Keywords

1 Introduction

In excitable cells voltage-gated calcium channels (CaVs; also termed voltage-dependent or voltage-operated calcium channels) mediate and regulate a variety of functions ranging from muscle contraction, secretion, synaptic function to gene regulation. Calcium entering through voltage-gated calcium channels operates as a local second messenger by activating downstream signalling proteins localized in the close vicinity of the channel pore. In neurons CaVs contribute to the specific action potential firing pattern, presynaptic CaVs regulate neurotransmitter release (Stanley 1993) and postsynaptic CaVs are involved in the transcriptional regulation of CREB (cAMP-responsive element-binding protein) and NFAT (nuclear factor of activated T cells) (Deisseroth et al. 2003; Dolmetsch 2003) and thus likely play a crucial part in the formation of new memory (Moosmang et al. 2005). Over the recent years a detailed picture on the distribution and function of pre- and postsynaptic calcium channel types has begun to emerge. Thus, in the central nervous system CaVs of the CaV2 family, namely P/Q-type (CaV2.1), N-type (CaV2.2), and R-type (CaV2.3) channels, are the major presynaptic pore forming subunits triggering synaptic release. The L-type channels CaV1.2 and CaV1.3 are mainly involved in postsynaptic functions including plasticity and gene transcription. The importance of the pre- and postsynaptic CaV pore-forming subunits is emphasized by the existence of channelopathies caused by loss-of-function as well as gain-of-function mutations (Pietrobon 2010; Striessnig et al. 2010). For example, dysregulation of presynaptic P/Q-type and postsynaptic L-type channels is involved in the etiology of migraine (Pietrobon 2010) and autism disorders (Splawski et al. 2004), respectively. In contrast, there is little to no evidence for a function of the skeletal muscle CaV1.1 isoform in the nervous system (Sinnegger-Brauns et al. 2009). The CaV1.4 isoform appears to be specifically expressed in the retina and its mutation causes congenital stationary night blindness type 2 (Wycisk et al. 2006a, b). Low-voltage activated calcium channels (T-type channels) CaV3.1, 3.2 and 3.3 are critical regulators of neuronal excitability. They are prominently expressed both in the central and peripheral nervous system and are involved in neurological disorders such as absence epilepsy and neuropathic pain (Iftinca 2011).

CaVs operate in heteromultimeric complexes with the auxiliary β (also termed CaVβ) and α2δ subunits, calmodulin and other calcium binding and regulating proteins. The pore-forming α1 subunit of voltage-gated calcium channels defines the basic biophysical, pharmacological and physiological properties of the channels. A plethora of studies within the last 20 years have extensively demonstrated their roles in the localization, trafficking and stabilization of the channel complex (reviewed in Arikkath and Campbell 2003; Obermair et al. 2008; Dolphin 2009; Buraei and Yang 2010). The great majority of these studies was performed with different channel subunit combinations heterologously expressed in Xenopus laevis oocytes or mammalian expression systems such as human embryonic kidney (HEK) cells. Therefore the informative value of these studies regarding the role of the auxiliary calcium channel subunits in native cell systems like neurons remained limited. Whereas studies in heterologous expression systems are ideally suited to investigate effects and mechanisms for the interaction of specific coexpressed subunit partners in isolation, such studies do not predict as to whether the same protein-protein interactions indeed occur in signaling complexes of differentiated cells. Neither can it be assumed that in the complex with additional up- and downstream interacting proteins in differentiated cells the properties and effects of such interactions are the same as in heterologous expression systems. The development of powerful neuronal expression systems and the analysis of calcium channel knock-out animal models (see box) in recent years have helped to reveal the physiological importance of auxiliary β and α2δ subunits in neuronal/synaptic function. With respect to the role of auxiliary calcium channel subunits in synaptic function the principal questions that now can be addressed include:

  • What is the complement of specific calcium channel isoforms expressed in synaptic compartments?

  • Do different subunit isoforms serve distinct functions and to what degree can they be compensated by other isoforms?

  • Do the auxiliary subunits exclusively function in the context of the calcium channel (i.e., regulate its expression and targeting, or modulate its gating properties) or do auxiliary calcium channel subunits also serve functions independent of the channel?

2 Structure and Function of Auxiliary Calcium Channel Subunits

2.1 The α2δ Subunit

A total of four genes (Cacna2d1-4) encode for α2δ subunits (α2δ-1 to α2δ-4), which display distinct tissue distribution and out of which three isoforms (α2δ-1 to -3) are strongly expressed in the central nervous system (CNS) (Arikkath and Campbell 2003; Schlick et al. 2010). α2δ-1 and α2δ-2 subunits are the primary targets for the anti-epileptic and anti-allodynic drugs gabapentin (GBP) and pregabalin (PG), which have also proven clinical efficacy in the treatment of generalized anxiety disorders (Bryans and Wustrow 1999; Rickels et al. 2005). Mature α2δ subunits consist of posttranslationally cleaved α2 and δ peptides, which are associated to each other by a disulfide bond (Calderon-Rivera et al. 2012). Until recently it had been suggested that the δ subunit constitutes a single-pass membrane protein, and the α2 subunit a highly glycosylated extracellular protein. However, this classical view has recently been challenged by the observation that α2δ subunits can form GPI-anchored proteins and that this posttranslational modification may be crucial for α2δ function (Davies et al. 2010). In either way the vast majority of the α2δ protein is extracellular, ideally situated to interact with constituents of the extracellular matrix or extracellularly exposed proteins. Consistent with a role in extracellular signaling is the domain structure of α2. A von Willebrand factor type A (VWA) domain and two Cache domains were identified by sequence homology in all α2δ subunits (Anantharaman and Aravind 2000; Canti et al. 2005; Davies et al. 2007). VWA-domains are found in a variety of extracellular matrix proteins and integrin receptors and are well known for their role in cell-cell adhesion (Whittaker and Hynes 2002) involving a metal ion-dependent adhesion site (MIDAS). The integrity of the MIDAS motif in α2δ-2 has been shown to be necessary for calcium current enhancement and CaV channel trafficking (Canti et al. 2005). Cache domains were named after their presence in calcium channels and chemotaxis receptors and have been suggested to be involved in small molecule interactions (Anantharaman and Aravind 2000). Thus, it has been hypothesized that these domains may be regulated by small endogenous ligands, such as the amino acid isoleucine (reviewed in Dooley et al. 2007), and that they are involved in GBP and PG binding (Davies et al. 2007). α2δ subunits also contain a conserved N-terminal α-helical domain found in several methyl-accepting chemotactic receptors and mutations within this domain have been shown to interfere with GBP and PG binding (Anantharaman and Aravind 2000).

2.2 The β Subunit

The entirely cytoplasmic β subunit consists of a conserved SH3 protein interaction domain and a nucleotide kinase-like domain (Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004) and thus resembles in structure the membrane-associated guanylate kinase proteins (Dolphin 2003; Takahashi et al. 2005). However, the SH3 domain of β subunits differs from that of canonical polyprolin-binding pockets and the guanylate kinase fold is modified so that it lacks kinase activity. Instead it binds the intracellular I-II linker of α1 subunits at the so-called α interaction domain (AID) with nanomolar affinity (De Waard et al. 1995; Van Petegem et al. 2008). The SH3 and the GK-like domains are highly conserved among the four genes encoding β subunits (Cacnb1-b4). The sequences connecting these domains and the N- and C-termini vary between isoforms and are subject to alternative splicing (Colecraft et al. 2002; Dolphin 2003). In the channel complex β subunits serve two roles: They have a chaperon function regulating the export of the calcium channel from the endoplasmic reticulum and thus membrane expression of functional channels (Fang and Colecraft 2011). Moreover, they modulate gating properties of the channel directly as well as by interaction with other regulatory proteins like Rab binding proteins or G-proteins. β itself is subject to PKA mediated phosphorylation (reviewed in Buraei and Yang 2010). The β2a isoform is palmitoylated at two N-terminal cysteines and therefore membrane-associated even in the absence of an α1 subunit. Nevertheless, the association of β subunits with the channel complex entirely depends on their binding to the AID in the α1 subunit. This binding site in the cytoplasmic loop between repeats I and II of the α1 subunit is a unique feature of the CaV1 and CaV2 subclasses of CaVs. Accordingly, at least in heterologous expression systems all β subunits can associate with any of the CaV1 or CaV2 members. However, the low-voltage activated calcium channels of the CaV3 subclass do not associate with β subunits (Dolphin 2003). Because of their central role in regulating functional expression and biophysical properties of calcium channels, and because of the well defined interaction site, interfering with the AID-β interaction is an attractive strategy for designing specific calcium channel antagonists. So far, such endeavors have not been successful. However, the high efficacy of members of the small G-protein Rem/Gem/Kir family in blocking calcium currents by interacting with the β subunit holds great promise for these calcium channel subunits as drug targets (Yang et al. 2007).

3 Neuronal Voltage-Gated Calcium Channel Complexes

In differentiated cells calcium channels do not function in isolation, rather they exert their functions in the context of multimolecular signalling complexes. The short range of the second messenger calcium necessitates that downstream signalling proteins and effectors are anchored in the close vicinity of the channel pore. Similarly, increasing evidence shows that upstream modulators, like protein kinases and phosphatases achieve substrate specificity and increased signalling efficiency if they pre-exist in a complex with the channel. Accordingly, a voltage-gated calcium channel complex is composed of the calcium channel subunits proper, upstream modulators and downstream effectors, and the adapter and scaffold proteins, assembling the complex.

In neurons two such complexes have been subject to extensive investigations: First, the synaptic vesicle fusion apparatus and second, the postsynaptic calcium channel complex mediating excitation-transcription coupling.

The presynaptic calcium channel complex: Calcium influx through CaVs transduces membrane depolarization into the chemical signal triggering fusion of neurotransmitter vesicles. Here CaVs of the CaV2 subclass (P/Q- and N-type) associate with the SNARE proteins of the synaptic core complex either directly by an interaction of the SYNPRINT domain within the II-III loop of the α1 subunits with syntaxin, SNAP-25 and synaptotagmin-1 (reviewed in Sheng et al. 1998; Zamponi 2003; Catterall 2011), or via the β subunit and the Rab interacting protein (RIM) (Kiyonaka et al. 2007). These interactions are believed to anchor the CaV channel close to the calcium sensor synaptotagmin and conversely to functionally modulate the calcium current, both enhancing the channel’s efficacy to activate vesicle fusion. Indeed a low number of channels and in the extreme even a single channel opening is sufficient for triggering vesicle fusion (Stanley 1993; Bucurenciu et al. 2010). Neurotransmitter release is commonly modulated by neuropeptides and hormones. Therefore G-protein coupled receptors (GPCRs), G-proteins, phospholipases, adenylate cyclases, and protein kinases may coexist with presynaptic calcium channel complexes.

The postsynaptic calcium channel complex: The postsynaptic CaV complex mediates excitation-transcription coupling. Here activation of L-type channels (CaV1) initiates a signalling cascade to the nucleus that regulates gene expression. To this end scaffold proteins like AKAP79/150 recruit protein kinases and the calcium/calmodulin dependent protein phosphatase calcineurin to the channel. Upon activation by the local calcium signal this signalling cascade leads to the translocation of NFATc4 into the nucleus and (Oliveria et al. 2007; Ma et al. 2011). Since overexpression of AKAP79/150 also enhances the L-type calcium currents (Altier et al. 2002) it seems that at least some of these signalling proteins are shared with upstream signalling cascades mediating GPCR-induced phosphorylation of the channel. Indeed β adrenergic receptors, AKAP, PKA and calcineurin were all detected in the CaV1.2 signaling complex in neurons (Davare et al. 2001; Dai et al. 2009).

The subunit composition of the pre- and postsynaptic CaVs is expected to influence the function of these signalling complexes in several ways. Modulation of current properties by auxiliary subunits will affect the signalling power of the complex. A participation of the auxiliary calcium channel subunits in scaffolding will affect the composition of the signalling complex and thus the signalling specificity. By targeting the channel into the close proximity of effector proteins the efficacy of the signalling process will be enhanced. As different subunits differ with respect to their modulatory properties, protein-protein interactions and subcellular targeting, as well as the molecular diversity of the auxiliary subunits may be important for the proper assembly and function of the different calcium channel complexes in neurons. In other words, the distinct molecular organizations and functions of different calcium channel signalling complexes may require a specific subunit composition of the channel. In turn, the distinct molecular compositions of pre- and postsynaptic signalling complexes may favour the incorporation of channels with specific subunit compositions.

3.1 Potential Mechanisms for Establishing Specific Neuronal CaV Complexes

Understanding the specificity of CaV subunit interactions in native differentiated cell systems is key for resolving the physiological neuronal functions of calcium channels and their auxiliary subunits. Three distinct mechanisms may explain the assembly of specific α1/β/α2δ subunit complexes in neurons (see Fig. 2.1):

Fig. 2.1
figure 1

Model explaining loss-of-function scenarios of distinct mechanisms for the assembly of specific α1/β/α2δ subunit complexes in neurons. (a) Different cell types may express exclusive α1/β/α2δ complexes (either orange or blue). Knockout of one subunit (orange α 2 δ) will lead to a loss of function (e.g. synaptic function) in one cell type, whereas the functions of other cell types are not affected. (b) Specific α1/β/α2δ complexes may form by differential targeting into distinct subcellular compartments (soma-blue, synapse-orange). The consequence of knockdown of one subunit (orange α 2 δ) depends on the uniqueness of the targeting properties of the remaining isoforms. Thus, if axonal targeting of the blue isoform is not possible, knockout will ultimately lead to a loss of function (upper). Otherwise the blue isoform may compensate for the loss of the orange isoform and restore normal synaptic function (lower). (c) Specificity of α1/β/α2δ complexes may be determined by distinct affinities or distinct interaction partners in a macromolecular complex. Putative compensatory mechanisms will then depend on the stringency of the individual complexes. If subunit exchange between the individual complexes is excluded, functional compensation is limited (upper). If subunits can associate/dissociate with and from individual complexes, at least partial compensation will occur (lower)

Full size image
  1. 1.

    Different affinities of auxiliary β and α2δ subunits for specific α1 isoforms may favour the preferential association of specific subunit combinations.

  2. 2.

    Limiting the number of isoforms expressed in a specific cell type at a given time will also favour the formation of a specific CaV complex.

  3. 3.

    Distinct subcellular targeting properties of individual subunits as well as protein-protein interactions with other proteins may yield complex specificity.

Alternatively, specific stable complexes may not exist in all neuronal compartments and CaV channels could be regulated by reversible interactions with pools of functionally distinct cytoplasmic β subunits or membrane anchored α2δ subunits.

3.1.1 The Importance of Affinity for Specific CaV Complex Formation

Heterologous coexpression studies have demonstrated that all four β subunits as well as all four α2δ subunits can enhance the trafficking and modulate the current properties of all high-voltage activated calcium channel α1 subunits, indicating a great promiscuity of subunit interactions (reviewed in Arikkath and Campbell 2003; Dolphin 2003; Obermair et al. 2008; Buraei and Yang 2010). In line with this observation low neuronal α1-β selectivity was suggested by immunoprecipitation experiments showing similar β subunit compositions of neuronal L-type, P/Q-type and N-type channels (Liu et al. 1996; Scott et al. 1996; Pichler et al. 1997). Furthermore, biochemical analysis revealed similarly high affinities of different β subunits to the AID (De Waard et al. 1995; Van Petegem et al. 2008). Therefore, if differences in the strength of interactions contribute to the formation of preferential subunit compositions, these may be determined by low affinity secondary interaction sites rather than by the AID, and/or by indirect interactions involving additional components of the signalling complex. In fact, recent experiments in skeletal muscle indicate that, although non-muscle β subunits can successfully compete with the skeletal muscle β1a for association with the channel, complexes including the heterologous isoforms are less stable than those consisting of all skeletal muscle isoforms (Campiglio et al. 2013). These experiments for the first time demonstrate the formation of complexes with preferential subunit composition in a native calcium channel complex, and suggest that isoform specific differences in the association with the α1 subunit underlie the distinct complex stabilities. For α2δ subunits no high-affinity binding sites have been identified in the α1 subunit. Interestingly, although in the initial biochemical purification of neuronal CaV channels α2δ reliably co-purified with the α1 and β subunits (McEnery et al. 1991; Witcher et al. 1994; Martin-Moutot et al. 1995; Liu et al. 1996), a recent quantitative proteomics approach of mammalian CaV2 channel complexes in brain extracts did not identify α2δ subunits as a core components of the complex (Müller et al. 2010).

Also upon coexpression in nerve or muscle cells, α2δ subunits appear more widely expressed throughout the plasma membrane than the other channel subunits (Schredelseker et al. 2005) (Schöpf, Obermair et al., unpublished observation). If distinct affinities of auxiliary subunit isoforms to preferential α1 subunit partners contribute to the formation of specific complexes, this may involve low-affinity binding sites, which so far eluded biochemical detection. Such low affinity interaction would allow dynamic exchange of subunits in response to changes in expression levels or local concentration of the auxiliary subunits.

3.1.2 Spatial and Temporal Separation of CaV Subunit Expression

Distinct cellular expression levels of calcium channel isoforms indeed provide an important determinant of complex specificity. They have been identified in non-neuronal cells such as skeletal muscle (CaV1.1/β1a/α2δ-1) and cardiac myocytes (CaV1.2/β2/α2δ-1) (Arikkath and Campbell 2003) and specialized neuronal cell types like retina photoreceptor cells (CaV1.4/β2/α2δ-4) (Ball et al. 2002; Barnes and Kelly 2002; Wycisk et al. 2006a; Neef et al. 2009). Similarly, the cerebellum shows a strong preference towards expression of one subunit combination (CaV2.1/β4/α2δ-2) (Ludwig et al. 1997; Brodbeck et al. 2002). We have recently shown that murine cortex, hippocampus and cerebellum simultaneously express mRNA of five out of seven high-voltage activated α1 subunits, all four β subunit isoforms, and three of four α2δ subunits at physiologically relevant levels (Schlick et al. 2010). Surprising was our finding that also a single cell type, cultured hippocampal pyramidal cells expresses all the same CaV subunit isoforms as hippocampus. This clearly suggests that in cultured hippocampal pyramidal cells, a restricted expression of auxiliary subunit isoforms is not the strategy to achieve specific CaV subunit compositions. Consequently other mechanisms, like specific targeting properties and interactions with anchoring proteins in pre- and postsynaptic compartments, must be responsible for assembling channels with distinct subunit compositions.

In neurons expression patterns of channels and signaling proteins are not static. Many receptors, channels and transport proteins serve different functions or have different properties during distinct phases in the life cycle of a neuron. Consequently their expression patterns change during development (Schlick et al. 2010) and in some cases they undergo an isoform switch during differentiation (Vance et al. 1998). During development CaVs serve in the regulation of neuronal mobility, path-finding and synapse formation (Pravettoni et al. 2000; Zheng and Poo 2007). These functions likely require different subunit isoforms or splice variants than in differentiated neurons. Moreover, neurons possess the unique ability to alter expression and composition of synaptic proteins in an activity-dependent manner. It can be expected that any of these changes also go hand in hand with altered subunit compositions or splice variant expression of CaVs. At present hardly any information on the differential expression and subunit composition of CaVs during neurogenesis and synaptic plasticity is available. Once the tools for analyzing expression patterns in specific populations of neurons are in place, the study of changing channel subunit combinations during differentiation and synaptic plasticity will be an important and fruitful undertaking.

3.1.3 Differential Targeting and Localization of Auxiliary CaV Subunits

Neurons are compartmentalized and structurally and functionally polarized more than any other cell type. Accordingly the composition of membrane proteins differs greatly between the input side, the somato-dendritic compartment, and the output side, the axonal compartment. Moreover, pre- and postsynaptic membranes differ from extrasynaptic membrane domains in composition of the lipid and protein content. Such compartmentalization requires complex targeting mechanism for most of the membrane proteins. However, CaVs serve important functions in both the pre- and postsynaptic compartment. Whereas the functional expression and targeting of calcium channels is unique in each neuronal cell type, an overall preference exists of CaV2 and CaV1 channels for the pre- and postsynaptic compartment, respectively. The specific interaction of CaV2 channels with the presynaptic fusion apparatus has been shown to contribute to this differential targeting (Mochida et al. 2003; Szabo et al. 2006; Simms and Zamponi 2012). It can be expected that the auxiliary calcium channel subunits also display differential targeting properties in neurons. If they formed preferential complexes with specific α1 subunits, their targeting mechanisms would drag the auxiliary subunits along. α2δ and β subunits would not require independent targeting mechanisms. However, if the auxiliary subunits possessed targeting mechanisms independent of α1 subunits, they could contribute to the differential distribution of the channels in the pre- and postsynaptic compartments.

Previously we have employed the expression of epitope-tagged β subunits and subsequent immunofluorescence labeling with a single antibody as a powerful approach to analyze their targeting behavior (Obermair et al. 2010). On the one hand all four examined β subunit isoforms and two β1 and β2 splice variants were found in the somato-dendritic as well as in the axonal compartment. Importantly, the β1 splice variants were less efficiently targeted to the distal axon, indicating a preferential role of in the postsynaptic/somatodendritic compartment (Fig. 2.2). However, even though the β1 isoform was poorly targeted to the distal axon, it could be incorporated into the nerve terminal like β2, β3, and β4. This observation is consistent with the great promiscuity of α1/β interactions observed upon heterologous coexpression and indicates that in neurons the affinities of specific β-AID pairs to each other (De Waard et al. 1995) by themselves do not determine the specificity of α1/β assemblies.

Fig. 2.2
figure 2

Isoform-specific localization of V5-tagged β subunits in cultured hippocampal neurons. Center: A 17 DIV cultured hippocampal neuron cotransfected with a V5-tagged β1b and an eGFP-tagged β4b subunit. The distribution of β1b-V5 is confined to the somatodendritic compartment (yellow neurites) whereas β4b-eGFP expression is high throughout the axon and the axonal branches (green neurites). Left: V5-tagged β1b co-localizes with membrane expressed HA-tagged CaV1.2 channel clusters along the dendrites and in dendritic spines (arrowheads). Note that some β1b clusters (arrow) do not colocalized with CaV1.2, indicative of an association with a different CaV α1 subunit. Right: Example of the presynaptic localization of a β subunit in a triple-labelling experiment. Fluorescence of soluble eGFP allows to morphologically identify axons with their short axonal branches with presynaptic terminals identified by staining for the vesicular glutamate transporter (vGlut1). The V5-tagged β4b isoform specifically accumulates in presynaptic terminals (examples indicated by arrowheads). For more details see Obermair et al. (2010)

Full size image

The great promiscuity in the interaction of β subunits with distinct α1 subunits allows for the differential modulation of CaVs by the association and dissociation with different β subunits (Obermair et al. 2008). Such a dynamic exchange of neuronal β isoforms with CaV1 channels has recently been demonstrated in differentiated skeletal myotubes (Campiglio et al. 2013). Thus, in neurons which express all β isoforms, shifts in the relative expression or local concentration of functionally distinct channel subunits may change the equilibrium between the subunit partners and thus the subunit composition of the channel complexes.

Neuronal α2δ subunits display regional differences in their expression levels in brain (Cole et al. 2005; Taylor and Garrido 2008). Nevertheless, similar to β subunits three out of four α2δ subunits are simultaneously expressed in neurons of the CNS (Schlick et al. 2010; Nimmervoll et al. submitted). α2δ subunits are abundantly expressed in the neuronal plasma membrane (Bauer et al. 2009; Müller et al. 2010) and one common feature is their localization in presynaptic terminals. Accordingly α2δ-1 to -3 isoforms localize to synapses upon overexpression in hippocampal neurons and can interact with presynaptic calcium channels (Hoppa et al. 2012; Nimmervoll et al. submitted). In the hippocampus immunostaining with a monoclonal antibody revealed a preferential localization of α2δ-1 in mossy fibre terminals in the CA3 region of the hippocampus (Taylor and Garrido 2008). Recently characterized α2δ-3 mutants in C. elegans (unc-36) and Drosophila (straightjacket) suggest a primary role in presynaptic calcium channel trafficking (Dickman et al. 2008; Saheki and Bargmann 2009). In addition α2δ subunits are synaptically expressed in specialized nerve cells including retinal photoreceptor cells (Mercer et al. 2011), dorsal root ganglion neurons (Bauer et al. 2009) and in synapses along the hearing pathway (Pirone et al. 2009). Apart from presynaptic localizations in the cerebellum α2δ-2 is concentrated in lipid rafts, suggestive of a restricted expression in microdomains, which may be important for its interaction with CaV2.1 channels (Davies et al. 2006). Nevertheless, it is currently not known whether α2δ subunits display an isoform specific targeting pattern and to which extent their localization depends on the interaction with α1 subunits and vice versa.

4 Neuronal Functions Related to Auxiliary Calcium Channel Subunits

4.1 Channel Trafficking and Current Modulation by β Subunits

The auxiliary α2δ and β subunits are important factors for trafficking CaVs to the plasma membrane and possibly for stabilizing them in functional signaling complexes. Because an increased density of functional channels in the synapse is expected to raise the efficacy of synaptic release, regulating membrane expression may act as an efficient mechanism to modulate synaptic function. Especially the β subunit has long been known to strongly increase calcium current density upon coexpression in HEK cells or X. laevis oocytes (reviewed in Dolphin 2003; Buraei and Yang 2010). Recent studies provided insight into the role of β subunits in membrane targeting of calcium channels in native cells and tissues (Tab. 2.1; reviewed in Buraei and Yang 2010).

For example, antisense knockdown of β subunits in cultured rat dorsal root ganglion neurons strongly decreased barium currents through endogenous calcium channels (Berrow et al. 1995). We could recently show that mutation of the AID in CaV1.2 channels completely prevented the surface expression of CaV1.2 in cultured hippocampal neurons (Obermair et al. 2010). Furthermore overexpression of β subunits substantially increased surface expression of CaV1.2 channels, indicating that the abundance of β subunits may present a limiting factor for the membrane expression of CaVs, and that the number of channels in the neuronal membrane can indeed be regulated by the amount of available β subunits.

The mechanism by which β subunits enhance membrane expression has long been a matter of discussion. Previously it has been suggested that β enables ER export of α1 subunits by masking an ER retention signal within the I-II intracellular loop of CaV1 α1 subunits (Bichet et al. 2000). Using an elegant combination of electrophysiology and quantification of channel surface expression in HEK cells, Fang and Colecraft have systematically characterized the contribution of all intracellular domains of CaV1.2 for the β-mediated surface expression (Fang and Colecraft 2011). These experiments clearly demonstrated that the I-II linker contains a putative ER export motif and that the β-dependent increase in surface expression may require a C-terminus-dependent rearrangement of intracellular domains, thereby overcoming retention signals within the other cytoplasmic loops. Furthermore, two recent studies demonstrated that association with a β subunit prevents the proteasomal degradation of the respective α1 subunits, thereby stabilizing and increasing the surface expression (Altier et al. 2011; Waithe et al. 2011).

In addition to effects of β subunits on membrane targeting, β subunits are powerful modulators of the channel’s gating properties. Upon coexpression in heterologous cells β subunits enhance the voltage-dependent activation and inactivation. The most notable isoform-specific effect is the strong inhibition of voltage-dependent inactivation by the palmitoylated β2a (Qin et al. 1998).

There are multiple lines of evidence demonstrating that β subunits modulate calcium channel functions in neurons. For example overexpression of β2a and β4b in hippocampal neurons induce depression and paired-pulse-facilitation of autaptic synapses, most likely by a differential modulation of the current properties of presynaptic CaVs (Xie et al. 2007). Moreover, CaV2 channels are subject to presynaptic inhibition by hormones and neurotransmitters through G-protein coupled receptors linked to Gi/o via Gβγ. This inhibition, which may be involved in short-term synaptic plasticity, is voltage-dependent and depends on the presence of the β subunit in an isoform-specific manner (reviewed in Dolphin 2003). It appears that G-protein βγ association with CaV2 channels antagonizes the effects of the β subunit on voltage-dependent activation. The larger the hyperpolarizing effect of the β subunit, the larger the G-protein induced inhibition. Conversely, the β subunits increase the dissociation of Gβγ and thus relieve inhibition during paired pulse facilitation (Canti et al. 2000; Feng et al. 2001).

A similarly strong β subunit dependence on GPCR modulation of CaVs via Gq-proteins has been reported. Both potential mechanisms, inhibition of CaV channels by phosphatidylinositol 4,5-bisphosphate (PIP2) depletion or arachidonic acid generation, are strongly abated upon coexpression of the palmitoylated β2a isoform (Heneghan et al. 2009; Suh et al. 2012). Thus, lipid modulation together with the nature of the CaV-associated β subunit emerges as a powerful modulator of neuronal excitability or neurotransmitter release (Striessnig 2009).

The RGK (Rad, Rem, Rem2, Gem/Kir) family of small monomeric GTP-binding proteins are potent inhibitors of neuronal CaVs; both when heterologously expressed and in native cells including neurons (Chen et al. 2005; reviewed in Buraei and Yang 2010). Multiple inhibitory mechanisms have been suggested including inhibition of membrane expression due to binding to and sequestration of the β subunit and current inhibition of channels preexisting in the membrane. Although calcium current inhibition by RGK proteins absolutely depends on the β subunit and its properties are reminiscent of Gβγ inhibition (see above), recent mutagenesis studies indicate that they use distinct mechanisms (Fan et al. 2010). Whether this potent inhibitory mechanism actually is in effect in synapses, and if so, how it would be activated in neurons remains to be investigated.

CaVs can be regulated by phosphorylation of the α1 subunits as well as the β subunits. PKA, PKC, CaMKII, PI3K/Akt and MAPK have all been shown to phosphorylate β subunits and modulate calcium currents in a β-dependent and isoform-specific manner (Dolphin 2003). For some of these protein kinases the phosphorylation sites in the β subunit have been identified and mutation thereof has been demonstrated to abolish the modulatory effects. If active in the synapse, any of these mechanisms might be fit to modulate synaptic transmission. Moreover, isoform-specific differences in phosphorylation add to the functional heterogeneity and potential specificity of modulatory mechanisms in synapses expressing channels of different subunit composition. However, whereas the physiological role of calcium channel phosphorylation in the fight-or-flight response is well established in the heart (Fuller et al. 2010), a similar role in presynaptic function, and particularly the involvement of β subunits is still elusive.

CaVs functionally interact directly and indirectly via the β subunit with a number of other ion channels and signaling proteins including calcium-activated K+ channels, bestrophin, the ryanodine receptor, dynamin, synaptotagmin I, and the Rab interacting protein RIM1. Most of these proteins can be found in synapses and therefore could potentially function as up- or downstream modulators of synaptic function. As of today the best candidate for such a modulation is RIM1, which is essential for synaptic transmission and plasticity and binds to β subunits with high affinity (Kiyonaka et al. 2007). This interaction appears to affect presynaptic function in two ways. First it is important for docking neurotransmitter vesicles to CaV2 channels, and secondly it modulates voltage-dependent inactivation of the channel. In heterologous expression systems this interaction was observed with any of the β isoforms coexpressed with CaV2 channels. Whether in the context of the synapse the RIM1-β interaction displays more isoform specificity remains to be investigated.

4.2 Channel Trafficking and Current Modulation by α2δ Subunits

The roles of α2δ subunits in synaptic function are less well defined than those of the β subunits. When heterologously expressed all α2δ subunit isoforms can modulate the trafficking and/or the current properties of CaV α1 subunits (reviewed in Arikkath and Campbell 2003; Davies et al. 2007; Obermair et al. 2008). In skeletal and cardiac muscle, for example, α2δ-1 determines the typical current properties of the respective L-type CaVs (Obermair et al. 2005, 2008; Tuluc et al. 2007; Gach et al. 2008). Therefore α2δ-1 is an important determinant of action potential duration in cardiac myocytes (Tuluc et al. 2007; Templin et al. 2011). When coexpressed with neuronal P/Q- or N-type channels all three neuronal α2δ subunits cause an increase in current density (e.g., Davies et al. 2007). Conversely shRNA depletion of α2δ-1 in the skeletal muscle expression system strongly reduced heterologously expressed CaV2.1 and CaV2.2 currents (Obermair et al. 2008).

Based on these results a role of α2δ subunits in triggering neurotransmitter release, which is directly related to the number of presynaptic CaVs (Schweizer et al. 2012), was to be expected. Nevertheless, conflicting results have been reported on the effects of GBP administration on synaptic functions. Whereas acute application of these drugs has only mild effects (if any) on calcium currents (Alden and Garcia 2001; Kang et al. 2002; Micheva et al. 2006; Davies et al. 2007; Dooley et al. 2007) chronic application of GBP has been shown to reduce both native N-type and heterologously expressed P/Q-type calcium currents by about 50 % (Hendrich et al. 2008). Thus it is meanwhile well established that chronic GBP treatment interferes with calcium channel trafficking to the cell surface (Tran-Van-Minh and Dolphin 2010; Dolphin 2012). The importance of α2δ subunits in presynaptic functions (see Table 2.1 and Fig. 2.3) related to their role in CaV targeting is further supported by the upregulation of α2δ subunits in animal models of neuropathic pain (Bauer et al. 2009; Lu et al. 2010) and impaired CaV trafficking after chronic GBP and PG treatment (Bauer et al. 2009; Tran-Van-Minh and Dolphin 2010). Also the recently identified interaction of α2δ-1 with mutant prion protein was shown to impair proper membrane trafficking of the calcium channel complex and consequently reduced glutamatergic transmission in CGNs (Senatore et al. 2012). Indeed chronic treatment with GBP significantly reduced synaptic release efficacy as measured by high KCl-induced FM-dye release in cultured hippocampal neurons (Nimmervoll et al. submitted). This GBP mediated inhibition of synaptic release was augmented in cultures from α2δ-3 null neurons, indicating that α2δ-3 partially compensated for the effects of GBP on α2δ-1 and -2. A similarly strong effect of chronic application of PG on synaptic transmission between dorsal root ganglion and dorsal horn neurons, which primarily express α2δ-1, has been observed (Hendrich et al. 2012).

Table 2.1 Effects of auxiliary calcium channel subunits on synaptic functions
Full size table

shRNA knockdown of α2δ-1 in hippocampal neurons reduced presynaptic expression of CaV2.1 and concomitantly synaptic release probability induced by single action potentials (Hoppa et al. 2012). Conversely, α2δ subunit overexpression increased presynaptic calcium channel density and release probability. However, at the same time the presynaptic calcium signal was significantly reduced. This suggests that α2δ subunits may be involved in linking presynaptic CaVs to the release site. Surprisingly and in contrast to increased release probability upon single action potentials, in our own experiments we observed a slight reduction of synaptic FM-dye release during sustained depolarization upon α2δ-1 overexpression (Nimmervoll et al. submitted). Thus, it is possible that α2δ-1 overexpression one the one hand inhibits calcium influx and consequently synaptic release during prolonged depolarization like trains of action potentials or high KCl. On the other hand this might increase release probability due to a tighter association of calcium channels with the releases site upon single action potentials (Hoppa et al. 2012). Alternatively the reduction in the presynaptic calcium influx may be a consequence of a reduction of the action potential duration upon α2δ subunit overexpression (Hoppa et al. 2012). Thus, besides regulating synaptic transmission, α2δ subunits may control neuronal excitability, for example by increasing somatodendritic calcium channels and thus the coupling to calcium-activated potassium channels as previously characterized for BK channels (Berkefeld et al. 2006). Apart from regulating surface expression the specific functions of α2δ subunits on somatodendritic calcium channels have so far not been studied. While chronic treatment with GBP strongly reduced synaptic FM-dye release (see above), we found release kinetics to be unaffected in a double α2δ-1 knockdown/α2δ-3 knockout model. This strongly implicated the remaining α2δ-2 subunit to compensate for the loss of α2δ-1 or α2δ-3 dependent trafficking and modulation functions. However, it may also suggest that α2δ-2 is chiefly involved in regulating transmitter release, likely by an association with presynaptic P/Q-type channels, which solely determine the kinetic properties of KCl-dependent FM-dye release in cultured hippocampal neurons (Nimmervoll et al. submitted).

A preferential correlation of CaV2.1 and α2δ-2 expression was observed when quantifying overall CaV mRNA abundance (Schlick et al. 2010). Importantly, these recent findings on the differential synaptic roles of α2δ isoforms are further in agreement with the phenotypes of isoform specific α2δ subunit null mice. While α2δ-1 and α2δ-3 knockout mice display only mild overall CNS phenotypes (Fuller-Bicer et al. 2009; Neely et al. 2010), α2δ-2 knockout or mutant (ducky, see box) mice display epilepsy and ataxia and show severely impaired cerebellar development (Brodbeck et al. 2002; Ivanov et al. 2004).

The altered release probabilities observed in hippocampal neurons from double knockout/knockdown cultures provided the first indirect evidence, that in a presynaptic bouton, which simultaneously expresses three α2δ isoforms, a single α2δ isoform may preferentially associate with a specific CaV α1 subunit partner. However, it did not allow any conclusion on the nature or stability of this interaction and many studies described above favor a nonselective interaction of α1 and α2δ subunits, similar to the promiscuity of the interaction with β subunits. In mouse chromaffin cells, for example, PG treatment blocked exocytosis by non-selectively inhibiting CaV1, CaV2.1 and CaV2.2 channels (Hernandez-Vivanco et al. 2012), further indicating the interaction of α2δ-1 with distinct α1 subunits. In agreement with these observations both indirect and direct evidence accumulated over the last years suggesting that α2δ subunits may not be tightly associated with channel complexes and also exist independent of the complex. In our own studies in skeletal muscle cells we could show that free α2δ exists in the plasma membrane without α1 subunits, and that membrane expression of α2δ subunits appears to be independently regulated (Flucher et al. 1991; Obermair et al. 2005, 2008; Schredelseker et al. 2005). Similarly, not all cerebellar CaV2.1 α1 subunits seem to be associated with an α2δ-2 subunit (Davies et al. 2006). As mentioned above, proteomics of mammalian CaV2 channels did not identify α2δ subunits as core components of the complex (Müller et al. 2010). Thus, with the exception of cellular model systems that express an exclusive or at least preferential set of CaV α1, β and α2δ subunits, little information exists on which and how α2δ subunits interact with the CaV complex.

4.3 Functions Independent of the Calcium Channel Complex

4.3.1 α2δ Subunits and Synapse Formation

Traditionally the auxiliary CaV subunits α2δ and β have been envisioned as stable components of the CaV complex in a 1:1 ratio with α1 subunits. However, recently experimental evidence accumulated that suggests cellular function of these two proteins that are in part or entirely independent of the CaV complex. As for the α2δ subunits several studies point towards a major calcium channel independent contribution to synapse formation, likely by interaction with components of the extracellular matrix (Fig. 2.3). α2δ-1 has been shown to act as a receptor for thrombospondin, an astrocyte-secreted protein that promotes CNS snaptogenesis (Eroglu et al. 2009). Overexpression of α2δ-1 strongly promoted, and shRNA knockdown inhibited excitatory synapse formation in cultured retinal ganglion cells. GBP treatment also inhibited synapse formation and the mechanism was shown to involve the α2δ-1 VWF domain. Using a forward genetic screen, Drosophila mutants for the α2δ-3 (straightjackt; stj) isoform have been identified which show defects in presynaptic CaV localization and synaptic function (Dickman et al. 2008). By further analyzing the phenotypes of the α2δ-3 (stj) null mutants, it became evident that motoneurons failed to develop normal synapses (Kurshan et al. 2009). Interestingly, this phenotype was independent of the Drosophila pore forming α1 subunit (cacophony) since cacophony null mutants showed no defect in synapse formation. Mutant (du, du 2J , entla) and targeted knockout mice for α2δ-2 display altered morphology and reduced calcium currents in Purkinje cells (Barclay et al. 2001; Brill et al. 2004; Ivanov et al. 2004; Donato et al. 2006), also suggesting a defect in synapse formation. Finally, it has been shown that the spontaneous mouse mutant of α2δ-4 (Cacna2d4) causes structural and functional abnormalities of retinal ribbon synapses associated with the loss of rods (Wycisk et al. 2006a).

Fig. 2.3
figure 3

Model summarizing the putative effects of the auxiliary α2δ and β subunits in the presynaptic compartment: (1) Trafficking: export from the endoplasmic reticulum (ER) by β; trafficking from recycling endosomes (RE) by α2δ. (2) CaV current modulation: modulation by distinct β isoforms either directly, or by mediating modulatory mechanism in a β-isoform dependent manner (e.g. GPCR modulation); modulation by distinct α2δ subunits (indicated in blue/orange) by association/dissociation. (3) Linking calcium channels to the release site (SV, synaptic vesicle): β subunits via binding to RIM and SNARE proteins; α2δ subunits by association with a potential extracellular ligand and/or the extracellular matrix (ECM). Thereby individual α2δ isoforms (indicated in orange) may link calcium channels better to the release site by interaction with special ECM components or a specialized lipid domain in the synaptic membrane (indicated by blue double line)

Full size image

All these strong effects of loss-of-α2δ-function on synapse structure and formation where revealed in model systems that primarily express only one α2δ isoform, such as retinal ganglion cells, Drosophila motoneurons, and mammalian photoreceptors. In cellular systems which express more than one α2δ isoform, such as CNS neurons, both calcium channel dependent and independent functions of α2δ subunits appear to be subject to compensation by other α2δ isoforms. Recently we analyzed the density of functional synapses of α2δ loss-of-function models (Nimmervoll et al. submitted). We found that synapse formation was still close to normal in α2δ-3 deficient cultured hippocampal neurons in which α2δ-1 was knocked down or α2δ-1 and α2δ-2 were chronically blocked with GBP. Thus, it seems that the contribution of individual α2δ isoforms to synapse formation is limited in neurons expressing three different α2δ isoforms. This is further supported by the recent characterization of α2δ-3 knockout mice, which did not reveal overall effects on synapse formation (Neely et al. 2010). To answer this question, it will ultimately be necessary to study synapse formation in CNS neurons lacking all α2δ isoforms.

4.3.2 β Subunits and Transcriptional Regulation

The first indication of calcium channel independent functions of β subunits came from isolated observations of heterologously expressed β subunits localized in the cell nuclei (Colecraft et al. 2002; Hibino et al. 2003; Beguin et al. 2006). Interestingly, the truncated chicken β4c isoform associated with heterochromatin protein 1 (HP1) a nuclear protein involved in gene silencing (Hibino et al. 2003). In 2009 we localized the endogenous β4 isoform in the nuclei of cerebellar granule and Purkinje cells and demonstrated in a skeletal muscle expression system that nuclear targeting of heterologous β subunits is isoform and splice variant specific (β4b>>β4a=β3>β1a=β1b=β2a=β2b) and negatively regulated by electrical activity and calcium influx into nerve and muscle cells (Subramanyam et al. 2009). The finding that in immature and quiescent cells β4b accumulated in the nucleus and upon the onset of electrical activity it was released from the nuclei suggested a possible role in activity dependent gene regulation. Very recently Tadmouri et al. reported that β4b associates with the regulatory subunits of protein phosphatases 2A, translocates into the nucleus in an activity dependent manner, where it associates with the tyrosine hydroxilase promoter and histone H3 in complex with HP1 (Tadmouri et al. 2012). Importantly, a truncated β4 mutant associated with juvenile myoclonic epilepsy failed to complex with B56δ and consequently did not translocate into the nucleus. These findings suggest that the neurological disease phenotype in humans and that of the β4 knockout mouse are at least in part related to the nuclear function of the β4b subunit, whereas its calcium channel dependent functions may be compensated by other β isoforms. Also β3 subunits may function in transcriptional regulation. Recently the specific interaction of β3 with a novel Pax6(S) transcriptional regulator has been described (Zhang et al. 2010). Upon coexpression in Xenopus oocytes β3 is translocated into the nucleus and suppresses the transcriptional activity of Pax6(S). As Pax6 transcriptional regulators are important during development, a role of this calcium channel independent activity of β in developmental regulation has been suggested. Consistent with function in early development, morpholino knockdown of β4 in zebrafish embryos blocked epiboly, a reorganization of cells during gastrulation (Ebert et al. 2008). Importantly, this effect could be rescued by coexpression of a β4a isoform with mutated AID binding pocket, again indicative of a calcium channel independent mechanism. So far no direct link of any of these nuclear functions of β subunits to synaptic function has been established. However, because these novel pathways for transcriptional regulation are activity dependent and affect developmental processes, a mechanism by which β subunit signaling provides a feedback loop from overall synaptic activity to synapse efficacy analogous to homeostatic plasticity can be envisioned.

4.4 Auxiliary β and α2δ Subunits and Neuronal Disease

There is little evidence for an involvement of calcium channel β subunits in neurological disease. Also, with the exception of lethargic (β4-null) mice, mouse mutant and knockout models of β subunits show little to no neurological defects (see box). Loss of function phenotypes can be observed in cell types predominantly expressing a single β isoform like skeletal (β1) and cardiac muscle (β2) or the retina and inner hair cells (β2). In other cells, including most neurons, expression of other β isoforms seems to compensate the loss of the respective isoform. The β4 subunit is the notable exception. Mutations resulting in a truncated protein have been linked to juvenile myoclonic epilepsy (Escayg et al. 2000) and the lethargic β4-null mutant mouse develops severe ataxia and epileptic seizures (Burgess et al. 1997). The similarity of this phenotype to that of CaV2.1 (tottering, leaner) (Doyle et al. 1997) and α2δ-2 (ducky) (Barclay et al. 2001) mutants and their predominant expression in cerebellum (see above) indicates that in some cerebellar neurons this set of subunits forms an essential channel complex. Loss of any one of its components cannot be fully compensated by other isoforms. Alternatively, the neurological β4 phenotype could arise from an exclusive nuclear function of this subunit in gene regulation.

As mentioned above, α2δ-dependent functions can be exerted as calcium channel subunits on the one hand, and independent of the CaV complex on the other. For example meanwhile it is well established that α2δ-1 is strongly upregulated in dorsal root ganglion neurons in animal models of neuropathic pain (Luo et al. 2001). The beneficial effect of GBP and PG in neuropathic pain (Field et al. 2006) most likely results from impairing excess α2δ subunit trafficking (Bauer et al. 2009). As a possible mechanism inhibiting recycling of α2δ subunits from the endosomes has been described (Tran-Van-Minh and Dolphin 2010). straightjacket mutants also display altered heat nociception and CACNA2D3 (α2δ-3) single nucleotide polymorphisms (SNPs) in humans have been linked to central pain processing (Neely et al. 2010). This phenotype, which is likely caused by a change in local CNS excitability, could both be explained by a defect in CaV trafficking and synapse formation. Mutant (du, du 2J , entla) and knockout mice for α2δ-2 display altered morphology and reduced calcium currents in Purkinje cells as well as cerebellar ataxia and absence epilepsy (Barclay et al. 2001; Brill et al. 2004; Ivanov et al. 2004; Donato et al. 2006). In humans the α2δ-2 gene (CACNA2D2) has been discussed as a potential tumor suppressor gene (Hesson et al. 2007) and in the context of childhood absence epilepsy (Chioza et al. 2009). Indeed, very recently Edvardson et al. identified the first human mutation in the CACNA2D2 gene associated with an early infantile epileptic encephalopathy (Edvardson et al. 2013). A spontaneous mouse mutant of α2δ-4 (Cacna2d4) causes structural and functional abnormalities of retinal ribbon synapses associated with the loss of rods (Wycisk et al. 2006a) and a human CACNA2D4 mutation underlies a slowly progressing cone dystrophy associated with night blindness (Wycisk et al. 2006b). Finally, clinical applications of GBP and PG provide an important link between α2δ subunits and neuronal disease. Besides their effectiveness in neuropathic pain conditions, which is most likely mediated by binding to α2δ-1 (Field et al. 2006), both drugs have proven efficacy in epilepsy and generalized anxiety disorders (Bryans and Wustrow 1999; Johannessen Landmark 2008). The recently identified interaction of mutant prion protein with α2δ-1 may provide an essential disease mechanism in the pathophysiology of prion diseases, namely by disrupting cerebellar glutamatergic neurotransmission (Senatore et al. 2012).

5 Conclusion

Moving the focus of calcium channel research from heterologous expression systems to differentiated cells including neurons and to the study of animal models have greatly advanced our understanding of the physiology of auxiliary CaV subunits. However, many of the new functional insights have also revealed our limited ability to associate their specific functions to particular molecular entities. This deficit has been further exacerbated by the growing molecular diversity of calcium channel subunits brought about by posttranscriptional modifications like splicing and RNA editing. Therefore, future research first and foremost needs to uncover how specific CaV complexes are established in neurons expressing many different isoforms. As outlined above, this will require the detailed study of their expression patterns, their targeting mechanisms and their protein-protein interactions. To uncover these aspects in the context of calcium channel signalling complexes like the presynaptic compartment high- and superresolution microscopy approaches will be necessary. Finally, the static picture of molecular complexes needs to be replaced by one of highly dynamic signalosomes, in which all the mechanisms mentioned above contribute to an equilibrium of multiple protein-protein interactions that ultimately determines the functional properties of the signalling complex. In the synapse such dynamic calcium channel complexes are critical for the activity-dependent regulation of synaptic strength and ultimately for the ability of our nervous system to learn and store new information.