The Directed Targeting Paradigm

Posttranscription and posttranslation, the proteins that form cardiac ion channels are modified and usually oligomerize in the Golgi apparatus, where they are inserted into membrane vesicles for delivery to the plasma membrane. These channel-laden vesicles are transported using the cytoskeleton as a path to the plasma membrane, with most studies focusing on the role of microtubule-based forward transport.16–21 A critical aspect of forward transport is localization to membrane subdomains. It is entirely possible that channels may be delivered to random regions of the plasma membrane, only to then laterally diffuse within the membrane to their appropriate subdomain.¹⁷ However, the temporal and stochastic inefficiency of random channel insertion, together with unexplained mechanisms of subsequent lateral localization other than chance interaction with a subdomain-specific anchor protein, suggests that specificity of delivery from the Golgi apparatus to the surface submembrane may also occur. Still to be fully explored, this directed targeting paradigm of ion channel delivery may be generalizable to all cardiac ion channels and explored in terms of other cytoskeletal elements and anchor proteins (discussed next).

This section of the chapter focuses on mechanisms that govern the forward trafficking of two major cardiac ion channels that are indispensable for cellular excitability, cell-cell coupling, and excitation-contraction coupling: (1) the L-type calcium channel Ca_v1.2 and (2) the dominant cardiac ventricular gap junction, connexin43 (Cx43).22–26 We also discuss in detail a protein with a newly described role in the heart that is essential for Ca_v1.2 trafficking and delivery, bridging integrator 1 (BIN1). BIN1 is rapidly emerging as a multifunctional cardiac player beyond its previously understood function as a membrane scaffold.

Forward Trafficking of Cx43

Gap junctions are intercellular channels that form low-resistance pathways, allowing ions and metabolites to flow from cell to cell. In the heart, gap junctions electrically couple cardiomyocytes to orchestrate spatial propagation of action potentials.²⁷ A gap junction consists of a pair of abutting hexameric hemichannels or connexons in adjacent cell membranes. Each connexon consists of six connexin proteins. More than 20 Cxs have been identified and each is usually named according to its respective molecular mass.²⁸ All connexins contain four transmembrane domains. Connexin43 is the most abundant connexin in the ventricular myocardium. In individual ventricular cardiomyocytes, Cx43 is localized at the intercalated disc at longitudinal ends of the cell, where Cx43 gap junctions provide rapid action potential transmission and synchronized cardiac excitation.^29,30

Cardiac-specific conditional deletion models with greater than 50% Cx43 loss develop increased arrhythmia susceptibility and sudden cardiac death.31–33 In general, these studies reveal that severe arrhythmia phenotypes can manifest when Cx43 coupling drops to less than 20% in the heart, which correlates with theoretic predictions.³⁴ Thus, maintenance of proper Cx43 expression and localization has a critical role in appropriate electrical coupling and ventricular function.

Data exist for multiple, but not incompatible, models of Cx43 trafficking to cell-cell borders of cardiac intercalated discs.17,18,20,25 There is almost universal agreement that microtubules help deliver Cx43 to the plasma membrane. A landmark series of two papers in 2002 found evidence that newly-formed Cx43 appear at the perimeter of the Cx43 plaques and then diffuse into central plaque regions.^17,20 Taken together with microtubule delivery, the model developed that Cx43 hemichannels are inserted into the general plasma membrane, rapidly diffuse to the edge of dense plaques, and then more slowly diffuse into the plaque center. Subsequent to these studies, it has been observed that Cx43 can be inserted directly at the edge and into plaques, and membrane fluidity exits within the plaque region.^18,35 It has also been found that the plaques are internalized not necessarily from the center, but that different segments of plaque can be internalized at any time³⁵ and full plaque internalization can occur in one step.³⁶ More recently, it has been found that Cx43 occurs in regions surrounding Cx43 plaques, the “perinexus,” at higher density than in general membrane where Cx43 proteins may interact with scaffolding proteins and other ion channels.³⁷ These studies are providing evidence that the gap junction plaque and surrounding regions are highly dynamic with complex behavior of targeted insertion and internalization.

Given the low density of Cx43 hemichannels in membrane well away from Cx43 plaque regions, and the technical difficulty of distinguishing Cx43 inserted in membrane from submembranous and still cytoplasmic collections of protein, it is difficult to find studies that can quantify the lateral diffusion coefficient of membrane-bound Cx43. It may be that free hemichannels rapidly diffuse within the plasma membrane before stopping at plaque regions, but direct evidence for this phenomenon is lacking. The directed targeting paradigm is based on the observation that de novo intracellular Cx43 hemichannels can arrive directly in the gap junction plaque region. The hemichannels are targeted to plaque regions with specificity obtained from the channel protein (Cx43), microtubule plus-end tracking proteins (EB1 and p150 [glued], and a membrane anchor [adherens junction structure]).¹⁸ By this model, the dynamic microtubule highways are anchored and terminate at adherens junction structures, allowing directed delivery of Cx43 hemichannels to adherens junction–containing membrane to occur. It is probable that once inserted into plaque regions, local hemichannel diffusion occurs within the plaque and between the plaque and the plaque perinexus.

Although most studies of Cx43 forward trafficking focus on the microtubule cytoskeleton, the actin cytoskeleton is also involved in delivery of membrane proteins and ion channels. Dye transfer studies revealed the dependence of Cx43 insertion into plaque on actin, as tested by pharmacologic actin disruption and anti-actin antibodies.38,39 Actin is also important for Cx43 plaque internalization, although we recently found that when internalization is blocked, forward delivery is still actin dependent.²⁶ Furthermore, a remarkable greater than 80% of cytoplasmic post-Golgi Cx43 is slow-moving or stationary and actin-associated. These findings indicate that microtubules work in concert with actin to deliver Cx43 to the plasma membrane. It may be that post-Golgi Cx43 vesicles exist in actin-associated reservoirs within the cytoplasm, waiting for either the right microtubule to permit membrane delivery or mass acute delivery in the case of a metabolic stress.

Forward Trafficking of Cav1.2

Although connexins electrically couple cardiomyocytes, the membrane ion channel most responsible for calcium entry and excitation-contraction coupling is the α1C pore-forming subunit (Ca_v1.2) of the L-type voltage-gated (Ca_v) Ca²⁺ channel (LTCC). In response to membrane depolarization by sodium currents, LTCCs open to allow inward Ca²⁺ entry during the plateau phase of the cardiac action potential. The L-type currents (I_Ca,L) that are generated initiate calcium-induced Ca²⁺ release from the sarcoplasmic reticulum (SR) via ryanodine receptors, which leads to muscle contraction. The Ca_v1.2 protein is large and consists of 24 transmembrane segments organized into four homologous domains.⁴⁰ Auxiliary subunits of the LTCC include the β-subunit, the α₂δ-subunit, and the γ subunit, which help regulate trafficking of Ca_v1.2 to the cell membrane and regulate the voltage dependence of channel gating.⁴¹

T-tubule invaginations of ventricular cardiomyocyte plasma membrane are enriched with LTCCs. This enrichment is necessary for calcium-induced calcium release with nearby ryanodine receptors, which is important for beat-to-beat excitation-contraction coupling in the heart. Trafficking of the LTCCs is not as extensively studied as that of Cx43, probably owing to a combination of the large size of the α-subunit, Cav1.2, the multiple β-subunits, the difficulty of manipulating cardiomyocytes with T-tubules in culture, and less dense enrichment of the channel on the plasma membrane, presenting difficulty to clearly characterize by cytochemistry a distinct phenotype. Subunits are important for LTCC trafficking. For instance, the β-subunit is necessary for surface expression.⁴² Moreover, the α2δ-subunit synergizes with the β-subunit to promote surface channel expression.⁴³ These auxiliary subunits enhance surface expression by promoting channel export from the endoplasmic reticulum and overall channel stability.⁴⁴

Despite the necessary localization of LTCC to T-tubules, it was not understood until recently how the localization occurs. In 2010, it was found that Cav1.2-based channels adhere to the directed targeted paradigm.²³ In particular, the membrane-scaffolding protein BIN1 was found to provide a membrane anchor that attaches dynamic microtubules, allowing delivery of Cav1.2 channels directly to T-tubule membranes. Using HL-1 cells that express Cav1.2 but do not form T-tubules, and non-muscle cell lines that neither express Cav1.2 nor naturally form T-tubules, it was found that exogenous BIN1 caused the formation of deep invaginations in the cell membrane enriched with endogenous or overexpressed Cav1.2, suggesting that BIN1-containing membrane is sufficient to recruit Cav1.2 channels. To test the possibility that BIN1 serves as an anchoring site for microtubules on which Cav1.2 channels are trafficked, the study tracked growing microtubules extending toward BIN1 clusters and found that the microtubule plus end-tracking proteins pause and associate with BIN1 clusters at the cell periphery. Moreover, it was determined that the non-BAR cytoskeleton–anchoring domain of BIN1 is required for this activity because truncation mutants lacking this domain failed to cluster Cav1.2 at cell surface invaginations. The microtubule plus end-tracking protein that may aid in microtubule anchoring to BIN1, analogous to EB1 binding to adherens junctions,^18,25 has not yet been identified.

Subcellular Localization of Kv Channels Within Myocytes

Most tissues, and even single cells, express multiple Kv channel types belonging to one or more subfamilies. Importantly, the subcellular localization of these different channel isoforms is necessary for proper function and signaling. For example, the regional and cell-specific distribution of Kv channels contributes to local variations in the shape and duration of the cardiac action potential. At the subcellular level, Kv channels can be expressed within T-tubules, at the intercalated disc, or on the lateral membrane, in an isoform specific pattern. Although it is known that Kv channel expression differs in different regions of the heart—for example the atria versus ventricles or endocardium versus epicardium—it is unclear whether the subcellular localization varies within the heart itself. Recently it has been shown that ectopically expressed Kv2.1 shows a different subcellular localization between atria and ventricles. It is not known whether this chamber-specific localization applies to endogenous channel localization. It is also not known how different Kv channel isoforms affect localization.

Historically, Kv localization is believed to primarily involve protein-protein interactions among channel proteins and PDZ-domain–containing scaffolding proteins or the actin cytoskeleton. Recent work shows that Kv2.1 channels are immobilized on the lateral membrane of atrial myocytes, where their diffusion is likely limited by actin cytoskeletal corrals as in hippocampal neurons.⁴⁵ In addition, Kv1.5 has been shown to directly couple to α-actinin-2 in HEK 293 cells via a specific sequence in the amino terminus of the channel.⁴⁶ In addition, Kv1.5-cytoskeleton interactions appear to play a role in modulation of the channel by protein kinase A in oocytes.⁴⁷ Further, the actin binding protein, cortactin, regulates Kv1.5./N-cadherin interactions at the intercalated disc. Kvβ-subunits are also likely to interact with the cytoskeleton and thus immobilize channels on the cell surface.⁴⁸ In addition, some Kvβ-subunits behave as chaperone proteins, promoting the cell surface expression of a subset of co-expressed α-subunits and perhaps influencing localization.^49,50 A role for PDZ domain–containing scaffolding proteins, such as SAP97, in the regulation of Kv1.5 surface localization has also been described.^51,52