The intracellular C-terminus confers compartment-specific targeting of voltage-gated Ca2+ channels

To achieve the functional polarization that underlies brain computation, neurons sort protein material into distinct compartments. Ion channel composition, for example, differs between axons and dendrites, but the molecular determinants for their polarized trafficking remain obscure. Here, we identify the mechanisms that target voltage-gated Ca2+ channels (CaVs) to distinct subcellular compartments. In hippocampal neurons, CaV2s trigger neurotransmitter release at the presynaptic active zone, and CaV1s localize somatodendritically. After knockout of all three CaV2s, expression of CaV2.1, but not of CaV1.3, restores neurotransmitter release. Chimeric CaV1.3 channels with CaV2.1 intracellular C-termini localize to the active zone, mediate synaptic vesicle exocytosis, and render release fully sensitive to blockade of CaV1 channels. This dominant targeting function of the CaV2.1 C-terminus requires an EF hand in its proximal segment, and replacement of the CaV2.1 C-terminus with that of CaV1.3 abolishes CaV2.1 active zone localization. We conclude that the intracellular C-termini mediate compartment-specific CaV targeting.


Introduction
Neurons are polarized cells with a defined signaling directionality from dendrites to soma to axon 1 .To achieve this morphological and functional polarization, neurons sort protein material into specific subcellular compartments 2,3 .Voltage-gated Ca 2+ channels (CaVs), which couple electrical activity to changes in intracellular Ca 2+ signaling, are a prototypical example of sorting specificity.They are a large protein family, and individual members localize to distinct subcellular domains in the dendrites, the soma and the axon 4,5 .However, CaV subtypes exhibit limited differences in their sequences, and the molecular determinants that target CaVs to specific subcellular compartments remain elusive.
The mechanisms that distinguish CaV1s from CaV2s and sort them into the somatodendritic and axonal compartments, respectively, remain unclear.Starting from their primary site of synthesis in the soma, CaVs likely undergo a series of interactions that target each subtype to its respective subcellular domain 2,26 .However, CaVs are highly similar in structure 5,27,28 , and notable overlap exists within the CaV1 and CaV2 interactome.For example, interactions with CaVβ, CaVα2δ, and calmodulin have been implicated in CaV trafficking [29][30][31][32][33][34] , but these proteins interact indiscriminately with CaV1s and CaV2s and are thus unlikely to encode specific sorting information.The intracellular CaV C-termini might mediate targeting specificity.CaV C-termini include a proximal segment with two EF hands and an IQ motif, and a distal segment containing binding sites for scaffolding proteins (Figs.S1A+B).The CaV2 C-terminus binds to the PDZ domain of the active zone protein RIM, and it contains a proline-rich sequence (which is also present in CaV1s) that binds to RIM-BP 24,35,36 .Together, these interactions help tether CaV2s to the presynaptic active zone 20,24,[37][38][39][40][41][42] .Analogous sequences in CaV1.3 bind to the postsynaptic scaffold Shank, and overall, CaV1 C-termini support cell surface expression and the assembly of CaV1 into dendritic clusters 43,44 .An additional poly-arginine motif specific to CaV2.1 may also contribute to its localization 20,45 .Sequences outside the C-terminus could also be involved.For example, binding of the CaV2 cytoplasmic II-III loop to SNARE proteins [46][47][48] and CaV interactions with material in the synaptic cleft may mediate anchoring at presynaptic sites 49,50 .
Taken together, multiple interactions have been implicated in CaV trafficking and targeting, but how these interactions direct CaV1s and CaV2s to opposing compartments has remained unclear.
Here, we found that the CaV C-termini are the primary determinants of channel localization in hippocampal neurons.Swapping the CaV2.1 C-terminus onto CaV1.3 targets the channel to the presynaptic active zone in CaV2 knockout neurons.This chimeric CaV1.3 channel mediates Ca 2+ entry for neurotransmitter release and renders synaptic vesicle exocytosis sensitive to L-type CaV blockers.In contrast, the inverse swap prevents active zone localization of CaV2.1.Within the CaV2.1 proximal C-terminus, an EF hand is required for presynaptic targeting, and its removal leads to loss of CaV2.1 from the active zone.We conclude that the C-terminus specifies CaV localization, and we identify the EF hand as an essential trafficking motif.

Lentivirally expressed CaV2.1, but not CaV1.3, localizes to active zones and mediates neurotransmitter release after CaV2 ablation
To determine the CaV sequences important for active zone localization, we expressed various CaVs using lentiviruses in cultured hippocampal neurons that lack CaV2.1, CaV2.2 and CaV2.3.Specifically, we transduced neurons that contain "floxed" conditional knockout alleles for these three channels (Fig. 1A) with lentiviruses that express cre recombinase under a synapsin promoter to generate CaV2 cTKO neurons 20 .Control neurons (CaV2 control) were identical except for transduction by a lentivirus expressing a truncated, recombination-deficient version of cre.In addition, we transduced CaV2 cTKO neurons with either a lentivirus expressing HAtagged CaV2.1 or with a lentivirus expressing HA-tagged CaV1.3.The tags were inserted near the CaV N-terminus in a position shown previously to not interfere with the expression (Figs. 1B, S1A-1E), targeting and function of CaV2.1 20,51 .We then used stimulated emission depletion (STED) microscopy (Fig. 1C-H), confocal microscopy (Fig. S1F-I), and electrophysiology (Fig. 1I-L) to assess CaV localization and synaptic transmission.
For morphological analyses, neurons were stained with antibodies against CaV2.1 or HA to detect CaVs, PSD-95 to mark postsynaptic densities, and synapsin to label synaptic vesicle clusters.For STED analyses (Fig. 1C-H), we selected synapses in side-view through the presence of a vesicle cloud (imaged with confocal microscopy) and an elongated PSD-95 structure (STED) at one edge of the vesicle cloud, as established previously 20,25,38,39,52 .We assessed CaV distribution and levels (STED) in these side-view synapses using line profiles drawn perpendicular to the PSD-95 structure, and we plotted the average line profiles (Fig. 1D+G) and peak intensities (Fig. 1E+H).
Endogenous and re-expressed CaV2.1 formed elongated structures apposed to PSD-95 with a maximal intensity within tens of nanometers of the PSD-95 peak (Fig. 1C-H).We have established before that this distribution is characteristic of active zone localization 20,25,39,53 .Furthermore, a strong PSD-95 peak was present in all conditions, matching our previous work that did not find morphological defects following CaV2 triple knockout 20 .Exogenously expressed CaV1.3, monitored via the HA-tag, was not detected at the active zone (Fig. 1F-H).Consistent with the STED analyses, robust levels of CaV2.1, but not CaV1.3,were present in synaptic regions of interest (ROIs) defined by synapsin (Fig. S1F-I).Independent of their synaptic targeting, both CaV2.1 and CaV1.3 were effectively expressed in the somata of transduced CaV2 cTKO neurons and in transfected HEK293T cells (Fig. S1C-E).
These morphological experiments were complemented with analyses of synaptic transmission in the same conditions (Fig. 1I-L).A focal stimulation electrode was used to evoke action potentials, and inhibitory or excitatory postsynaptic currents (IPSCs or EPSCs) were isolated pharmacologically.EPSCs were monitored via NMDA receptors because network excitation confounds the interpretation of EPSC amplitudes when AMPA receptors are not blocked.CaV2 cTKO nearly abolished synaptic transmission, as characterized in detail before 20 .Reexpression of CaV2.1 restored EPSCs and IPSCs effectively, but exogenous expression of CaV1.3 failed to produce any recovery (Fig. 1I-L), in agreement with the absence of CaV1.3 from presynaptic sites (Fig. 1F-H).Taken together, these results establish that CaV2.1, but not CaV1.3,localizes to the active zone and gates neurotransmitter release when expressed in CaV2 cTKO neurons.
We then assessed the localization of these chimeric channels in the experimental setup described above and compared them side-by-side with CaV2.1 and CaV1.3.Strikingly, translocating the CaV2.1 C-terminus onto CaV1.3 efficiently targeted the resulting chimeric CaV1.3 2.1Ct channel to the active zone in CaV2 cTKO neurons, as assessed with STED microscopy (Fig. 2B-D).The distribution profile of CaV1.3 2.1Ct and its abundance at the active zone recapitulated those of re-expressed CaV2.1 (Fig. 2B-D).In contrast, the inverse swap abolished active zone localization of CaV2.1 1.3Ct (Fig. 2B-D) despite effective somatic expression (Fig. S2B+C).Confocal microscopic analyses of CaV levels in synaptic ROIs corroborated these findings by revealing robust synaptic localization of CaV1.3 2.1Ct but not of CaV2.1 1.3Ct (Fig. 2E+F).
These results establish that CaV1.3 is targeted to the presynaptic active zone when its Cterminus is replaced with that of CaV2.1.Conversely, CaV2.1 loses its active zone localization following the reverse swap.We conclude that the CaV C-termini contain sufficient information to define CaV compartment specificity, and these and previous data lead to two predictions.First, because removing known scaffolding motifs in the distal C-terminus only partially impaired active zone localization 20,45 , there must be essential targeting motifs in the CaV C-terminus that have not yet been identified.Second, if the chimeric CaV1.3 2.1Ct channel is appropriately coupled to primed vesicles within the active zone, then CaV1.3 2.1Ct expression should restore synaptic transmission in CaV2 cTKO neurons and render neurotransmitter release sensitive to L-type channel blockade.We next tested both predictions.

An EF hand in the proximal C-terminus is necessary for CaV2 active zone targeting
Removal of the known active zone scaffolding motifs in the CaV2.1 C-terminus produces a partial defect in CaV2.1 active zone targeting, but truncation of the entire C-terminus fully abolishes active zone localization 20 .To define C-terminal sequences that contain unidentified targeting motifs, we segregated the CaV2.1 C-terminus into a distal segment containing the active zone scaffolding motifs, and the complementary proximal segment (Fig. S1A+B)..1ProxCt ) or only the CaV2.1 distal C-terminus (CaV1.3 2.1DistCt ) onto CaV1.3.Both CaV1.3 2.1ProxCt and CaV1.3 2.1DistCt were expressed efficiently in HEK293T cells after transfection (Fig. S3A) and in neuronal somata after lentiviral transduction (Fig. S3B+C).With STED microscopy, we detected CaV1.3 2.1ProxCt at the active zone (Fig. 3B-D) of CaV2 cTKO neurons.Active zone levels of CaV1.3 2.1ProxCt were reduced compared to CaV1.3 2.1Ct and resembled those of a mutant CaV2.1 that lacks the active zone scaffolding motifs in the distal Cterminus 20 .Hence, active zone targeting of chimeric CaV1.3s operates in part through these distal sequences.Accordingly, CaV1.3 2.1DistCt exhibited strong active zone localization in CaV2 cTKO neurons and was indistinguishable from CaV1.3 2.1Ct (Fig. 3B-D).Confocal analyses of protein levels in synaptic ROIs matched these findings (Fig. S3D+E).
CaV1.3 2.1ProxCt demonstrates that translocation of the CaV2.1 proximal C-terminus onto CaV1.3 suffices to mediate some active zone localization (Fig. 3B-D) and indicates that the proximal Cterminal sequences are important for presynaptic trafficking.The CaV proximal C-termini (Fig. S1A+B) contain two EF hands 54,55 .The first EF hand has been implicated in calmodulindependent modulation of CaV function [56][57][58] , though no evidence to date establishes a role in CaV trafficking.We tested whether the first EF hand mediates active zone targeting by deleting the first EF hand from CaV2.1 (CaV2.1 ΔEF1 , Fig. 3E).CaV2.1 ΔEF1 was readily expressed in transfected HEK293T cells and detected in somata of lentivirally transduced neurons (Fig. S3F-H).
In summary, the CaV2.1 distal C-terminus needs to be paired with proximal C-terminal elements to effectively localize CaVs to the active zone.Our data establish that the proximal EF hand is required for active zone targeting of CaV2.1.

CaV2 ablation
Efficient neurotransmitter release requires that CaVs are coupled to fusion-competent synaptic vesicles.Having demonstrated that translocation of the CaV2.1 C-terminus directs CaV1.3 to the active zone, we next asked whether the chimeric CaV1.3 2.1Ct channel provides Ca 2+ for action potential-triggered release (Fig. 4A).CaV1.3 2.1Ct expression in CaV2 cTKO neurons indeed resulted in EPSCs (Fig. 4B+C) and IPSCs (Fig. 4D+E) that were indistinguishable from those measured from CaV2 cTKO neurons with re-expressed CaV2.1.In contrast, and consistent with the loss of active zone targeting (Fig. 2B-D), CaV2.1 1.3Ct failed to restore synaptic transmission (Fig. 4B-E).
It is possible that the presynaptic targeting and function of CaV1.3 2.1Ct results from removal of a dendritic targeting sequence rather than addition of an axonal targeting motif.To address this possibility, we generated a CaV1.3 lacking the entire C-terminus (CaV1.3ΔCt ).CaV1.3 ΔCt was effectively expressed (Fig. S4A-D) but was not targeted to synapses (Fig. S4E+F) or active zones (Fig. S4G-I).Furthermore, CaV1.3 ΔCt did not mediate neurotransmitter release (Fig. S4J-M).We conclude that active zone targeting of CaV1.3 2.1Ct arises from an instructive role of the CaV2.1 C-terminus.
At central synapses, neurotransmitter release is insensitive to L-type CaV blockade (Fig. S5) 17 .

Discussion
Voltage-gated Ca 2+ channels are a prototypical protein family to illustrate neuronal polarization: distinct CaVs are sorted effectively to dendritic, somatic and axonal compartments.Here, we establish that the CaV C-termini contain the necessary and sufficient information to sort CaVs into specific subcellular compartments.Within the C-terminus of CaV2.1, the proximal EF hand is essential for presynaptic targeting and it operates in concert with distal scaffolding motifs.Multiple cargo selectivity filters converge within the endoplasmic reticulum, the Golgi apparatus, the axon initial segment, and the presynaptic bouton that together permit the targeting of a limited subset of proteins to the active zone while deflecting other cargo 60,61 .Sequence motifs within these proteins may dictate compartment sorting at two major checkpoints: (1) they may mediate protein recruitment into cargo vesicles that are directed to the axon, and (2) they may stabilize proteins at the active zone following their delivery 2,62 .Our work establishes that the CaV2.1 C-terminus encodes necessary and sufficient information to navigate these two checkpoints and implies a cooperative relationship between the proximal and distal elements.The CaV2.1 distal C-terminus efficiently localizes chimeric CaV1.3s to the active zone, indicating that the distal C-terminal sequences permit both CaV sorting into presynaptic cargo and CaV tethering at the active zone, so long as a proximal EF hand is present.The distal motifs that bind to active zone proteins likely fulfill these roles as disrupting their interactions with RIM and RIM-BP leads to targeting defects 20,24,36,37,45 similar to those exhibited by chimeric CaV1.3s with the CaV2.1 proximal C-terminus and the CaV1.3 distal C-terminus (Fig. 3).
The efficiency with which the chimeric CaV1.3 2.1Ct and CaV1.3 2.1DistCt channels are targeted to the active zone establishes that the proximal C-termini of both CaV1.3 and CaV2.1 contain necessary information for active zone CaV delivery.This is in line with the high homology of the EF hands and IQ-motif across CaV proximal C-termini and with the presence of these sequences in other voltage-gated channels 55,63 .The proximal C-terminus might include multiple instructive signals that together inform CaV targeting.The EF hand binds to AP-1 and possibly Ca 2+ , which could provide for a trafficking control checkpoint 64,65 .Calmodulin binds to the IQ motif and might regulate channel trafficking and function 10,33,34,56 .Other unknown interactions with these sequences or with sequences elsewhere in the proximal C-terminus might be involved in targeting as well.Altogether, we posit that the proximal EF hand is necessary for passing a trafficking checkpoint that permits incorporation of these CaVs into axon-bound cargo, but likely has no role in stabilizing CaVs within the active zone.
Our work on CaVs provides mechanistic insight into the polarized trafficking of protein material in neurons and raises multiple questions.First, some synapses depend on only a single CaV2 subtype while others redundantly use multiple CaV2s, and some synapses experience developmental switches in their CaV2 usage 66,67 .Whether there are specific trafficking and anchoring mechanisms or whether these properties are determined wholly by switches in gene expression remains to be determined.Second, the proximal sequences we identified as important for targeting are also present in other ion channels that undergo polarized trafficking, for example in neuronal Na + channels 55,63 .It is possible that the mechanisms we describe for CaVs are broadly employed across channel proteins.The example of CaVs forms an ideal framework to build on and further define mechanisms that sort proteins into specific neuronal compartments.

Primary neuronal cultures
Primary mouse hippocampal cultures were generated from newborn mice as described previously 20,38,39 .Hippocampi were dissected out from newborn mice within 24 h after birth.
HEK293T cell batches were typically replaced after 20 passages by thawing a fresh vial from the expanded stock.

Lentiviruses
Lentiviruses used to transduce primary hippocampal neurons were produced in HEK293T cells.
HEK293T cells were transfected with the Ca 2+ phosphate method with REV (p023), RRE (p024) and VSVG (p025), as well as a lentiviral plasmid encoding the protein of interest.For CaV proteins of interest, these were plasmids p789, p947, p1077, p1078, p1079, p1080, p1083, and p1084.To produce lentiviruses expressing EGFP-tagged Cre recombinase (to generate CaV2 cTKO neurons), pFSW EGFP-Cre (p009) was used.For lentiviruses expressing a truncated, enzymatically inactive EGFP-tagged Cre (to generate CaV2 control neurons), pFSW EGFP-ΔCre (p010) was used.Plasmids were transfected at a 1:1:1:1 molar ratio and with a total amount of 6.7 μg DNA.Approximately 24 h after transfection, the medium was switched to neuronal growth medium (described above), and the HEK293T cell supernatant was harvested 24-36 h later by centrifugation at 700 x g.For expression of EGFP-Cre and EGFP-ΔCre, neurons were infected by adding HEK293T cell supernatant at DIV5.For expression of CaVs, neurons were infected at DIV1.CaV2 control neurons were additionally infected with a virus made using a pFSW plasmid (p008) lacking a cDNA in the multiple cloning site in place of an expression virus.Neuronal protein expression from these lentiviruses was driven by a human synapsin promoter 38,70 .
Neurons cultured on 0.17 mm thick 12 mm diameter (#1.5) coverslips were washed two times with PBS warmed to 37 °C, and then fixed in 2% PFA + 4% sucrose (in PBS) at room temperature.After fixation, coverslips were rinsed three times in PBS + 50 mM glycine, then permeabilized in PBS + 0.1% Triton X-100 + 3% BSA (TBP) for 1 h at room temperature.
Coverslips were stained with primary antibodies diluted in TBP for ~48 h at 4 °C.The following primary antibodies were used: mouse IgG1 anti-HA (1:500, RRID: AB_2565006, A12), rabbit anti-CaV2.1 (1:200, RRID: AB_2619841, A46), guinea pig anti-PSD-95 (1:500, RRID: AB_2619800, A5), rabbit anti-synapsin (1:500, RRID: AB_2200097, A30), and mouse IgG1 antisynapsin (1:500, RRID_2617071, A57).After primary antibody staining, coverslips were rinsed twice and washed three times for 5 min in PBS + 50 mM glycine at room temperature.Alexa Fluor 488 (to detect HA-tagged CaVs or endogenous CaV2.1; anti-mouse IgG1, RRID: AB_2535764, S7; or, anti-rabbit, RRID: AB_2576217, S5), Alexa Fluor 555 (to detect the postsynaptic marker PSD-95; anti-guinea pig, RRID: AB_2535856, S23), and Alexa Fluor 633 (to detect the synaptic vesicle cloud; anti-rabbit, RRID: AB_2535731, S33; or, anti-mouse IgG1, RRID: AB_2535768, S29) conjugated antibodies were diluted in TBP at 1:200 (for Alexa Fluor 488 and 555) or 1:500 (for Alexa Fluor 633), and coverslips were incubated with the secondary antibody solution for ~24 h at 4 °C.Coverslips were then rinsed twice with PBS + 50 mM glycine and once with deionized water, air-dried and mounted on glass slides in fluorescent mounting medium.Confocal and STED images were acquired on a Leica SP8 Confocal/STED 3X microscope with an oil immersion 100x 1.44 numerical aperture objective and gated detectors as described previously 20,72 .58.14 x 58.14 μm 2 areas were acquired using 2x digital zoom (4096 x 4096 pixels, pixel size of 14.194 x 14.194 nm 2 ).Alexa Fluor 633, Alexa Fluor 555, and Alexa Fluor 488 were excited at 633 nm, 555 nm and 488 nm using a white light laser at 1-10% of 1.5 mW laser power.The Alexa Fluor 633, Alexa Fluor 555, and Alexa Fluor 488 channels were acquired first in confocal mode.For the Alexa Fluor 555 and Alexa Fluor 488 channels, the same areas were then sequentially acquired in STED mode using 660 nm and 592 nm depletion lasers, respectively.Identical imaging and laser settings were applied to all conditions within a given biological repeat.For analyses of presynaptic CaV distribution in STED images, synapses were selected in side-view.Side-view synapses were defined as synapses that contained a synaptic vesicle cluster labeled with synapsin and were associated with an elongated PSD-95 structure along the edge of the vesicle cluster as described previously 20,39,52,72,74 .For intensity profile analyses, a ∼1000 nm long, 200 nm wide, rectangular ROI was drawn perpendicular and across the center of the PSD-95 structure, and the intensity profiles were obtained using this ROI for both the protein of interest and PSD-95.To align individual profiles, the PSD-95 signal only was smoothened using a rolling average of 5 pixels, and the smoothened signal was used to define the peak position of PSD-95.The profiles for the protein of interest (CaV or HA) and smoothened PSD-95 were aligned to the PSD-95 peak position, averaged across synapses, and then plotted.Peak intensities were also analyzed by extracting the maximal value from the line profiles of the protein of interest (CaV or HA) and smoothened PSD-95 within a 200 nm window around the PSD-95 peak.Peak intensity values were plotted for each synapse and averaged.For quantification of confocal images, a custom MATLAB program (https://github.com/hmslcl/3D_SIM_analysis_HMS_Kaeser-lab_CL) was used to generate masks of the presynaptic marker (synapsin), with the threshold determined by automatic twodimensional segmentation (Otsu algorithm) 75 .Regions of interest (ROIs) were defined as synapsin-positive areas formed by contiguous pixels of at least 0.05 μm 2 in size.Each image typically contained between 500 and 1500 synapsin ROIs.Levels of HA or CaV2.1 within these ROIs were measured and the average intensity across all ROIs within an image was calculated and plotted.Representative images in figures were cropped, rotated with bi-linear interpolation, and then brightness and contrast adjusted to facilitate inspection.Brightness and contrast adjustments were made for display in figures and were done identically for images within an experiment, but image quantification was performed on raw images without these adjustments.
The experimenter was blind to the condition/genotype for image acquisition and analyses for STED and confocal microscopic experiments.

Confocal imaging of neuronal somata
Neurons cultured on 0.17 mm thick 12 mm diameter (#1.5) coverslips were washed with PBS warmed to 37 °C and fixed in 2% PFA + 4% sucrose for 10 min at room temperature.Coverslips were then rinsed three times in PBS + 50 mM glycine at room temperature, permeabilized in TBP for 1 h at room temperature, and incubated in primary antibodies at for ~48 h at 4 °C.The following primary antibodies were used: mouse IgG1 anti-HA (1:500, RRID: AB_2565006, A12) and mouse IgG2b anti-NeuN (1:500, RRID: AB_101711040, A254).After staining with primary antibodies, coverslips were rinsed twice and washed three times for 5 min in PBS + 50 mM glycine at room temperature.Alexa Fluor 555 (to detect HA; anti-mouse IgG1, RRID: 2535769, S19), and 633 (to detect neuronal somata; anti-mouse IgG2b, RRID: AB_1500899, S31) conjugated secondary antibodies were used at 1:500 dilution in TBP.Secondary antibody staining was carried out for ~24 h at 4 °C.Coverslips were rinsed twice in PBS + 50 mM glycine, once in deionized water, air-dried and then mounted on glass slides using fluorescent mounting medium.Confocal images of neuronal somata were acquired on a Leica Stellaris 5 microscope with a 63x oil-immersion objective.Single section, 92.65 x 92.65 μm 2 areas were acquired using 2x digital zoom (1024 x 1024 pixels, pixel size of 90.2 x 90.2 nm 2 ).Imaging and laser settings were identical for all conditions within a given biological repeat.For analyses of somatic HA signals, the NeuN signal was used to mark the neuron somata, and EGFP-Cre or EGFP-ΔCre was used to define nuclei.Somatic ROIs were drawn as donut shapes by using the outer edge of the NeuN profile along the main somatic compartment not including neurites, and by excluding the EGFP-labeled nucleus.The average pixel intensity within the somatic ROI was then calculated for HA and plotted for each cell.Representative images in figures were cropped and adjusted for brightness and contrast to facilitate inspection.Brightness and contrast adjustments were made for display in figures and were done identically for images within an experiment, but image quantification was performed on raw images without these adjustments.
The experimenter was blind to the condition/genotype for image acquisition and analyses.

Electrophysiology
Electrophysiological recordings in cultured hippocampal neurons were performed as described previously 20,39,74  Action potentials were elicited with a bipolar focal stimulation electrode fabricated from nichrome wire.To evaluate the CaV blocker sensitivity of synaptic transmission, ω-agatoxin IVA (to block CaV2.1) or isradipine (to block CaV1s) were used.Blockers were pipetted into the recording chamber as concentrated stocks diluted in extracellular solution for a final working concentration of 200 nM for ω-agatoxin IVA and 20 µM for isradipine.For wash-in, cells were incubated after blocker addition for 5 min.IPSCs were recorded first in the absence of CaV blockers.Then, IPSCs were measured after wash-in of 200 nM ω-agatoxin IVA and again after wash-in of 200 nM ω-agatoxin IVA and 20 µM isradipine (Fig. 4F-I), or after wash-in of 20 µM isradipine (Fig. S5).Data were acquired at 5 kHz and lowpass filtered at 2 kHz with an Axon 700B Multiclamp amplifier and digitized with a Digidata 1440A digitizer.Data acquisition and analyses were done using pClamp10.For electrophysiological experiments, the experimenter was blind to the genotype throughout data acquisition and analyses.
Homogenates were centrifuged at 16,200 x g for 10 min at room temperature, run on 6% (for CaVs) or 12% (for β-actin) polyacrylamide gels, and transferred onto nitrocellulose membranes for 6.5 h at 4 °C in transfer buffer (containing per L, 200 mL methanol, 14 g glycine, 3 g Tris).
Finally, the membranes were exposed to films, and films were developed and scanned.
Corresponding western blots of CaVs and β-actin were run simultaneously, on the same day, and on separate gels using the same samples.For illustration in figures, blots were rotated with bilinear interpolation and cropped for display.
Together, the CaV2.1 C-terminal sequences are sufficient to re-direct somatodendritic CaV1 channels to the active zone.Conversely, the CaV1.3C-terminal sequences disrupt CaV2.1 active zone localization.Our work establishes mechanisms for compartment-specific targeting of a protein family central to the polarized organization of neurons.

Figure 3 .
Figure 3.An EF hand in the proximal C-terminus is essential for CaV2 active zone

Quantification and statistical analyses
Data are displayed as mean ± SEM.Statistics were performed in GraphPad Prism 9, and significance is presented as *p < 0.05, **p < 0.01, and ***p < 0.001.Sample sizes and statistical tests for each experiment are included in each figure legend.For electrophysiological experiments, the sample size used for statistical analyses was the number of recorded cells.For STED microscopic data, the sample size used for statistical analyses was the number of synapses.For confocal microscopic data, the sample size used for statistical analyses was the number of images for analyses of synapsin ROIs, or the number of neurons for analyses of somata.Single factor, multiple group comparisons were conducted using Kruskal-Wallis tests followed by Dunn's multiple comparisons post-hoc tests for proteins of interest (HA or CaV2.1) and for current amplitudes (EPSCs, IPSCs).To compare the efficacy of blockade of synaptic transmission by different pharmacological agents in Fig. 4H, Friedman tests and Dunn's multiple comparisons post-hoc tests were used.To compare the effects of CaV blockers on synaptic transmission across genotypes in Fig. 4I, two-way, repeated-measures ANOVA and Dunnett's multiple comparisons post-hoc tests were used.In Fig. S5, the Wilcoxon matched-pairs signed rank test was used.