The interplay between splicing of exons 18a and 47 differentially affects membrane targeting and function of human CaV2.2

N-type calcium channels (CaV2.2) are predominantly localized in presynaptic terminals, and are particularly important for pain transmission in the spinal cord. Furthermore, they have multiple isoforms, conferred by alternatively-spliced or cassette exons, which are differentially expressed. Here we have examined alternatively-spliced exon47 variants that encode a long or short C-terminus in human CaV2.2. In the Ensembl database, all short exon47-containing transcripts were associated with the absence of exon 18a, therefore we also examined effect of inclusion or absence of exon18a, combinatorially with the exon47 splice variants. We found that long exon47, only in the additional presence of exon18a, results in CaV2.2 currents that have a 3.6-fold greater maximum conductance than the other three combinations. In contrast, cell surface expression of CaV2.2 in both tsA-201 cells and hippocampal neurons is increased ∼4-fold by long exon47 relative to short exon47, in either the presence or absence of exon18a. This surprising discrepancy between trafficking and function indicates that cell surface expression is enhanced by long exon47, independently of exon 18a. However, in the presence of exon47, exon18a mediates an additional permissive effect on CaV2.2 gating. We also investigated the SNP in exon47 that has been linked to schizophrenia and Parkinson’s disease, which we found is only non-synonymous in the short exon47 C-terminal isoform, resulting in two minor alleles. This study highlights the importance of investigating the combinatorial effects of exon inclusion, rather than each in isolation, in order to increase our understanding of calcium channel function. Graphical abstract


Graphical abstract Introduction
Synaptic transmission in the central and peripheral nervous system is dependent on voltagegated calcium (Ca V ) channels that provide Ca 2+ for vesicular release from presynaptic terminals.The particular channels involved are predominantly Ca V 2.1 and Ca V 2.2 1,2 . These channels are associated with auxiliary α 2 δ and β subunits which optimise their trafficking and function [3][4][5][6] .
CACNA1A, the gene encoding Ca V 2.1, is associated with multiple genetic disorders 7,8 ; in contrast there are very few identified human or animal pathological variants of CACNA1B, encoding Ca V 2.2 which underlies N-type Ca V channels [9][10][11] .Nevertheless, Ca V 2.2 knockout mice are associated with altered pain sensation and other phenotypes 12 .These channels are particularly important for neurotransmission at primary afferents, including nociceptor terminals 13,14 , and also in the sympathetic nervous system 15 .A human variant in CACNA1B was linked to myoclonus-dystonia syndrome, although this was then disputed 10,11 .A recent study identified novel risk loci in a genome-wide association (GWAS) analysis of Parkinson's disease and schizophrenia, including a single nucleotide polymorphism (SNP), Rs2278973, in CACNA1B 16 .However, upon analysis, we identified that this SNP was only non-synonymous in an alternatively-spliced exon47, which encodes a shorter C-terminal sequence than long exon47 (Figure 1A, B).The predominant residue at this SNP is arginine (Arg), and the two minor SNPs encode leucine (Leu) and histidine (His) (Supplementary Figure 1).
We further noted from a BLAST search (NCBI, Ensembl 17 ) of the short exon47 splice variant that all reported human Ca V 2.2 sequences containing this short exon47 variant were also associated with absence of cassette exon18a (Figure 1C), whereas exon18a appeared to undergo alternative splicing in the Ca V 2.2 variant with the long C terminus.
It has previously been reported that the inclusion of exon18a, which encodes a 21 amino acid sequence in the intracellular linker connecting Domains II and III of Ca V 2.2, resulted in two-fold increase in Ca V 2.2/β3 currents, when these channel combinations were expressed in a cell line 18 .The mechanism controlling insertion of exon18a was found to involve the RNA binding protein Rbfox2, which suppressed exon18a inclusion 18 .In further support of the effect of exon18a inclusion, ω-conotoxin GVIA-sensitive N-type Ca V currents were larger in sympathetic neurons from mice engineered to express exon18a, compared to those not expressing it 18 .
In the present study, we first examined the effect of inclusion of the long or short exon47 variants on Ca V 2.2 expression, function and trafficking.Since our analysis suggests there is differential expression of exon18a associated with the exon47-containing isoforms, we also examined effect of the inclusion or absence of exon18a combinatorially with the two exon47 splice variants.We then determined whether there was any effect of the SNPs in short exon47 ofCACNA1B 16 .
Our results reveal a surprising discrepancy between Ca V 2. Thermo Fisher Scientific).Neurons were kept in a 5 % CO2 incubator at 37 °C and half the medium was replaced every 3-4 days.At 7 days in vitro and 2 h before transfection, half of the medium was removed, and fresh medium was added.

Expression constructs and mutagenesis
The following expression constructs were used: human Ca V 2.2 (huCaV2.
For immunocytochemistry, secondary anti-rat-Alexa Fluor 594 and anti-rat-Alexa Fluor 647 antibodies (1:500, Life Technologies) were used to visualise the HA-tagged channels.I=G max *(V−V rev )/(1+exp(−(V−V 50, act )/k)) where I is the current density (in pA/pF), G max is in nS/pF, V rev is the apparent reversal potential, V 50, act is the midpoint voltage for current activation, and k is the slope factor.

Immunoblotting
Immunoblotting was carried out on tsA-201 cells expressing the cDNAs as described.At 48 h after transfection, cells were rinsed with phosphate buffered saline (PBS, pH 7.4), scraped into cold PBS and centrifuged at 1,000 x g at 4⁰C for 10 minutes.Cell pellets were homogenised in PBS containing 1% Igepal, 0.1% SDS and protease inhibitors (PI, cOmplete, Sigma-Aldrich), pH 7.4, and then incubated on ice for 30 min to allow cell lysis.Whole cell lysates (WCL) were centrifuged at 20,000 x g for 25 min at 4⁰C and supernatants were assayed for total protein (Bradford assay, BioRad).Aliquots of WCL (30 µg total protein, per sample) were diluted with 2x Laemmli buffer supplemented with 100 mM dithiothreitol and incubated at 60°C for 10 min
Imaging was performed on Zeiss LSM 780 confocal microscope.

Image analysis
The tsA-201 cell images were obtained using a 63 × objective at a resolution of 1024 × 1024 pixels and an optical section of 0.5 μm.After choosing a region of interest containing transfected cells, the 3 × 3 tile function of the microscope was used to allow a larger area to be selected without bias.Images were then analyzed using Fiji.Every transfected cell was included.Surface labelling (HA staining) was measured using a freehand line tool of 5 pixels

Data analysis and statistics
Quantitative data are presented by GraphPad Prism 8 software (v8.0.0;GraphPad Software, Inc.) or Origin-Pro 2021 as the mean ± standard error of the mean (SEM) and individual data points as described.One-way analysis of variance (ANOVA) followed by a Šidák post hoc test for multiple comparisons were used.

Results
In this study we have used human Ca V 2.2 containing the following alternatively-spliced exons: +exon 10a, +exon18a, Δexon24a, +exon31a, +exon37b and +long exon47 , and a GFP tag to the N terminus 23,24 (Figure 1A), to form GFP_Ca V 2.2-HA-C-long.It contains the alternatively-spliced 21 amino acid exon18a in the II-III linker (Figure 1C).We used this construct to generate GFP_Ca V 2.2-HA-C-short, which uses an alternative splice site part-way through exon47 (Supplementary Figure 1).This creates a frame-shift, resulting in an alternative, shorter C terminus (Figure 1B).The sequence for exon18a was then deleted from both these constructs to give GFP_Ca V 2.2-HA-C-long-Δexon18a and GFP_Ca V 2.2-HA-C-short-Δexon18a.
The combination of C-terminal long exon47 together with exon18a in the II-III linker results in an increase in Ca V 2.2 currents.
We next examined the currents generated by the GFP_Ca V 2.2-HA constructs expressed in tsA-201 cells, together with β1b and α 2 δ-1 subunits.For GFP_Ca V 2.2-HA-C-long, the additional presence of exon18a resulted in a ~3.6-fold increase of the maximum conductance (G max ), compared to the absence of exon18a (Figure 2A-C).Surprisingly, for GFP_Ca V 2.2-HA-C-short (containing short exon47), the inclusion of exon18a produced no increase in calcium currents (Figure 2A-C), which were of very similar magnitude to those produced by GFP_Ca V 2.2-HA-C-long-Δexon18a.
The presence of C-terminal long exon47 is important for cell surface expression of Ca V 2.2 In order to determine whether the marked increase in Ca V 2.2 current amplitude observed for the long exon47 variant of Ca V 2.2, in conjunction with the presence of exon18a, was due to an increase in cell surface expression for this exon combination, we next expressed the GFP_Ca V 2.2-HA constructs, together with β1b and α 2 δ-1 subunits, in tsA-201 cells and used confocal microscopy to visualize expression of the GFP_Ca V 2.2-HA subunit at the cell surface.Transfected cells were identified by expression of the GFP tag on GFP_Ca V 2.2-HA.
Ca V 2.2 on the plasma membrane was measured by anti-HA antibody binding to the extracellular HA tag of GFP_Ca V 2.2-HA, in non-permeabilised conditions.We found that cell-surface expression of the GFP_Ca V 2.2-HA-C-long(+exon18a) variant was increased by 4.5-fold compared with GFP_Ca V 2.2-HA-C-short(+exon18a) (Figure 3A, B i).However, in contrast to the calcium current measurements (Figure 2), deletion of exon18a had no significant effect on HA expression at the plasma membrane for either the C-long or C-short Ca V 2.2 variants.Cell-surface expression of the GFP_Ca V 2.2-HA-C-long(Δexon18a) variant was still increased by 3.2-fold compared with GFP_Ca V 2.2-HA-C-short(Δexon18a) (Figure 3A, B i).
There was a smaller increase in intracellular GFP_Ca V 2.2-HA, as determined by GFP signal associated with the presence of C-long exon47, compared to C-short exon47 (1.7-fold in the presence of exon18a and 1.2-fold in its absence) (Figure 3A, B ii).From the ratio of cell surface/intracellular expression of the four Ca V 2.2 variants, there was only a significant effect of the long exon47 to increase trafficking to the cell surface, and no effect of exon18a (Figure 3B iii).
These results suggest that the presence or absence of exon18a does not affect cell surface expression or trafficking of GFP_Ca V 2.2-HA in tsA-201 cells but that trafficking to the plasma membrane is highly influenced by the presence of the long exon47 variant in the Cterminus.
We then performed immunoblotting with these constructs to determine whether there were similar effects on total Ca V 2.2 protein expression in the tsA-201 cells (Figure 3C, Supplementary Figure 2).This showed that the level of full length GFP_Ca V 2.2-HA-C-long protein was more than 2-fold greater than that of GFP_Ca V 2.2-HA-C-short, regardless of the inclusion or absence of exon18a (Figure 3D), in a clear parallel with the immunocytochemistry results.Together these data indicate that the presence of short exon47 in Ca V 2.2 confers a defect in cell surface trafficking relative to long exon47, and a reduction in the total full-length channel level.
The presence of long exon47 is important for trafficking of Ca V 2.2 in hippocampal neurons We then examined the effect of both exon47 splice variants and the inclusion of exon18a on Ca V 2.2 expression in hippocampal somata and neurites.In cell bodies, cell-surface expression of the GFP_Ca V 2.2-HA-C-long (+exon18a) variant was increased by 4.3-fold compared with GFP_Ca V 2.2-HA-C-short(+exon18a) (Figure 4A, B i).For the same pair of constructs there was a smaller increase (1.9-fold) in total Ca V 2.2 expression attributable to long exon47, measured by GFP signal (Figure 4A, B ii).In contrast with the data in tsA-201 cells, the inclusion of exon18a also produced a small increase of HA expression at the plasma membrane (1.4-foldincrease) and GFP signal (1.3-fold increase) for the long exon47 variant, but it had no effect for the short exon47 variant (Figure 4A, B i and B ii).However, from the ratio of cell surface/total expression of Ca V 2.2, there was only a significant effect of the long exon47 to increase trafficking to the cell surface, and no effect of the inclusion of exon18a (Figure 4B iii).
Very similar results were observed in the neurites of these hippocampal neurons, in that the presence of the long exon47 variant induced a large increase (4.9-fold) of cell surface Ca V 2.2 in the neurites relative to the short exon47 variant (Figure 4C, D i).There was also an increase (3.1-fold increase) in total Ca V 2.2 expression for the long exon47 variant compared to the short exon47 variant, measured by GFP signal (Figure 4C, D ii).The presence of exon18a also induced a small increase in membrane expression of Ca V 2.2 in the neurites, but only for the long exon47 variant (Figure 4D i).However, from the ratio of cell surface/total expression of Ca V 2.2, there was only a significant effect of the long exon47, but not of exon 18a, to increase trafficking to the cell surface of the neurites (Figure 4D iii).
Examination of the effect of SNP variants in the short exon47 of Ca V 2.2 The Rs2278973 SNP gives no change in amino acid in the long exon47 variant of Ca V 2.2.In the short exon47 variant, however, this SNP results in three alternative amino acids, Arg, Leu or His, at this position.Unlike the reference gene, Ensembl ENSG00000148408 (NCBI Reference Sequence: NM_000718.4),which would produce a Leu residue at this position, the Addgene plasmid #62574 that we used produces an Arg.Arg has a frequency of 91.7% in the population and, from comparison with other species, appears to be the ancestral gene (gnomAD genomes v3.1.2database (Ensembl) https://www.ensembl.org/Homo_sapiens/Variation/Population?db=core;r=9:138121310-138122310;v=rs2278973;vdb=variation;vf=729552325).
We created two additional constructs in the GFP_Ca V 2.2-HA-C-short (which contains exon18a and has an Arg at position 2236 of Ca V 2.2, see Supplementary Figure 1), to give GFP_Ca V 2.2-HA-C-short-Leu and GFP_Ca V 2.2-HA-C-short-His, corresponding to the two minor alleles, which have population frequencies of 8.27% and 0.003% respectively.
Expression of these constructs, together with β1b and α 2 δ-1, in tsA-201 cells revealed that the HA signal at the plasma membrane was not significantly different for any of the amino acid substitutions in the short C-terminus (Figure 5A, B).Plasma membrane expression of GFP_Ca V 2.2-HA-C-long was markedly increased when compared to all the C-short constructs (by 4-fold compared to GFP_Ca V 2.2-HA-C-short-Arg, and by 4.6-fold compared to both the Leu and His-containing variants).Intracellular GFP expression of GFP_Ca V 2.2-HA-C-long was increased by a smaller amount (1.7-fold) compared to all C-short variants (Figure 5C).This result was paralleled by an increase in the total GFP_Ca V 2.2-HA protein level, determined by immunoblotting for GFP_Ca V 2.2-HA-C-long, relative to all the C-short variants (Figure 5D, E).
These GFP_Ca V 2.2-HA SNP variants (containing exon18a), together with β1b and α 2 δ-1, were then expressed in tsA-201 cells, and Ca V 2.2 currents were recorded to examine whether there was any effect of the different SNP variants (Figure 6A-C).The data for GFP_Ca V 2.2-HA-C-long, and GFP_Ca V 2.2-HA-C-short are repeated from Figure 2 for comparison, as all experiments were performed contiguously.In agreement with the cell surface expression data, there was no difference between the Ca V 2.2 G max or other biophysical properties examined for the GFP_Ca V 2.2-HA-C-short constructs containing Arg, Leu or His (Figure 6A-C).

Discussion
N-type calcium channels (Ca V 2.2) are important in primary afferent transmission, including at nociceptor terminals in the dorsal horn.For this reason they are a drug target in pain therapy 13,14,[25][26][27] .It is therefore particularly important to understand their trafficking and function.All Ca V channels and their auxiliary subunits, like most proteins, have multiple isoforms, conferred by one or more alternatively-spliced or cassette exons, each of which may be differentially expressed in diverse tissues and developmental stages .
In the present study, our initial rationale for examining the effect of the C-terminal short exon47 related to a GWAS study which reported that non-synonymous SNPs in alternatively-spliced exon47 were associated with schizophrenia and Parkinson's disease 16 . Because we noted that all reported full length human Ca V 2.2 transcripts which include short exon47 were missing exon18a, we first examined the effect of inclusion of the long or short exon47 variants on human Ca V 2.2 function and cell surface expression, in the presence or absence of exon18a.
Exon18a encodes 21 residues which are inserted at the proximal end of the intracellular II-III linker.Distal to this region, the II-III linker contains a domain that binds certain synaptic proteins including syntaxin and synaptotagmin 34,35 , although the importance of this "synprint" domain for presynaptic targeting of calcium channels and for neurotransmitter release is still unclear 36,37 .
We found that Ca V 2.2 containing the long exon47, in conjunction with exon18a, supports a large increase in Ca V 2.2 current amplitude (~ 4-fold), compared to when exon18a is skipped, or compared to Ca V 2.2 containing short exon47.In contrast, for Ca V 2.2 containing short exon47, the inclusion of exon 18a has no effect on calcium current amplitude.Surprisingly, this result is not mirrored by a similar pattern of changes in cell surface expression.Here, the presence of long exon47 supports an ~4-fold increase in Ca V 2.2 cell surface trafficking in both tsA-201 cells and in hippocampal neurons, compared to Ca V 2.2 containing short exon47, irrespective of the presence or absence of exon18.In parallel there is an ~2-fold increase in the full-length Ca V 2.2 protein level, conferred by long exon47, as determined by immunoblotting (Figure 3 and Supplementary Figure 2).This result suggests the possibility that if the channel does not reach the cell surface, or is not anchored there, it may be diverted to a degradation pathway, which would be the case for Ca V 2.2 constructs containing short exon47.In support of this, a lower molecular weight band, which may be a degradation product, is more evident for Ca V 2.2 containing short exon47 (see Supplementary Figure 2).
Taking these experiments together, it appears that the effect of exon18a inclusion to increase current amplitude, which has been noted previously 18 , is not mediated primarily by an increase in cell surface expression of the channel.It must rather be mediated by a permissive effect on gating that requires the additional presence of long exon47, since Ca V 2.2 containing long exon47 mediates a similar increase in cell surface expression irrespective of the presence or absence of exon18a.A corollary of this finding is that the increase in cell surface expression seen for the Δexon18a / long exon47 combination does not result in any increase in Ca V 2.2 currents, suggesting that the absence of exon18a may exert an inhibitory role for Ca V 2.2 function.Of great interest, exon18a has been found to be differentially expressed in certain cholecystokinin-containing interneurons 28 , which depend on N-type Ca V channels to mediate GABA release 38 .
Several previous studies have highlighted the importance of domains in the C-terminus of Ca V 2.2 in its trafficking and function.In particular, the proximal C-terminus is encoded by exon37, which has two differentially expressed alternative forms, 37a and 37b, that affect Ca V 2.2 current properties 33,39 .When exon37a is present, Ca V 2.2 currents, cell membrane expression and forward trafficking are all increased 23,33 , via interactions with adaptor protein complex-1 (AP-1)

23
. All constructs in our study contained exon 37a.Of further relevance to the present study, it has also been found previously that protein-protein interaction motifs in the long C-terminal human Ca V 2.2 splice variant contain sequences, including a C-terminal Post synaptic density protein, Drosophila disc large tumor suppressor and Zonula occludens-1 protein (PDZ) motif, that are important for its synaptic targeting in cultured hippocampal neurons .In an attempt to mimic the short human C-terminal Ca V 2.2, a hybrid human-rat short C-terminal construct was also generated in that study 40 , although the rat sequence is widely divergent from the human short C-terminus (see .However, in our study a reduction in cell surface expression and trafficking was observed in the non-neuronal tsA-201 cells to the same extent as in hippocampal neuron cell bodies and neurites, indicating that the effect of protein-protein interaction domains in long exon47 is a more general one to enhance cell surface trafficking or anchoring in the cell membrane, rather than (or in addition to) any specific effect to enhance synaptic targeting in neurons.
In this study we observed no effect on Ca V 2.2 function, expression and trafficking of the SNP variants in short exon47, which encode three different amino acids.This indicates that the observed association of these variants with Parkinson's disease and schizophrenia 16 is not mediated by any grossly altered function.Instead, it is possible that the SNP variants might influence the splicing of exon47, and therefore the relative inclusion of the short and long exon47 variants in Ca V 2.2 transcripts.In this regard, it is of great interest that Ca V 2.2 channels play an important role in mediating dopamine release in striatum
Future studies will be needed to determine the relative expression of the long and short exon47-containing isoforms in different human tissues and disease states, and whether the SNP alleles influence these properties.for half-maximal activation (V 50,act ) (mV) was -6.07 ± 0.76, -7.51 ± 0.71, -4.1± 0.73 and -5.D Immunoblot of whole-cell lysates (WCL) from tsA-201 cells transfected with either GFP_Ca V 2.2-HA C-long, C-short-Arg, C-short-Leu or C-short-His, together with α 2 δ-1 and β1b.A Examples of whole-cell patch-clamp recordings of human GFP_Ca V 2.2-HA C-long (blue), C- short-Arg (red), C-short-Leu (green) and C-short-His (pink) variants.All conditions are in the presence of α 2 δ-1 and β1b.Holding potential −80 mV, steps between −50 and +60 mV for 50 ms (applies to all traces).

( 0 .
65 μm) width and manually tracing the surface of the cells.Intracellular GFP staining was measured by drawing around the cell (omitting the nucleus and the plasma membrane).The value of the mean pixel intensity in different channels was measured separately and background was subtracted by measuring the intensity of an imaged area without transfected cells.Hippocampal neurons were imaged using a 20 × objective with a 5 μm optical section.The fluorescence intensity along neuronal projections was assessed as follows: two concentric circles of 100 μm and 150 μm diameter were drawn around each neuronal cell body.A freehand line tool of 3 μm width tracing the neuronal processes (3 to 5 per neuron) between the circles was drawn in the mCherry images and used as template for GFP and HA images.Hippocampal somata were imaged at 63 x objective with a 1 μm optical section.For each neuron selected as positive for mCherry, immunofluorescence (HA and GFP) was measured by two ROIs, one ROI manually traced around the cell body, using a freehand line tool (0.4 μm width) to analyse the cell surface staining and another ROI drawn between the cell surface and the nucleus to determine the intracellular staining.Total staining was calculated per each cell combining the 2 ROIs.The mean pixel intensity in the different channels were measured and the background was subtracted from each image.

Figure 2 Long
Figure 2Long exon47 increases Ca V 2.2 currents, only in the presence of exon18a.

Figure 3
Figure 3 Long exon47 increases cell surface expression of Ca V 2.2 irrespective of exon18a inclusion.A Representative tsA-201 cells expressing GFP-Ca V 2.2-HA having the long (left) or short (right) variant of exon47 in the presence (top row, +18a) or absence (bottom row, ∆18a) of alternatively-spliced exon18a.All conditions are in the presence of α 2 δ-1 and β1b.Cells were incubated with anti-HA antibody in non-permeablised conditions to show HA staining on the extracellular side of the plasma membrane (panels i and iv), to be compared with intracellular GFP fluorescence (ii and v).Merged images (nuclei stained with DAPI in blue and HA in red) are shown in panels iii and vi.Scale bar = 10 μm.B Quantification of HA staining at the plasma membrane (i) or intracellular GFP fluorescence (excluding the nucleus and plasma membrane, ii), showing mean ± SEM for GFP-CaV2.2-HAC-long containing exon18a (blue solid) or without exon18a (blue striped), C-short with exon18a (red solid) or without exon18a (red striped).Individual data points show the mean of 30-100 cells from 7-9 different transfections in 3 independent experiments.The ratio of HA/intracellular GFP for each individual cell was calculated; mean ± SEM for different transfections is shown in iii.Statistical significance was determined using one-way ANOVA followed by a Šidák post hoc test for multiple comparisons: **** p<0.0001, *** p<0.001, ns: non-significant.C Immunoblot of whole-cell lysates (WCL) from tsA-201 cells transfected with either GFP_Ca V 2.2-HA C-long or C-short in the presence (+18a) or absence (∆18a) of alternatively-

Figure 4
Figure 4 Effects of alternate C-terminal exon47 with or without exon18a on Ca V 2.2 expression in soma and neurites of hippocampal neurons.

Figure 5
Figure 5 Non-synonymous SNPs in short exon47 do not influence cell surface or total expression of Ca V 2.2.A Representative tsA-201 cells expressing GFP-Ca V 2.2-HA having the long (i-iii) or short-Arg ancestral (iv-vi), short-Leu (vii-ix) or short-His (x-xii) Rs2278973 variants of C-short-exon47.All variants contain exon18a and are in the presence of α 2 δ-1 and β1b.Cells were incubated with anti-HA antibody in non-permeablised conditions to show HA staining on the

Figure 6
Figure 6Non-synonymous SNPs in short exon47 do not influence function of CaV2.2.

4 ± 1 .
21 and -5.56 ± 1.13 for C-long, C-short-Arg, C-short-Leu and C-short-His variants respectively.C G max (nS/pF) from the IV relationships shown in B. GFP_Ca V 2.2-HA C-long and C-short-Arg conditions data are replotted from Fig. 2 for comparison.Individual data (same symbols as in B) and meanD±DSEM are plotted.Statistical significance was determined using one-way ANOVA followed by Šidák post hoc test for multiple comparisons: ****PD<D0.0001,ns: nonsignificant.