|Home | About | Journals | Submit | Contact Us | Français|
The β-subunit of voltage-gated Ca2+ channels is essential for trafficking the channels to the plasma membrane and regulating their gating. It contains a Src homology 3 (SH3) domain and a guanylate kinase (GK) domain, which interact intramolecularly. We investigated the structural underpinnings of this intramolecular coupling and found that in addition to a previously described SH3 domain β strand, two structural elements are crucial for maintaining a strong and yet potentially modifiable SH3-GK intramolecular coupling: an intrinsically weak SH3-GK interface and a direct connection of the SH3 and GK domains. Alterations of these elements uncouple the two functions of the β-subunit, degrading its ability to regulate gating while leaving its chaperone effect intact.
High voltage-activated (HVA) Ca2+ channels are composed of the pore-forming α1 subunit and ancillary α2δ, β and, in skeletal muscles, γ subunits . The β-subunit (Cavβ) is essential for transporting the channel complexes to the plasma membrane and modulating their gating properties. There are four subfamilies of Cavβs, each with multiple splicing isoforms and unique modulatory functions . Previous studies show that Cavβ has a modular structure consisting of five distinct domains or regions: the N-terminus, an Src homology 3 (SH3) domain, a HOOK region, a guanylate kinase (GK) domain and the C-terminus [3–6]. The SH3-HOOK-GK module is also present in membrane-associated guanylate kinases (MAGUKs), a large family of PDZ domain-containing scaffold proteins important for the synaptic localization and clustering of various membrane receptors and ion channels [7,8].
Previous studies indicate that Cavβ also functions modularly. The GK domain binds tightly to a region called the α interaction domain (AID) in the cytoplasmic loop connecting the first two repeats of the α1-subunit [9–11]. This AID-GK domain interaction is necessary for trafficking HVA Ca2+ channels to the cell surface [9,10,12–18]. The SH3-HOOK-GK core module is responsible for modulating the activation properties  and the HOOK and N-terminus play a predominant role in modulating inactivation, either directly by interacting with the α1-subunit and/or indirectly through palmitoylation [16,19–29].
A feature revealed by the Cavβ crystal structures is that the SH3 and GK domains are coupled intramolecularly [3,5,6]. This structural feature is shared by the MAGUK SH3-HOOK-GK module [7,8]. Disruption of the SH3-GK interaction in MAGUKs has severe functional consequences [30,31]. The SH3-GK intramolecular coupling is also important for Cavβ functions, including gating modulation [32–35] and surface trafficking . The β5 strand of the SH3 domain has been found to be critical for the SH3-GK intramolecular coupling [6,32–36].
In this work we investigated additional structural underpinnings of the SH3-GK intramolecular coupling in Cavβ. Taking full advantage of the Cavβ crystal structures, we made specific point mutations, deletions and insertions to create various mutant forms or fragments of Cavβ and tested their functions. Our results shed new light on understanding the structural design and function of the SH3-GK intramolecular coupling in Cavβ.
P/Q-type Ca2+ channels containing wild-type (WT) or mutant β2a were expressed in Xenopus oocytes. The boundaries of the five domains/regions of β2a are: N-terminus: M1-P59; SH3 domain: V60-S120 and P219-P224; HOOK region: P121-P218; GK domain: S225-T410; C-terminus: H411-Q604. Currents were recorded with cell-attached patch-clamp or two-electrode voltage-clamp recording. Full details are available in Supplementary Materials and Methods.
The functional properties we examined included the voltage-dependence (the midpoint V1/2 and slope factor K) of activation and steady-state inactivation (protocols illustrated in Fig. 1A–D) and the kinetics of inactivation. All experiments were performed on P/Q-type Ca2+ channels produced in Xenopus oocytes expressing Cav2.1, α2δ, and a WT or mutant β2a, using the cell-attached patch-clamp recording with Ba2+ as the charge carrier. β2a was used because the difference in the inactivation properties is the largest between the WT and the GK-C module . This allowed us to more easily detect changes caused by the point mutations, deletions and insertions.
The Cavβ structures show that SH3 and GK domains associate with each other through limited interactions [3,5,6] [Supplementary Fig. S1]. To examine the role of the SH3-GK interface interactions in Cavβ functions, we simultaneously mutated seven amino acids in β2a that are directly involved in these interactions to alanine (this mutant is referred to as β2a_Mut7) (Fig. 1A). These residues are highly conserved among all four Cavβ subfamilies (Supplementary Fig. S2) and the hepta-mutation is very likely to severely disrupt the SH3-GK domain association.
β2a_Mut7 clearly stimulated channel surface expression (Supplementary Fig. S3). However, compared with channels containing WT β2a, those containing β2a_Mut7 exhibited much faster inactivation (Fig. 1B and C, and Table 1). The voltage-dependence of steady-state inactivation was shifted toward more hyperpolarized potentials (Fig. 1D and Table 1). However, there was no change in the voltage-dependence of activation (Table 1). The fact that the hepta-mutation specifically affected β2a’s modulation of inactivation without compromising its ability to regulate channel expression and activation suggests that the mutation did not significantly perturb protein folding. Instead, the mutation probably altered the quaternary arrangement of the N-SH3-HOOK module with respect to the GK-C module. This change may render the N-terminus and the HOOK region less efficient in regulating inactivation [16,19–29].
Are the SH3-GK interface interactions alone sufficient for maintaining a strong SH3-GK intramolecular coupling? To address this question, we divided β2a into two fragments, one containing the N-SH3-HOOK module (β2a_N-SH3-HOOK) and the other containing the GK-C module (β2a_GK-C) (Fig. 2A). We then tested whether these two parts could coassemble into a complex that could fully recapitulate the functions of WT β2a.
We first examined the functionality of each individual component. β2a_N-SH3-HOOK was completely ineffective in stimulating channel expression on the plasma membrane, as determined by two-electrode voltage-clamp (data not shown). This is expected since the AID-GK interaction is required for this effect [9,10,12–17]. In contrast, β2a_GK-C fully retained the chaperone function, as we previously observed . This contradicts the lack of trafficking effect of the GKC construct of Takahashi et al. [34,35]. The reasons underlying this discrepancy are unclear. One possibility is the difference in the expression system (Xenopus oocytes vs. HEK 293 cells). Channels containing β2a_GK-C inactivated rapidly (Fig. 2B and Table 1) and exhibited a ~27-mV negative shift in the half-inactivation voltage compared to WT (Fig. 2C and Table 1). These results confirm that the GK-C module cannot bestow WT inactivation and reinforce the critical role of the N-terminus and the HOOK in regulating inactivation [16,19–27,29].
We next expressed β2a_N-SH3-HOOK and β2a_GK-C together. Surprisingly, the resultant channels behaved exactly like those obtained when β2a_GK-C was expressed alone (Fig. 2D and E, and Table 1). This result could come about if the β2a_N-SH3-HOOK protein was not synthesized efficiently or folded properly. This possibility seemed highly unlikely since a construct similar to β2a_N-SH3-HOOK but missing the last 9 amino acids including the β5 strand of the SH3 domain (and hence was more likely to be defunct) was synthesized efficiently and folded properly (see below). Our hypothesis was that the SH3-GK interface interactions alone were intrinsically too weak to hold β2a_N-SH3-HOOK and β2a_GK-C together, so that the surface channels contained only β2a_GK-C; meanwhile, β2a_N-SH3-HOOK itself could not bind α1 tightly enough to exert its influence on inactivation.
To test this hypothesis, we examined the SH3-GK interface interactions biochemically. Indeed, purified β2a_N-SH3-HOOK protein (tagged with glutathione S-transferase (GST)) failed to bind β2a_GK-C (tagged with maltose binding protein (MBP)) in an in vitro pull down assay (Fig. 2F, lane 5).
So what other structural elements are also involved in maintaining the strong SH3-GK intramolecular coupling? It was noted in the Cavβ structure work that unlike canonical SH3 domains, which consist of five contiguous β strands, the Cavβ SH3 domain has a split configuration in which the fifth β strand (β5) is separated from the rest of the SH3 domain by a HOOK region [3,5,6]. The β5 strand interacts weakly with the GK domain but strongly with other parts of the SH3 domain (Supplementary Fig. S4). Since the β5 strand is directly connected to the GK domain in the primary amino acid sequence, these interactions bring the SH3 and GK domains in close proximity and thereby greatly increase the effective affinity between the two domains (Fig. 1A).
Previous studies suggest that the β5 strand plays a crucial role in the SH3-GK intramolecular coupling [6,32–36]. To further investigate its role, we divided β2a into two other fragments, referred to as β2a_N-SH3-HOOK (−β5) and β2a_GK-C (+β5) (Fig. 2A). As expected, β2a_GK-C (+β5) increased Ca2+ channel current (Supplementary Fig. S3) and β2a_N-SH3-HOOK (−β5) did not (data not shown). The channels containing β2a_GK-C (+β5) behaved the same as those containing β2a_GK-C did, exhibiting rapid inactivation and negatively shifted half-inactivation voltage (Fig. 2D and F, and Table 1), indicating that the β5 strand itself does not change the inactivation properties. However, when β2a_N-SH3-HOOK (−β5) was expressed together with β2a_GK-C (+β5), the resulting channels behaved similarly as channels containing WTβ2a did (Fig. 2D and F, and Table 1), indicating the association of β2a_N-SH3-HOOK (−β5) and β2a_GK-C (+β5) and functional reconstitution. Consistent with these results, purified β2a_N-SH3-HOOK (−β5) protein was able to pull down β2a_GK-C (+β5) (Fig. 2F, lane 3).
Taken together, these results indicate that in addition to the direct contact between the SH3 and GK domains, the β5 strand of the SH3 domain is indispensable for forging a strong SH3-GK intramolecular coupling. Biochemical association and functional reconstitution have already been demonstrated for several other singly cleaved Cavβ fragments [6,32–36]. It is notable that in all the cases of successful complementation, the β5 strand was always included together with the GK-C module.
It is evident from the Cavβ crystal structures that the SH3 and GK domains are juxtaposed without a flexible linker (Fig. 1A). We suspected that this structural feature might also be an important element in maintaining a strong SH3-GK intramolecular coupling. To test this hypothesis, we inserted a flexible linker of variable length (containing 3, 6, 12, 21 and 30 amino acids) at the SH3-GK junction, between P224 and S225 (Fig. 3A). These constructs were called β2a_linker3, β2a_linker6, etc. Our preceding results show that the SH3-GK interface interactions are too weak to glue the two domains together. Thus, when a flexible linker is inserted in between the two domains, the N-SH3-HOOK module is likely to separate from the GK-C module, which is anchored to the α1 subunit via its high-affinity binding to the AID. With increasingly longer linkers, the effective local concentration of the N-SH3-HOOK module near the α1-subunit would decrease gradually and the properties affected by this module would change incrementally.
The results came out as expected. With a linker as short as 3 amino acids, inactivation of the resultant channels became significantly faster than that of WT channels (Fig. 3B and Table 1) and the half-inactivation voltage was shifted to more hyperpolarized potential (Fig. 3C and Table 1) (compare β2a_WT and β2a_linker3). With longer linkers, inactivation became progressively faster (Fig. 3D and Table 1) and the half-inactivation voltage was shifted progressively to more hyperpolarized potentials (Fig. 3E and Table 1). This graded change and the fact that the voltage-dependence of activation was unaffected (Table 1) suggest that the effects on inactivation are unlikely nonspecific effects produced by the linkers themselves.
Furthermore, when a linker was inserted into β2a_Mut7 (the resulting construct was referred to as β2a_Mut7_linker3, β2a_Mut7_linker6, etc.), the changes in inactivation were even more dramatic. This was the case for both short (Fig. 3F and G, and Table 1) and long linkers (Fig. 3H and I, and Table 1). The graded change in the kinetics and voltage-dependence of inactivation with increasingly longer linkers was also more robust in the β2a_Mut7 background (Fig. 3J and K, and Table 1). Indeed, channels containing β2a_Mut7_linker30 behaved almost the same as channels containing β2a_GK-C did (Fig. 3H-K, and Table 1). Thus, it appears that with a 30-amino acid flexible linker, the effective concentration of the N-SH3-HOOK module near the surface channels is too low to exert its modulatory function, even though it is still tethered to the GK-C module. These results not only demonstrate the importance of a direct physical connection of the SH3 and GK domains, but also further support the notion of an intrinsically weak SH3-GK interface.
We next performed surface biotinylation and Western blot to confirm that β2a_Mut7_linker30 was associated with surface Ca2+ channels and was intact. Indeed, β2a_Mut7_linker30 was pulled down together with biotinylated surface α1- and/or α2δ-subunits by streptavidin, indicating β2a_Mut7_linker30 was present in surface Ca2+ channel complexes (Fig. 3L). Furthermore, the molecular weight of the pull-down β2a_Mut7_linker30 remained as expected (Fig. 3L), indicating that the N-SH3-HOOK module was not cleaved off by proteolysis as a result of linker insertion.
Our results indicate that the strong SH3-GK intramolecular coupling in Cavβ is orchestrated by three structural elements: a direct but weak SH3-GK interface, a split SH3 domain with a β strand (β5) that engages in both intradomain (SH3) as well as interdomain (SH3-GK) interactions, and a direct linkage of SH3 and GK domains in the primary amino acid sequence. This intramolecular coupling is not necessary for the chaperone function of Cavβ-every mutant construct with disrupted SH3-GK coupling was still able to traffic Ca2+ channels to the surface membrane as long as it contained the GK domain. This notion is consistent with recent findings that the GK domain alone is sufficient to carry out the chaperone function [16,18]. On the contrary, a strong SH3-GK intramolecular coupling is essential for the modulation of gating (especially inactivation) by Cavβ. Through this coupling, the N-SH3-HOOK module is tightly bound to the GK-C module and as such is anchored to the α1-subunit. This enables the N-terminus and the HOOK region to interact with the α1-subunit, interactions that are of intrinsic low affinity but are essential for the modulation of inactivation.
A unique property of β2a (from rat and human) is that its N-terminus can be palmitoylated, which anchors β2a to the plasma membrane and contributes partly to its depolarizing shift of the steady-state inactivation voltage and slowing of the inactivation kinetics [23–25,29]. Our results show that the N-SH3-HOOK module was functionally mute when it was expressed alone, or coexpressed with β2a_GK-C (Fig. 2B and C, and Table 1), or tethered to the GK-C module via a 30-amino acid linker (Fig. 3 and Table 1). These observations suggest that palmitoylation alone is insufficient to promote the SH3-GK intramolecular coupling and confer functionality to the N-SH3-HOOK module. On the other hand, it has been reported that in tsA-201 cells palmitoylation of β2a is able to endow it the ability to modulate gating of a mutated Cav2.2 with a severely weakened AID-GK interaction . The discrepancy between our work and this work is probably because in tsA-201 cells, due to its relatively small size, overexpression and palmitoylation of β2a greatly increases its effective concentration near the channels on the plasma membrane, and thus, enabling it to modulate gating.
Three lines of biochemical and, more important, functional evidence indicate that the SH3-GK interface interactions are intrinsically weak: β2a_N-SH3-HOOK and β2a_GK-C did not bind each other in vitro (Fig. 2F) and did not complement each other functionally when expressed together (Fig. 2B and C) or when linked together via a long linker (Fig. 3B–E). An intrinsically weak SH3-GK interface may confer additional versatility to Cavβ modulation of HVA Ca2+ channel gating. Under normal conditions, the SH3 and GK domains are likely to act as a rigid body owing to the strong SH3-GK intramolecular coupling forged synergistically by the above three structural elements. This rigid body structure is essential for the gating modulation. However, if any region of the N-SH3-HOOK module were engaged in a high-affinity interaction with another protein, it might generate enough force to disrupt the intrinsically weak SH3-GK interface, which would result in dramatic changes in the inactivation of the channels. Similar functional consequences could also be produced by naturally occurring mutations that severely disrupt the SH3-GK intramolecular coupling. It would be interesting to identify such proteins or mutations in future studies.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.