Autism-Linked Mutations in α2δ-1 and α2δ-3 Reduce Protein Membrane Expression but Affect Neither Calcium Channels nor Trans-Synaptic Signaling

Abstract

Background: α₂δ proteins regulate membrane trafficking and biophysical properties of voltage-gated calcium channels. Moreover, they modulate axonal wiring, synapse formation, and trans-synaptic signaling. Several rare missense variants in CACNA2D1 (coding for α₂δ-1) and CACNA2D3 (coding for α₂δ-3) genes were identified in patients with autism spectrum disorder (ASD). However, the pathogenicity of these variants is not known, and the molecular mechanism by which α₂δ proteins may contribute to the pathophysiology of autism is, as of today, not understood. Therefore, in this study we functionally characterized two heterozygous missense variants in α₂δ-1 (p.R351T) and α₂δ-3 (p.A275T), previously identified in patients with ASD.

Methods: Electrophysiological recordings in transfected tsA201 cells were used to study specific channel-dependent functions of mutated α₂δ proteins. Membrane expression, presynaptic targeting, and trans-synaptic signaling of mutated α₂δ proteins were studied upon expression in murine cultured hippocampal neurons.

Results: Homologous expression of both mutated α₂δ proteins revealed a strongly reduced membrane expression and synaptic localization compared to the corresponding wild type α₂δ proteins. Moreover, the A275T mutation in α₂δ-3 resulted in an altered glycosylation pattern upon heterologous expression. However, neither of the mutations compromised the biophysical properties of postsynaptic L-type (Ca_V1.2 and Ca_V1.3) and presynaptic P/Q-type (Ca_V2.1) channels when co-expressed in tsA201 cells. Furthermore, presynaptic expression of p.R351T in the α₂δ-1 splice variant lacking exon 23 did not affect trans-synaptic signaling to postsynaptic GABA_A receptors.

Conclusions: Our data provide evidence that the pathophysiological mechanisms of ASD-causing mutations of α₂δ proteins may not involve their classical channel-dependent and trans-synaptic functions. Alternatively, these mutations may induce subtle changes in synapse formation or neuronal network function, highlighting the need for future α₂δ protein-linked disease models.

Keywords: autism spectrum disorder; auxiliary subunit; calcium current; cultured hippocampal neurons; electrophysiology; trans-synaptic function; voltage-gated calcium channels.

Figure 1

The highly conserved R351 (α₂δ-1) and A275 (α₂δ-3) residues are predicted to be important for the integrity of the mature protein. Amino acid sequence alignments of (A) α₂δ-1 and (B) α₂δ-3 between different species. Positions corresponding to arginine 351 (R351) and alanine 275 (A275) in human α₂δ-1 and α₂δ-3, respectively, are highlighted in bold red. (C) Schematic overview of α₂δ protein illustrating the positions of the R351 and A275 residues within the VWA domain. (D) Cryo-EM structure of human α₂δ-1 protein (color code as in C) in complex with the human Ca_V2.1 channel (gray, PDB code: 7MIY). On the left panel, the enlarged image shows the ionic interactions (dashed lines) of R351 with tyrosine 347 (Y347) and leucine 344 (L344) (upper image), which stabilize the loop. Substituting R351 with threonine prevents the formation of these stabilizing interactions (lower image). On the right panel, the enlarged image of the VWA domain of α₂δ-3 suggests that A275 participates in a hydrophobic pocket with surrounding residues in the VWA domain (gray, surface representation) to stabilize the fold of the VWA domain (upper image). Substituting A275 with threonine, an amino acid with a polar side chain, may disrupt the hydrophobic pocket (lower image). (Color code: α₂ peptide in purple, VWA domain in orange, δ peptide in pink, and α₁ subunit in gray). AlphaFold per-residue model confidence score (pLDDT) for A275 is very high with a value of 94.1.

Figure 2

Strongly reduced membrane expression of α₂δ-1_R351T and α₂δ-3_A275T in differentiated cultured hippocampal neurons. (A,C) Representative images of primary cultured hippocampal neurons (DIV 21) transfected with soluble eGFP together with either HA-tagged WT (2HA-α₂δ-1 or 2HA-α₂δ-3) or mutated (2HA-α₂δ-1_R351T or 2HA-α₂δ-3_A275T) α₂δ. Anti-HA live-cell labelling demonstrates a reduced staining intensity of both α₂δ-1_R351T and α₂δ-3_A275T in the soma, dendrites, and axons (indicated by arrows) compared to the respective WT α₂δ. (B,D) Quantification of the average HA fluorescent intensities in the different compartments shows a reduced surface expression of mutated α₂δ compared to WT α₂δ. Graphs show values for individual cells (dots) and means ± SEM (lines). All values were normalized to the mean fluorescence intensity of the WT 2HA-α₂δ within each culture preparation. Statistics: (B) Data were obtained from three independent culture preparations; 45 and 57 cells expressing WT or mutated HA-tagged α₂δ-1 were analyzed, respectively. Unpaired t test, soma: t₍₁₀₀₎ = 4.5; *** p < 0.0001, dendrite: t₍₁₀₀₎ = 6.5; *** p < 0.0001, axon: t₍₁₀₀₎ = 6.3; *** p < 0.0001. (D) Data were obtained from four independent culture preparations; 52 and 60 cells expressing WT or mutated HA-tagged α₂δ-3 were analyzed, respectively. Unpaired t test, soma: t₍₁₁₀₎ = 3.3; ** p = 0.0013, dendrite: t₍₁₁₀₎ = 2.9; ** p = 0.0035, axon: t₍₁₁₀₎ = 4.6; *** p < 0.0001. Scale bars, 10 µm.

Figure 3

Comparable overall expression levels of mutated and corresponding WT α₂δ proteins. Primary cultured hippocampal neurons were transfected with soluble eGFP together with either HA-tagged WT or mutated α₂δ. (A,C) Representative images of permeabilized anti-HA immunostaining, eGFP fluorescence, and anti-HA/eGFP overlay. Axons are indicated by arrows. (B,D) Analysis of whole-cell fluorescence intensities. Values for individual cells (dots) and means ± SEM (lines). All values were normalized to the mean fluorescence intensity of the WT 2HA-α₂δ within each culture preparation. Statistics: (B) Data were obtained from three independent culture preparations; 28 cells expressing WT or mutated HA-tagged α₂δ-1 were analyzed. Unpaired t test, t₍₅₄₎ = 1.38; p = 0.17. (D) Data were obtained from two independent culture preparations; 17 and 16 cells expressing WT or mutated HA-tagged α₂δ-3 were analyzed, respectively. Unpaired t test, t₍₃₁₎ = 0.7; p = 0.49. Scale bars, 20 µm.

Figure 4

Strongly reduced membrane expression of α₂δ-1_R351T in tsA201 cells. tsA201 cells were co-transfected with soluble eGFP together with either WT or mutated α₂δ-1 (R351T), both of which are tagged with double HA tag at the N-terminus. (A) Representative images of anti-HA live-cell-labelled tsA201 cells. The membrane localization of 2HA-α₂δ-1 (WT) is characterized by a smooth, fine-dotted pattern on the surface of tsA201 cells. In contrast, the labelling of 2HA-α₂δ-1_R351T (R351T) exhibits a sparsely dotted localization pattern, accompanied by a reduced overall fluorescence intensity. (B) Quantification of the membrane expression of WT and mutated α₂δ-1. Log-transformed anti-HA live-cell staining intensities (arbitrary units) are shown for individual cells (dots) and means ± SEM (lines). Data were obtained from three independent experiments, and 68 and 64 cells transfected with WT or mutated HA-tagged α₂δ-1 were analyzed, respectively. (C) Immunoblot of whole-cell lysates obtained from tsA201 cells transfected with 2HA-α₂δ-1 (left lane) or 2HA-α₂δ-1_R351T (right lane). α₂δ-1 was detected with an anti-HA antibody (upper panel), anti-α-tubulin labelling was used as a loading control (middle panel), and anti-GFP labelling was used for comparing the transfection efficiency (lower panel). (D) Quantification of protein expression levels of α₂δ relative to GFP expression and normalized to WT expression. Statistics: (B) unpaired two-tailed t-test, t₍₁₃₀₎ = 5.5; *** p < 0.0001; (D) unpaired two-tailed t-test, t₍₄₎ = 0.05; p = 0.96. Scale bars, 10 μm.

Figure 5

The p.A275T mutation does not affect the overall expression levels of α₂δ-3 but alters the glycosylation pattern. (A) Schematic representation of “pro” form of α₂δ (left) showing the approximate position of the signal peptide (SP), double HA tags (2HA), and VWA domain. Post-translational processing α₂δ (mature form, right) includes glycosylation (Y), glycosyl-phosphatidyl inositol (GPI)-anchoring, proteolytic cleavage, and formation of multiple disulfide bonds between and within α₂ and δ peptides. (B,C) Immunoblot of whole-cell lysates obtained from tsA201 cells transfected with 2HA-tagged WT or mutated α₂δ-3 together with eGFP. α₂δ-3 protein was detected with an anti-HA antibody (upper panel), anti-tubulin labelling was used as a loading control (middle panel), and anti-GFP for controlling transfection efficiency (lower panel). (B) Glycosylated and reduced α₂δ-3 (lanes 1 and 2), glycosylated and unreduced α₂δ-3 (lanes 3 and 4). (C) Glycosylated and reduced α₂δ-3 (lanes 1 and 2), de-glycosylated and reduced α₂δ-3 (lanes 3 and 4). Proteins were de-glycosylated with Peptide N-glycosidase (PNGase-F) and reduced with dithiothreitol (DTT). (D) Quantification of protein expression levels of α₂δ-3 relative to GFP expression and normalized to WT expression. (E) The ratio of the intensities of the upper and lower anti-HA-immunoreactive bands under reducing conditions. Data were obtained from three independent transfections. Values of individual transfections (dots) and mean bars ± SEM (lines) are shown. Statistics: (D) unpaired two-tailed t-test, t₍₄₎ = 0.69; p = 0.53; (E) unpaired two-tailed t-test, t₍₄₎ = 5.4; p = 0.006 (**).

Figure 6

Mutated α₂δ proteins increase current densities of Ca_V2.1 channels similar to WT α₂δ. (A–D) Calcium current properties of Ca_V2.1 recorded from tsA201 cells co-transfected with Ca_V2.1 and β₄ alone as a control (control, gray triangles) or together with WT (WT, blue circles), or mutated α₂δ-1 (R351T, red rectangles), and (E–H) from tsA201 cells transfected with Ca_V2.1 and β₄ alone as a control (control, gray triangles) or together with WT (WT, yellow circles), or mutated α₂δ-3 (A275T, orange rectangles). 50 msec test pulses from a holding potential of −80 mV to +80 mV were applied in 5 mV increments. (A,E) Representative whole-cell Ca²⁺ current traces obtained at V_MAX. Current–voltage relationships (B,F), peak current densities (C,G), and half-maximal activation potentials (D,H) are shown. Statistics: (C,D) One-way ANOVA with Tukey’s post hoc multiple comparison was performed on 10–33 recordings per condition obtained from three independent experiments. (C) Maximal current density, F_{(2, 7)} = 10.7; p < 0.0001, (D) half-maximal activation potential, F_{(2, 67)} = 0.25; p = 0.78. (G,H) ANOVA with Tukey’s post hoc multiple comparison was performed on 6–19 recordings per condition obtained from three independent experiments. (G) Maximal current density, F_(2,39) = 6.0; p = 0.006, (H) half-maximal activation potential, F_{(2, 39)} = 2.6; p = 0.09. Significances of post hoc tests between conditions are indicated in the graphs by asterisks (*** p < 0.001, ** p < 0.01, * p < 0.05).

Figure 7

Mutated α₂δ proteins modulate current properties of Ca_V1.3 channels similar to WT α₂δ. (A–D) Calcium current properties of Ca_V1.3 channels recorded from tsA201 cells transfected with Ca_V1.3 and β₃ alone as control (control, gray triangles) or together with WT (WT, blue circles), or mutated α₂δ-1 (R351T, red rectangles), and (E–H) recordings from tsA201 cells transfected with Ca_V1.3 and β₃ alone as control (control, gray triangles) or together with WT (WT, yellow circles), or mutated α₂δ-3 (A275T, orange rectangles). 50 msec test pulses from a holding potential of −80 mV to +90 mV were applied in 5 mV increments. (A,E) Representative whole-cell Ca²⁺ current traces obtained at V_MAX. Current–voltage relationships (B,F), peak current densities (C,G), and half-maximal activation potentials (D,H) are shown. Statistics: (C,D) One-way ANOVA with Tukey’s post hoc multiple comparison was performed on 10–23 recordings per condition obtained from three independent experiments. (C) Maximal current density, F_{(2, 51)} = 6.6; p = 0.0028, (D) half-maximal activation potential, F_{(2, 51)} =20.6; p < 0.0001. (G,H) ANOVA with Tukey’s post hoc multiple comparison was performed on 6–16 recordings per condition obtained from two independent experiments. (G) Maximal current density, F_{(2, 34)}= 4.5; p = 0.019, (H) half-maximal activation potential, F_{(2, 34)} =34.9; p < 0.0001. Significances of post hoc tests between conditions are indicated in the graphs by asterisks (*** p < 0.001, ** p < 0.01, * p < 0.05).

Figure 8

α₂δ-1_R351T modulates current properties of Ca_V1.2 channels like WT α₂δ-1. (A–D) Calcium current properties of Ca_V1.2 channels recorded from tsA201 cells transfected with Ca_V1.2 and β₁ alone as control (control, gray triangles) or together with WT (WT, blue circles), or mutated α₂δ-1 (R351T, red rectangles). 50 msec test pulses from a holding potential of −80 mV to +90 mV were applied in 5 mV increments. (A) Representative whole-cell Ca²⁺ current traces obtained at V_MAX. Current–voltage relationships (B), peak current densities (C), and half-maximal activation potentials (D) are shown. Statistics: One-way ANOVA with Tukey’s post hoc multiple comparison was performed on 18–24 recordings per condition obtained from three independent experiments. (C) Maximal current density, F_{(2, 61)} = 14.9; p < 0.0001, (D) half-maximal activation potential, F_{(2, 61)} =34.7; p < 0.0001. Significances of post hoc tests between conditions are indicated in the graphs by asterisks (*** p < 0.001).

Figure 9

Strongly reduced presynaptic targeting of α₂δ-1_R351T and α₂δ-3_A275T. (A,C) Representative presynaptic boutons of cultured hippocampal neurons transfected with eGFP together with WT or mutated α₂δ. Presynaptic boutons are identified by the clustering of synapsin proteins (blue) within axonal varicosities as visualized by the eGFP fluorescence (outlined by a dashed line). Live-cell staining reveals that mutated α₂δ-1 and α₂δ-3 proteins show a lower expression at the surface of presynaptic boutons compared to WT α₂δ proteins (red). (B,D) Average fluorescence intensity measurements of the HA signal in transfected boutons revealed a strong reduction in presynaptic membrane expression of mutated α₂δ compared to WT α₂δ proteins. Graph shows mean values of minimum five synapses of individual cells (dots) and means ± SEM (lines). Data were obtained from at least three independent culture preparations. Statistics: (B) 29 and 35 cells expressing WT or mutated HA-tagged α₂δ-1 were analyzed, respectively. Unpaired t-test, t₍₆₂₎ = 3.5; *** p = 0.0008. (D) 43 and 44 cells expressing WT or mutated HA-tagged α₂δ-3 were analyzed, respectively. Unpaired t-test, t₍₈₅₎ = 2.6; * p = 0.01. Scale bars, 1 µm.

Figure 10

Unaltered trans-synaptic coupling of α₂δ-1_R351T_ΔE23 to postsynaptic GABA_AR. Synapses in hippocampal neurons transfected with soluble eGFP as control (green triangles), or together with either WT (blue circles) or mutated (red rectangles) α₂δ-1_ΔE23, were identified by immunofluorescent labelling of presynaptic vGLUT1 and postsynaptic GABA_A-receptors. (A) Both overexpression of WT and mutated α₂δ-1_ΔE23 leads to the formation of mismatched synapses as detected by postsynaptic GABA_A receptor clusters opposite vGLUT1 positive glutamatergic terminals (A, α₂δ-1, α₂δ-1_R351T). Quantifications of immunofluorescence intensities of GABA_A receptor (B) and vGLUT1 (C) labelling show values for individual cells (dots) and means ± SEM (lines). Cells were obtained from three independent culture preparations. Values were normalized to the fluorescent intensities of the control condition within each culture preparation. Statistics: ANOVA with Tukey’s post hoc multiple comparison was performed on 41–52 cells per condition. GABA_AR: F_{(2, 136)} = 29.5; p < 0.0001, vGLUT1: F_{(2, 136)} = 0.4; p = 0.67. Significances of post hoc tests between the control and the α₂δ-1_ΔE23 conditions are indicated in the graphs by asterisks (*** p < 0.001). Scale bars, 1 µm.