The Calcium Channel β_4a Subunit: A Scaffolding Protein Between Voltage-Gated Calcium Channel and Presynaptic Vesicle-Release Machinery?

Neuronal voltage-gated calcium channels represent a major pathway for calcium entry into nerve termini, where they control neurotransmitter release. The channels are composed of a pore-forming subunit (Ca_v2.x) and auxiliary subunits (Ca_vβ_1–4, α₂Δ_1–4 and γ_1–8). The cytoplasmic Ca_vβ subunits belong to the membrane-associated guanylate kinase (GK) family of proteins and regulate trafficking to the plasma membrane and gating properties of Ca_v2.x channels. Their five domains (A–E), include hypervariable A, C, and E domains linked to the highly conserved Src homology 3 (SH3) (B) and GK (D) domains. The crystal structure of the core domains (B–D) of several β subunits was recently elucidated (Chen et al., 2004; Opatowsky et al., 2004; Van Petegem et al., 2004), but less is known about the structure of the A and E domains. A recent report published in The Journal of Neuroscience provides new information on the Ca_vβ_4a subunit hypervariable A domain (Ca_vβ_4a-A) (Vendel et al., 2006a). The authors previously showed that alternative splicing of the Ca_vβ₄-A domain generates two distinct proteins, Ca_vβ_4a and Ca_vβ_4b (Helton and Horne, 2002), and they recently provided the solution structure of the Ca_vβ_4a-A domain (Vendel et al., 2006b). Vendel et al. now address the expression pattern of the Ca_vβ_4a splice variant and the functional importance of the Ca_vβ_4a-A domain.

Using immunohistochemistry, the authors show that the Ca_vβ_4a splice variant was expressed as punctuate structures throughout the molecular layer of the cerebellum [Vendel et al. (2006a), their Fig. 3D,E (http://www.jneurosci.org/cgi/content/full/26/10/2635/F3)]. In contrast, Ca_vβ_4b was expressed in basket cells surrounding Purkinje cell bodies as well as in the Bergmann glia [Vendel et al. (2006a), their Fig. 3F (http://www.jneurosci.org/cgi/content/full/26/10/2635/F3)]. To determine the functional importance of the Ca_vβ_4a-A domain, the authors performed two electrode voltage-clamp recording in Xenopus laevis oocytes expressing Ca_v2.1 channels (P/Q type currents) in combination with α₂Δ and Ca_vβ_4a subunits with (ABCDE) or without (BCDE) the N-terminal A domain. The Ca_vβ_4a-A domain was not essential for the expression of the Ca_v2.1 channel at the plasma membrane and did not influence gating properties. Indeed, no difference was observed in current amplitude [Vendel et al. (2006a), their Fig. 2A (http://www.jneurosci.org/cgi/content/full/26/10/2635/F2)], nor in the voltage dependence of activation or inactivation [Vendel et al. (2006a), their Fig. 2C,D (http://www.jneurosci.org/cgi/content/full/26/10/2635/F2)]. These results are, however, in agreement with published studies (Bichet et al., 2000; Van Petegem et al., 2004) because (1) enhanced Ca_v2.x subunit trafficking to the plasma membrane occurs after binding of the Ca_vβ subunit to the channel (Bichet et al., 2000) via the α interaction domain (AID)-binding pocket within the GK domain (Van Petegem et al., 2004), and (2) the molecular determinants by which the Ca_vβ subunit modulates channel gating remain unclear but are probably carried by the conserved core domains because all Ca_vβ subtypes are able to modulate the biophysical properties of the channel. However, it is more surprising that channel inactivation kinetics was not influenced by deletion of the Ca_vβ_4a-A domain [Vendel et al. (2006a), their Fig. 2B (http://www.jneurosci.org/cgi/content/full/26/10/2635/F2)]. Indeed, it is known that a deletion of the N-terminal of Ca_vβ subunits results in a drastic slowing of channel inactivation kinetics (Olcese et al., 1994). It seems not to be the case for the Ca_vβ_4a splice variant. Thus, further investigations will provide interesting structural information on how the N-terminal of Ca_vβ subunits controls channel inactivation.

Because the Ca_vβ_4a-A domain was not a key determinant in the regulation of Ca_v2.1 channel gating, Vendel et al. looked for a role of this domain in protein–protein interactions. Using the yeast two-hybrid system, the authors screened a human cerebellum cDNA library with the Ca_vβ_4a-A domain (amino acids 1–58). Synaptotagmin I (Syt I) (amino acids 95–337 including the entire C2A domain and a half of the C2B domain) as well as the microtubule-associated protein 1A (MAP1A) (amino acids 2508–2775 corresponding to the complete LC2 domain) interacted specifically with the N-terminal A domain of the Ca_vβ_4a splice variant but not with the Ca_vβ_4b-A domain [Vendel et al. (2006a), their Fig. 4B (http://www.jneurosci.org/cgi/content/full/26/10/2635/F4)]. The authors then focused their study on Syt I and confirmed its interaction with the Ca_vβ_4a-A domain by in vitro pull-down experiments. The interaction did not occur in the presence of 10 mm Ca²⁺ and could also be disrupted by adding Ca²⁺ to the medium [Vendel et al. (2006a), their Fig. 5C (http://www.jneurosci.org/cgi/content/full/26/10/2635/F5)]. These data certainly represent the most important findings of this study.

In conclusion, Vendel et al. (2006a) provide evidence that alternative splicing of the N-terminal A domain of the Ca_vβ₄ auxiliary subunit confers functions other than modulations of channel gating and trafficking. Because Ca_vβ_4a splicing variant can bind synaptotagmin I, an important protein for presynaptic vesicle release, the β subunit, could conceivably act as a scaffolding element to facilitate coupling of calcium signaling with neurotransmitter release (Fig. 1). However, the authors do not provide evidence that Ca_vβ_4a can bind Ca_v2.1 channel and Syt I simultaneously. Pull-down experiments performed by preincubating the full-length Ca_vβ_4a with the AID peptide (the molecular determinant of the Ca_v2.x channel, which interacts with the AID-binding pocket of the Ca_vβ subunit) before adding Syt I could partially answer this question. Finally, the fact that Ca_vβ_4a–Syt I interaction is disrupted by Ca²⁺ is interesting. In this context, we speculate that at basal Ca²⁺ levels, Ca_vβ_4a interacts both with Ca_v2.1 and Syt I to organize the vesicle-release machinery and that calcium entry into cells via voltage-gated calcium channels breaks this interaction, thus releasing the vesicle to allow fusion with the plasma membrane. FRET experiments using tagged Syt I and Ca_vβ_4a in the presence of Ca_v2.1 channels before and during membrane depolarization might address such a possible mechanism.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

A putative model of coupling between voltage-gated calcium channels, calcium signaling, and presynaptic vesicle-release machinery. A, In basal condition with a low Ca²⁺ level, Ca_vβ_4a interacts both with Ca_v2.1 channel and Syt I to position the synaptic vesicle in front of the channel to better couple calcium signaling and neurotransmitter release. B, Calcium entry into the cell via voltage-gated calcium channels and after membrane depolarization disrupts Ca_vβ_4a–Syt I interaction. The content of the synaptic vesicle is thus released by the fusion of the vesicle with the plasma membrane after soluble N-ethylmaleimide-sensitive factor attachment protein (SNAP) receptors v-SNAREs (synaptobrevin and synaptotagmin I) and t-SNAREs (syntaxin and SNAP 25) interactions. PM, Plasma membrane.