Ca_vβ surface charged residues contribute to the regulation of neuronal calcium channels

Abstract

Voltage-gated calcium channels are essential regulators of brain function where they support depolarization-induced calcium entry into neurons. They consist of a pore-forming subunit (Ca_vα₁) that requires co-assembly with ancillary subunits to ensure proper functioning of the channel. Among these ancillary subunits, the Ca_vβ plays an essential role in regulating surface expression and gating of the channels. This regulation requires the direct binding of Ca_vβ onto Ca_vα₁ and is mediated by the alpha interacting domain (AID) within the Ca_vα₁ subunit and the α binding pocket (ABP) within the Ca_vβ subunit. However, additional interactions between Ca_vα₁ and Ca_vβ have been proposed. In this study, we analyzed the importance of Ca_vβ₃ surface charged residues in the regulation of Ca_v2.1 channels. Using alanine-scanning mutagenesis combined with electrophysiological recordings we identified several amino acids within the Ca_vβ₃ subunit that contribute to the gating of the channel. These findings add to the notion that additional contacts besides the main AID/ABP interaction may occur to fine-tune the expression and properties of the channel.

Main text

Neuronal high-voltage-activated (HVA) calcium channels are multisubunits complexes that support depolarization-induced calcium entry and downstream cellular functions [1]. They are composed of a pore-forming subunit (Ca_vα₁) that consists of four homologous membrane domains, each composed of six transmembrane helices, connected via cytoplasmic linkers (I–II, II–III, and III–IV loops), and cytoplasmic amino- and carboxy termini. They require the co-assembly with ancillary subunits to ensure the proper functioning of the channel. Among these ancillary subunits, the cytoplasmic Ca_vβ regulates several aspects of HVA channels including their gating properties and expression at the cell surface (for review see [2]). Ca_vβ subunits are encoded by four different genes (Ca_vβ_1–4) and belong to the family of membrane-associated guanylate kinase (MAGUK). They consist of a conserved core region formed by Src homology 3 (SH3) and guanylate kinase (GK) domains connected by a HOOK region, flanked by non-conserved amino- and carboxy-termini (Fig. 1a). The molecular assembly of the Ca_vα₁/Ca_vβ complex relies on a conserved 18 residue sequence within the I–II loop of Ca_vα₁ called α₁ interaction domain (AID) [3] that binds into a hydrophobic groove within the GK domain of Ca_vβ termed AID-binding pocket (ABP) [4,5,6] (Fig. 1a). This high-affinity interaction is critical for Ca_vβ-mediated enhancement of Ca_vα₁ surface expression and gating. Mutation of key residues within the ABP that weakens or abolishes AID-ABP interaction severely alters the functional influence of Ca_vβ [7]. Besides the AID/ABP interaction, additional low-affinity contacts between Ca_vα₁ and Ca_vβ that do not involve the ABP have been proposed to confer essential Ca_vβ modulatory properties [8,9,10]. In this study, we aimed to assess the functional importance of Ca_vβ surface charged residues in the regulation of Ca_v2.1 channels. To do so, we generated a number of Ca_vβ₃ mutants where surface charged residues, most belonging to the GK domain, were replaced with an alanine (Fig. 1b), and recombinant Ca_vβ₃ were expressed in Xenopus oocytes with Ca_v2.1 for electrophysiological analyses in the presence of 40 mM barium as charge carrier. Ca_vβ₃ was chosen over Ca_vβ₄ because it induces a more pronounced phenotype on Ca_v2.1 with faster inactivation kinetics, and also because according to our experience, the association of Ca_vβ₃ with Ca_v2.1 in expression experiments is more complete than of Ca_vβ₄ which would have made the interpretation of Ca_vβ₄ variants slightly more difficult overall. As expected, the maximal macroscopic conductance (G_max) in cells expressing Ca_v2.1 was increased by 3.3-fold (p = 0.0001) in the presence of wild-type (WT) Ca_vβ₃ compared to cells expressing Ca_v2.1 alone (Fig. 1c, d, Additional file 1: Fig. S1, Additional file 2: Table S1). Similarly, all Ca_vβ₃ variants, except the E347A mutant, produced a significant increase of Ca_v2.1 conductance indicative of the proper expression of Ca_vβ₃ mutants (Fig. 1d, Additional file 1: Fig. S1, Additional file 2: Table S1). However, Ca_vβ₃-dependent potentiation of Ca_v2.1 currents was significantly reduced when residues D343, D344, E347, E354 (located in the GK domain), and R358 (located in the N-terminus) were mutated (ranging from 1.4-fold reduction for Ca_vβ₃ R358A to 2.0-fold reduction for Ca_vβ₃ E347A compared to WT Ca_vβ₃) (Fig. 1d, Additional file 1: Fig. S1, Additional file 2: Table S1). While the exact underlying mechanisms have not been further investigated in this study, this alteration is likely to have resulted from either a decreased surface expression of the channel, or from a decreased Ca_vβ-dependent potentiation of the single channel gating (channel open probability and latency to first channel opening). Consistent with the latest, we observed that while co-expression of WT Ca_vβ₃ produced a 10.7 mV hyperpolarizing shift (p = 0.0001) of the mean-half activation potential of Ca_v2.1, this effect was significantly reduced when the channel was co-expressed with Ca_vβ₃ D343A, D344A, and E347A (Fig. 1e and f, Additional file 1: Fig. S2, Additional file 2: Table S1). In contrast, mutation of residues E354 and R358 did not alter Ca_vβ₃-mediated hyperpolarization of the voltage-dependence of activation of Ca_v2.1 suggesting that the effect of Ca_vβ₃ mutants on Ca_v2.1 conductance may have resulted from distinct gating alteration. In that respect, we note that while Ca_vβ₃ H206A did not alter the maximal macroscopic conductance of Ca_v2.1-expressing cells (Fig. 1d, Additional file 1: Fig. S1, Additional file 2: Table S1), it reduced the hyperpolarizing shift of the voltage-dependence of activation produced by WT Ca_vβ₃ (Fig. 1f, Additional file 1: Fig. S2, Additional file 2: Table S1). Finally, we assessed the effect of Ca_vβ₃ surface charged residues on the voltage-dependence of inactivation of the channel. Co-expression of WT Ca_vβ₃ produced a 16.7 mV hyperpolarizing shift (p = 0.0001) of the mean-half inactivation potential of Ca_v2.1 (Fig. 1g and h, Additional file 1: Fig. S3, Additional file 2: Table S1). Although this effect was significantly altered upon mutation of Ca_vβ₃ surface charged residues, the magnitude of this alteration remained modest and all Ca_vβ₃ mutants retained their ability to significantly enhance the voltage-dependence of inactivation of the channel (Fig. 1h, Additional file 1: Fig. S3, Additional file 2: Table S1). Indeed, the weakest enhancement was observed with Ca_vβ₃ E339A and H348A which still produced a 9.1 mV 9.4 mV hyperpolarized shift, respectively, suggesting that Ca_vβ₃ surface charged residues have minimal influence on the voltage-dependence of inactivation of Ca_v2.1 channels. These data however allow us to conclude that for the Ca_vβ₃ mutations for which there is a reduced G_max (Fig. 1d), the channels under study remain in the Ca_v2.1 / Ca_vβ₃ complex form.

Fig. 1

Ca_vβ₃ surface charged residues contribute to the modulation of Ca_v2.1 channels. a Cartoon representation of secondary structural elements of the rat Ca_vβ₃ subunit in complex with the Ca_vβ₁ interacting domain (AID) (PDB 1VYT). b Position of surface charged residues within the Ca_vβ₃ subunit. Positively (H, histidine; R, arginine) and negatively (E, glutamic acid; D, aspartic acid) charged residues are shown in blue and red, respectively. c Mean current–voltage (I/V) relationship for Ca_v2.1 expressed alone (filled circles) and in combination with wild-type Ca_vβ₃ (open circles). d Corresponding mean maximal macroscopic conductance (G_max) obtained from the fit of the I/V curves with the modified Boltzmann function (1) for Ca_v2.1 alone and in combination with WT and mutant Ca_vβ₃ (ANOVA results: F = 26.6; p < 0.0001 and F = 10.15; p < 0.0001 for Ca_v2.1 expressed alone versus in the presence of Ca_vβ₃ variants and Ca_v2.1 expressed with Ca_vβ₃ wild-type versus with Ca_vβ₃ variants, respectively. e Mean normalized voltage-dependence of activation for Ca_v2.1 expressed alone (filled circles) and in combination with WT Ca_vβ₃ (open circles). f Corresponding mean half-activation potential values obtained from the fit of the activation curves with the modified Boltzmann function (1) for Ca_v2.1 alone and in combination with WT and mutant Ca_vβ₃ (ANOVA results: F = 34.29; p < 0.0001 and F = 9.965; p < 0.0001 for Ca_v2.1 expressed alone versus in the presence of Ca_vβ₃ variants and Ca_v2.1 expressed with Ca_vβ₃ wild-type versus with Ca_vβ₃ variants, respectively). g Mean normalized voltage-dependence of inactivation for Ca_v2.1 expressed alone (filled circles) and in combination with WT Ca_vβ₃ (open circles). h Corresponding mean half-inactivation potential values obtained from the fit of the activation curves with the two-state Boltzmann function (3) for Ca_v2.1 alone and in combination with WT and mutant Ca_vβ₃ (ANOVA results: F = 47.9; p < 0.0001 and F = 27.84; p < 0.0001 for Ca_v2.1 expressed alone versus in the presence of Ca_vβ₃ variants and Ca_v2.1 expressed with Ca_vβ₃ wild-type versus with Ca_vβ₃ variants, respectively. Statistical analysis (ANOVA followed by Dunnett’s post hoc multiple comparisons test) was performed for all Ca_vβ₃ variants either against Ca_v2.1 expressed alone (black statistical symbols) or Ca_v2.1 expressed with WT Ca_vβ₃ (red statistical symbols): * p < 0.05. The exact p values of the Dunnett’s post hoc analysis are provided in Additional file 3: Table S2

While AID-ABP interaction is a prerequisite for Ca_vβ-dependent modulation of HVA channels, additional interactions are expected to contribute to Ca_vβ modulatory properties [8, 9]. Here, we reported that Ca_vβ surface charged residues located outside of the ABP play a significant role in Ca_vβ₃-dependent modulation of Ca_v2.1 channels. In particular, residues D343, D344, and E347 appear to form a hot-spot at the surface of the GK domain to influence activation of the channel, with limited effect on its inactivation. It is of interest that this cluster of residues is in close proximity to the AID sequence itself (Fig. 1B). These data are consistent with previous studies showing that the effect of Ca_vβ on the voltage-dependence of Ca_v2.1 channel activation is largely reconstituted by the core region of Ca_vβ [11]. The question then arises as to how surface charged residues regulate channel gating. One possibility is via enabling additional low affinity interactions between Ca_vβ and other parts of Ca_vα₁. For instance, the amino- and carboxy-termini as well as the III–IV loop of Ca_vα₁ have been shown to interact directly with Ca_vβ [8, 9, 12, 13]. In addition, it was reported that the orientation of Ca_vβ relative to Ca_vβ₁ is essential for Ca_vβ-mediated regulation of the channel activation [14, 15]. Therefore, it is a possibility that surface charged residues, by supporting low affinity interactions, may contribute to the proper positioning of Ca_vβ. Inherent to our study are a number of limitations that will need to be addressed in future studies. First, in addition to Ca_vβ, Ca_v2.1 associated with Ca_vα₂δ that on the one hand mediates its own effects on the channel, and on the other hand influences the modulatory input of Ca_vα₂δ. For that reason, Ca_vα₂δ was purposely left out of our experiments to simplify the mechanistic analysis of Ca_vβ₃ variants. However, given the important role of Ca_vα₂δ in the modulation of Ca_v2.1, the present findings will need to be confirmed in the presence of Ca_vα₂δ where it can be expected that allosteric modulations will add another level of complexity to the regulation described in the present study. Second, in this study we used Ca_vβ₃ because it produces a more pronounced phenotype on Ca_v2.1 evidenced by faster inactivation kinetics compared for instance to Ca_vβ₄, and also because according to our experience the associated of Ca_vβ₃ with Ca_v2.1 in expression experiments is more complete than of Ca_vβ₄ which would have made the interpretation of the data more complicated. However, and although Ca_vβ₃ represents a legitimate subunit that does associate with Ca_v2.1 in native condition, Ca_vβ₄ remains the major isoform found co-associated with Ca_v2.1 in the brain and therefore it will be important to confirm our findings in the presence of Ca_vβ₄. And third, another potential limitation inherent to our experimental settings is the use of Xenopus oocytes where trace levels of endogenous Ca_vβ have been reported. While such an endogenous Ca_vβ is unlikely to have played a major role in the regulation of recombinant Ca_v2.1 since otherwise we would not have observed any effect of the co-expression of Ca_vβ₃, it would nevertheless be important to reproduce these findings in a mammalian cell line.