Different Subcellular Populations of L-type Ca²⁺ Channels Exhibit Unique Regulation and Functional Roles in Cardiomyocytes

Abstract

Influx of Ca²⁺ through L-type Ca²⁺ channels (LTCCs) contributes to numerous cellular processes in cardiomyocytes including excitation-contraction (EC) coupling, membrane excitability, and transcriptional regulation. Distinct subpopulations of LTCCs have been identified in cardiac myocytes, including those at dyadic junctions and within different plasma membrane microdomains such as lipid rafts and caveolae. These subpopulations of LTCCs exhibit regionally distinct functional properties and regulation, affording precise spatiotemporal modulation of L-type Ca²⁺ current (I_Ca,L). Different subcellular LTCC populations demonstrate variable rates of Ca²⁺-dependent inactivation and sometimes coupled gating of neighboring channels, which can lead to focal, persistent I_Ca,L. In addition, the assembly of spatially defined macromolecular signaling complexes permits compartmentalized regulation of I_Ca,L by a variety of neurohormonal pathways. For example, β-adrenergic receptor subtypes signal to different LTCC subpopulations, with β₂-adrenergic activation leading to enhanced I_Ca,L through caveolar LTCCs and β₁-adrenergic stimulation modulating LTCCs outside of caveolae. Disruptions in the normal subcellular targeting of LTCCs and associated signaling proteins may contribute to the pathophysiology of a variety of cardiac diseases including heart failure and certain arrhythmias. Further identifying the characteristic functional properties and array of regulatory molecules associated with specific LTCC subpopulations will provide a mechanistic framework to understand how LTCCs contribute to diverse cellular processes in normal and diseased myocardium.

1. Introduction

In the heart, voltage-dependent L-type Ca²⁺ channels (LTCCs) are essential to numerous cellular processes including excitability, excitation-contraction (EC) coupling, hormone secretion, and regulation of gene expression. Participation in such diverse functions demands that the influx of Ca²⁺ through L-type channels (L-type Ca²⁺ current, I_Ca,L) is tightly controlled and compartmentalized within the cardiac myocyte. It has long been recognized that discrete clusters of LTCCs exist along the sarcolemma, and studies in recent years have greatly extended our understanding of how specific subcellular localization impacts channel function and regulation by a variety of neurohormonal and second messenger pathways [1-6].

A number of important LTCC subpopulations have been identified in cardiomyocytes that associate with unique macromolecular signaling complexes and scaffolding proteins, which enables spatiotemporal modulation of I_Ca,L. These include channels that are localized to dyadic junctions as well as extradyadic channels that reside in biochemically distinct regions of surface membrane known as membrane microdomains. Plasma membrane microdomains, including lipid rafts and caveolae, exhibit unique lipid composition and protein components and coordinate numerous cellular functions including various signal transduction pathways and protein recycling [7-9]. Numerous signaling molecules have been localized to caveolae including components of the β₂-adrenergic receptor/adenylyl cyclase/protein kinase A (PKA) cascade [5,6]. This review will highlight the evolving understanding of distinct subcellular populations of LTCCs in cardiomyocytes and their differing regulation and contributions to Ca²⁺ signaling in the heart.

2. LTCCs in the heart

2.1. Molecular composition of cardiac LTCCs

LTCCs are multimeric complexes consisting of a pore forming α₁ subunit and auxiliary β, α₂δ, and α subunits [10]. The α₁ subunit serves as the main functional component of the channel complex and consists of four homologous domains (I-IV) each containing six transmembrane segments (S1-S6). Ca_v1.2 (α_1C, encoded by the CACNA1C gene) is the predominant α₁ subunit in ventricular myocardium, whereas both Ca_v1.2 and Ca_v1.3 (α_1D, encoded by CACNA1D) are expressed in atrial tissue as well as nodal cells, where I_Ca,L contributes to automaticity [11-15]. Extensive alternative splicing of Ca_v1.2 has been reported, and these splice variants play unique roles in cardiovascular physiology, pharmacology, and disease [16,17]. One important example is alternative splicing of Ca_v1.2 within transmembrane segment IS6, which impacts sensitivity to the dihydropyridine class of LTCC blockers. Differential expression of these splice variants lead to higher or lower sensitivity in smooth and cardiac muscle, respectively [18,19].

Ca²⁺ channel auxiliary subunits further add to the functional diversity of LTCCs. The cytosolic β subunits promote trafficking of the channel complex to the plasma membrane and modulate gating properties of the channel [20-22]. The β subunits are encoded by four distinct genes (CACNB1-4), each of which undergoes alternative splicing to generate a total of 18 or more unique β subunit isoforms in human myocardium [23]. The α₂δ subunits arise from a common precursor protein that is post-translationally cleaved and relinked via a disulfide bridge. The extracellular α₂ peptide is heavily glycosylated and the δ peptide contains a single transmembrane domain [24]. Of the four α₂δ subunits (encoded by CACNA2D1-4), α₂δ-1-3 are expressed in atrial tissue whereas α₂δ-1 and α₂δ-2 are present in ventricular myocardium [25,26]. The α₂δ subunits modify both channel gating properties and surface membrane expression of the L-type channel complex [20,27]. Ca²⁺ channel α subunits, of which eight exist (encoded by CACNG1-8), were originally demonstrated to associate with voltage-dependent Ca²⁺ channels in skeletal muscle and brain [28,29]. However, recent evidence suggests several α subunits including α4, α6, α7, and α8 are present in cardiac muscle and associate with the cardiac Ca_v1.2 channel complex, altering both activation and inactivation properties of the channel [30].

2.2. Distinct LTCC subpopulations in cardiac myocytes

2.2.1. Dyads

A critical subpopulation of LTCCs is that which participates in EC coupling. A number of studies applying immunoconfocal and electron microscopy techniques have demonstrated that a subset of LTCCs form dyadic complexes with Ca²⁺-release channels (ryanodine receptors) on apposing junctional sarcoplasmic reticulum (SR) [1,2,31,32]. Upon membrane depolarization, activation of these LTCCs leads to an influx of Ca²⁺ into the dyadic cleft space, which triggers the opening of ryanodine receptors and subsequent release of SR Ca²⁺ stores. This Ca²⁺-induced Ca²⁺ release (CICR) mechanism underlies the rise in free intracellular Ca²⁺ concentration ([Ca²⁺]_i) that activates myofilament proteins leading to muscle contraction [33].

Studies have estimated that approximately 75% of LTCCs reside at dyad junctions in cardiac myocytes [34]. In mammalian ventricular cardiomyocytes, dyadic couplings occur predominantly within the transverse (T)-tubule network, which represents a complex system of interconnected membrane structures continuous with the extracellular space [35]. T-tubules extend deep into the myocyte at each myofibrillar Z-line, bringing the plasma membrane in close proximity with junctional SR through the width of the cell, thereby allowing synchronous Ca²⁺ release throughout the cell with each depolarization. However, as many as 25% of dyads are found at the surface sarcolemma in ventricular myocytes [35,36]. Studies in which the T-tubule system is isolated from the surface membrane using osmotic shock led to an 75-80% decrease in I_Ca,L without significantly altering SR Ca²⁺ load; therefore, it is hypothesized that although the majority of LTCCs reside in the T-tubular membrane, surface membrane LTCCs may play a particularly important role in SR Ca²⁺ loading [37].

Dyadic LTCCs are the target of numerous neurohormonal signaling pathways, most prominently the β-adrenergic/PKA signaling pathway, which increases contractility partially through enhancing I_Ca,L [38]. Whereas extradyadic LTCCs do not contribute directly to EC coupling, these channels likely play roles in several other cellular processes. These extradyadic channels are not found randomly throughout the sarcolemma, but associate with important signaling molecules in biochemically-distinct membrane microdomains [3,4,6,39].

2.2.2. Lipid rafts

The plasma membrane is a heterogeneous mixture of lipids, cholesterol, and proteins that can form localized regions that compartmentalize cellular processes. The tight packing of sphingolipids and cholesterol forms liquid-ordered microdomains popularly termed “lipid rafts”, which serve a number of important cellular functions including signal transduction and membrane trafficking [7,8]. Although the small (10-200 nm), heterogeneous, and dynamic nature of lipid rafts has made their visualization and characterization challenging, one common feature is that these microdomains resist solubilization with non-ionic detergents such as Triton X-100 at 4°C and can be found in low density fractions on density gradients [8,40,41]. A number of studies utilizing this biochemical methodology have demonstrated the presence of Ca_v1.2 in low density fractions from ventricular myocardium [5,42,43], strongly suggesting LTCCs localize to lipid raft domains.

2.2.3. Caveolae

Caveolae represent a subset of lipid rafts that are morphologically distinct structures appearing in electron micrographs as flask-shaped invaginations of membrane that are 50-100 nm in diameter. Caveolae are rich in cholesterol and are associated with the integral membrane protein caveolin [44]. Three genes (CAV1-3) encode six known caveolin subtypes (caveolin-1α and −1β; caveolin-2α, −2β, and −2α; and caveolin-3) that have varying tissue distribution [45]. Caveolin (Cav)-1 and Cav-2 are expressed in most cell types, whereas expression of Cav-3 is restricted to cardiac, skeletal, and smooth muscle types, as well as some types of neurons [45-47].

The lack of morphologically distinct caveolae in skeletal muscle from Cav-3 knockout mice highlights the importance of caveolin in caveolae formation [48]. Cholesterol is an additional component necessary for caveolar assembly and maintenance, because depletion of membrane cholesterol dramatically reduces caveolae number [49]. More recently, the cytosolic protein Polymerase I and Transcript Release Factor (PTRF), also known as Cavin, was also demonstrated to be required for caveolar biogenesis [50].

Several methods have been used to demonstrate the presence of LTCCs within caveolae in cardiomyocytes, exploiting the presence of Cav-3 as a marker to help differentiate the localization of proteins to caveolae versus non-caveolar lipid rafts. First, co-immunoprecipitation from ventricular myocytes using anti-Cav-3 or anti-Ca_v1.2 showed that the two proteins are associated [5,51]. Immunofluorescence studies of isolated ventricular myocytes using confocal microscopy also demonstrated extensive colocalization between Cav-3 and Ca_v1.2 [5,43]. Perhaps most strikingly, immunogold electron microscopy showed colabeling of Cav-3 and Ca_v1.2 in close association within morphologically distinct caveolae (Fig. 1A,B) [5,51].

Figure 1. Cardiomyocyte L-type Ca²⁺ channels localized to caveolae are regulated by β₂-adrenergic receptor activation.

(A,B) Immunogold electron micrographs of mouse neonatal ventricular myocytes demonstrate co-localization of the Ca_v1.2 subunit of L-type Ca²⁺ channel (large particle, arrows) with Cav-3 (small particle, arrowheads) within caveolae. Scale bars represent 200 nm. (C) β₂-adrenergic receptor activation with 10 μM salbutemol (Sal) plus 10 μM atenolol (Aten) increases peak I_Ca,L in neonatal mouse ventricular myocytes under control conditions. (D) β₂-adrenergic stimulation of I_Ca,L is lost in cells in which caveolae are disrupted using Cav-3 siRNA. Copyright (2006) National Acadamy of Sciences, USA [5].

2.2.4. Caveolin-3 scaffolds

The presence of caveolins in cell types lacking morphologically distinct caveolae such as neurons has sparked interest in understanding the physiological role of caveolins independent of caveolae. It has been suggested that extra-caveolar caveolin forms scaffolds that regulate protein trafficking and signal transduction in some cell types [52]. Cav-3 scaffolds may be particularly important in cardiac myocytes, which have relatively few caveolae compared to other cell types such as endothelial cells and smooth muscle myocytes, yet express Cav-3 at high levels [52,53]. In adult rat cardiomyocytes, a significant amount of Cav-3 is found in heavy (non-buoyant) fractions, and immunoprecipition studies showed that Cav-3 associates with a number of signaling proteins including G-protein coupled receptors, adenylyl cyclase, and heterotrimeric G proteins even in heavy fractions [54]. These findings suggest Cav-3 may regulate signal transduction in sarcolemmal regions outside of caveolae in cardiac myocytes. However, no data yet directly demonstrate localization of LTCCs to noncaveolar Cav-3 scaffolds and additional approaches such as immunogold electron microscopy are needed to carefully investigate this possibility.

2.2.5. Nucleus

The carboxyl (C)-terminus of Ca_v1.2 undergoes proteolytic cleavage and serves as an autoinhibitory domain by reassociating with the truncated Ca_v1.2 subunit at the surface membrane [55-57]. Recently, the Ca_v1.2 C-terminus was also localized to the nucleus in cardiac myocytes, where it autoregulates transcription [58]. Western blotting of nuclear fractions showed expression of the cleaved Ca_v1.2 C-terminus in nuclear fractions, and GFP-tagged Ca_v1.2 C-terminus showed nuclear localization [58]. Interestingly, expression of various β subunits fused to GFP in ventricular myocytes also revealed significant localization of β₄ in the nucleus; however, the significance of this is currently unclear [59].

2.2.6. Other subcellular compartments

Additional specialized compartments may exist that could prove important for LTCC modulation. For example, a recent study using electron tomography identified electron-dense bridges between dyadic clefts and the mitochondrial outer membrane in mouse ventricular myocardium, and I_Ca,L has been demonstrated to alter mitochondrial function either directly or through association with actin filaments [60-62]. This raises the possibility of a functional coupling between dyadic LTCCs and mitochondria. Similarly, a close spatial relationship exists between the T-tubule membrane, dyadic cleft, and the nuclear envelope [63]. However, the detailed identity and function of these different subcellular populations of LTCCs are largely unknown at this time.

Furthermore, the subcellular localization of LTCCs may be more complex than currently appreciated. There is a dynamic range in the size of dyadic junctional spaces, for instance, which could differentially impact LTCC function at individual dyads [60]. Likewise, subcompartmentalization of membrane microdomains into smaller nanodomains may be necessary to more precisely define their function and impact on LTCCs. For example, it is currently unknown whether subpopulations of LTCCs coupled to different signaling pathways coexist within the same caveola. Addressing these possibilities will require refinement of available biochemical and microscopy techniques.

3. Subcellular localization impacts LTCC function

With each heart beat, I_Ca,L activates ryanodine receptors to release SR Ca²⁺ stores, leading to a transient rise in global [Ca²⁺]_i from ~100 nM during diastole to ~1 μM during systole, which activates myofilament proteins, producing contraction. How then, is specificity of Ca²⁺ signaling achieved within the cardiomyocyte? It is increasingly recognized that the localization of LTCCs to structurally or biochemically distinct subcellular regions affords functional and physical compartmentalization. Within these spatially defined areas, local [Ca²⁺] may differ significantly from bulk cytosolic [Ca²⁺]_i, which modulates LTCC function and differentially regulates a number of Ca²⁺-dependent cellular processes [64,65].

3.1. Local effects on Ca²⁺-dependent inactivation of LTCCs

Upon opening, LTCCs undergo rapid voltage-dependent and Ca²⁺-dependent inactivation that limits the amount of Ca²⁺ entry during each action potential. The mechanism of Ca²⁺-dependent inactivation (CDI) involves tethering of the Ca²⁺ sensor protein calmodulin (CaM) to the carboxyl tail of Ca 1.2 [66,67]. Ca²⁺v ions near the mouth of the channel bind to CaM, which subsequently changes conformation to bind a nearby IQ motif, accelerating channel inactivation. The degree of CDI observed depends on local [Ca²⁺] and thus may differ among channels localized to various subcellular environments [68].

At the dyadic cleft, a spatially restricted area estimated between 4.39 × 10⁵ nm³ and 1.5 × 10⁶ nm³ in size, I_Ca,L and subsequent SR Ca²⁺ release combine to dramatically increase local [Ca²⁺] from approximately 100 nM during diastole to 100-600 μM during systole [60,69]. Elevated [Ca²⁺] in the dyadic space during CICR greatly enhances CDI of LTCCs, which impacts action potential duration in ventricular myocytes (Fig. 2) [70-75]. This was demonstrated in rat ventricular myocytes by depleting SR Ca²⁺ stores with thapsigargin to abolish CICR, which slowed CDI, resulting in a 40-70% increase in I_Ca,L during the action potential waveform and prolonging action potential duration [72].

Figure 2. Schematic of L-type Ca²⁺ channel subpopulations and associated macromolecular signaling complexes that contribute to diverse processes within the cardiac myocyte.

Within T-tubules, some L-type Ca²⁺ channels (LTCCs) form dyadic junctions with ryanodine receptors (RyR2) on the apposing sarcoplasmic reticulum (SR). A small influx of Ca²⁺ through LTCCs triggers release of SR Ca²⁺ stores through RyR2, leading to the intracellular [Ca²⁺]_i transient, which is essential for activation of myofilament proteins leading to contraction. β-adrenergic receptor (β-AR) stimulation of [Ca²⁺]_i transients involves a signaling complex consisting of LTCCs, β-AR, adenylyl cyclase (AC), protein kinase A (PKA), caveolin-3 (Cav-3), and A-kinase anchoring proteins (AKAPs) [85]. Multiple AKAPs are important for regulating LTCC function in the heart including AKAP5 and AKAP15 [85,131]. Additional LTCC subpopulations in the surface sarcolemma are implicated in signaling to the nucleus to regulate the transcription of genes involved in cell survival or cardiac hypertrophy and may also contribute to membrane excitability. Within caveolae, LTCCs associate in a macromolecular protein complex with β₂-AR, AC, PKA, protein phosphatase 2A (PP2A), and Cav-3. Caveolar LTCCs are locally stimulated by β₂-AR activation [5]. Coupled gating of closely neighboring LTCCs results in persistent Ca²⁺ sparklets, which are enhanced by AKAPs and protein kinase C alpha (PKCα) and may serve as a local source of Ca²⁺ to stimulate calcineurin (PP2B) signaling to the nucleus [83,97].

Interestingly, LTCCs outside of T-tubules have been shown to exhibit slower inactivation kinetics, indicating surface membrane channels are less sensitive to SR Ca²⁺ release [74,76]. It is not clear whether subpopulations of surface membrane LTCCs localized to membrane microdomains such as lipid rafts and caveolae have a characteristic ‘signature’ of CDI. However, experiments using fluorescence-based Ca²⁺ sensors have indicated that subcaveolar [Ca²⁺] is higher than levels associated with the surrounding plasma membrane in endothelial cells [77]. Given the known clustering of LTCCs in lipid raft and caveolar microdomains in cardiomyocytes, it is possible that locally elevated subcaveolar [Ca²⁺] may contribute to unique CDI among caveolar LTCCs relative to channels residing outside of caveolae [5,43,51].

There is evidence that the conformational change of the CaM-bound C-terminus of Ca_v1.2 during CDI directly impacts Ca²⁺-dependent cellular processes such as excitation-transcription coupling and SR Ca²⁺ release [78]. For instance, activation of cAMP-responsive element-binding protein (CREB)-mediated gene transcription upon depolarization is dependent on a freely moving Ca_v1.2 C-terminus as well as the binding of CaM. This was demonstrated through experiments in which the Ca_v1.2 C-terminus was either immobilized to the plasma membrane by fusing it to a pleckstrin homology domain or made insensitive to CaM, both of which diminished CDI and abolished CREB activation despite an influx of Ca²⁺ and corresponding rise in [Ca²⁺]_i [79]. These findings suggest at least one mechanism of excitation-transcription coupling is attributed more to the movement of CaM-bound Ca_v1.2 C-terminus than to I_Ca,L or changes in [Ca²⁺]_i [78, 79]. Thus, overall differences in the degree and kinetics of CDI among subpopulations of LTCCs provide a mechanism to differentially regulate Ca²⁺ signaling processes within distinct subcellular domains of cardiac myocytes.

3.2. Coupled gating of LTCCs

Ca²⁺ influx through single LTCCs or small clusters of channels can be measured with high spatiotemporal resolution using total internal reflection fluorescence (TIRF) microscopy [80,81]. Whereas single L-type channels normally exhibit random, infrequent openings that permit small rises in submembrane [Ca²⁺] termed “Ca²⁺ sparklets”, some channels associate in highly active clusters that generate localized regions of significantly elevated [Ca²⁺] [82,83]. These “persistent Ca²⁺ sparklets” arise from the coupled opening of as many as six adjacent LTCCs [83]. Coupled gating of neighboring LTCCs associated with persistent Ca²⁺ sparklets is enhanced by the scaffold protein A-kinase anchoring protein (AKAP) 5 (also known as AKAP 79/150) as well as protein kinase C alpha (PKCα) [83]. Thus, subpopulations of LTCCs associating with AKAP5 and PKCα may serve as specialized Ca²⁺-signaling domains that are coupled to discrete cellular functions requiring locally high [Ca²⁺] [83,84]. Whether LTCCs giving rise to persistent Ca²⁺ sparklets localize to specific subcellular compartments remains unclear; however, AKAP5 has been shown to associate with Cav-3, indicating caveolae or Cav-3 scaffolds may serve an important role in modulating persistent Ca²⁺ sparklet activity [85]. Future work is needed to define the precise subcellular localization and physiological role of persistent Ca²⁺ sparklets in cardiac myocytes.

3.3. Membrane microdomains and excitation-transcription coupling

LTCCs provide a critical link between cellular excitability and gene regulation. Considering transcriptional regulation is on a time course of minutes to hours, it has remained puzzling how cardiac myocytes regulate Ca²⁺-dependent gene regulation given the rhythmic cycling of cytosolic [Ca²⁺]_i with each heart beat. This has led some to propose that specific membrane microdomain compartments are responsible for Ca²⁺-dependent signal transduction to the nucleus [86,87].

CaM plays a significant role in coupling I_Ca,L to transcriptional regulation in cardiac myocytes, particularly in the activation of hypertrophic signaling [88,89]. Ca²⁺-CaM activates a number of transcriptional pathways, and the two most prominent in cardiac muscle involve either Ca²⁺-CaM-dependent protein kinases (CaMK) or the phosphatase calcineurin (protein phosphatase 2B, PP2B) [88]. The importance of I_Ca,L in stimulating hypertrophic remodeling was recently demonstrated in transgenic mice with cardiac-specific overexpression of the LTCC β_2a subunit to increase I_Ca,L, which develop hypertrophy that is inhibited using blockers of LTCCs, CaMK, or calcineurin [90].

Relatively little is currently known regarding whether specific LTCC populations are responsible for signaling to the nucleus through the CaM-CaMK pathway. Interestingly, CaMKIIα has been localized to lipid rafts in neurons, and disruption of lipid rafts using methyl-β-cyclodextrin reduced phosphorylation of CaMKII substrates [91,92]. However, several studies have shown that overexpression of the nuclear localized CaMKIV and CaMKII-δ_B isoforms stimulate cardiomyocyte hypertrophy whereas the nonnuclear CaMKIIα does not seem to play a significant role in hypertrophic remodeling [93,94]. There is evidence that I_Ca,L triggers translocation of CaM from the cytosol to the nucleus, where it activates nuclear CaMKIV [95]. In cardiomyocytes, CaMKIV induces a hypertrophic response through activation of the transcription factor MEF2 [94].

Activation of the Ca²⁺-CaM-calcineurin pathway leads to dephosphorylation of members of the nuclear factor of activated T cells (NFAT) transcription factor family within the cytosol, which then translocate to the nucleus to regulate genes involved in cardiac hypertrophy [96]. In arterial smooth muscle cells and neurons, a macromolecular signaling complex between Ca_v1.2, AKAP5, and calcineurin couples I_Ca,L with NFAT-mediated gene transcription [97,98]. Recently, calcineurin has been demonstrated to associate with Ca_v1.2, AKAP5, and Cav-3 in ventricular myocytes [85]. This raises the possibility that LTCCs within caveolae or associated with Cav-3 scaffolds specifically couple to calcineurin-NFAT-mediated gene regulation (Fig. 2); however, little experimental evidence currently exists to directly support this hypothesis. It is worth noting that Cav-3 has a known role in cardiac hypertrophy, as knockout of Cav-3 leads to hypertrophic cardiomyopathy in mice [99].

3.4. Membrane microdomains and excitation-secretion coupling

LTCCs provide a critical influx of Ca²⁺ necessary for secretion of hormones in some cell types [100,101]. Ca_v1.2 associates with several soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins involved in exocytosis [102,103]. In pancreatic β-cells, it has been shown that the coupling of LTCCs to SNARE proteins permits locally elevated [Ca²⁺] near exocytotic machinery to trigger insulin secretion [102].

In atrial myocytes, LTCCs regulate secretion of atrial natriuretic peptide (ANP), a hormone with diuretic, natriuretic, and vasodilatory properties [104]. However, the precise role of I_Ca,L remains unclear. Some studies have shown that I_Ca,L stimulates ANP release, but other studies have suggested that I_Ca,L inhibits ANP secretion [105-112]. ANP release is also largely controlled by mechanical stretch, and I_Ca,L is coupled to the regulation of stretch-induced ANP secretion, suggesting there may be different roles for Ca²⁺ in basal versus stimulated release [112-114].

Little is known regarding the subcellular localization of ANP release machinery in atrial myocytes. However, several aspects of caveolae make this microdomain an attractive candidate for coordinating secretion of ANP. First, ANP has been localized to caveolae in atrial myocytes using electron microscopy [115]. Second, the localization of SNARE proteins to cholesterol-rich microdomains has been reported in various cell types and methyl-β-cyclodextrin treatment reduces exocytosis of secretory vesicles [116-118]. Finally, caveolae contain a subpopulation of LTCCs and are essential for translating mechanical stretch to activation of downstream signaling pathways, both of which impact the rate of ANP release [5,119]. Thus, it is possible that ANP release from atrial myocytes occurs in specialized membrane microdomains such as caveolae where LTCCs have preferred access to secretory machinery.

3.5. Modulation of I_Ca,L by membrane cholesterol and lipids

Relative to the bulk plasma membrane, lipid rafts and caveolae are enriched in sphingolipids, cholesterol, and phosphoinositides, which have been shown to modulate the functional properties of a number of ion channels [39,120]. Experimental approaches to investigate the role of cholesterol or lipids on ion channel function are often complicated because these interventions also target a large number of signaling molecules that regulate ion channel activity. However, there is some evidence that cholesterol directly modulates I_Ca,L. Exposure of smooth muscle cells to cholesterol-enriched liposomes led to a gradual increase in I_Ca,L and a positive shift in the voltage dependence of inactivation that was speculated to result from changes in membrane physical properties [121]. Importantly, lipid-altering strategies such as statin drugs or dietary supplementation with n-3 polyunsaturated fatty acids have shown promise as anti-arrhythmic therapy. These interventions are believed to exert effects via several mechanisms including modulation of ion channel expression and function [122-124]. The antiarrhythmic action of polyunsaturated fatty acids such as eicosapentaenoic acid has been partially attributed to a decrease in I_Ca,L [122,125]. Likewise, simvastatin has been shown to suppress I_Ca,L in mouse ventricular myocytes [126]. Studies have also shown that statins and diets rich in n-3 polyunsaturated fatty acids alter lipid raft and caveolar composition [127-130]. Future work is needed to investigate whether these lipid-altering therapies specifically result in changes in the localization, function, or regulation of LTCCs within membrane microdomains.

4. Unique regulation of LTCC subpopulations

Neurohormonal regulation of LTCCs is central to the ability of the heart to adapt to changing physiological needs by altering heart rate and contractility [131,132]. Membrane microdomains such as lipid rafts and caveolae are particularly well-suited for regulation of I_Ca,L because of the targeting of proteins involved in a variety of signaling cascades to these domains. Co-localization of signaling molecules with LTCCs enables highly localized and specific regulation of the channels.

4.1. Localization of LTCCs to a caveolar macromolecular signaling complex is required for β₂-adrenergic regulation

The sympathetic nervous system drives the fight-or-flight response through activation of β-adrenergic receptors, and one of the essential downstream targets of β-adrenergic regulation is the LTCC [38]. Three subtypes of β-adrenergic receptors exist (β_1-3) in myocardium and notable differences exist in the manner by which each receptor affects downstream pathways. For example, chronic activation of β₁-adrenergic receptors results in cardiomyocyte apoptosis, whereas β₂-adrenergic receptor stimulation is cardioprotective [133]. Furthermore, β₁ receptor activation causes a global cellular response that includes PKA-mediated phosphorylation of multiple proteins involved in EC coupling including LTCCs and the SR protein phospholamban, yet β₂ receptors signal locally to increase I_Ca,L without affecting other PKA-dependent processes such as phospholamban phosphorylation [134,135].

A number of recent studies have helped explain these divergent physiological responses observed with β₁ versus β₂-adrenergic receptor activation. A subpopulation of LTCCs was identified as part of a caveolar macromolecular complex with the β₂-adrenergic receptor, adenylyl cyclase, Gα_s, Gα_i, and protein phosphatase 2A (PP2A) in ventricular myocytes [5]. Disruption of caveolae in neonatal mouse ventricular myocytes with either methyl-β-cyclodextrin or short interfering RNA (siRNA)-mediated knockdown of Cav-3 abolished β₂-adrenergic stimulation of I_Ca,L (Fig. 1C,D) without effecting β₁-adrenergic enhancement of I_Ca,L[5]. Similar studies utilizing methyl-β-cyclodextrin or intracellular application of Cav-3 antibody in adult rat ventricular myocytes showed disruption of normal β₂ regulation of I_Ca,L [42]. These results demonstrate specific coupling of LTCCs with β₂-adrenergic receptors within caveolae or in a complex scaffolded by Cav-3 and suggest that channels coupled to β₁-adrenergic signaling molecules localize outside of cholesterol-rich membrane microdomains. Interestingly, I_Ca,L-mediated cardiomyocyte apoptosis resulting from β₁-adrenergic activation involves CaMKII, which has not been reported in caveolae [136]. Thus, the specific coupling of caveolar LTCCs with β₂-adrenergic receptors may stimulate pro-survival signaling (Fig. 2) [86,137].

Activation of Gα_s-coupled receptors activates adenylyl cyclase, resulting in production of cAMP. Monitoring subcellular localization of cAMP pools using fluorescence resonance energy transfer (FRET)-based biosensors in ventricular myocytes has demonstrated that stimulation of β-adrenergic receptors using isoproterenol leads to cAMP accumulation within cytosolic and caveolar compartments, whereas activation of another Gα_s-coupled receptor, the E-type prostaglandin receptor 4, results only in cytoplasmic cAMP production [138]. Thus, restricted production of cAMP within the caveolar compartment following β₂-adrenergic receptor activation may explain the localized signaling between β₂ receptors and LTCCs. This idea is consistent with studies in which disruption of caveolae with methyl-β-cyclodextrin led to more diffuse cAMP regulation after β₂-adrenergic activation that included PKA phosphorylation of phospholamban [139]. Spatially restricted cAMP generation associated with β₂ adrenergic activation may also involve coupling of β₂ receptors with Gα_i or localized expression of phosphodiesterase 4D isoforms [140,141].

4.2. AKAP5 essential for β-adrenergic stimulation of [Ca²⁺]i transient

Whereas selective activation of β₂-adrenergic receptors within caveolae does not have far-reaching regulatory effects on phospholamban or contractile proteins, β₁-adrenergic stimulation leads to global signaling events that increase contractility and speed relaxation. Recently, a macromolecular signaling complex involving the scaffolding protein AKAP5 and Cav-3 was demonstrated to be essential for β-adrenergic regulation of EC coupling proteins including LTCCs, ryanodine receptors and phospholamban [85]. Knockout of AKAP5 in mice abolished isoproterenol (a nonselective β-adrenergic receptor agonist) stimulation of [Ca²⁺]_i transients, as well as phosphorylation of ryanodine receptor and phospholamban. Intriguingly, enhancement of whole cell I_Ca,L by isoproterenol remained intact in AKAP5 knockout mice. Further work demonstrated that in wild type mice, Ca_v1.2 channels associated with Cav-3 were preferentially phosphorylated at Serine 1928 (S1928) after treatment with isoproterenol whereas the fraction of Ca_v1.2 not associating with Cav-3 did not undergo significant phosphorylation. In AKAP5 knockout mice, however, the converse was observed as isoproterenol predominantly led to phosphorylation of non-Cav-3-associated Ca_v1.2 channels. Thus, loss of AKAP5 leads to β-adrenergic stimulation of I_Ca,L acting on a different subpopulation of LTCCs that does not communicate directly with ryanodine receptors. However, the physiological relevance of S1928 phosphorylation is debated as recent studies have demonstrated that S1928 is not necessary for the PKA-mediated increase in I_Ca,L [131,142,143]. Nevertheless, the authors provided important evidence to propose a model in which AKAP5 serves as a scaffold to assemble proteins involved in EC coupling with a β-adrenergic signaling complex within T-tubules, and interestingly this macromolecular signaling complex included Cav-3 (Fig. 2).

Questions remain regarding the precise subcellar localization of this AKAP5/Ca_v1.2/Cav-3 complex. In ventricular myocytes, caveolae are found in both surface and T-tubular sarcolemma but are absent from dyadic junctions [53,144]. A number of studies have demonstrated Cav-3 expression within the T-tubule system in ventricular myocytes, where Cav-3 has some degree of colocalization with Ca_v1.2 and ryanodine receptor [54,145]. Considering caveolar structures are excluded from dyads, the observation that Cav-3 colocalizes with both Ca_v1.2 and ryanodine receptor within the cell interior may suggest the presence of Cav-3 scaffolds at or very close to dyadic junctions. A recent study using a triple immunolabeling strategy revealed that a small but significant percentage (~3-5%) of Cav-3 colocalizes with both Ca_v1.2 and ryanodine receptor in atrial myocytes, suggesting Cav-3 is expressed at or adjacent to select dyads; however, atrial myocytes lack the highly developed T-tubule system seen in adult ventricular myocytes, which may lead to differences in dyad architecture between the two cell types [146,147]. Another study using confocal microscopy has reported little to no Cav-3 localization at dyad junctions in rat ventricular myocytes [148]. Because perturbation of cholesterol rich microdomains reduces [Ca²⁺]_i transient amplitude and knockout of Cav-3 in mice leads to reduced myocardial contractility measured by fractional shortening, it seems possible that Cav-3 is expressed sufficiently close to dyadic junctions to play an important role in regulating EC coupling (Fig. 2) or alternatively influx of Ca²⁺ through distinct caveolar LTCCs contributes significantly to loading SR stores [42,99]. Additional strategies such as immunogold electron microscopy may yield insight into the localization of Cav-3 relative to dyadic Ca_v1.2 channels.

5. Targeting LTCCs to subcellular compartments

Considering spatially defined subpopulations of LTCCs regulate discrete cellular functions, understanding the mechanisms responsible for appropriate channel targeting to various subcellular compartments in cardiac myocytes is important, yet these pathways are largely undefined. Recent advances in our understanding of LTCC trafficking in cardiomyocytes will be briefly discussed.

5.1. Influence of Ca²⁺ channel auxiliary subunits on subcellular localization of LTCCs

The LTCC β subunits may play a significant role in defining subcellular L-type Ca²⁺ channel populations. Early evidence that β subunits can directly impact subcellular Ca²⁺ channel targeting came from studies in polarized epithelial cells, in which Ca_v2.1 subunits were trafficked to the basolateral membrane by β_1a and β₄ subunits but exhibited apical membrane localization when coexpressed with β_2a [149]. Disparities in the localization of various β subunits have been also observed in cardiomyocytes. Western blots of canine ventricular membrane fractions and immunoconfocal microscopy of isolated ventricular myocytes demonstrated that β_1b, β₂, and β₃ isoforms localize predominantly to T-tubule membranes while expression of β_1a and β₄ was detected more strongly at the surface sarcolemma [23]. Furthermore, adenovirus-mediated expression of various β subunits fused to GFP showed marked differences in subcellular localization among β subunit isoforms in cultured adult rat ventricular myocytes [59]. There is mounting evidence that caveolar localization of Ca_v1.2 channels requires an association between β₂ subunits and Cav-3, suggesting β subunits play an active role in targeting L-type Ca²⁺ channel complexes in cardiac myocytes [150].

A role for α₂δ subunits in subcellular targeting of LTCCs has also been proposed. The α₂δ subunits undergo posttranslational modification to add glycosylphosphatidyl inositol (GPI) anchors, which are known to localize proteins to highly ordered microdomains such as lipid rafts and caveolae [151-152]. Another possibility is that trafficking of cardiac LTCCs to caveolae may involve Cav-3. Cav-1 has been demonstrated to direct trafficking of the voltage-dependent K⁺ channel K_v1.5 to cholesterol-rich rafts [153]. The skeletal muscle Ca_v1.1 LTCC subunit has been reported to interact directly with Cav-3; however, this has not been demonstrated with Ca_v1.2 [154]. Future work will be necessary to carefully address whether other auxiliary subunits are important for targeting LTCC complexes to specific membrane microdomains such as caveolae.

5.2. BIN1 targets channels to T-tubules in cardiac myocytes

Localization of LTCCs to dyadic junctions within T-tubule structures is essential for EC coupling, and mislocalization of channels caused by remodeling in heart failure is thought to contribute to impaired contractility. T-tubule biogenesis involves the membrane scaffolding protein BIN1, and recent studies have implicated BIN1 in targeting LTCCs to T-tubules [155,156]. BIN1 localized to T-tubules in human and mouse ventricular myocytes in that BIN1 tethers microtubule structures to coordinate anterograde trafficking of Ca_v1.2 channels to T-tubule structures [156]. Importantly, knockdown of BIN1 reduced plasmalemmal expression of Ca_v1.2 channels and diminished [Ca²⁺]_i transients, suggesting fewer dyadic LTCCs [156].

5.3. Internalization and degradation of LTCCs

Although BIN1 plays an important role in anterograde trafficking of LTCCs to the plasma membrane, other mechanisms likely regulate channel retrieval and intracellular trafficking. There is some evidence for regulated internalization of LTCC complexes in cardiomyocytes; exposure of ventricular myocytes to isoproterenol for several minutes led to β-arrestin-1-mediated internalization of Ca_v1.2 into clathrin-coated vesicles [157]. However, it was unclear whether a specific subpopulation of LTCCs was targeted for internalization (i.e. bulk plasma membrane vs. caveolar); however, caveolar endocytosis is thought to be largely clathrin-independent, which suggests internalized channels are extracaveolar [158]. Furthermore, the destination of these internalized channels was not defined. Evidence exists for endocytic recycling of some cardiac ion channels to the surface membrane, and this process is normally regulated by the small GTPase Rab11 [159-161]. However, Rab11b has been demonstrated to target surface membrane LTCCs for degradation in cardiomyocytes [162]. Thus, recycling of cardiac LTCCs to the plasma membrane may occur via an alternative pathway, or internalized Ca_v1.2 channels may be preferentially degraded rather than recycled. In neurons, L-type Ca_v1.2 and Ca_v1.3 channels internalized in the presence of glutamate are targeted to the lysosome, which is thought to protect against excitotoxicity [163,164].

6. Altered microdomains disrupt LTCC function in cardiac disease

Dysregulation of LTCCs contributes to the pathophysiology of numerous heart diseases including heart failure, atrial fibrillation, and long and short QT syndromes [165-170]. Several reports have suggested that the geometry and protein composition of subcellular compartments associated with LTCC activity are altered in some cardiac diseases [171-174]. These changes could directly impact the function and regulation of LTCCs and contribute to defects in Ca²⁺ signaling, dysregulated EC coupling, and electrical instability [175].

6.1 Mutations in Ca_v1.2 and LTCC auxiliary subunits

Mutations in the genes encoding Ca_v1.2 and LTCC auxiliary subunits including β_2b and α₂δ-1 have been reported in patients with inherited arrhythmia disorders [165,169,176]. A number of mutations in Ca_v1.2, β₂, and α₂δ-1 lead to a loss-of-function phenotype characterized by dramatically reduced I_Ca,L. Although in most cases the reduction in I_Ca,L was attributed to changes in biophysical properties of the channel, one mutation in Ca_v1.2 (A39V) associated with short QT syndrome and sudden death disrupted LTCC trafficking [165,176]. It is unclear whether any of these reported loss-of-function mutations alter the specific subcellular localization of LTCCs. However given the potential role of auxiliary β and α₂δ subunits in the subcellular targeting of LTCCs (see Section 5.1), it is possible these phenotypes could result from the mislocalization of critical LTCC subpopulations.

Timothy syndrome (TS) is a multisystem disorder characterized by ventricular arrhythmias and autism resulting from gain-of-function mutations (G406R, G402S) in Ca_v1.2 that interfere with voltage-dependent channel inactivation [169]. Interestingly, TS mutant G406R Ca_v1.2 channels have recently been demonstrated to exhibit increased propensity for coupled gating and persistent Ca²⁺ sparklets [83]. TS Ca_v1.2 channels in transgenic ventricular myocytes exhibited prominent localization at the surface sarcolemma and intercalated discs relative to T-tubules, suggesting these mutant channels represent a distinct subcellular population [167]. TS Ca_v1.2 channels appear to be abnormally coupled to AKAP5, undergoing frequent and prolonged openings that are eliminated by knockout of AKAP5 [167]. AKAP5 ablation also protects TS transgenic mice from arrhythmias, indicating persistent Ca²⁺ sparklet activity plays a significant role in the TS phenotype. Precisely where TS Ca_v1.2 channels target remains unclear, but caveolae represent an intriguing candidate due to the known association between AKAP5 and Cav-3 [85].

6.2. CAV3 mutations associated with arrhythmia and cardiomyopathy

Congenital long QT syndrome (LQTS) is a potentially lethal disorder associated with delayed cardiac repolarization, prolonged QT interval, and ventricular arrhythmias and arises due to mutations in a number of ion channels and scaffolding proteins [178]. Recently, mutations in the CAV3 gene were identified in a subset of LQTS patients (designated LQT9). Initial studies showed that the CAV3 mutations resulted in increased late Na⁺ current through Na_v1.5 channels, which could prolong action potential duration [179]. Because LTCCs and important regulatory molecules such as β₂-adrenergic receptors associate with Cav-3, alterations in I_Ca,L may also contribute to the pathophysiology of LQT9. However, additional information regarding the electrophysiological effects of LQTS CAV3 mutations is not available. Other mutations in CAV3 have been linked to familial hypertrophic cardiomyopathy in one family and dilated cardiomyopathy with AV conduction defects in another [180,181]. The molecular mechanisms leading to these phenotypes are not well understood.

In addition to CAV3, mutations in the gene encoding the caveolar protein PTRF/Cavin have also been linked to LQTS, sinus bradycardia, and supraventricular and ventricular tachycardias [182]. Further work will be required to understand the detailed mechanisms underlying cardiac arrhythmias and cardiomyopathies induced by mutations in caveolar proteins.

6.2. Remodeling in failing heart

Heart failure is characterized by weakened myocardial contractile force, partially due to abnormal EC coupling resulting in reduced SR Ca²⁺ release [171,183,184]. A number of studies have demonstrated extensive remodeling of the T-tubular system in failing heart, which likely contributes to inefficient EC coupling [167,185-187]. T-tubule structural remodeling in heart failure is associated with a significant (~50%) reduction in the density of LTCCs, and Ca_v1.2 channels within T-tubules may be specifically decreased compared to channels in the surface membrane [167,188]. It has also been suggested that T-tubule remodeling may alter the geometry of the dyadic cleft microdomain [171,174,189]. Alternatively, changes in T-tubular architecture and composition could potentially disrupt macromolecular signaling complexes and lead to dysregulation of LTCCs. Blunted β-adrenergic regulation of I_Ca,L has been observed in animal heart failure models and in human heart failure [167,190]. Changes in the relative density, localization, and coupling of β-adrenergic receptor subtypes have been observed [190-193]. For example, in a canine tachycardia-induced heart failure model, enhanced β₂-adrenergic signaling through Gα_i strongly blunted the increase in I_Ca,L by β₁-adrenergic stimulation [190].

6.4. Atrial fibrillation

Chronic atrial fibrillation involves important structural and electrical changes including significant downregulation of I_Ca,L that results in shortened action potential duration and reduced atrial contractility [194,195]. Some studies have attributed the decrease in I_Ca,L to a reduction in transcription of Ca_v1.2 or various LTCC auxiliary subunits genes resulting in reduced mRNA and protein levels, whereas other reports have suggested impaired posttranslational modifications such as dephosphorylation of Ca_v1.2, altered channel trafficking, or enhanced degradation [196-200]. A decrease in T-tubule density was observed in atrial myocytes from a sheep model of atrial fibrillation, which may underlie dysynchronous CICR and reduced contractility [201]. However, it remains unclear whether specific populations of LTCCs are altered in atrial fibrillation.

7. Conclusions and future directions

In cardiac myocytes, LTCCs are localized to multiple distinct subcellular compartments that impact their function and regulation (Fig. 2). The significance of dyadic LTCCs in EC coupling has long been recognized, but other subpopulations, such as those localized to caveolae, are increasingly implicated in a variety of cellular functions and signaling pathways. Many cardiac diseases involve changes in subcellular architecture and organization, thus altered subcellular localization of LTCCs with associated changes in channel function properties can produce aberrant electrophysiology with resulting arrhythmias as well as dysregulation of various Ca²⁺-dependent cellular processes. Many gaps in our knowledge remain. For example, technical limitations still preclude detailed single channel studies of critical populations of LTCCs localized to T-tubules and thus out of reach from the patch pipette or TIRF microscope. Future studies will undoubtedly reveal additional details regarding how LTCCs and associated Ca²⁺-signaling molecules are assembled and targeted to defined subcellular compartments. Furthermore, a better understanding of the various subcellular populations of LTCCs may enable new therapeutic approaches for prevalent forms of heart disease such as heart failure and atrial fibrillation.

Highlights.

• >L-type Ca²⁺ channels regulate diverse cellular processes in the heart.
• >Different L-type Ca²⁺ channel subpopulations exist in cardiomyocytes.
• >Function and regulation of L-type Ca²⁺ channels depend on subcellular localization.
• >Altered localization of L-type Ca²⁺ channels plays a role in heart disease.

Abbreviations

LTCC

L-type Ca²⁺ channel

EC

excitation-contraction

I_Ca,L

L-type Ca²⁺ current

SR

sarcoplasmic reticulum

T-tubules

transverse tubules

CICR

Ca²⁺-induced Ca²⁺ release

[Ca²⁺]_i

free intracellular Ca²⁺ concentration

PKA

protein kinase A

Cav

caveolin

PTRF

Polymerase I and Transcript Release Factor

C-terminus

carboxyl terminus

CDI

Ca²⁺-dependent inactivation

CaM

calmodulin

CREB

cAMP-responsive element-binding protein

TIRF

total internal reflection fluorescence

AKAP

A-kinase anchoring protein

PKCα

protein kinase C alpha

CaMK

Ca²⁺-CaM-dependent kinase

PP2B

protein phosphatase 2B

NFAT

nuclear factor of activated T cells

SNARE

Soluble N-ethylmaleimide-sensitive factor attachment protein receptor

ANP

atrial natriuretic peptide

PP2A

protein phosphatase 2A

FRET

fluorescence resonance energy transfer

GPI

glycosylphosphatidyl inositol

TS

Timothy syndrome

LQTS

congenital long QT syndrome