact site of in vivo proteolytic processing (Ca_V1.2Δ1800) (21), distal_1801–2171 (DCT), α₂δ₁, and β_1b to produce a functional auto-inhibitory Ca_V1.2 channel complex. We assessed channel activity by measuring peak Ba²⁺ current and by determining the coupling efficiency of pore opening to gating charge movement (28). Gating charge movement was measured by integrating the gating current during depolarization to the reversal potential, and pore opening was measured from tail currents upon repolarization from the reversal potential. The functional properties of Ca_V1.2Δ1800 channels and their inhibition by DCT were identical to those of Ca_V1.2Δ1821 and distal_1822–2171 [compare (28) and fig. S1].

Transfection of cells with increasing ratios of cDNA encoding DCT to cDNA encoding Ca_V1.2Δ1800 revealed increasing inhibition of channel activity, with maximal inhibition at a ratio of 0.75 (Fig. 1, A and B). Coexpression of the DCT with the truncated channel does not alter the density of Ca_V1.2 channels on the cell surface, as determined by gating current measurements (28); therefore, reduction of the tail current provides a direct measure of the decrease in the coupling efficiency of the gating charge movement to opening of the pore: 43.6 ± 2.4 nA/pC (n = 13) for Ca_V1.2Δ1800 channels versus 15.2 ± 1.0 nA/pC for Ca_V1.2Δ1800 + DCT (n = 19; P < 0.01) (Fig. 1A).

Fig. 1.

Inhibition of Ca_V1.2Δ1800 channels by the distal C terminus. (A) Coupling efficiency for Ca_V1.2Δ1800 channels in the presence and absence of the indicated molar ratios of DCT/Ca_V1.2Δ1800 plasmids expressed in tsA-201 cells. *P < 0.05 or **P < 0.01 versus Ca_V1.2Δ1800 (I_Tail, peak tail current; Q, total gating current; V_rev, reversal potential). Red symbols indicate 0.75:1 molar ratio. n values and ± SEM are indicated. Significance was determined by ANOVA. Inset: Representative records for Ca_V1.2Δ1800 and Ca_V1.2Δ1800 + DCT channels. (B) Representative Ba²⁺ currents through Ca_V1.2Δ1800 and Ca_V1.2Δ1800 + DCT channels elicited by a test pulse to +10 mV from a holding potential of −80 mV. (C and D) Mean current-voltage (C) and conductance-voltage (D) relationships for Ca_V1.2Δ1800 and Ca_V1.2Δ1800 + DCT channels.

Inward Ba²⁺ currents conducted by Ca_V1.2 channels were decreased when DCT was present, similar to previous results (Fig. 1C) (28). However, the shift in voltage-dependent activation to more depolarizing potentials observed with large amounts of distal_1822–2171 (28) was barely detectable (but still significant) with smaller amounts of DCT (Fig. 1D). These results suggest that expression of an optimal 0.75:1 molar ratio of the cDNAs encoding DCT and Ca_V1.2Δ1800 causes a specific reduction in coupling efficiency of gating charge movement to pore opening.

Reduction of basal Ca_V1.2 channel activity by inhibition of protein kinases

The increase in L-type Ca²⁺ current caused by basal phosphorylation in unstimulated cardiac myocytes (37) is inhibited by 1 μM RO 31-8220 (38), a potent protein kinase C (PKC) inhibitor (39) that also inhibits PKA and other protein kinases at 1 μM and higher concentrations (40). To determine whether protein kinase inhibition would also decrease L-type current through Ca_V1.2 channels in tsA-201 human embryonic kidney cells, we applied RO 31-8220 (50 nM or 1 μM) to the extracellular surface of cells transiently transfected with cDNAs encoding Ca_V1.2Δ1800 + DCT for 5 min before recording channel activity. RO 31-8220 (1 μM) significantly decreased inward Ba²⁺ currents and coupling efficiency in cells expressing Ca_V1.2Δ1800 + DCT (Fig. 2, brown). In contrast, 50 nM RO 31-8220, an effective concentration for inhibiting PKC but not PKA (40), had no effect on channel activity. We also observed a significant decrease in basal channel activity when cells expressing Ca_V1.2Δ1800 + DCT were treated for 1 hour with the specific, membrane-permeant PKA inhibitor myristoylated PKI 14-22 amide (5 μM) before recording channel activity with an intracellular solution containing 2 μM PKI 6-22 amide, the same PKA inhibitor in membrane-impermeant form (Fig. 2, C and D). Collectively, these results indicate that PKA-dependent phosphorylation increases basal Ca_V1.2 channel activity. The coupling efficiency during PKA inhibition defines the lower limit of Ca_V1.2 channel activity compared to the maximal coupling efficiency for Ca_V1.2Δ1800 channels in the absence of DCT.

Fig. 2.

Basal activity of Ca_V1.2 is reduced by inhibiting kinase activity. (A) Current-voltage relationships for Ca_V1.2Δ1800 + DCT in the presence and absence of 50 nM or 1 μM RO 31-8220 (RO). (B) Coupling efficiency (nA/pC) for Ca_V1.2Δ1800 + DCT channels in the presence and absence of 50 nM or 1 μM RO 31-8220. **P < 0.01. Significance was determined by Student’s t test (I_Tail, peak tail current; Q, total gating current; V_rev, reversal potential). (C) Current-voltage relationships for Ca_V1.2Δ1800 + DCT in the presence and absence of 1 μM myristoylated PKI 14-22 or 5 μM myristoylated PKI 14-22 with 2 μM PKI 6-22. (D) Coupling efficiency (nA/pC) for Ca_V1.2Δ1800 + DCT channels in the conditions described in (C). **P < 0.01. Significance was determined by Student’s t test.

Requirement for AKAP15 for PKA-dependent increase in Ca_V1.2 channel activity

AKAPs act as molecular scaffolds that enable the formation of signaling complexes (41). AKAP15 directly associates with Ca_V1.1 and Ca_V1.2 channels through a modified leucine zipper motif and promotes their co-localization with PKA (29–33). Moreover, physiological regulation of Ca_V1.2 channels by β-adrenergic stimulation in dissociated cardiac myocytes requires PKA anchored to AKAP15 (33). To test the requirement for AKAP15, we applied 5 μM forskolin to cells transfected with cDNA encoding Ca_V1.2Δ1800 + DCT for 5 min before recording channel activity. No increase in coupling efficiency was observed in the absence of AKAP15 (Fig. 3A). However, titration of the cDNA encoding AKAP15 downward from a molar ratio of 1:1 relative to that encoding Ca_V1.2Δ1800 revealed a substantial increase in coupling efficiency by forskolin at low amounts of AKAP15 (Fig. 3A, blue). With the optimum amount of AKAP15, forskolin increased Ca_V1.2Δ1800 + DCT coupling efficiency to 28.0 ± 2.0 nA/pC versus 15.2 ± 1.0 nA/pC for control without forskolin (n = 19 to 20, P < 0.01), with a corresponding increase in inward Ba²⁺ current (Fig. 3C), suggesting that AKAP15 is required for PKA-dependent Ca_V1.2 channel regulation. Comparison of the results with forskolin treatment (Fig. 3) to the results with kinase inhibitors (Fig. 2) revealed a factor of 3.6 dynamic range of modulation of Ca_V1.2 channel activity, similar to that seen in dissociated cardiac myocytes (25, 34, 38). Acute application of 5 μM forskolin to tsA-201 cells expressing Ca_V1.2Δ1800 + DCT with the optimum amount of AKAP15 gave a similar dynamic range of modulation (Fig. 3B). To ensure maximum activation of PKA in an undisturbed cellular context, we performed all subsequent experiments on intact Ca_V1.2-expressing cells treated with 5 μM forskolin for 5 min before recording channel activity.

Fig. 3.

Regulation of Ca_V1.2 channel activity by optimal expression of cDNA encoding AKAP15. (A) Coupling efficiency (nA/pC) for Ca_V1.2Δ1800 and Ca_V1.2Δ1800 + DCT channels with and without AKAP15, and 5 μM forskolin. **P < 0.01 versus control without forskolin. Dashed black line indicates mean current for unstimulated Ca_V1.2Δ1800 + DCT (I_Tail, peak tail current; Q, total gating current; V_rev, reversal potential). (B) Time course of peak Ba²⁺ current during perfusion with 5 μM forskolin (Fsk). Pulses to 10 mV before and during application of forskolin. Inset: current traces indicated by a and b. (C) Current-voltage relationships of Ca_V1.2 channels from (A). (D) Coupling efficiency (nA/pC) for Ca_V1.2Δ1800 and Ca_V1.2Δ1800 + DCT channels with PKA Cα catalytic subunit, PKA RIIα regulatory subunit, AKAP15, and 5 μM forskolin. ***P < 0.001 versus control. Dashed black line indicates mean current for unstimulated Ca_V1.2Δ1800 + DCT. Right, coupling efficiency (nA/pC) of individual experiments in each condition. Red line indicates the maximum current observed with Ca_V1.2Δ1800 + DCT. (E) Current-voltage relationships of Ca_V1.2 channels from (D). Significance was determined by ANOVA.

Overexpressed AKAP15 is expected to have a dominant-negative effect on regulation of Ca_V1.2 channel activity when its concentration exceeds that of endogenous PKA, so that AKAP15 lacking bound PKA occupies the Ca_V1.2 channel binding sites unproductively. This necessitated the use of low molar ratios of the cDNA encoding AKAP15 to those encoding Ca_V1.2Δ1800 + DCT. Analysis of coupling efficiency in individual cells indicated that 65% of cells cotransfected with AKAP15 and Ca_V1.2Δ1800 + DCT showed higher coupling efficiencies when treated with 5 μM forskolin than the maximum observed in its absence (fig. S2, A and B). In contrast, no individual cells had coupling efficiencies in the same range when Ca_V1.2Δ1800 + DCT channels lacking AKAP15 were treated with 5 μM forskolin (fig. S2A). The mean coupling efficiency in cells responsive to forskolin treatment was 33.5 ± 1.6 nA/pC (n = 13), whereas unresponsive cells had a mean coupling efficiency of 17.9 ± 0.8 nA/pC (n = 7), similar to that observed for Ca_V1.2Δ1800 + DCT in the absence of forskolin (fig. S2C). These findings show that PKA-dependent modulation of Ca_V1.2 channel activity can be reconstituted in non-muscle cells, but only in the presence of optimal amounts of AKAP15. To circumvent the dominant-negative effect of AKAP15, we coexpressed cDNA encoding PKA together with those encoding Ca_V1.2Δ1800 + DCT and AKAP15. Under these conditions, each individual cell exhibited a coupling efficiency in the presence of forskolin higher than that seen in its absence (Fig. 3, D and E). These results suggest that PKA overexpression ensures that an AKAP15-PKA complex associates with each Ca_V1.2Δ1800 + DCT channel to mediate forskolin-dependent modulation. However, all subsequent experiments were performed in the absence of exogenous PKA to avoid unwanted effects of kinase overexpression that might result from phosphorylation of secondary sites.

Modulation of full-length Ca_V1.2 channels by AKAP15

Earlier attempts to reconstitute PKA-dependent modulation of Ca_V1.2 channels in non-muscle cells used full-length channels (Ca_V1.2FL) and showed only small increases in Ca_V1.2 channel activity in response to PKA activation [for instance, (31, 35, 36, 42)]. To determine whether the effects of PKA on Ca_V1.2FL channel activity were enhanced in the presence of optimal amounts of AKAP15, we recorded Ca_V1.2FL channel activity in the presence and absence of 5 μM forskolin in cells coexpressing AKAP15. In contrast to Ca_V1.2Δ1800 + DCT, 5 μM forskolin failed to elicit a significant increase in coupling efficiency for Ca_V1.2FL at AKAP15 cDNA ratios of 0.003:1 and 1:1 (Fig. 4A, gray and blue), although it produced a small but significant increase in inward Ba²⁺ current (Fig. 4B, blue; 5.0 ± 0.6 nA/pC versus 7.4 ± 0.6 nA/pC, n = 7; P < 0.05). Inhibition of protein kinases with 1 μM RO 31-8220 decreased coupling efficiency of Ca_V1.2FL from 17.8 ± 1.5 to 7.0 ± 1.0 (Fig. 4A, brown; n = 5; P < 0.01) and also decreased Ba²⁺ current (Fig. 4B, brown; n = 5; P < 0.05). These results show that regulation of basal activity of Ca_V1.2FL is similar or greater than Ca_V1.2Δ1800 + DCT, but up-regulation by PKA is substantially reduced in channels with the distal C terminus covalently attached.

Fig. 4.

Regulation of full-length Ca_V1.2 channels. (A) Coupling efficiency (nA/pC) for Ca_V1.2FL channels in the presence or absence of 1 μM RO 31-8220 or Ca_V1.2FL channels with AKAP15 (1:1 or 0.003:1 molar cDNA ratio) and 5 μM forskolin. **P < 0.01 (I_Tail, peak tail current; Q, total gating current; V_rev, reversal potential). (B) Current-voltage relationships for the Ca_V1.2 channels studied in (A). Significance was determined by Student’s t test.

Phosphorylation of sites at the regulatory interface between the distal and the proximal C-terminal domains of Ca_V1.2 channels

Successful reconstitution of Ca_V1.2 channel modulation provided the opportunity to identify phosphorylation sites that are important for PKA-dependent potentiation of L-type calcium current. Ser¹⁹²⁸ in the distal C terminus is phosphorylated in response to β-adrenergic stimulation in cardiac myocytes (20, 25, 43). However, its phosphorylation is not required for the increase in L-type channel activity elicited by β-adrenergic stimulation of cardiac myocytes (34, 44). We identified two previously unrecognized phosphorylation sites in the C terminus of the Ca_V1.1 channel by mass spectrometry (MS) (45), and these sites are conserved in the Ca_V1.2 at Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴ (Fig. 5A). Ser¹⁷⁰⁰ is predicted to be a substrate for PKA and calcium/calmodulin-dependent protein kinase II (CaMKII), whereas Thr¹⁷⁰⁴ is predicted to be a substrate of casein kinase II (http://www.phosphosite.org). Both of these sites reside at the interface between the PCRD and the DCRD, where introduction of negatively charged phosphates could disrupt ionic interactions and relieve autoinhibition (Fig. 5A) (28). Analysis of in vitro phosphorylation of glutathione S-transferase (GST)–labeled Ca_V1.2(1670–1731) by MS showed that Ser¹⁷⁰⁰ was a good substrate for both PKA and CaMKII with 27% and 77% phosphorylation, respectively, whereas 10% of Thr¹⁷⁰⁴ was phosphorylated by casein kinase II (Fig. 5B). The extents of protein phosphorylation observed in these in vitro experiments reflect both the amino acid sequence contexts of the substrate sites and the local secondary and tertiary structures of the GST-labeled C-terminal protein, which may not retain native conformation in the absence of the complete Ca_V1.2 channel and its associated proteins.

Fig. 5.

Phosphorylation of Ser¹⁷⁰⁰ and Ser¹⁹²⁸ in Ca_V1.2 channels. (A) Top: docking model of the PCRD-DCRD complex in ribbon representation. Bottom: amino acid sequence surrounding Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴. A, Ala; I, Ile; S, Ser; G, Gly; D, Asp; L, Leu; T, Thr; E, Glu; R, Arg. (B) Purified GST-Ca_V1.2(1670–1731) was phosphorylated with indicated kinases, digested with trypsin, and subjected to LC-MS analysis. MH²⁺ extract ions corresponding to the unphosphorylated (0P) or monophosphorylated (1P) forms of the peptides are indicated. CK2, casein kinase II. (C) Cells expressing Ca_V1.2Δ1800 + DCT channels were treated as indicated, fixed, probed with antibody against Ca_V1.2-pS¹⁷⁰⁰, and visualized by indirect immunofluorescence. OA, okadaic acid; CsA, cyclosporin A; Iono, ionomycin. (D) Membranes from cells expressing Ca_V1.2FL, Ca_V1.2Δ1800, or Ca_V1.2Δ1800 + DCT channels were solubilized and separated by SDS-PAGE, and immunoblots were probed with anti-bodies against Ca_V1.2, Ca_V1.2-pS¹⁷⁰⁰, or Ca_V1.2-pS¹⁹²⁸. For quantitation, band intensities were normalized to those of the unstimulated control samples (n = 3). (E) Cells expressing Ca_V1.2Δ1800 + DCT channels, plus cDNA encoding the brain CaMKIIN inhibitor peptide as indicated, were pretreated with ionomycin and autocamtide-2 inhibitory peptide (AIP) and subjected to immunoblot analysis as in (D). For quantitation, band intensities were normalized to those of the unstimulated control samples (n = 3).

We used phosphospecific antibodies and immunocytochemistry to assess PKA phosphorylation of Ser¹⁷⁰⁰ and Ser¹⁹²⁸ in Ca_V1.2 channels expressed in tsA-201 cells. Cells expressing Ca_V1.2Δ1800 + DCT were treated with 10 μM forskolin to activate adenylyl cyclase and PKA, or 5 μM ionomycin to increase intracellular Ca²⁺ concentration and thereby activate CaMKII. Cells were fixed and antibody against pS¹⁷⁰⁰ immunoreactivity was visualized by indirect immunofluorescence. We observed an increase in phosphorylation of Ser¹⁷⁰⁰ after treatment with forskolin or ionomycin in the presence of 1 μM okadaic acid to inhibit phosphoprotein phosphatases 1 and 2A (Fig. 5C). When tested alone, neither okadaic acid (10 nM) nor the calcineurin inhibitor cyclosporin A (10 nM) substantially increased phosphorylation of Ser¹⁷⁰⁰ (Fig. 5C). These results indicate that both PKA and CaMKII can phosphorylate Ser¹⁷⁰⁰ of Ca_V1.2 channels in tsA-201 cells.

Immunoblot analysis showed measurable basal phosphorylation of Ser¹⁷⁰⁰ and small increases in antibody against pS¹⁷⁰⁰ immunoreactivity when cells expressing Ca_V1.2Δ1800 channels were treated with ionomycin in the presence of okadaic acid or with forskolin in the presence of okadaic acid and cyclosporin A (Fig. 5D). A more substantial increase in phosphorylation at Ser¹⁷⁰⁰ was observed for Ca_V1.2Δ1800 + DCT after treatment with forskolin (factor of 1.8 ± 0.4, P < 0.05, n = 3; Fig. 5D) or ionomycin (factor of 2.5 ± 0.04, P < 0.01, n = 3; Fig. 5D) in the presence of okadaic acid or cyclosporin A or both. A small increase in phosphorylation was also observed with a phosphospecific antibody directed against Ser¹⁹²⁸ with both ionomycin and forskolin treatment (Fig. 5D). The Ca²⁺-dependent increase in phosphorylation of Ser¹⁷⁰⁰ was blocked by inhibition of CaMKII with either CaMK inhibitory protein or autocamtide-2–related inhibitor peptide II (1 μM) (Fig. 5E). In contrast to these results with Ca_V1.2Δ1800 + DCT, we observed no change in phosphorylation of Ser¹⁷⁰⁰ of Ca_V1.2FL channels with drug treatment (forskolin: factor of 0.8 ± 0.1, P > 0.2; ionomycin: factor of 1.5 ± 0.7, P > 0.05; Fig. 5D). However, we did observe a robust increase in phosphorylation at Ser¹⁹²⁸ in Ca_V1.2FL channels (Fig. 5D). These results demonstrate a difference in PKA phosphorylation of Ser¹⁷⁰⁰ between Ca_V1.2FL and Ca_V1.2Δ1800 + DCT channels. Evidently, Ser¹⁷⁰⁰ is a substrate for basal PKA-dependent phosphorylation of Ca_V1.2 channels in transfected cells, but the increase in PKA phosphorylation of Ser¹⁷⁰⁰ is blocked by covalent association of the distal C terminus in Ca_V1.2FL.

Phosphorylation sites required for basal Ca_V1.2 channel activity

To test the role of individual phosphorylation sites in control of basal Ca_V1.2 channel activity, we made alanine substitutions for Ser¹⁷⁰⁰, Thr¹⁷⁰⁴, and Ser¹⁹²⁸ and recorded channel activity. The S1700A and T1704A mutations had no significant effects on the high activity of Ca_V1.2Δ1800 channels expressed without DCT (fig. S3). However, both S1700A and T1704A significantly reduced the basal coupling efficiency of Ca_V1.2Δ1800 + DCT and decreased basal Ba²⁺ currents (Fig. 6A, gray, blue), whereas S1928A did not have a significant effect (Fig. 6A, orange). The only double-alanine mutation to significantly reduce both basal coupling efficiency and Ba²⁺ currents of Ca_V1.2Δ1800 + DCT was S1700A, T1704A (Fig. 6B, cyan). These results indicate that phosphorylation of Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴ increases basal activity of Ca_V1.2Δ1800 + DCT channels, whereas phosphorylation of Ser¹⁹²⁸ does not.

Fig. 6.

Requirements for phosphorylation of Ser¹⁷⁰⁰, Thr¹⁷⁰⁴, and Ser¹⁹²⁸ for basal Ca_V1.2 channel activity. (A and B) Coupling efficiency (nA/pC) and current-voltage relationships for wild-type Ca_V1.2Δ1800 + DCT channels or channels with alanine substitutions at Ser¹⁷⁰⁰, Thr¹⁷⁰⁴, and Ser¹⁹²⁸ individually or simultaneously. *P < 0.05 or **P < 0.01 versus wild-type Ca_V1.2Δ1800 + DCT channels (I_Tail, peak tail current; Q, total gating current; V_rev, reversal potential). (C) Coupling efficiency (nA/pC) and current-voltage relationships for Ca_V1.2Δ1800 + DCT channels with triple-alanine substitution (S1700A, T1704A, and S1928A) or wild-type channels in the presence or absence of 1 μM RO 31-8220. **P < 0.01 versus Ca_V1.2Δ1800 + DCT channels in control conditions. (D) Coupling efficiency (nA/pC) and current-voltage relationships for Ca_V1.2FL channels with triple-alanine substitution (S1700A, T1704A, and S1928A) or wild-type channels in the presence or absence of 1 μM RO 31-8220. **P < 0.01 versus wild-type Ca_V1.2FL channels in control conditions. Significance was determined by ANOVA.

The largest reduction of basal Ba²⁺ currents and coupling efficiency was observed with Ca_V1.2Δ1800 + DCT channels containing alanine substitutions at each phosphorylation site (Fig. 6C, bright green); both the low coupling efficiency and the Ba²⁺ current for the triple-alanine mutant were similar to those observed for wild-type Ca_V1.2Δ1800 + DCT during inhibition of protein kinase activity (Fig. 6C, brown). Analogous results were observed for Ca_V1.2FL channels (Fig. 6D). These results show that the basal Ca_V1.2 channel activity depends on phosphorylation of Thr¹⁷⁰⁴ and Ser¹⁷⁰⁰, but not on phosphorylation of Ser¹⁹²⁸.

Phosphorylation sites required for PKA stimulation of Ca_V1.2 channel activity

We tested the role of phosphorylation of Ser¹⁷⁰⁰, Thr¹⁷⁰⁴, and Ser¹⁹²⁸ in PKA-dependent stimulation of Ca_V1.2 channels by stimulating adenylyl cyclase with 5 μM forskolin. No forskolin-induced stimulation of the triple-alanine mutant was observed (Fig. 7, A and B, bright green) compared to wild-type channels treated with RO 31-8220 (1 μM) (Fig. 7A, brown). These results indicate that both PKA-stimulated activity and basal activity of Ca_V1.2 channels require phosphorylation of one or more of these three sites.

Fig. 7.

Requirement for phosphorylation of Ser¹⁷⁰⁰, Thr¹⁷⁰⁴, and Ser¹⁹²⁸ for PKA-mediated stimulation of Ca_V1.2 channel activity. (A) Coupling efficiency for the indicated constructs of Ca_V1.2Δ1800 + DCT in the presence or absence of 1 μM RO 31-8220, AKAP15, and 5 μM forskolin. Dashed black line indicates the mean current in unstimulated Ca_V1.2Δ1800 + DCT. *P < 0.05 or **P < 0.01 versus AKAP15 and forskolin (I_Tail, peak tail current; Q, total gating current; V_rev, reversal potential). (B and C) Current-voltage relationships of Ca_V1.2Δ1800 + DCT from (A) as indicated. Significance was determined by ANOVA.

Restoration of Ser¹⁹²⁸ had no effect (Fig. 7, A and B, cyan). Restoration of Thr¹⁷⁰⁴ restored basal channel activity but not forskolin-stimulated activity (Fig. 7, A and B, yellow). Restoration of Ser¹⁷⁰⁰ yielded substantial PKA-dependent stimulation of coupling efficiency and Ba²⁺ currents (Fig. 7, A and B, purple), similar to the effects of forskolin on wild-type Ca_V1.2Δ1800 + DCT. Together, these results define distinct roles for phosphorylation of Ser¹⁷⁰⁰, Thr¹⁷⁰⁴, and Ser¹⁹²⁸ in regulation of Ca_V1.2 channels: Phosphorylation of Thr¹⁷⁰⁴ contributes primarily to basal regulation of channel function, whereas phosphorylation of Ser¹⁷⁰⁰ is uniquely required for PKA-dependent enhancement of channel activity. A similar pattern emerged when Ca_V1.2 channels containing only a single-alanine substitution were studied (Fig. 7, A and C). Mutant S1700A lost PKA-dependent modulation of Ca_V1.2 channel activity but retained basal activity (Fig. 7, A and C, gray-green). In contrast, single-alanine substitutions at Ser¹⁹²⁸ or Thr¹⁷⁰⁴ had no significant effect on coupling efficiency or Ba²⁺ currents (Fig. 7, A and C, blue and orange). Analyses of the coupling efficiency of individual cells supported these conclusions (fig. S2B). Five of eight (62.5%) cells coexpressing Ca_V1.2Δ1800(T1704A) + DCT with AKAP15 and 5 of 10 (50%) cells coexpressing Ca_V1.2Δ1800 + DCT(S1928A) with AKAP15 showed a higher coupling efficiency in the presence of 5 μM forskolin than the maximum observed with Ca_V1.2Δ1800 + DCT in the absence of forskolin (fig. S2D). In contrast, no individual cells showed a coupling efficiency in the same range when cells coexpressing Ca_V1.2Δ1800(S1700A) + DCT and AKAP15 were treated with 5 μM forskolin. Collectively, these results indicate that phosphorylation of Ser¹⁷⁰⁰ is required for PKA-mediated stimulation of Ca_V1.2 channels and that phosphorylation of Thr¹⁷⁰⁴ plays a primary role in regulation of basal activity. We did not identify a contribution for phosphorylation of Ser¹⁹²⁸ to short-term regulation of Ca_V1.2 channel activity.

DISCUSSION

Reconstitution of physiological regulation of Ca_V1.2 channels in non-muscle cells

PKA-mediated regulation of Ca_V1.2 channels in cardiac myocytes has several enigmatic aspects. The C-terminal domain is proteolytically processed in vivo (20, 46), but is required for regulation of Ca_V1.2 channels in cardiac myocytes (34). Despite the high concentration of PKA, AKAP anchoring is required for β-adrenergic regulation in ventricular myocytes (33); however, the AKAP15 binding site is located in the proteolytically processed distal C-terminal domain (33). This unexpected set of properties has confounded previous attempts to reconstitute PKA-dependent regulation of Ca_V1.2 channels in non-muscle cells, as required to define the molecular mechanism of this process. Indeed, only slight PKA-dependent modulation of full-length Ca_V1.2 channels has been previously observed, even with coexpression ofAKAP150 [for instance, (31, 35, 36, 42)], raising the possibility that key molecular components of the in vivo Ca_V1.2 channel complex were missing from these reconstituted systems. Our results resolve this conundrum. Here, we show that a reconstituted autoinhibitory Ca_V1.2 channel signaling complex containing AKAP15 and the noncovalently associated distal C-terminal domain is both necessary and sufficient to reconstitute physiological levels of PKA-dependent regulation of Ca_V1.2 channels, and we identify key phosphorylation sites that are required for basal regulation of Ca_V1.2 channel activity and for the PKA-dependent increase of channel activity. Because our reconstitution system incorporates all of the previously described characteristics of regulation of cardiac Ca_V1.2 channels in vivo, and resolves the apparent paradoxes in the previous in vivo studies, these results provide a conceptual and molecular framework for calcium channel regulation in the fight-or-flight response, in which PKA phosphorylation mediates disinhibition of an autoinhibitory signaling complex containing Ca_V1.2Δ1800 + DCT and AKAP15.

Requirement for AKAP15 for reconstitution of PKA regulation

AKAP15 was identified biochemically in purified preparations of skeletal muscle Ca_V1.1 channels (29) and cloned from skeletal and cardiac muscle (30, 31), where it was also designated AKAP18 (31). It anchors PKA to both Ca_V1.1 and Ca_V1.2 channels (29–31), and its interaction with the distal C-terminal domain of Ca_V1.2 channels is required for β-adrenergic regulation through the PKA pathway in ventricular myocytes (33). Consistent with this requirement for AKAP in ventricular myocytes, we find that AKAP15 is required for reconstitution of PKA-dependent modulation in non-muscle cells, supporting the conclusion that our in vitro reconstitution mimics this aspect of PKA regulation in vivo.

Requirement for noncovalent association of the distal C-terminal domain for PKA regulation of Ca_V1.2 channels

A surprising conclusion from our results is that the distal C-terminal domain is required for Ca_V1.2 channel regulation, but is effective only when it is noncovalently associated. Most Ca_V1.2 channels in cardiac and skeletal muscle are truncated near the center of the C terminus (18–20, 46), and about half of Ca_V1.2 channels in neurons are also truncated (23). The distal C-terminal domain is not degraded and is retained in noncovalent association with the channel (21, 28). Furthermore, the distal C-terminal domain is required for β-adrenergic regulation of truncated Ca_V1.2 channels in cardiac myocytes (34). Therefore, the requirement for noncovalent association of the distal C-terminal domain observed in our reconstitution system mimics this requirement in cardiac myocytes.

It is unexpected that the distal C-terminal domain is required for anchoring AKAP15 and PKA and yet must be truncated and noncovalently associated to mediate channel regulation. Our biochemical studies provide an explanation for this paradox. When full-length Ca_V1.2 channels are studied, activation of PKA does not increase phosphorylation of Ser¹⁷⁰⁰ (Fig. 5D). In contrast, activation of PKA increases phosphorylation of Ser¹⁷⁰⁰ substantially for Ca_V1.2Δ1800 + DCT (Fig. 5D). Evidently, covalent association of the distal C-terminal in the full-length Ca_V1.2 channel occludes Ser¹⁷⁰⁰ and reduces or prevents its phosphorylation by PKA. These results support the conclusion that the noncovalently associated complex of Ca_V1.2 channels with DCT is the primary physiological substrate for PKA-dependent enhancement of channel activity.

PKA-mediated enhancement of Ca_V1.2 channel activity results from disinhibition

The distal C-terminal inhibits the activity of Ca_V1.2 channels when it is noncovalently associated (28). Our results show that PKA phosphorylation reverses this inhibition and returns the activity of the noncovalently associated complex of Ca_V1.2Δ1800 + DCT to nearly that observed for Ca_V1.2Δ1800 alone. These results illustrate an additional essential property of the distal C-terminal domain—it must inhibit Ca_V1.2 channel activity so that PKA-mediated phosphorylation can relieve that inhibition and increase channel activity. Molecular models suggest that Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴ are located at the interface between the DCRD and the PCRD, where we have shown that charge-neutralizing mutations can block the inhibitory effects of the DCRD (Fig. 5A) (28). These results predict that phosphorylation at Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴ would disrupt the interaction between these two domains and lead to disinhibition of the channel. Thus, PKA-dependent regulation of Ca_V1.2 channels proceeds through disinhibition of an autoinhibited signaling complex of Ca_V1.2Δ1800, non-covalently associated DCT, and AKAP15. This mode of regulation by dis-inhibition of an autoinhibited signaling complex is unique for ion channels studied to date.

Differential regulation of basal and PKA-dependent Ca_V1.2 channel activity by protein phosphorylation

The basal activity of Ca_V1.2 channels is decreased by inhibition of PKA and by mutation of Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴ (Figs. 2 and 6). These results implicate phosphorylation of Ser¹⁷⁰⁰ by PKA and Thr¹⁷⁰⁴ by casein kinase II in regulation of basal Ca_V1.2 channel activity in unstimulated cells. In contrast, the forskolin-dependent increase in Ca_V1.2 channel activity required phosphorylation of Ser¹⁷⁰⁰, and little or no contribution of phosphorylation of Thr¹⁷⁰⁴ or Ser¹⁹²⁸ was detected. Because the response to activation of PKA was completely blocked for mutant S1700A, it is likely that Ser¹⁷⁰⁰ is the primary site of regulation of Ca_V1.2 channels in response to the β-adrenergic receptor–PKA signaling pathway.

In contrast to Ser¹⁷⁰⁰ and Thr¹⁷⁰⁴ in the proximal C-terminal domain, Ser¹⁹²⁸ is located far from the interface between the DCRD and the PCRD. Although Ser¹⁹²⁸ is phosphorylated after β-adrenergic activation in ventricular myocytes (25), we found no effect of Ser¹⁹²⁸ phosphorylation in regulation of Ca_V1.2 channels in our reconstituted regulatory system. In this respect, our results fit closely with those observed in vivo, where phosphorylation of Ser¹⁹²⁸ was found not to be required for β-adrenergic regulation of Ca_V1.2 channels in cardiac myocytes in studies using viral transduction or mouse knock-in mutation methods (34, 44). Because Ser¹⁹²⁸ is robustly phosphorylated in response to β-adrenergic stimulation, it seems likely that its phosphorylation serves an unidentified regulatory function.

Dual regulatory role of the distal C-terminal domain of Ca_V1.2 channels

The distal C-terminal domain of Ca_V1.2 channels has been implicated in regulation of gene expression in heart (47) and brain (48). In cardiac myocytes, some of the proteolytically cleaved distal C-terminal domain is found in the nucleus and can regulate the transcription of Ca_V1.2 messenger RNA (mRNA) (47). In brain neurons, the distal C-terminal domain is proteolytically processed in response to calcium entry (24). The distal C-terminal protein is localized in the nuclei of a small fraction of brain neurons and can regulate the transcription of neuronal genes (48). The distal C-terminal domain also binds the calcium-dependent phosphatase calcineurin through anchoring by AKAP150 (36), and anchored calcineurin dephosphorylates the transcription factor NFAT (nuclear factor of activated T cells) and regulates gene expression (36).

Ca_V1.2 channel regulation and disease

Heart failure is one of the most debilitating diseases of the cardiovascular system. Increasing evidence indicates that heart failure involves, at least in part, a maladaptive regulation of calcium signaling in the heart, in which normal β-adrenergic regulation of Ca_V1.2 channels fails (49, 50). The importance of precise regulation of Ca_V1.2 channels is also illustrated by Timothy syndrome, a multifaceted disease in which mutations that impair voltage-dependent inactivation cause arrhythmia, developmental abnormalities, and autism spectrum disorder (51). Our results provide the molecular basis for future investigations of the role of misregulation of Ca_V1.2 channels by the PKA pathway in heart failure, Timothy syndrome, and other diseases.

MATERIALS AND METHODS

Antibodies and cDNA constructs

cDNA constructs (28), antibody against Ca_V1.2, and antibody against pS¹⁹²⁸ (20) were described previously. Phosphospecific antibody against Ca_V1.2-pS¹⁷⁰⁰ was generated against residues ¹⁶⁹⁴EIRRAIpSGDLTAEEEL in the proximal C terminus (Pacific Immunology). Mutants S1700A, T1704A, and S1928A were constructed as described in the Supplementary Materials. GST-Ca_V1.2(1670–1731) peptides were expressed and purified from BL21-STAR Escherichia coli.

Electrophysiology

tsA-201 cells were transfected, whole-cell voltage clamp recordings were performed, and data were analyzed as previously described (28) (see Supplementary Materials). All data are expressed as means ± SEM of n cells. Statistical significance was tested with Student’s t test for pairwise analysis, and analysis of variance (ANOVA) followed by Dunnett’s test for comparison of multiple conditions.

Immunofluorescence, immunoblots, and MS

For immunofluorescence experiments, cells were rinsed with phosphate buffered saline (PBS) (pH 7.4) and fixed for 45 min in 4% paraformaldehyde. Indirect immunofluorescence detection was performed with antibody against Ca_V1.2-pS¹⁷⁰⁰ and biotinylated goat antibody against rabbit. Samples for all images were prepared identically and viewed at a common gain to permit qualitative comparisons. For immunoblot analysis, cells were washed with PBS 24 hours after transfection and treated with drugs as indicated; membranes were collected, solubilized, and separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE); and the resulting blots were probed with the indicated antibodies. For MS, phosphorylated GST-Ca_V1.2(1670–1731) peptides were digested with trypsin and subjected to liquid chromatography–MS (LC-MS) analysis. See Supplementary Materials and Methods for details.