Functional Characterization of Four Known Cav2.1 Variants Associated with Neurodevelopmental Disorders

Cav2.1 channels are expressed throughout the brain and are the predominant Ca2+ channels in the Purkinje cells. These cerebellar neurons fire spontaneously, and Cav2.1 channels are involved in the regular pacemaking activity. The loss of precision of the firing pattern of Purkinje cells leads to ataxia, a disorder characterized by poor balance and difficulties in performing coordinated movements. In this study, we aimed at characterizing functional and structural consequences of four variations (p.A405T in I-II loop and p.R1359W, p.R1667W and p.S1799L in IIIS4, IVS4, and IVS6 helices, respectively) identified in patients exhibiting a wide spectrum of disorders including ataxia symptoms. Functional analysis using two major Cav2.1 splice variants (Cav2.1+e47 and Cav2.1−e47) in Xenopus laevis oocytes, revealed a lack of effect upon A405T substitution and a significant loss-of-function caused by R1359W, whereas R1667W and S1799L caused both channel gain-of-function and loss-of-function, in a splice variant-dependent manner. Structural analysis revealed the loss of interactions with S1, S2, and S3 helices upon R1359W and R1667W substitutions, but a lack of obvious structural changes with S1799L. Computational modeling suggests that biophysical changes induced by Cav2.1 pathogenic mutations might affect action potential frequency in Purkinje cells.


Introduction
By mediating spatially and temporarily regulated Ca 2+ entry, voltage-gated Ca 2+ channels participate in many biological processes such as muscle contraction, hormone or neurotransmitter release, and gene expression [1]. The precise Ca 2+ signaling is achieved thanks to the existence of three families (Cav1-3) of voltage-gated Ca 2+ channels with distinctive biophysical and pharmacological properties. While Cav3 Ca 2+ channels (three members Cav3.1 to 3.3) can form functional channels alone, typical Cav1 (four members Cav1.1 to 1.4) and Cav2 (three members Cav2.1 to 2.3) Ca 2+ channels include with the pore-forming Cavα subunit (Cav1.X or Cav2.X), one of the four Cavβ (Cavβ1 to 4) and one of the four Cavα2δ (Cavα2δ to 4) auxiliary subunits, which modulate properties, expression and regulation of Cavα [2]. The diversity of Ca 2+ channels resulting from the different Cavα/β/α2δ combinations of subunits can be further accentuated by the alternative splicing of Cavα and/or their auxiliary subunits. For example, alternative splicing produces Cav2.1 subunits with distinct cytoplasmic C-terminus, Cav2.1−e47 and Cav2.1+e47 [3]. The Cavα subunit has 24 transmembrane helices organized into four repeats (I to IV), each with 6 transmembrane helices (S1 to S6) forming two functional domains: the voltage sensing domain (VSD: S1-S4) and the pore domain (S5-S6) ( Figure 1). The S4 helices contain several positively charged residues, either R or K, at every third or fourth position (Figure 1), which are known as the gating charges. In fact, S4 helices constitute the actual voltage-sensor able to move across the electric field of the membrane potential [4]. The linker connecting the S5 and S6 helices forms a , and S1799L. The Cavα is formed of four repeats (I to IV) each containing 6 transmembrane helices (S1-S6). The S1-S4 helices constitute the voltagesensing domain (VSD) able to move in response to membrane potential changes. The four re-entrant P loops between segments S5 and S6 carry the selectivity filter and delineate the extracellular end of the channel pore. The S5 and S6 helices of the four repeats delineate the cytoplasmic end of the , and S1799L. The Cavα is formed of four repeats (I to IV) each containing 6 transmembrane helices (S1-S6). The S1-S4 helices constitute the voltage-sensing domain (VSD) able to move in response to membrane potential changes. The four re-entrant P loops between segments S5 and S6 carry the selectivity filter and delineate the extracellular end of the channel pore. The S5 and S6 helices of the four repeats delineate the cytoplasmic end of the channel pore. (Bottom) Amino-acid sequences of the I-II loop, and of the IIIS4, IVS4, and IVS6 helices. Cav2.1 variants referenced in the NCBI ClinVar database are indicated with their accession number. References in italics correspond to variants 'of uncertain significance' when subjected to in silico analysis. The 'R' above the sequences of the IIIS4 and IVS4 helices indicates the position of a basic residue. The Cav2.1 sequence is numbered according to the NCBI Genbank ® sequence AAB64179.
Next-generation sequencing approaches (whole-genome, whole-exome, or targeted genes sequencing) are an increasingly practical tool for the molecular diagnosis of neurological disorders [23,24]. ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/ (accessed on 28 October 2022)) includes hundreds of CACNA1A variants, among which 309 are categorized as pathogenic or likely pathogenic. Many disease-causing mutations seem to be preferentially located in the S4 and the S5-S6 helices of the different repeats [22,25]. In this study, we aimed to characterize the functional consequences of four CACNA1A mutations identified in patients exhibiting a wide spectrum of disorders. The first of these mutations, A405T, is located in the intracellular loop connecting repeat I and II ( Figure 1); R1359W and R1667W affect one of the gating charges in IIIS4 and IVS4 helices, respectively; and finally, S1799L is located in the midst of S6 helix in repeat IV. Such analyses are important, not only to understand the pathophysiological mechanism of the diseases but also to shed new light on Ca 2+ channel structure-function relationships. We found that A405T substitution does not impact Cav2.1 voltage-dependent properties, whereas R1359W causes a channel loss-of-function by stabilizing the inactivated state. On the other hand, R1667W and S1799L mutations cause both channel loss-of-function and gain-of-function, in a splice variant-dependent manner.

Electrophysiology
Expressed currents were recorded at room temperature using the two-electrode voltage clamp method on day 2-3 after injection, as previously described [27]. Electrodes were pulled from borosilicate glass and filled with 3M KCl. Currents were recorded with a Geneclamp 500 amplifier (Molecular Devices, San Jose, CA, USA) and digitized with a Digidata 1200 converter (Molecular Devices) using the Clampex software (Version 7.0, Molecular Devices). The external solution (10 mM BaOH, 10 mM TEAOH, 2 mM CsOH, 50 mM N-methyl-D-glucamine, 10 mM HEPES, pH 7.2 with methane sulfonic acid) was continuously perfused in the recording chamber at the rate of 1 mL/min. Currents were elicited by a two-pulse protocol from a holding potential of −100 mV. The first pulse lasted for 2.5 s and varied from −80 to +50 mV in 10 mV increments. The second pulse, separated from the first pulse by a 10 ms interval at −100 mV, lasted for 400 ms and was set at +10 mV. The isochronal inactivation curves were obtained by measuring the peak amplitude of the current evoked by the second pulse and by plotting the normalized current values (I/Imax) as a function of the voltage reached during the first pulse. They were fitted using the modified Boltzmann equation: where I is the current amplitude measured during the second pulse following a first depolarization to V, Imax is the current amplitude measured during the second pulse following a first pulse at −80 mV, Vi is the half-maximal inactivation potential, k is a slope factor and Rin is the portion of non-inactivating current.
Currents were also elicited by single 400 ms-long depolarizations varying from −60 mV to +40 mV, from a holding potential of −100 mV. Current-voltage curves were obtained by measuring the peak amplitude and by plotting the normalized current values (I/Imax) as a function of the voltage. They were fitted using the modified Boltzmann equation: where I is the current amplitude measured during the depolarization to V, Imax is the peak current amplitude measured at the maximum of the current-voltage curve, G is the normalized macroscopic conductance, Erev is the apparent reversal potential, Va is the potential for half activation, and k is a slope factor. The inactivation velocity (R400) was estimated as the ratio of the peak current that disappeared after 400 ms-long depolarizations to +10 mV or 0 mV.
Recovery from inactivation was determined using a protocol with two pulses at +10 mV for 2.5 s or 100 ms, respectively, from a holding potential of −100 mV, separated by an interval of time between 100 ms and 8 s. Traces were normalized to the maximum current measured during the first pulse. The percentage of current recovered during the second pulse was plotted as a function of the inter-pulse interval. Traces were fitted with a double exponential equation.

Expression in HEK Cells
For immunostaining experiments, the HEK293T [28] adherent cell line (#ACC 635, DSMZ, Braunschweig Germany) was used. Cells were cultivated in high-glucose Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 5% fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA), and Penicillin-Streptamycin, in a 5% CO 2 atmosphere at 37 • C. Cells were seeded at a concentration of 70.000 cells by cm 2 24 h before transient transfection. Appropriate volumes of plasmids (the total quantity of cDNA was 2 µg) and Lipofectamin ® reagent (Thermo Fisher Scientific, Waltham, MA, USA) were diluted in 100 µL of OptiMEM ® medium (Thermo Fisher Scientific, Waltham, MA, USA) and incubated for 20 min at room temperature. Complete medium exchange was carried out before transfection, and the Lipofectamin complex was added to medium. Cells were fixed with paraformaldehyde (3.7%) and kept in PBS at 4 • C for immunolabeling. Cells were treated with blocking solution (ADB, 3% BSA, 2% goat serum in PBST) for 2 h at room temperature. They were then incubated with anti-Cav2.1 (1:100, ACC-001, Alomone, Jerusalem, Israel), and anti-ZO-1 (1:100, Thermo Fisher Scientific, Waltham, MA, USA) antibodies at 4 • C overnight, washed twice with PBS 1X and then incubated with appropriate secondary antibodies for 1 h at room temperature. Cells were counterstained with 0.5 µg/mL Hoechst33342 in Vectashield ® antifade mounting medium (Vector Laboratories, Newark, CA, USA). Immunofluorescence was detected using an inverted Fluorescence Microscope (Montpellier Ressources Imagerie, Montpellier, France). Cav2.1 channel localization was assessed using ImageJ software. The outer and inner membrane perimeter was delineated using ZO-1 staining, and the fluorescence within the membrane and the cytoplasm were integrated as previously described [29].

Homology Modeling of Human Cav2.1
The model of the human channel Cav2.1 was built using the homology modeling module of BIOVIA Discovery Studio (Biovia Discovery Studio Modeling Environment, release 2021, Vélizy-Villacoublay, France). Briefly, based on the alignment of the human Cav2.1 sequence (NCBI GenBank ® : AAB64179.1) with proteins of the Protein Data Bank [30], the structure of the human N-type voltage-gated calcium channel Cav2.2 was selected as the template (PDB ID: 7MIX-A [6]) (amino acid identity: 81.4%, and similarity: 90.4%). Twenty homology models of Cav2.1 bearing four disulfide bridges (C256-C281, C272-C287, C1406-C1417, C1771-C1782) and one calcium ion in complex were generated for the sequence: K89-D410, F469-A793, and L1233-L1939. The model with the best PDF total energy score (−43.850) was selected and evaluated with QMEANBrane to be within the expected range for a transmembrane structure [31]. Then, four mono-mutated models, A405T, R1359W, R1667W, and S1799L were generated using the Side-Chain Refinement [32] module of Discovery Studio. Details of interactions for wild and mutated residues were analyzed with the Receptor-Ligand Interactions module of BIOVIA Discovery Studio.

Computational Modeling
Effects of the Cav2.1 mutations on Purkinje cell firing were predicted using the Purkinje cell model developed in the NEURON simulation environment (http://www.neuron.yale. edu (accessed on 28 October 2022)) by Luthman et al. [33]. Activities of ion channels were modeled using Hodgkin-Huxley equations as previously described [33]. Six cases were analyzed here: the wild-type situation (WT Cav2.1), a hyperpolarizing shift of Va (as obtained with R1359W and R1667W mutants), a depolarizing shift of Va (as obtained with S1799L mutant), an increase of ka (as obtained with R1359W, R1667W and S1799L mutants), or both changes of Va and ka.

Data Analysis
Results were expressed as the mean ± SEM. The statistical significance of differences between two groups was determined using the non-paired Student's t-test. Significance was set at 0.05.

Clinical Features Associated with A405T, R1359W, R1667W and S1799L Variants
The A405T variant (ClinVar68419) has been identified in three females from the same family with a spectrum of disorders ranging from episodic ataxia, progressive cerebellar ataxia, and familial hemiplegic migraine [34]. However, since two males carrying the same variation were healthy, it was hypothesized that additional genetic and/or environmental factors might play a role in the pathophysiology of the disease [34]. Computer-based analysis does not categorize A405T as pathogenic but rather as 'uncertain significance'. The Cav2.1 R1359W variant (ClinVar448996) was first identified in a 20-year-old female who exhibited developmental delay, intellectual disability, epileptic seizures, and cerebellar signs [35]. The same variation was also found in a 4-year-old girl who belonged to a large cohort of patients recruited for clinical exome sequencing [36]. In both cases, the patients showed progressive cerebellar atrophy. Moreover, the R1359W variant is suspected to cause Congenital Ataxia [22], and this variation is predicted to be pathogenic or likely pathogenic by in silico algorithms. Mutation of the equivalent residue (R1358G with mouse Cav2.1 numbering) in the mouse Cacna1a gene has been identified in the rolling Nagoya mutant mice (tg rol ). Homozygous tg rol /tg rol mice display a phenotype of severe gait ataxia and motor dysfunction [37]. Several variations of the Cav2.1 R1667 residue have been reported [38][39][40][41][42]. The R1667W (ClinVar68433), R1667Q (ClinVar585575), and R1667P (Clin-Var 638582) variations are predicted to be pathogenic whereas the R1667F (ClinVar1496838) variation is of uncertain significance. The R1667W was identified during the screening for CACNA1A mutations in patients suffering hemiplegic migraine with or without cerebellar signs [38], in a family in which the individuals displayed progressive cerebellar ataxia either pure or associated with episodic ataxia and hemiplegic migraine [39], and in a large cohort of 412 individuals recruited to identify genetic variants in patients with hereditary cerebellar ataxia [40]. On the other hand, the R1667P variation was found in two female children with congenital ataxia and fatal cerebral edema [41] or global developmental delay [42]. The R1799L variation has been identified in an 8-year-old boy exhibiting developmental delay and showing hypotonia and ataxia [35]. The R1799L variation (ClinVar420055) and another variation of the same residue, R1799P (ClinVar624124), are predicted to be pathogenic or likely pathogenic.

Functional Expression of Cav2.1 Variants
The mutations were introduced into human Cav2.1 with (+e47) or without (−e47) the coding sequence of exon 47. This exon encodes the Cav2.1 cytoplasmic C-terminus and includes the CAG repeat that encodes the polyglutamine expansion associated with SCA6 [43]. The alternative splicing at the exon 46-47 boundary produces a frame-shift and introduces a premature stop codon at the beginning of exon 47, and it has been shown that Cav2.1+e47 and Cav2.1−e47 splice variants are differently affected by pathogenic mutations [44,45]. The effects of A405T, R1359W, R1667W, and S1799L substitutions on Cav2.1 channel properties were investigated after the expression of the wild type and variants with Cavβ4a and Cavα2δ1 auxiliary subunits in X. laevis oocytes.
All the Cav2.1 variants (A405T, R1359W, R1667W, and S1799L) gave rise to functional channels irrespective of the splice variant background (+e47 or −e47), and families of inward Ba 2+ currents recorded with the Cav2.1+e47 variants are displayed on Figure 2A. Only few studies have reported a lack of expression with Cav2.1 variants carrying a missense mutation. The E1756K and F1490S mutations identified in families suffering from EA2 impair functional expression of Cav2.1 channels in heterologous expression systems [46,47]. The former mutation is located at the EEEE locus and the latter in the IIIS6 helix. More interestingly, the R174W mutation associated with Malignant Hyperthermia susceptibility ablates functional expression of Cav1.1 channels [48]. The mutation affects the innermost basic residue of the Cav1.1 IS4, similar to R1359W in the Cav2.1 IIS4. This suggests variable tolerance for the mutation between different Cavα subunits or between different S4 helices in a Cavα subunit. . Asterisks and number signs denote significant differences vs. WT (*** p < 0.001) and vs. Cav2.1+e47 splice variants (### p < 0.001), respectively (nonpaired Student's t-test). n.s. = non-significant.
Since it was virtually impossible to compare mean current amplitudes of ten different Ca 2+ channels in the same batch of oocytes and the same day after injection due to technical constraints (cDNA injection, inhibition of endogenous Cl − current with BAPTA), we evaluated the effects of the mutations on Cav2.1 membrane expression by immunostaining of HEK cells transfected with WT or mutant channels together with Cavβ4a and Cavα2δ1 subunits ( Figure 2B). We measured a weaker membranous expression for Cav2.1−e47 compared to Cav2.1+e47 ( Figure 2C), reminiscent of previous observations realized in Figure 2. (A) Representative current traces were recorded from X. laevis oocytes that expressed the indicated Cav2.1+e47 variants with Cavβ4a and Cavα2δ1. Currents were elicited from a holding potential of −100 mV by a two-pulse protocol illustrated on top, and consisting of a 2.5 s-long depolarization from −80 mV to +40 mV, followed by a 400 ms-long depolarization to 0 mV (WT and R1359W) or +10 mV (A405T, R1667W, and S1799L). Scale bars, 200nA. (B) Representative images of HEK cells immunostained with anti-Cav2.1 (red) and anti-ZO-1 (green). Anti-ZO-1 was used to visualize the cell membrane. Scale bars, 10 µm. (C) Quantification of fluorescence intensity realized on individualized cells (n = 10 for each variant). Asterisks and number signs denote significant differences vs. WT (*** p < 0.001) and vs. Cav2.1+e47 splice variants (### p < 0.001), respectively (non-paired Student's t-test). n.s. = non-significant.
Since it was virtually impossible to compare mean current amplitudes of ten different Ca 2+ channels in the same batch of oocytes and the same day after injection due to technical constraints (cDNA injection, inhibition of endogenous Cl − current with BAPTA), we evaluated the effects of the mutations on Cav2.1 membrane expression by immunostaining of HEK cells transfected with WT or mutant channels together with Cavβ4a and Cavα2δ1 subunits ( Figure 2B). We measured a weaker membranous expression for Cav2.1−e47 compared to Cav2.1+e47 ( Figure 2C), reminiscent of previous observations realized in HEK cells [45]. This decrease was also found for Cav2.1 splice variants containing the A405T substitution but not the R1359W, R1667W, or S1799L substitutions ( Figure 2C). Compared to WT channels, R1667W and S1799L induced a weaker membranous expression with the Cav2.1+e47 splice variant but not with the Cav2.1−e47 splice variant, and R1359W induced a weaker or a stronger expression in the Cav2.1+e47 or Cav2.1−e47 slice variant background, respectively. In most cases, Cav2.1 pathogenic mutations decrease current density (but see A713T in IIS6 [29] for counterexample), and the reduced influx of Ca 2+ during membrane depolarization may be involved in the pathophysiological mechanism of the disease. The decrease of current density can be linked to a weaker expression of channels at the plasma membrane (for example G230 in IS5 helix [29]), or to a biophysical effect altering channel conductance (for example Y1384C in IIIS6 helix [45]). However, it is noteworthy that current density can vary between heterologous expression systems and neuronal cells, and even between different types of neurons (R592Q in IS4 [49,50]). Moreover, we show here that Cav2.1 membranous expression, at least in HEK cells, is splice variant-dependent ( Figure 2C). It is important to keep these observations in mind to avoid over-interpretation. That said, our results show that the studied mutations, as many others, influence Cav2.1 channel expression at the cell membrane.
Given the presence of healthy carriers in the family in which the A405T mutation was found [34], the lack of noticeable effects on Cav2.1 properties is not unexpected. However, it questions the pathophysiological mechanism of the disorders found in the three individuals of the family. Interestingly, another and similar mutation in the Cav2.1 I-II loop, A454T (ClinVar68420, categorized as pathogenic, or likely pathogenic), has been found in two sisters with progressive ataxia [55], and more intriguingly, in a family of migraineurs (including FHM) in which it appeared that A454T substitution was not the cause of migraine but modified the disease phenotype [56]. In HEK cells, voltage-dependent activation (Va, ka), inactivation kinetics, and recovery from inactivation of Cav2.1 A454T variant were similar to the WT channel; however, Vi was shifted towards hyperpolarized or depolarized potentials depending on which Cavβ (Cavβ2 or Cavβ3) subunit was coexpressed with the mutant channel [56]. Moreover, A454T mutation decreased Cav2.1 modulation by Syntaxin1A or SNAP25, two members of the protein family mediating vesicle docking and fusion (SNAREs) which are known to interact with the intracellular linker connecting repeat II and III [57]. Even if phenotypes associated with A454T mutation differ from those associated with A405T mutation, a similar modification of the regulation by different Cavβ subunits (Cavβ1 and Cavβ2 instead of Cavβ4) or by SNARE could not be excluded, but these possibilities were not assessed in the present study.
Overall, the hyperpolarizing shift of Vi (Figure 4), the faster inactivation kinetics ( Figure 5), and the decreased rate of recovery from inactivation ( Figure 6) observed with R1359W seem to indicate a stabilization of the inactivated state. R1359W may be therefore defined as a loss-of-function mutation. On the other hand, R1667W causes both channel gain-of-function (hyperpolarizing shift of Va, Figure 3) and loss-of-function (faster inactivation kinetics, Figure 5) in a splice variant-dependent manner. Therefore, our results show that R1667W, like many other pathological mutations (see [42] and references included), cannot be strictly defined as a channel gain-of-function or loss-of-function.
To analyze the functional impact of the mutations on Cav2.1 channel structure, we generate a homology model based on the atomic resolution structure of human Cav2.2 displaying a closed pore, and the IS4, IIIS4, and IVS4 helices in the depolarized state (S4s are 'up'), and the IIS4 helix exhibiting a 'down' conformation [6]. R1359 and R1667 are perfectly conserved in all voltage-gated Ca 2+ channels [22], and R1359 is the IIIS4 innermost gating charge (R5) while R1667 is in the midst of IVS4 (R3) ( Figure 7A). R1359 residue is attracted towards S2 and S3 helices thanks to electrostatic interactions with E1291 (IIIS2) and D1317 (IIIS3), and a repulsive interaction with the neighboring K1358 (IIIS4) ( Figure 7C). The salt bridges are lost upon replacement of the arginine residue with the neutral tryptophane ( Figure 7D). Indeed, the novel W1359 residue lies between hydrophobic interactions with I1250 and L1251 (IIIS1), and a potential pi-cation interaction with K1358. These new interactions impact probably the mobility of the IIIS4, which may explain the stabilization of the inactivated state. As noted previously [42], R1667 is positioned just extracellular to F1609 (IVS2), which constitutes the VSDII hydrophobic constriction site that prevents transmembrane movement of water and ions [58]. R1667 makes hydrogen bonds with S1641 (IVS3) and N1579 (IVS1), potentially reinforced by repulsive interaction with K1670 (IVS4) ( Figure 7E). The interactions are lost upon R1667 substitution, and the novel W1667 residue is surrounded by highly stabilizing hydrophobic (K1670) and pi-cation (F1609) interactions ( Figure 7F). Pathogenic mutations affecting the gating charges can be found in the four Cav2.1 S4 helices (see Figure 1 for IIIS4 and IVS4 helices). One of the first mutations associated with FHM1 was an R192Q substitution of the outermost (R1) basic residue of the IS4 helix [59]. The functional consequences of this specific Cav2.1 mutation have been the subject of several studies, using heterologous expression system [44,[60][61][62], transfected hippocampal [63] or trigeminal ganglion [64] neurons from Cav2.1 knockout mice, and moreover, an Pathogenic mutations affecting the gating charges can be found in the four Cav2.1 S4 helices (see Figure 1 for IIIS4 and IVS4 helices). One of the first mutations associated with FHM1 was an R192Q substitution of the outermost (R1) basic residue of the IS4 helix [59]. The functional consequences of this specific Cav2.1 mutation have been the subject of several studies, using heterologous expression system [44,[60][61][62], transfected hippocampal [63] or trigeminal ganglion [64] neurons from Cav2.1 knockout mice, and moreover, an R192Q knockin mouse strain has been generated [50]. In X. laevis oocytes, the R192Q mutation did not change the Cav2.1 voltage dependence of activation and inactivation [60], and the Cav2.1 voltage dependence of activation is affected by R192Q mutation only in a subpopulation of trigeminal ganglion neurons [50]. However, in HEK cells, the R192Q substitution in Cav2.1 produced a hyperpolarizing shift of both Va and Vi and did not modify ka and ki [44]. Similar effects were produced by other substitutions: R582Q (R2 in IIS4), associated with FHM1 [65], and R1349Q (R2 in IIIS4, Figure 1) found in the tg-5J ataxic mouse strain [66]. In our study, R1359W (IIIS4) displaced also Va and Vi towards hyperpolarized voltages, but in a splice variant-dependent manner, and increased both ka and ki (Figures 3 and 4). In strike contrast, mutations of basic residues R3 (R1667P [42] and R1667W (this study)) or R5 (R1679P [67]) in the IVS4 helix, did not modify voltage dependence of inactivation (Vi and ki), although they produced an increased in ka and either a hyperpolarizing (R3 mutations) or a depolarizing (R5 mutation) shift of Va. Some discrepancies found between studies arise undoubtedly from the different Cav2.1 variants and/or the different heterologous expression systems, but they also suggest that equivalent substitutions in different S4 helices or different mutations of the same residue induce specific alterations of the channel gating properties. Functional differences between the four S4 helices of the Cavα subunit are suggested by the specific length of the helix and/or the various number of charged residues in each S4 (see Figure 1 for IIIS4 and IVS4) and are demonstrated by accumulating evidence obtained with Cav1 channels [68]. Neutralization of S4 charges revealed the critical role played by IS4 and IIIS4 helices in Cav1.2 channel activation whereas IIS4 and IVS4 helices have less or no impact [69]. Charged residues in helix IS4 are also critical for the voltage dependence of Cav1.2 inactivation, whereas neutralization of charges in IIS4 and IIIS4 has less effect [70] (see also [4]). Using voltage-clamp fluorometry, which allows the tracking of individual S4 movement, Pantazis et al. [71] found that the four Cav1.2 S4 helices exhibit distinct voltagedependent properties and that IIS4 and IIIS4 helices are critical for channel activation. Using the same approach, Savalli et al. [72] highlighted the prominent role played by IS4 in Cav1.1 activation. Similar studies on Cav2 channels are lacking, and it is highly speculative to think that the biophysical mechanisms leading to channel activation are strictly the same for Cav1 and Cav2 channels. The charged residues in S4 helices of Cav1 and Cav2 channels are not perfectly conserved, and the cryo-electron microscopy of Cav1.1 [73,74] and Cav2.2 [6,7] has highlighted fundamental differences between the VSDs of the two channels. Nevertheless, studies with pathogenic variants (including ours) suggest that S4 helices of the four repeats are involved in the voltage dependence of Cav2.1 activation, whereas only the IS4, IIS4, and IIIS4 helices are involved in the voltage dependence of channel inactivation.
In striking contrast to VSDs, the intracellular gate formed by S6 helices is nearly identical in Cav1.1 and Cav2.2 structures [6,7,73,74], and insights on Cav2 structure-function relationships might come from pathological mutations found on Cav1 channels. The Cav1.4 I754T variant is associated with a severe visual disorder and induces a remarkable hyperpolarizing shift of Va and a significant decrease in inactivation kinetics [75]. The residue, located at the cytoplasmic end of IIS6, is perfectly conserved in all Cavα subunits (I711 in Cav2.1), and, when introduced to Cav1.2, the substitution produces the same effects as in Cav1.4 [76]. On the other hand, Timothy syndrome is a multi-systemic disorder with dysmorphic features, long-QT syndrome, and autism spectrum disorders caused by a mutation in Cav1.2 channels [77]. The first mutation associated with Timothy syndrome, G406R in IS6, produces a hyperpolarizing shift of half-maximal activation potential and a dramatic decrease in inactivation kinetics. Again, this residue is perfectly conserved in all Cavα subunits (G361 in Cav2.1), and when introduced in Cav2.3 [78] or Cav2.1 [27], the effects produced by the mutation are similar to those produced in Cav1.2. These two examples suggest that the pore of both Cav1 and Cav2 share common steps during channel activation and inactivation. Until now, it has been difficult to have functional insights into channel gating from studies focused on Cav2.1 pathogenic mutations located in S5 and S6 helices. Indeed, several Cav2.1 variants, associated with FMH1 (V714A [29,60,63,79] and D715E in IIS6 [65], V1695I in IVS5 [80], and I1810L in IVS6 [60,61,81]), with Epileptic Encephalopathy (G230V in IS5, A713T in IIS6 and V1396M in IIIS5 [29]), with episodic ataxia and epilepsy (L621R in IIS5 [82]), and with hemiplegic migraine and developmental delay (Y1384C in IIIS5 [45]) are located in the pore domain and have been studied in heterologous expression systems or transfected neurons. Given the diversity of phenotypes, residue location in helix, and their substitution, the functional consequences of these mutations on the voltage dependence of activation and inactivation are very different. Moreover, the decrease of Ca 2+ influx caused by the weaker expression of functional channels at the plasma membrane and/or by the reduced conductance of the mutated pore might cause the pathophysiology without any modification of voltage-dependent properties. Nevertheless, these disease-causing mutations in the pore domain uncover the structural rearrangements undergone by S5 and S6 helices upon channel activation and inactivation.

Computational Simulation
Cav2.1 channels are the predominant Ca 2+ channels expressed in Purkinje cells, where they play a fundamental role in cellular excitability [83,84]. Purkinje cells fire spontaneously, and their activity is required to perform planned and coordinated movements. Consequently, malfunction of Purkinje cells often results in motor dysfunctions [85]. To seek the effects of Cav2.1 mutations on neuronal electrical activity, we used a computer model of the human Purkinje cell where all the ionic conductances were modeled with the Hodgkin and Huxley formalism [33]. Six simulations were run with: the wild-type electrophysiological conductances (run 1), a 3.5 mV hyperpolarizing shift of Va (as found with the R1359W variant, run 2), an 8.9 mV depolarizing shift of Va (as found with the S1799L variant, run 3), a 2 mV increase of ka (as found with the R1359W variant, run 4), and both a hyperpolarizing shift of Va and increase of ka (run 5) or a depolarizing shift of Va and increase of ka (run 6) ( Figure 8A). We did not observe any modification of action potential frequency and area upon decreasing the steepness of ka ( Figure 8B,C). However, the hyperpolarizing shift of Va decreased the frequency and increased the area, while the depolarizing shift produced the opposite effect ( Figure 8B,C). Area changes resulted from a modification of the rates of the rising or decaying phases of action potential (not shown). A modification of frequency may be explained by the variation of the Ca 2+ influx caused by the Va shift, and consequently an altered amplitude of the KCa was induced after hyperpolarization [85]. The additional modification of ka tends to increase the effects produced by Va changes (Figure 8B,C). These results suggest that the biophysical changes produced by Cav2.1 pathogenic mutations, such as those studied herein, can dramatically alter the firing pattern of neuronal cells.
Membranes 2023, 13, x FOR PEER REVIEW 19 of 24 variant, run 3), a 2 mV increase of ka (as found with the R1359W variant, run 4), and both a hyperpolarizing shift of Va and increase of ka (run 5) or a depolarizing shift of Va and increase of ka (run 6) ( Figure 8A). We did not observe any modification of action potential frequency and area upon decreasing the steepness of ka ( Figure 8B,C). However, the hyperpolarizing shift of Va decreased the frequency and increased the area, while the depolarizing shift produced the opposite effect ( Figure 8B,C). Area changes resulted from a modification of the rates of the rising or decaying phases of action potential (not shown). A modification of frequency may be explained by the variation of the Ca 2+ influx caused by the Va shift, and consequently an altered amplitude of the KCa was induced after hyperpolarization [85]. The additional modification of ka tends to increase the effects produced by Va changes (Figure 8B,C). These results suggest that the biophysical changes produced by Cav2.1 pathogenic mutations, such as those studied herein, can dramatically alter the firing pattern of neuronal cells.

Conclusions
Next-generation sequencing has allowed the identification of hundreds of variations in CACNA1A. Their impact on Cav2.1 biophysical properties has been studied for only a few of them, occasionally highlighting discrepancies in the phenotype/genotype correlation. Nevertheless, such studies are necessary to elucidate the pathophysiological mechanism of the diseases associated with Cav2.1 mutations. Moreover, they shed new light on structure/function relationships and may help to understand the molecular rearrangements that lead to channel activation and inactivation. Of the four Cav2.1 mutations studied herein, which were associated with distinct phenotypes, one did not modify the biophysical properties of the channel (A405T), one stabilizes the inactivated state (R1359W), and the last two produce both the channel lost-of-function and gain-of-function in a splice variantdependent manner (R1667W and S1799L). Simulations suggest that altered biophysical properties may affect the firing of Purkinje cells. Apart from the Cav2.1 Ca 2+ channel, K + channels also play a crucial role in pacemaking activity [15,86], and it has been shown that the K + channel blocker 4-AP was able to restore the precision of Purkinje cell pacemaking in the tottering mice [87] and in a mouse model of SCA6 [88]. Moreover, the Ca 2+ -dependent K + channel activators EBIO and chlorzoxazone can rescue the firing regularity of Purkinje cells in leaner, ducky, and tottering mice [83,87]. We may therefore expect that future therapies targeting Cav2.1 channels would be as effective for improving motor functions in neurodevelopmental diseases [89].  Institutional Review Board Statement: The care and use of X. laevis conformed to institutional policies and guidelines. The experimental protocols were approved by the "Direction Départementale des Services Vétérinaires" (authorization C34.16).

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.