Reversal of Neuropathic Pain in Diabetes by Targeting Glycosylation of Ca_v3.2 T-Type Calcium Channels Open Access

Ethics approval was obtained for all experimental protocols from the University of Virginia Animal Care and Use Committee, Charlottesville, Virginia. All experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals adopted by the National Institutes of Health. Every effort was made to minimize animal suffering and the number of animals used. We used our standard procedure for testing mechanical and thermal sensitivity as we previously described (5). Statistical comparisons were made using one- and two-way repeated ANOVAs (paw and time postinjection) followed by Holm-Sidak multiple comparison with statistical significance accepted at P < 0.05. All drug injections were performed in a blinded manner. ECN [(3β,5α,17β)-17-hydroxyestrane-3-carbonitrile] was dissolved in 15% β-cyclodextrin ([2-hydroxypropyl]-β-cyclodextrin) (Cyc) solution (Sigma-Aldrich), and 750 μL i.p. solution containing ECN or vehicle alone was injected.

Patch-clamp recordings from acutely dissociated DRG neurons and human embryonic kidney (HEK)-293 cells were described in detail in our previous publication (6). The external solution for voltage-clamp experiments in HEK-293 cell experiments contained (in millimoles) 152 TEA-Cl, 2–10 BaCl₂, and 10 HEPES, adjusted to pH 7.4 with tetraethyl ammonium hydroxide (TEA-OH). For voltage-clamp experiments in DRG cells, we used 2 mmol/L Ca²⁺ in external solution instead of Ba²⁺. The external solution for current-clamp experiments and recordings of voltage-gated sodium currents contained (in millimoles) 140 NaCl, 4 KCl, 2 MgCl₂, 2 CaCl₂, 10 glucose, and 10 HEPES, adjusted to pH 7.4. The internal solution for voltage-clamp experiments with DRG neurons contained (in millimoles) 135 TMA-OH, 40 HEPES, 10 EGTA, and 2 MgCl₂, adjusted to pH 7.2 with hydrogen fluoride (7). The internal solution for voltage-clamp experiments with HEK-293 cells contained (in millimoles) 110 Cs-MeSO₄ 14 creatine phosphate, 10 HEPES, 9 EGTA, 5 Mg-ATP, and 0.3 Tris-GTP, adjusted to pH 7.3 with CsOH. The internal solution for current-clamp experiments contained (in millimoles) 130 KCl, 40 HEPES, 5 MgCl₂, 2 Mg-ATP, 1 EGTA, and 0.1 Na₃GTP, adjusted to pH 7.3 with KOH.

Statistical comparisons were made using paired or unpaired t tests where appropriate. All data are expressed as means ± SEM; P values are reported only when statistically significant (<0.05).

cDNA encoding the human Cav3.2 gene was subcloned into the mammalian expression vector pDoubleTrouble (8), resulting in hexahistidine and FLAG tags at the NH₂-terminal of the channel.

HEK-293 cells were cultured in Dulbecco’s modified Eagle’s medium containing 10% FBS. Stable cell lines expressing the epitope-tagged Cav3.2 channel were selected using the above media containing 500 μg/mL G418. Alternatively, the pDoubleTrouble vector containing Cav3.2 was used to transiently transfect cells. Fifteen-centimeter plates were transfected with 35 μg each of Cav3.2 DNA and Lipofectamine 2000 at a ratio of 1:1 and harvested at 48 h.

Purified _6HIS/FLAGCav3.2 channel was separately incubated with PNGase F (New England Biolabs) and endoglycosidases (Endo) F1, F2, and F3 (Sigma) for 1 h at 37°C according to the manufacturers’ instructions. Reactions were terminated by the addition of sample buffer.

Methods for purification of Ca_V3.2 channels and SDS-PAGE and immunoblotting are described in the Supplementary Data.

Confocal images were acquired using a Zeiss 510UV confocal microscope and LSM ImageExaminer software. Single-plane images were taken using a 100× oil-immersion objective. For quantification of images, the colocalization plug-in for ImageJ (NIH) was used.

Ca_v3.2 N-glycosylation mutants (N192Q, N271Q, and N1466Q) were generated by mutating Asn¹⁹², Asn²⁷¹, and Asn¹⁴⁶⁶ residues of human Ca_v3.2 (GenBank accession no. AF051946) into Gln residues using two-step PCR methods. The specific methods are described in the Supplementary Data.

Recombinant Ca_V3.2 channels expressed heterologously in HEK-293 cells are typically cultured in a high-glucose medium to provide the cells’ energy supply. Interestingly, we determined that the glucose level in Invitrogen’s Dulbecco’s modified Eagle’s medium routinely used to grow these cells is 315 mg/dL (17.5 mmol/L) glucose—a value similar to blood glucose levels in diabetic ob/ob mice (5). We hypothesized that exposure of recombinant Ca_V3.2 channels to neuraminidase (NEU) (1.5 units/mL for 1–3 h at 37°C), an enzyme that deglycosylates proteins by removing sialic acid residues, would alter T-currents. We first characterized the effects of NEU on the current-voltage (I-V) relationship. We found that depolarization to different test potentials (V_t) from a holding potential (V_h) of −90 mV resulted in similar normalized I-V relationships in control (top traces, Fig. 1A) and NEU-treated (bottom traces, Fig. 1A) cells (untreated cells, n = 18; NEU-treated cells, n = 13 [Fig. 1B and C]). Nevertheless, we found a visible slowing of macroscopic current kinetics of activation (as measured by 10–90% rise time) and inactivation (as measured by inactivation τ) after treatment with NEU in comparison with control, untreated cells (Fig. 1C and D) with a dramatic twofold slowing of the τ of inactivation. We also examined the effects of PNGase-F (PNG) (20 units/cc for 12 h at 37°C), an enzyme that selectively cleaves N-glycosylated groups on proteins. Like NEU, PNG exposure slowed the kinetics of current activation and inactivation by twofold (gray columns, Fig. 1E). Additionally, we found that treatment with NEU or PNG also reduced peak current density by ~40% but that in combination, these agents were not more effective (Fig. 1F). Taken together, these data strongly suggest that NEU and PNG may share a common mechanism of modification of Ca_V3.2 channels.

Extracellular loops of ion channels may contain multiple potential motifs for glycosylation, which typically consist of an asparagine (Asn, N) residue separated from a serine (Ser, S) or threonine (Thr, T) residue by one amino acid (9). In Ca_V3.2 channels, conserved extracellular Asn residues in positions 192 (N192) and 271 (N271) of domain I and 1466 (N1466) of domain III are excellent candidate sites for N-glycosylation (Fig. 2A). We hypothesized that Ca_V3.2 channels with mutated critical Asn residues reared in high-glucose medium would have slower macroscopic kinetics of activation and inactivation than wild-type (WT) Ca_V3.2 channels. To test this hypothesis, we generated single-point mutants of Ca_V3.2 channels in which one Asn reside was mutated into a glutamine (Q), namely, N192Q, N1466Q, and N271Q. Using transiently transfected HEK-293 cells, we first compared the sensitivity of currents carried by N192Q mutant and WT Ca_V3.2 channels to nickel, a traditional Ca_V3.2 blocker. Traces depicted in Fig. 2B show that 100 μmol/L nickel applied in the external solution almost completely blocked inward currents carried by WT (top traces) and mutant (bottom traces) channels. Hence, N192Q mutation did not alter a signature pharmacological property of Ca_V3.2 channels. Nevertheless, consistent with a functional role for putative N192 glycosylation in channel activity, macroscopic current kinetics of activation and inactivation of N192Q mutant channels (n = 15) were significantly slower than those of WT channels (n = 22; P < 0.001) (Fig. 2C). Interestingly, current density (V_h −90 mV, V_t −30 mV) in N192Q channel mutant (17 ± 5 pA/pF, n = 13) was not different from WT Ca_V3.2 channel (15 ± 3 pA/pF, n = 7, P > 0.05). In contrast to the N192Q channel mutant, currents were not evident in recordings of N271Q and N1466Q channel mutants (n ≥ 10, data not shown).

We next hypothesized that glycosylation state may regulate plasma membrane expression of Ca_V3.2 channels. To test this hypothesis, we constructed green fluorescent protein–tagged WT Ca_V3.2 channels (enhanced green fluorescent protein [EGFP]-Ca_V3.2) and mutant channels containing disrupted putative glycosylation motifs, (EGFP-N192Q, EGFP-N271Q, and EGFP-N1466Q), expressed them in HEK-293 cells, and used confocal microscopy to quantify immunofluorescence. Figure 3 illustrates a representative experiment from each channel type (WT Ca_V3.2 [Fig. 3A] and mutant channels [Fig. 3B–D]) showing red immunofluorescence for the structural membrane protein concanavalin A (left panels, Fig. 3A–D) and green immunofluorescence for the EGFP-tagged channel protein (middle panels, Fig. 3A–D). As indicated by the merging of red and green fluorescence (yellow color on right panels of Fig. 3A–D), we found that EGFP colocalized well with concanavalin A for WT (Fig. 3A) as well as N192Q (Fig. 3B) and N271Q (Fig. 3C) channel mutants consistent with predominant plasma membrane expression of these channel constructs. In contrast, the EGFP-N1466Q mutant channel was expressed predominantly in cytoplasmic organelles in agreement with lack of colocalization of the fluorescent signals (right panel, Fig. 3D). Average data from similar experiments were quantified blinded with arbitrary colocalization values and are summarized in Fig. 3E. Colocalization values were not significantly different for N192Q and N271Q mutants compared with WT Ca_V3.2 channels. By contrast, there is a large decrement in colocalization for N1466Q mutant: 28% of that of WT Ca_V3.2 channels (n = 4, P < 0.01), corroborating the lack of T-currents measured in our patch-clamp experiments in HEK cells expressing the N1466Q mutant.

Since two of our point mutations are located in repeat I of Ca_V3.2 channels (Fig. 2A), we generated NH₂-terminal FLAG-taggedCa_V3.2 channels (_6HIS/FLAGCa_V3.2) to enable ensuing biochemical studies. We used the FLAG tag to immunoprecipitate the _6HIS/FLAGCa_V3.2 channel from HEK-293 cells grown in high glucose. Next, we treated the purified _6HIS/FLAGCa_V3.2 channel with enzymes such as PNG and Endo F1, F2, and F3. Deglycosylation of proteins can result in a faster electrophoretic mobility and, thus, a lower apparent molecular weight. In the case of the full-length Ca_V3.2 channel (~260 kDa), very little change in electrophoretic mobility was observed after treatment with these enzymes (top arrow, Fig. 4), likely due to the large size of the protein. However, a small NH₂-terminal fragment of the Ca_V3.2 channel recognized by the FLAG antibody shifted its electrophoretic mobility after treatment with PNG and Endo F1 consistent with a change in apparent molecular weight from 60 kDa to 50 kDa (bottom arrow, Fig. 4). In contrast, treatments of _6HIS/FLAGCa_V3.2 channel purified from HEK-293 cells with enzymes that are known to cleave complex carbohydrate molecules, such as Endo F2 and Endo F3, did not affect electrophoretic mobility of the NH₂-terminal fragment of Ca_V3.2 (bottom arrow, Fig. 4). Thus, our data strongly suggest that the Ca_V3.2 channel is glycosylated with a relatively simple sugar moiety within the first ∼500 residues (50 kDa).

Next, we asked whether glycosylation may modulate native Ca_V3.2 currents in DRG nociceptors and may participate in painful PDN. Based on our recent studies using ob/ob mice, the upregulation of T-currents in small DRG neurons coincided with significant hyperglycemia, morbid obesity, and the development of painful PDN in mice aged 10–16 weeks (5). Thus, we first compared the biophysical properties of T-currents in acutely dissociated small DRG cells from diabetic ob/ob and healthy WT mice at age 10–16 weeks. Representative traces of T-currents in DRG cells from control (Fig. 5A) and ob/ob mice (Fig. 5B) indicate marked enhancement of their amplitudes in ob/ob mice over the wide range of membrane potentials. We also found significant (up to twofold) speeding of the kinetics of macroscopic current inactivation and activation in ob/ob mice (n = 20 cells) compared with the WT group (n = 27 cells) at all potentials tested (n ≥ 6 mice) (Fig. 5C and D). In contrast, there was no apparent difference between the two groups in the voltage dependence of activation (Fig. 5E) or inactivation (Fig. 5F) of T-currents. Shown in Fig. 6 are representative T-currents in DRG cells treated with NEU isolated from WT (top traces, Fig. 6A) or ob/ob (bottom traces, Fig. 6A) mice. Bar graphs in Fig. 6 summarize average data from similar experiments. Interestingly, treatment with NEU (filled columns) did not significantly change T-current density in cells from WT mice but effectively decreased T-current density in the ob/ob group to WT levels (∼17 pA/pF) (Fig. 6B). Furthermore, NEU significantly slowed both WT and ob/ob control T-current kinetics to similar levels: τ of inactivation to ∼105 ms and 10–90% rise time of ∼28 ms (Fig. 6C and D).

We also tested the possibility that NEU may affect other parameters besides T-currents in ob/ob mice using voltage- and current-clamp recordings from small DRG cells of diabetic ob/ob mice. In these experiments, we blindly assigned saline-treated and NEU-treated dishes from the same preparations of DRG cells. Cells were first voltage clamped, and inward currents were evoked with a series of depolarizing pulses of 200 ms duration (V_h = −90 mV, V_t = −60 mV through 0 mV). We found no significant difference in the peak amplitude of voltage-gated sodium currents in two groups: saline 230 ± 28 pA/pF (n = 25 cells) and NEU 307 ± 46 pA/pF (n = 18 cells, P > 0.05, n ≥ 6 mice per group). In ensuing current-clamp experiments in some of these cells, we injected a series of incremental hyperpolarizing pulses and we measured passive membrane properties such as soma diameter, membrane capacitance, input resistance, and resting membrane potential. Supplementary Table 1 summarizes these data and indicates that NEU treatment did not have significant effects on any of these parameters (P > 0.05).

We previously demonstrated that selective T-channel blocker ECN (10) at the dose of 25 mg/kg i.p. effectively reversed hyperalgesia in diabetic ob/ob mice and diabetic WT mice but had no effect on nociception in diabetic Ca_V3.2 knockout (KO) mice (5). However, the effects of NEU on pain thresholds in whole animals are not known. Hence, we examined whether NEU modifies in vivo sensitivity to mechanical and noxious thermal (heat) stimuli. We first measured baseline paw withdrawal responses (PWRs) in the presence of a mechanical stimulus elicited by von Frey filament (time point 0) and then injected (arrows, Fig. 7) 10 μL NEU or vehicle in the hind paws of WT (n = 5) or diabetic ob/ob (n = 10) mice. We then measured PWRs in both paws at 30, 60, and 90 min after injection. As confirmed in Fig. 7A and B, ob/ob mice had increased baseline paw withdrawal latencies (PWLs) of ~60% compared with their age-matched WT counterparts (Fig. 7C and D), suggestive of prominent mechanical hyperalgesia (5,11). Importantly, intraplantar (i.pl.) injection of 1.5 units/mL of NEU, the same concentration used in our in vitro experiments, completely reversed mechanical hyperalgesia in the injected (right) paws of ob/ob mice (Fig. 7A). In contrast, saline injections did not cause any significant changes in PWRs (Fig. 7A). Interestingly, there was a transient hyperalgesic response to injections of NEU in WT mice, while saline injections did not cause any alterations in baseline PWRs (Fig. 7C). PWLs in uninjected (left) paws of ob/ob mice (Fig. 7B) and WT mice (Fig. 7D) remained stable throughout testing, indicating a lack of systemic effect. To test whether the antihyperalgesic effect of NEU in ob/ob mice is related to its effects on T-currents, we injected 25 mg/kg i.p. ECN ~60 min before i.pl. injections of NEU (n = 8) (Fig. 7E). As expected, the injections of ECN completely reversed mechanical hyperalgesia in ob/ob mice, as evidenced by a notable decrease of mechanical PWRs in both paws measured 60 min after injections (P < 0.001). Importantly, subsequent i.pl. injections of NEU did not further decrease PWRs in right paws. Unlike with ECN, the i.p. injections of Cyc (vehicle) did not change baseline mechanical PWRs in ob/ob mice and subsequent local injections of NEU into right paws significantly decreased PWRs up to ~50% (Fig. 7F). PWLs in uninjected (left) paws of the same ob/ob mice remained stable throughout testing, indicating a lack of systemic effect of NEU (Fig. 7F).

Next, we measured baseline PWL in the presence of a radiant heat stimulus (time point 0) and then injected (arrows, Fig. 8) 10 μL NEU or vehicle i.pl. in 16-week-old WT (n = 7) or ob/ob (n = 7) mice. ob/ob mice (top panel, Fig. 8A) had decreased baseline PWLs (4.0 ± 0.2 s) for ~20% compared with their age-matched WT counterparts (5.1 ± 0.3 s, P < 0.01) (bottom panel, Fig. 8A), indicating mild heat hyperalgesia (5). Importantly, injection of 1.5 units/mL NEU i.pl. produced a significant antihyperalgesia, as evident by an increase in PWLs only in ob/ob mice at 30 min after injection (top panel, Fig. 8A) but not in WT mice (bottom panel, Fig. 8A). In control experiments, the same volume of saline (vehicle) had no effect on thermal PWLs in either WT or ob/ob mice (n ≥ 7, data not shown). As depicted in Fig. 8B, the intraperitoneal injections of ECN had analgesic effects in ob/ob mice as evidenced by a notable increase of thermal PWLs in both paws (n = 4, P < 0.001). Importantly, subsequent i.pl. injections of NEU did not further prolong PWLs in right paws.

It has been established that T-channels can contribute to the hyperexcitability of sensory neurons manifested by hyperalgesia and allodynia, two frequent symptoms of chronic neuropathic pain (4,12). Several studies have validated that plasticity of T-channels is implicated in hyperalgesia and allodynia in animal models of PDN (5,12–17). Taken together, these studies identify an important pronociceptive role of Ca_V3.2 T-channels in neuropathic pain in animal models of type 1 and type 2 diabetes.

However, molecular mechanisms for the alteration of Ca_V3.2 channels in DRG cells from diabetic animals have not previously been described. Indeed, herein, we provide evidence for the first time that targeting glycosylation states of T-channels may be a promising new treatment for painful PDN. This conclusion is based on several observations from our study. First, we show that macroscopic current activation and inactivation kinetics as well as current density are drastically reduced when recombinant human Ca_V3.2 channels expressed in HEK-293 cells reared in hyperglycemic cell culture medium are exposed to NEU and PNG. Second, the effects of NEU are more prominent in DRG cells from diabetic ob/ob mice than in healthy WT mice. Third, NEU in vivo completely reversed thermal and mechanical hyperalgesia in diabetic ob/ob mice, whereas it was completely ineffective in age-matched healthy WT mice. The fact that NEU modified DRG T-current kinetics from healthy WT mice but to a lesser extent than in ob/ob mice suggests that the level of T-channel glycosylation is an important physiological mechanism that fine-tunes the activity of these channels in pain pathways. However, it appears that this process is maladaptive and leads to cellular hyperexcitability and, consequently, hyperalgesia in diseases like PDN. Hence, glycosylation of Ca_V3.2 channels is an important mechanism of sensitization of peripheral nociceptors that could be exploited for novel pain therapies.

Our molecular studies identify conserved extracellular asparagine residues, most notably N192 and N1466, as important regulators of Ca_V3.2 current kinetics and channel membrane expression, respectively. This is supported by our patch-clamp recordings that demonstrated slower current kinetics in N192Q Ca_V3.2 mutant with apparently normal membrane expression. In contrast, we could not consistently record T-currents in HEK-293 cells transfected with N1466Q Ca_V3.2 mutant, and our imaging studies with the EGFP-tagged mutant showed minimal membrane expression. Thus, different glycosylation sites in Ca_V3.2 channels may have distinct functional roles. Surprisingly, we could not record T-currents from N271Q Cav3.2 channels despite apparently normal membrane expression. It remains possible that N271Q channels trafficked to the membrane are nonfunctional. During the review of our study, another in vitro study using recombinant human Ca_V3.2 channels also examined the effect of glycosylation on T-current kinetics and surface membrane expression (18). Surprisingly, they found that treatments with PNG but not NEU decreased Ca_V3.2 current density and slowed kinetics of channel inactivation. Furthermore, their work suggests that asparagine N192 serves as a regulator of channel membrane expression and asparagine N1466 as a regulator of channel kinetics. It is possible that specific conditions of enzymatic deglycosylation or different levels of glucose in cell culture could have contributed to the different findings between the studies. However, regardless of observed differences, our study directly demonstrates that Ca_V3.2 -channels are indeed glycosylated within domain I of the channel protein and for the first time reveals the prominent effects of NEU on native T-currents in DRG cells and on pain perception in vivo using an animal model of PDN.

Several other in vitro studies have reported that glycosylation may modulate properties of other voltage-gated ion channels. For example, in embryonic DRG neurons, NEU affected steady-state inactivation of voltage-gated sodium channels (19). While future biophysical studies may reveal fine details of the effects of glycosylation upon Ca_V3.2 current kinetics, it is reasonable to propose that increased current density and increased kinetics of Ca_V3.2 current activation alone may contribute to the hyperexcitable state of DRG cells under hyperglycemic conditions. Similar to the results of our study, the findings of Tyrrell et al. (19) did not show any effects of NEU on voltage-gated sodium channels in small DRG cells from adult animals. Future extensive electrophysiological studies could be expanded to involve examination of other voltage-gated channels that are crucial for the control of cellular excitability of DRG cells from diabetic animals that might also be modulated by glycosylation (20–22).

Previous in situ hybridization studies (23) and electrophysiological studies using KO mice (6,24) have established that the Ca_V3.2 is the most prevalent isoform of the Cav3.0 family in small DRG cells. Thus, we have focused our investigation on the effects of glycosylation on this particular isoform and on the peripheral sensitization of pain responses in PDN. Our molecular studies have found that extracellular asparagine residues N192 and N1466 are likely putative substrates for glycosylation that alter T-channel membrane expression and current kinetics. Since these asparagine residues are conserved across all T-channel isoforms, it is likely that glycosylation may similarly modulate the other two T-channel isoforms, namely, Ca_V3.1 and Ca_V3.3. Interestingly, all three isoforms of T-type channels are expressed in dorsal horn neurons of spinal cord (23), and recent studies have shown that they all may support spinal nociceptive processing in different animal models of neuropathic pain (12,25,26). Thus, simultaneous glycosylation of multiple T-channel isoforms in spinal dorsal horn neurons may contribute to their hyperexcitability, which in turn may influence central sensitization of pain responses that is implicated in many pain disorders (27).

Overall, the results presented here fundamentally advance our understanding of the mechanisms of glycosylation underlying the posttranslational modification of Ca_V3.2 T-channels that has an important function in supporting peripheral nociceptive signaling. Our results strongly suggest that the manipulation of glycosylation states of peripheral nociceptors could be useful for the development of novel therapies for the treatment of painful PDN. This method may have an advantage over direct blockers of T-channels to suppress pain because NEU and related agents will correct the pathology of diabetes at its source rather than ameliorating the problem through separate pathways that also may be a source of unintended side effects. Our goal is to provide novel therapeutic modalities that would not only alleviate neuropathic pain in patients with diabetes but, even more importantly, also halt its progression without causing dangerous systemic side effects or creating the potential for drug abuse.

This research is supported by American Diabetes Association 7-09-BS-190 (to S.M.T.), Dr. Harold Carron Endowment Fund, Priority Research Centers Program through the National Research Foundation of Korea (2012-0006690 to J.H.L.), National Institutes of Health grants 7-HL-036977 (to P.Q.B.) and GM-47969 (to D.F.C.), and funds from the Department of Anesthesiology at InJe University.

No potential conflicts of interest relevant to this article were reported.

P.O. researched data, wrote the manuscript, and contributed to discussion. H.P.O. researched data. W.E.M. researched data and contributed to and reviewed and edited the manuscript. M.O.J., R.S., M.M.J., W.J.C., J.L., S.-S.L., K.E.R., N.P., M.R.D., and K.K. researched data. D.F.C. researched data, contributed to discussion, and reviewed and edited the manuscript. J.-H.L. contributed to discussion and reviewed and edited the manuscript. P.Q.B. researched data, wrote the manuscript, and reviewed and edited the manuscript. V.J.-T. researched data, contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript. S.M.T. researched data, wrote the manuscript, contributed to discussion, and reviewed and edited the manuscript. S.M.T. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in abstract form at the 39th Annual Society for Neuroscience Meeting, Chicago, Illinois, 17–21 October 2009; the 40th Annual Society for Neuroscience Meeting, San Diego, California, 13–17 November 2010; the 2nd Conference on Calcium Channel Research, Placencia, Belize, 28 March–3 April 2010; and the 3rd Conference on Calcium Channel Research, Krabi, Thailand, 24–29 March 2013.

The authors thank Mr. Damir Bojadzic for technical assistance and Dr. Jan Redick and Dr. Stacey Guillot of the Advanced Microscopy Core Facility (University of Virginia) for help with imaging studies.