Proteolytic maturation of α2δ represents a checkpoint for activation and neuronal trafficking of latent calcium channels

Abstract

The auxiliary α₂δ subunits of voltage-gated calcium channels are extracellular membrane-associated proteins, which are post-translationally cleaved into disulfide-linked polypeptides α₂ and δ. We now show, using α₂δ constructs containing artificial cleavage sites, that this processing is an essential step permitting voltage-dependent activation of plasma membrane N-type (Ca_V2.2) calcium channels. Indeed, uncleaved α2δ inhibits native calcium currents in mammalian neurons. By inducing acute cell-surface proteolytic cleavage of α₂δ, voltage-dependent activation of channels is promoted, independent from the trafficking role of α₂δ. Uncleaved α₂δ does not support trafficking of Ca_V2.2 channel complexes into neuronal processes, and inhibits Ca²⁺ entry into synaptic boutons, and we can reverse this by controlled intracellular proteolytic cleavage. We propose a model whereby uncleaved α₂δ subunits maintain immature calcium channels in an inhibited state. Proteolytic processing of α₂δ then permits voltage-dependent activation of the channels, acting as a checkpoint allowing trafficking only of mature calcium channel complexes into neuronal processes.

Keywords: Ca2+ channels; biochemistry; electrophysiology; mouse; neuron; neuroscience; rat.

Figure 1.. Effect of mutation of α₂δ-1 to disrupt the proteolytic cleavage site.

(a) Cartoon of uncleaved pro-α₂δ-1 and cleaved α₂δ-1, showing the approximate position of inserted HA tag and disulfide bonds between α₂ (grey) and δ (blue). (b) Rat α₂δ-1 sequence at the identified cleavage-site. The underlined sequence (including LEA//VEM) in α₂δ-1 was mutated to a V6 or 3C-protease motif. (c) Comparison of glycosylated (lanes 1, 2) and deglycosylated α₂δ-1 (lanes 3, 4), expressed in tsA-201 cells (lanes 1, 3) or present in the brain (lanes 2, 4), showing resolution between pro-α₂δ-1 and the cleaved form of α₂δ after deglycosylation. α₂δ* indicates glycosylated species, and pro-α₂δ and α₂ indicate deglycosylated proteins. Uncleaved pro-α₂δ-1 is observed in transfected cells, but not brain. Proteins visualized with α₂-1 Ab. (d) α₂δ-1-HA (left) and α₂(V6)δ-1-HA (right) expressed in tsA-201 cells; proteins deglycosylated with PNGase-F. Upper panel: HA-blots, lower panel: endogenous GAPDH loading control. (e) Normalized binding curves, using DRM fractions from transfected tsA-201 cells, for ³H-gabapentin binding to WT α₂δ-1 (■, n = 4) and α₂(V6)δ-1 (○, n = 4). Mean (± SEM) data are fit by hyperbolae with K_D of 82.3 and 68.3 nM, respectively. (f) Imunoblot analysis of deglycosylated α₂δ-1-HA and α₂(3C)δ-1-HA. WCL input (lanes 1, 2) and cell-surface biotinylated material (lanes 3, 4) are shown. Upper panel: HA-blot, lower panel: endogenous GAPDH. (g) Mean ± SEM (and individual data points) of α₂(3C)δ-1 (blue) cell surface levels, measured as a proportion of biotinylated: total protein, normalized to control (WT α₂δ-1; red), for 4 experiments including that shown in (f). p=0.5057, one sample t test. (h) Cell-surface expression of α₂δ-1-HA (left) and α₂(3C)δ-1-HA (right) in non-permeabilized tsA-201 cells, using HA Ab (red). Nuclei visualized with DAPI (blue). Scale bars 20 μm. (i) Box and whisker plots for quantification of α₂δ-1 on cell-surface, from HA fluorescence, for experiments including (h), for WT α₂δ-1 (red) and α₂(3C)δ-1 (blue). N = 290 and 317 cells, respectively, from 3 separate transfections, normalized to WT α₂δ-1 in each experiment. p=0.259, Student’s t test. DOI: http://dx.doi.org/10.7554/eLife.21143.002

Figure 1—figure supplement 1.. Alignment of the proteolytic cleavage site in α₂δ-1, showing species conservation.

/ indicates cleavage site previously identified in rabbit α₂δ-1 by N-terminal sequencing of delta Jay et al., 1991; De Jongh et al., 1990. DOI: http://dx.doi.org/10.7554/eLife.21143.003

Figure 1—figure supplement 2.. α₂(V6)δ-1 is localised in DRMs to a similar extent to α₂δ-1.

(a) Sucrose gradient fractions showing a comparison of DRM localisation (lanes 2–4) of α₂δ-1 in untransfected (UTR) tsA-201 cells (top panels, showing some endogenous α₂δ-1), α₂δ-1 transfected cells (middle panels) and α₂(V6)δ-1-transfected cells (bottom panels). The upper panel is a Western blot with α₂δ-1 mAb, and the lower panel of each set shows Flotillin-1 (Flot-1) a DRM marker. (b) Deglycosylation of the peak DRM fractions from the three conditions shown in A, demonstrating the absence of any proteolytic cleavage of α₂(V6)δ-1 (lane 3). The small amount of free α₂-1 indicated is identical to that in the UTR cells (lane 1). DOI: http://dx.doi.org/10.7554/eLife.21143.004

Figure 1—figure supplement 3.. α₂(3C)δ-1 is localised in DRMs to a similar extent to α₂δ-1.

(a) Comparison of DRM localisation of α₂δ-1-HA (left panels) and α₂(3C)δ-1-HA (right panels), isolated from transfected tsA-201 cells. Only peak DRM (4–7) and soluble (11–13) fractions are shown. The upper panel is a Western blot with HA Ab, and the lower panels show Flotillin-1 (Flot-1), a DRM marker. (b) Quantification of DRM localization for α₂δ-1 (red) and α₂(3C)δ-1 (blue), expressed as DRM/total for n = 3 experiments, normalized to the DRM localization for WT α₂δ-1 in each experiment. Mean ± SEM and individual data points. (c) Deglycosylation of the peak DRM fractions from the two conditions shown in A, demonstrating the absence of any proteolytic cleavage of α₂(3C)δ-1. Flot-1 (lower panel) is a loading control. DOI: http://dx.doi.org/10.7554/eLife.21143.005

Figure 2.. Effect of mutation of α₂δ-1 cleavage site to an HRV-3C site on cell-surface expression and functional properties of Ca_V2.2.

(a) Example traces (−30 to +10 mV in 5 mV steps) for Ca_V2.2/β1b-GFP and no α₂δ (black traces), WT α₂δ-1 (red traces) or α₂(3C)δ-1 (blue traces). Charge carrier 1 mM Ba²⁺. Scale bars refer to all traces. (b) Mean (± SEM) IV curves for Ca_V2.2/β1b-GFP and no α₂δ (black open circles, n = 14), WT α₂δ-1 (red squares, n = 34) or α₂(3C)δ-1 (blue triangles, n = 21). G_max: 0.25 ± 0.04, 1.91 ± 0.30 and 0.20 ± 0.03 nS/pF, respectively. V_50,act: 2.85 ± 0.68, 3.33 ± 0.46 and 3.89 ± 0.53 mV, respectively. (c) tsA-201 cells transfected with GFP-Ca_V2.2 (lanes 1 and 3) or GFP (lanes 2 and 4), plus β1b, and either WT α₂δ-1 (lanes 1 and 2) or α₂(3C)δ-1 (lanes 3 and 4). Immunoprecipitation of GFP-Ca_V2.2 with anti-GFP Ab; WB with Ca_V2.2 II-III loop Ab (upper panels, lanes 1 and 3) produced co-immunoprecipitation (lower panels) of WT α₂δ-1 (lane 1) and α₂(3C)δ-1 (lane 3), revealed by α₂δ-1 mAb. Right panels: WCL input for lanes 1–4: upper panels, Ca_V2.2-GFP input; lower panels, α₂δ-1 input. All samples deglycosylated. (d) Immunocytochemical detection of cell-surface expression of Ca_V2.2-BBS, with β1b, and empty vector (left), WT α₂δ-1-HA (middle) or α₂(3C)δ-1-HA (right) in N2A cells. Upper panel: Ca_V2.2-BBS cell-surface staining prior to permeabilization (grey-scale); lower panel: total Ca_V2.2 after permeabilization (II-III loop Ab, red). Scale bar 5 µm. (e) Quantification of Ca_V2.2-BBS cell-surface expression (box and whisker plots) with empty vector (open bar, n = 206), WT α₂δ-1 (red bar, n = 191) or α₂(3C)δ-1 (blue bar, n = 181). Statistical differences: ANOVA and Bonferroni post-hoc test; ***p<0.001, compared to no α₂δ. (f) Resting membrane potential of control N2A cells (white bar, n = 16) and following expression of TASK3 (gray bar, n = 12). Box and whisker plots; ***p<0.0001, Student’s t test. (g) Quantification of Ca_V2.2-BBS cell-surface expression in N2A cells co-expressing TASK3, with empty vector (gray bar, n = 70), WT α₂δ-1 (pink bar, n = 73) or α₂(3C)δ-1 (pale blue bar, n = 81). Box and whisker plots, statistical differences: ANOVA and Bonferroni post-hoc test, compared to no α₂δ; ***p<0.001. (h) Cartoon showing the ability of ‘latent’ Ca_V2.2 (cyan) plus α₂(3C)δ-1 (grey α₂, blue δ), to traffic to the plasma membrane (PM). DOI: http://dx.doi.org/10.7554/eLife.21143.006

Figure 2—figure supplement 1.. Examination of effect of α₂(3C)δ-1 on Ca_V2.2/β1b calcium channel currents in tsA-201 cells.

These experiments were performed using 10 mM Ba²⁺ to amplify any differences in current amplitude between the two conditions examined. (a) Example traces (−25 to +25 mV steps) for Ca_V2.2/β1b-GFP and either no α₂δ (black traces) or α₂(3C)δ-1 (blue traces). The scale bars refer to all traces. (b) Mean (± SEM) IV curves for experiments including those in (a), for Ca_V2.2/β1b-GFP and either no α₂δ (black open circles, n = 10) or α₂(3C)δ-1-HA (blue triangles, n = 14). Peak I_Ba at +25 mV was −37.9 ± 5.7 pA/pF and −27.2 ± 4.8 pA/pF, respectively. G_max values were 0.91 ± 0.19 and 0.76 ± 0.12 nS/pF, respectively. V_50,act values were 13.1 ± 1.6, and 14.9 ± 1.0 mV, respectively. DOI: http://dx.doi.org/10.7554/eLife.21143.007

Figure 2—figure supplement 2.. Examination of effect of α₂(V6)δ-1 on Ca_V2.2/β1b calcium channel currents in tsA-201 cells.

(a) Example traces (−25 to +10 mV in 5 mV steps) for Ca_V2.2/β1b-GFP and either no α₂δ (black traces), α₂δ-1 (red traces) or α₂(V6)δ-1 (blue traces). The scale bars refer to all traces. Charge carrier 1 mM Ba²⁺ (b) Mean (± SEM) IV curves for experiments including those in (c), for Ca_V2.2/β1b-GFP and either no α₂δ (black open circles, n = 8), WT α₂δ-1 (red squares, n = 7) or α₂(V6)δ-1 (blue triangles, n = 8). G_max values were 0.32 ± 0.04, 6.72 ± 1.09 and 0.83 ± 0.52 nS/pF, respectively. V_{50, act} values were 3.6 ± 2.2,,–0.7 ± 1.1 and 5.8 ± 1.1 mV, respectively. DOI: http://dx.doi.org/10.7554/eLife.21143.008

Figure 2—figure supplement 3.. Examination of effect of α₂(3C)δ-1 on Ca_V2.2/β1b cell surface expression in tsA-201 cells.

(a) Immunocytochemical detection of cell-surface expression of Ca_V2.2-BBS, with β1b, and empty vector (left), WT α₂δ-1-HA (middle) or α₂(3C)δ-1-HA (right) in tsA-201 cells. Upper panel: Ca_V2.2-BBS cell-surface staining prior to permeabilization (grey-scale); lower panel: total Ca_V2.2 after permeabilization (II-III loop Ab, red). Scale bar 5 µm. (b) Quantification of Ca_V2.2-BBS cell-surface expression in tsA-201 cells (box and 10%–90% whisker plots) with empty vector (open bar, n = 191), WT α₂δ-1 (red bar, n = 175) or α₂(3C)δ-1 (blue bar, n = 177). N = 3 experiments. Statistical differences: ANOVA and Bonferroni post-hoc test; ***p<0.001, compared to no α₂δ. DOI: http://dx.doi.org/10.7554/eLife.21143.009

Figure 3.. Effect of proteolytic cleavage of α₂δ-1 containing an HRV-3C cleavage site on cell-surface expression and functional properties of Ca_V2.2.

(a) Cartoon showing intracellular cleavage by 3C-protease (green) of α₂(3C)δ-1 (gray/blue) associated with Ca_V2.2 (cyan). (b) Deglycosylated, cell-surface biotinylated fractions for α₂δ-1-HA (lanes 1 and 2) and α₂(3C)δ-1-HA (lanes 3 and 4), expressed in tsA-201 cells without (lanes 1 and 3) or with (lanes 2 and 4) 3C-protease. Representative of n = 4 experiments. WCL in Figure 3—figure supplement 1 (c) Example traces (−30 to +10 mV steps) for Ca_V2.2/β1b-GFP/α₂(3C)δ-1-HA and no protease (black traces), 3C-protease (blue traces) or inactive mutant 3C-(C147V) protease (cyan traces). Charge carrier 1 mM Ba²⁺. Scale bars refer to all traces. (d) Mean (± SEM) IV curves for Ca_V2.2/β1b-GFP/α₂(3C)δ-1-HA and no protease (black open circles, n = 26), 3C-protease (blue squares, n = 22) or 3C-(C147V)-protease (cyan triangles, n = 23). G_max: 0.26 ± 0.04, 0.90 ± 0.22 and 0.22 ± 0.03 nS/pF, respectively. G_max values in the presence of the active 3C-protease were greater than in the absence of protease or presence of 3C-(C147V)-protease (Kruskal-Wallis test with Dunn’s multiple comparison post-hoc test, p<0.05). V_50,act: 6.05 ± 0.82, 5.18 ± 0.74 and 6.20 ± 1.20 mV, respectively. (e) Time constant of activation (τ_act) for I_Ba at +10 mV for Ca_V2.2/β1b-GFP with no α₂δ-1 (open bar, n = 14), WT α₂δ-1-HA (red bar, n = 25), α₂(3C)δ-1-HA (blue bar, n = 21) or α₂(3C)δ-1-HA + 3C-protease (blue hatched bar, n = 17). Box and whisker plots. Statistical significance determined by ANOVA and Bonferroni’s post-hoc test (*p<0.05). (f) Immunocytochemical detection of cell-surface Ca_V2.2-BBS, plus β1b, and α₂δ-1-HA (left panel) or α₂(3C)δ-1-HA (right panel), with 3C-protease. Upper panel: Ca_V2.2-BBS cell-surface staining (grey-scale), lower panel total Ca_V2.2 (II-III loop staining). Scale bar 5 µm. (g) Lack of effect of 3C-protease (hatched bars) on cell-surface expression of Ca_V2.2-BBS following expression of Ca_V2.2-BBS/β1b in N2A cells, with no α₂δ (open bar, n = 206), WT α₂δ-1 (red bar, n = 212), WT α₂δ-1 and 3C-protease (red hatched bar, n = 192), α₂(3C)δ-1 (blue bar, n = 181) or α₂(3C)δ-1 and 3C-protease (blue hatched bar, n = 200). Box and whisker plots. Statistical differences determined by ANOVA and Bonferroni’s post-hoc test (***p<0.001; ns: p>0.05). DOI: http://dx.doi.org/10.7554/eLife.21143.010

Figure 3—figure supplement 1.. The effect of 3C-protease on α₂δ-1 and α₂(3C)δ-1 expressed in tsA-201 cells.

(a) Deglycosylated WCL for the experiment shown in Figure 3b; α₂δ-1-HA (lanes 1 and 2) and α₂(3C)δ-1-HA (lanes 3 and 4), expressed in tsA-201 cells, without (lanes 1 and 3) or with (lanes 2 and 4) 3C-protease. Upper panel shows HA blot, lower panel: shows GAPDH blot loading control. (b) Cell surface biotinylation for α₂δ-1-HA (lanes 1 and 2) and α₂(3C)δ-1-HA (lanes 3 and 4), expressed in tsA-201 cells with inactive C147V (lanes 1 and 3) or (lanes 2 and 4) 3C-protease. Proteins were deglycosylated. WB was performed with HA Ab. DOI: http://dx.doi.org/10.7554/eLife.21143.011

Figure 3—figure supplement 2.. Lack of effect of 3C-protease on Ca_V2.2/β1b currents expressed in tsA-201 cells.

Top: example traces (−25 to +15 mV steps) for Ca_V2.2/β1b-GFP and either no protease (black traces), WT 3C-protease (blue traces) or inactive mutant 3C-protease (C147V, cyan traces). The charge carrier was 1 mM Ba²⁺. The scale bars refer to all traces. Bottom: Mean (± SEM) IV curves for experiments including those shown, for Ca_V2.2/β1b-GFP and either no protease (black open circles, n = 9), WT 3C-protease (blue circles, n = 11) or inactive mutant 3C-protease (C147V) (cyan triangles, n = 9). Peak I_Ba at +15 mV was −8.2 ± 1.8 pA/pF, −4.8 ± 1.7 pA/pF and −10.1 ± 2.8 pA/pF, respectively. G_max values were 0.32 ± 0.08, 0.20 ± 0.06 and 0.36 ± 0.09 nS/pF, respectively. V_{50, act} values were 4.1 ± 0.7, 1.2 ± 1.2 and 3.8 ± 1.0 mV, respectively. DOI: http://dx.doi.org/10.7554/eLife.21143.012

Figure 4.. Effect of proteolytic cleavage of α₂δ-3 containing an HRV-3C cleavage site on cell-surface expression and functional properties of Ca_V2.2.

(a) Images showing cell-surface Ca_V2.2-BBS (upper row, white), and total Ca_V2.2 (II-III loop Ab, lower row, red), for Ca_V2.2-BBS/β1b in N2A cells, with empty vector (panel 1) or α₂δ-3-HA (panel 2). Scale bar 5 µm. (b) Quantification (box and whisker plots) of effect α₂δ-3 on cell-surface Ca_V2.2-BBS following expression of Ca_V2.2-BBS/β1b with empty vector (open bar, n = 188) or WT α₂δ-3 (red bar, n = 188). Statistical difference determined by Student’s t test, **p=0.0028. (c) Alignment of α₂δ-3 sequence around the predicted cleavage site with α₂δ-1, showing weak homology. Underlined sequence (MTAKAQ) mutated to V6 or HRV-3C cleavage motif. (d) α₂δ-3-HA (lanes 1, 2) and α₂(3C)δ-3-HA (lanes 3, 4) expressed in tsA-201 cells, with either inactive (C147V, lanes 1, 3) or WT 3C-protease (WT, lanes 2, 4), cell-surface biotinylated and deglycosylated. Full WB and corresponding WCL in Figure 4—figure supplement 3 (e) Quantification of cell-surface expression of WT α₂δ-3-HA (red speckled bar) and α₂(3C)δ-3-HA (blue speckled bar), with 3C-protease, normalized relative to inactive 3C-protease (C147V) for n = 3 experiments. Data are mean (± SEM) and individual data points: p=0.4721 for WT α₂δ-3-HA and p=0.9513 for α₂(3C)δ-3 (1 sample t-test compared to respective control). (f) Example traces (−30 to +5 mV steps) for Ca_V2.2/β1b-GFP with no α₂δ (black traces), WT α₂δ-3 (red traces) or α₂(3C)δ-3 (blue traces). G_max: 0.24 ± 0.03, 1.46 ± 0.22 and 0.21 ± 0.03 nS/pF respectively. V_50,act: 0.91 ± 1.015, 1.01 ± 0.85 and 4.03 ± 1.04 mV, respectively. (g) Example traces (−30 to +10 mV steps) for Ca_V2.2/β1b-GFP/α₂(3C)δ-3 and no protease (black traces), 3C-protease (blue traces) or inactive 3C-protease (C147V) (cyan traces). For (f) and (g), charge carrier 1 mM Ba²⁺, scale bars refer to all traces. (h) Mean (± SEM) IV curves for Ca_V2.2/β1b-GFP/α₂(3C)δ-3 without protease (black open circles, n = 28), with 3C-protease (blue squares, n = 29) or inactive 3C-(C147V)-protease (cyan triangles, n = 24). G_max: 0.28 ± 0.05, 0.70 ± 0.11 and 0.30 ± 0.08 nS/pF, respectively. G_max for 3C-protease condition larger than other two conditions (Kruskal-Wallis test with Dunn’s post-hoc test, p<0.05). V_{50, act}: 4.0 ± 0.7, 0.3 ± 0.7 and 1.5 ± 0.6 mV, respectively. DOI: http://dx.doi.org/10.7554/eLife.21143.013

Figure 4—figure supplement 1.. Lack of cleavage of α₂(V6)δ-3.

The expression of α₂δ-3 compared to α₂(V6)δ-3. The constructs were expressed transiently in tsA-201 cells; peak lipid raft fractions were taken for western blotting. There was a complete loss of free δ-3 (lower panel) in α₂(V6)δ-3 (lane 2), compared to WT α₂δ-3 (lane 1). α₂δ-3 is shown in the upper panel. DOI: http://dx.doi.org/10.7554/eLife.21143.014

Figure 4—figure supplement 2.. Effect of 3C-protease on expression and cleavage of α₂(3C)δ-3.

Full data from experiment shown in Figure 4d. WT α₂δ-3-HA (lanes 1, 2 and 3) and α₂(3C)δ-3-HA (lanes 4, 5 and 6) were expressed in tsA-201 cells, either without or with inactive (C147V, lanes 2 and 5) or WT 3C-protease (WT, lanes 3 and 6), subjected to cell surface biotinylation and streptavidin pull-down. Proteins were deglycosylated with PNGase-F. (a) shows biotinylated proteins and (b) shows WCL. The upper blot with HA Ab, shows that WT, but not inactive mutant, 3C-protease cleaved α₂(3C)δ-3 (lane 6). The lower blot shows GAPDH, indicating that no intracellular proteins were biotinylated, and the bottom blot in (b) shows expression of the His-tagged 3C-proteases. DOI: http://dx.doi.org/10.7554/eLife.21143.015

Figure 4—figure supplement 3.. Lack of effect of purified 3C-protease on expression and cleavage of α₂(3C)δ-3.

Incubation of cells expressing α₂(3C)δ-3-HA with 3C-protease enzyme did not result in cleavage on the cell surface. Cell surface biotinylated deglycosylated α₂(3C)δ-3-HA, from cells incubated without (left) or with (right) 3C-protease. DOI: http://dx.doi.org/10.7554/eLife.21143.016

Figure 5.. Effect of thrombin on the properties and function of α₂δ-3 containing a thrombin proteolytic cleavage site.

(a) Sequence at α₂δ-3 cleavage site mutated to a thrombin cleavage site. (b) Cartoon of thrombin cleavage of α₂δ-3 on cell-surface. (c) Cell-surface biotinylation (left panel) shows efficient cleavage of cell-surface α₂(Th)δ-3-HA (lane 2), with no effect on total α₂(Th)δ-3-HA in WCL (right panel, lane 4). Samples were deglycosylated prior to loading. (d) Quantification of cell surface biotinylation experiments such as those shown in (c), indicating that thrombin does not decrease the amount of α₂(Th)δ-3-HA on the cell surface (hatched blue bar), normalized to vehicle application in each experiment (solid blue bar). Mean (± SEM) and individual data points for n = 4; p=0.2105, 1 sample t test. (e) Mean (± SEM) and individual data points of peak I_Ba at +10 mV, for Ca_V2.2/β1b with WT α₂δ-3 (red squares), α₂(Th)δ-3 (blue triangles) or no α₂δ (black circles) and either no protease (closed symbols), or 60 min thrombin incubation (open symbols). The charge carrier was 2 mM Ba²⁺. For data without or with thrombin, respectively, n = 12, 16 for WT α₂δ-3; 15, 24 for α₂(Th)δ-3 and 11, 7 without α₂δ, from at least 3 independent transfections. Statistical difference between thrombin and vehicle determined by Kruskal-Wallis ANOVA and Dunn’s multiple comparison post-hoc test, *p<0.05. (f) Mean (± SEM) full IV curves for the same conditions as in (e) (excluding the no α₂δ data), fitted with a modified Boltzmann equation to +50 mV. G_max values (nS/pF) were 2.80 ± 0.61 (WT α₂δ-3; n = 10), 2.60 ± 0.55 (WT α₂δ-3 plus thrombin, n = 15), 0.43 ± 0.08 (α₂(Th)δ-3; n = 14), 1.26 ± 0.21 (α₂(Th)δ-3 plus thrombin, n = 21). V_50,act values (mV) were +0.44 ± 1.71 (WT α₂δ-3), + 0.55 ± 1.28 (WT α₂δ-3 plus thrombin), +9.34 ± 0.08 (α₂(Th)δ-3), +0.87 ± 1.38 (α₂(Th)δ-3 plus thrombin). (g) Example Ba²⁺ currents (from −50 to +60 mV) for the four conditions shown in (f). DOI: http://dx.doi.org/10.7554/eLife.21143.017

Figure 5—figure supplement 1.. Controls for cleavage of α₂(Th)δ-3 by thrombin.

(a) Deglycosylated cell surface biotinylation for tsA-201 cells expressing WT α₂δ-3-HA and incubated either with vehicle or thrombin for 60 min (lanes 1 and 2), showing WT α₂δ-3 is not cleaved by thrombin. (b) Deglycosylated cell surface biotinylation for tsA-201 cells expressing WT α₂δ-3-HA or α₂(Th)δ-3-HA (lanes 1 and 2), showing α₂(Th)δ-3 reaches cell surface. (c) Deglycosylated cell surface biotinylation for tsA-201 cells expressing α₂(Th)δ-3-HA and incubated with thrombin for 0, 30 and 60 min (lanes 1–3), showing α₂(Th)δ-3 is cleaved on cell surface at 60 min. DOI: http://dx.doi.org/10.7554/eLife.21143.018

Figure 6.. Proteolytic processing of endogenous α₂δ-1 and effect of exogenous expression of α₂δ-1 and α₂(3C)δ-1 in DRGs.

(a) DRGs, spinal nerves and dorsal roots from rats, 4 days after SNL, were dissected and segmented according to the diagram. X marks site of ligation. Tissue was pooled from 4 rats. It was either treated or not with PNGase-F as indicated, and reduced with DTT; deglycosylation allows resolution of two α₂-immunoreactive bands. Unprocessed α₂δ-1 is present only in the cell body compartment (segment 3) and is distinct from processed α₂-1 (indicated by arrows). Lower blot is GAPDH loading control. (b) WCL for empty vector-transfected DRGs (U/T, lane 1); WT α₂δ-1-HA-transfected DRGs (lane 2); α₂(3C)δ-1-HA transfected DRGs (lane 3). WB: anti-HA. (c) Left: Example traces (−45 to +5 mV steps) for control (empty vector-transfected) DRG neurons (black traces, top) and DRGs transfected with α₂(3C)δ-1 (blue traces, bottom). The charge carrier is 1 mM Ba²⁺. The scale bars refer to all traces. Right: Mean (± SEM) IV curves for control DRG neurons (black squares, n = 12) and DRGs transfected with α₂(3C)δ-1 (blue circles, n = 14), from 3 independent experiments. G_max values were 2.20 ± 0.30 and 1.26 ± 0.14 nS/pF, respectively; p=0.0061 (Student’s t test). V_{50, act} values were −1.5 ± 1.2 and −1.5 ± 0.7 mV, respectively. (d) Comparison of normalized peak I_Ba in control DRGs (open red bar, n = 55) and WT α₂δ-1-transfected DRGs (closed red bar, n = 54) including data from Figure 2d in, or comparison of control DRGs (open blue bar, n = 12) with α₂(3C)δ-1 transfected DRGs (closed blue bar, n = 14). Box and whisker plots. Statistical differences, Student’s t test: *, p=0.048, **p=0.0067 compared to respective control. (e) Confocal optical sections (1 µm) showing GFP-Ca_V2.2-HA in non-permeabilized DRG neurons (top, white), when co-transfected with β1b and either empty vector (left), WT α₂δ-1 (middle) or α₂(3C)δ-1 (right). GFP fluorescence is shown in the merged lower panel. Scale bars: 10 µm. (f) Box and whisker plot of cell surface HA fluorescence density as a ratio of internal GFP density for Ca_V2.2-HA expression in DRG somata, transfected with empty vector (open bar, n = 81), WT α₂δ-1 (red bar, n = 133) or α₂(3C)δ-1 (blue bar, n = 159). ***p<0.001, 1 way ANOVA and Bonferroni post hoc test. DOI: http://dx.doi.org/10.7554/eLife.21143.019

Figure 7.. Effect of α₂δ-1 and proteolytic cleavage of α₂(3C)δ-1 on trafficking of Ca_V2.2 into hippocampal neurites.

(a) Images showing Ca_V2.2-HA in permeabilized hippocampal neurons (top, white), when co-transfected with β1b, mCherry (bottom, red) and either no α₂δ (left), WT α₂δ-1 (middle) or α₂(3C)δ-1 (right). Scale bar: 50 µm applies to all images. (b) Box and whisker plots for Ca_V2.2-HA expression in processes without α₂δ (open bar, n = 136 processes from 29 cells), with α₂δ-1 (red bar, n = 147 processes from 27 cells) or with α₂(3C)δ-1 (blue bar, n = 109 processes from 22 cells). ***p<0.001, 1 way ANOVA and Bonferroni post hoc test. (c) Images showing Ca_V2.2-HA in permeabilized hippocampal neurons (white), when co-transfected with β1b, α₂(3C)δ-1 and mCherry (transfection marker, not shown), either without (left) or with (right) 3C-protease. Scale bar: 50 µm applies to both images. (d) Box and whisker plots for Ca_V2.2-HA expression in processes with α₂(3C)δ-1, transfected without (solid blue bar, n = 191 processes), or with 3C-protease (blue hatched bar, n = 187 processes). ***p<0.001, 1 way ANOVA and Bonferroni post hoc test. (e) Images showing WT α₂δ-1-HA (left) or α₂(3C)δ-1-HA (right) expressed in permeabilized hippocampal neurons (white), co-transfected only with mCherry (transfection marker, not shown). Antigen retrieval was used prior to the HA Ab. Scale bar: 50 µm applies to both images. (f) Box and whisker plots for expression in the processes of WT α₂δ-1-HA (red bar, n = 248 processes from 52 cells) and α₂(3C)δ-1 (blue bar, n = 263 processes from 51 cells). (g) Images showing α₂δ-1 in hippocampal neurons (white), when transfected with Ca_V2.2-HA, β1b, mCherry (transfection marker, not shown) and either WT α₂δ-1 (left) or α₂(3C)δ-1 (right). Antigen retrieval was used prior to the α₂δ-1 mAb. Scale bar: 50 µm applies to both images. (h) Box and whisker plots for α₂δ-1 expression in hippocampal processes, for WT α₂δ-1 (red bar, n = 221 processes) and α₂(3C)δ-1 (blue bar, n = 184 processes). ***p<0.0001, Student’s t test. DOI: http://dx.doi.org/10.7554/eLife.21143.020

Figure 7—figure supplement 1.. Cartoon showing processing and trafficking of α₂δ in neurons.

Cartoon of the effect of proteolytic processing of α₂δ (dark gray α_2, blue δ) on Ca_V2.2 trafficking and voltage-dependent activation. The gray channel is one that cannot be activated by depolarization, whereas the purple channel is functional. The solid arrows indicate the likely trafficking pathway in neurons. The dotted lines show that trafficking of channels containing uncleaved α₂δ to the plasma membrane and into processes, and extracellular proteolytic cleavage of α₂δ indicates do not occur in native neurons studied here, although it is possible that this may occur in pathological conditions, for example following α₂δ-1 upregulation as a result of neuropathic nerve injury in DRG neurons. DOI: http://dx.doi.org/10.7554/eLife.21143.021

Figure 8.. Effect of proteolytic cleavage of α₂(3C)δ-1 on Ca²⁺ influx in presynaptic terminals of hippocampal neurons.

(a) Fluorescence changes in presynaptic terminals of hippocampal neurons expressing sy-GCaMP6f and VAMP-mOr2 in response to electrical stimulation. White arrows point to transfected boutons. Top three panels show sy-GCaMP6f fluorescence: at rest (top), after 1 AP (middle) and after 100 Hz stimulation for 1 s (bottom). The bottom panel shows VAMP-mOr2 fluorescence after 100 Hz stimulation for 1 s. Scale bar 5 µm. The pseudocolour scale is shown below the third panel. (b) Mean example traces from the same experiment of sy-GCaMP6f fluorescence changes in response to 5 single APs from individual presynaptic terminals of neurons co-transfected with either empty vector (black trace) or α₂(3C)δ-1 (blue trace). (c) Sy-GCaMP6f fluorescence changes (expressed as % of F_max) in response to 1 AP from boutons co-transfected with either empty vector (black filled circles) or α₂(3C)δ-1 (blue filled squares) (n = 7 independent experiments, *p=0.049, paired t test). (d) Mean example traces of sy-GCaMP6f fluorescence changes in response to 5 single APs from presynaptic terminal of neurons co-transfected with either α₂(3C)δ-1 (blue trace) or α₂(3C)δ−1 + 3C-protease (red trace). (e) Sy-GCaMP6f fluorescence changes (expressed as % of F_max) in response to 1 AP from boutons co-transfected with either α₂(3C)δ-1 (blue filled squares) or α₂(3C)δ−1 + 3C-protease (red open squares) (n = 8 independent experiments, ***p=0.0005, paired t test). DOI: http://dx.doi.org/10.7554/eLife.21143.022