A Ca(v)3.2/syntaxin-1A signaling complex controls T-type channel activity and low-threshold exocytosis.

Results: Syntaxin-1A interacts with the carboxy-terminus domain of Ca v 3.2 channels and modulates channel activity and low-threshold exocytosis. Conclusion: Low-threshold exocytosis relies to a syntaxin-1A/T-type calcium channel signaling the molecular ABSTRACT T-type calcium channels represent a key pathway for Ca 2+ entry near the resting membrane supports a unique role of these channels in fast and low-threshold exocytosis in an action potential-independent manner, but the underlying mechanisms the existence of calcium channel signaling complex that relies on molecular determinants that are distinct from the synaptic protein interaction site ( synprint ) found in synaptic high-voltage-activated calcium


INTRODUCTION
Ca 2+ entry through the high-voltageactivated (HVA) Ca v 2.2 (N-type) and Ca v 2.1 (P/Qtype) Ca 2+ channels into presynaptic nerve terminals supports a transient Ca 2+ microdomain that is essential for synaptic exocytosis (1)(2)(3). To ensure fast and efficient neurotransmitter release, the vesicle-docking / release machinery must be located near the source of Ca 2+ entry. In many cases, this close localization is achieved by direct interaction of SNARE proteins with the Ca 2+ channels (4)(5)(6). Indeed, syntaxin-1A/1B, SNAP-25 and synaptotagmin-1 specifically interact with Ca v 2.1 and Ca v 2.2 channels by binding to a synaptic protein interaction site (synprint) located within the intracellular II-III linker region of the channels (7)(8)(9)(10)(11), and disruption of Ca 2+ channel-SNARE coupling alters neurotransmitter release (6,12,13). A perhaps equally important consequence of SNARE protein-Ca 2+ channel interactions is the ability of SNARE proteins to modulate presynaptic Ca 2+ channel activity, thus fine tuning the amount of Ca 2+ that enters the synaptic terminal. Specifically, the binding of syntaxin-1A and SNAP-25 to Ca v 2.1 and Ca v 2.2 subunits shifts the voltage-dependence of inactivation toward more hyperpolarized membrane potentials to reduce channel availability (11,(14)(15)(16)(17)(18), and this effect is modulated by other members of the vesicle release complex (15,16,19,20).
In contrast with HVA channels, lowvoltage-activated (LVA) T-type Ca 2+ channels represent an important source of Ca 2+ entry near the resting membrane potential. It is established that T-type Ca 2+ currents mediate low-threshold burst discharges that occur during physiological neuronal rhythmogenesis as well as during pathological states such as epilepsy (21), and that they regulate various cellular functions such as neuronal development, dendritic integration, sensory transmission or pain signaling (for review see (22)). Furthermore, there is an increasing body of evidence that T-type channels can trigger fast and low-threshold exocytosis (23)(24)(25)(26)(27). In contrast with HVA Ca 2+ channels, all members of the Ttype channel family (i.e., Ca v 3.1, Ca v 3.2 and Ca v 3.3) lack the synprint site, and it is unclear as to how T-type channels are coupled to the exocytosis process at the molecular level.
Here, we reveal the existence of a syntaxin-1A/Ca v 3.2 T-type Ca 2+ channel complex in central neurons and show that this interaction involves different molecular determinants from those found in HVA channels. This interaction of syntaxin-1A with the Ca v 3.2 subunit potently modulates channel gating properties and appears essential for T-type channel-triggered lowthreshold exocytosis.

EXPERIMENTAL PROCEDURES
Plasmid cDNA constructs-The cDNAs constructs used in this study were human Ca v 3.1, Ca v 3.2 and Ca v 3.3 (28,29), mouse syntaxin-1A (Stx-1A) and SNAP-25 subcloned in pcDNA3 vector. The pcDNA3 construct encoding for the botulinium neurotoxin C1 (BoNT/C1) was generously provided by Dr. Robert G. Tsushima and Dr. Herbert Gaisano. The Stx-1A construct lacking the transmembrane domain (Stx1A TM , amino acids 1 to 265) in pCMV5 was kindly provided by Dr. Randy D. Blakely. The Stx1A"Open" in pMT2 was generated by mutating Leu 165 and Glu 166 to Ala as previously described (11).

Heterologous
Expression and Electrophysiology-Human embryonic kidney tsA-201 cells were grown in a Dulbecco's modified Eagle's culture medium containing 10% fetal bovine serum and 1% penicillin/streptomycin (all products were purchased from Invitrogen) and maintained under standard conditions at 37°C in a humidified atmosphere containing 5% CO 2 . Cells were transfected using the jetPEI™ transfection regent (Qbiogen, OH, USA) with either the Ca v 3.1, Ca v 3.2 or Ca v 3.3 channel, along with a green fluorescent protein (pEGFP, Clontech, CA, USA). Patch-clamp recordings were performed 48-72 hrs after transfection as previously described (30) in 3 the whole-cell configuration of the patch-clamp technique at room temperature (22-24°C) in a bathing medium containing (in millimolar): 5 BaCl 2 , 5 KCl, 1 MgCl 2 , 128 NaCl, 10 TEA-Cl, 10 D-glucose, 10 HEPES (pH 7.2 with NaOH). Patch pipettes were filled with a solution containing (in millimolar): 110 CsCl, 3 Mg-ATP, 0.5 Na-GTP, 2.5 MgCl 2 , 5 D-glucose, 10 EGTA, 10 HEPES (pH 7.4 with CsOH), and had a resistance of 2-4 M . Whole-cell patch-clamp recording were performed using an Axopatch 200B amplifier (Axon Instruments, Union City, CA). All traces were corrected on-line for leak currents, digitized at 10 KHz and filtered at 2 KHz. Intramembrane charge movement measurements were recorded using an HEKA-10 patch clamp amplifier (HEKA Electronic, Lambrecht, Germany). The extracellular solution contained (in millimolar): 105 CsCl, 10 HEPES, 10 D-glucose, 10 TEA-Cl, 2 CaCl 2 , 1 MgCl 2 (pH 7.4 with CsOH). The intracellular solution contained (in millimolar): 130 CH 3 SO 3 Cs, 10 EGTA, 5 MgCl 2 , 10 TEA-Cl, 5 Na-ATP, 10 HEPES (pH 7.4 with CsOH). 10 mM stock solution ErCl 3 (Sigma-Aldrich, St. Louis, MO, USA) was prepared daily in deionized water and diluted in the bath solution at a concentration of 30 M prior to the experiment. Only the cells in which series resistance was below 5 M were used for measurement of gating currents. Series resistance and capacitive transients were partly compensated by built-in circuits of the EPC 10 amplifier. Data were sampled at 10 kHz and filtered at 3 kHz. The linear component of the leak current and non-compensated capacitive transient was subtracted using the P/8 procedure. Total charge transferred during each pulse was evaluated by integrating the area below gating current trace at the beginning (Q on ) and after the end of each depolarizing pulse (Q off ). For capacitance measurements, MPC 9/3L-AH cells were transfected using Lipofectamine Plus (Invitrogen), as previously reported (6,31). The Ca v 3.2 channel was co-expressed with rat 2a and rabbit 2 subunits along with either the pEGFP, CD4-EGFP-Ca v 3.2 II-III or CD4-EGFP-Ca v 3.2 Cter construct, using a ratio of 1:1:1:0.6 respectively. Recordings were performed 48-72 hrs after transfection. The external solution contained (in millimolar): TEA-Cl 140, CaCl 2 5, HEPES 10, D-glucose 10 (pH 7.3; 300 mosmoles/l) (CaCl 2 was omitted and 1mM EGTA was added for studies in 0 Ca 2+ ). Patch pipettes were filled with a solution containing (in millimolar): 13 CsCl, 120 CH 3 CO 2 Cs, 2.5 MgCl 2 , 10 HEPES, 1 EGTA, 4 Na 2 ATP, 0.1 Na 3 GTP (pH 7.2; 295-300 mosmoles/l) and had a resistance of 2-4 M . Capacitance measurements were performed using an EPC-10 patch-clamp amplifier using the "sine + dc" software lock-in amplifier method implemented in PatchMaster software. MPC 9/3L-AH cells were held at -90 mV except during membrane depolarization. The sinusoid had amplitude of 25 mV and a frequency of 1 kHz. The stimulation protocol consisted in a single depolarization to -20 mV for 80 ms. Single-pulse current data were leak-subtracted by hyperpolarizing sweeps. Capacitance changes in response to ionomycin were recorded from MPC 9/3L-AH cells transfected with either pEGFP or CD4-EGFP-Ca v 3.2 Cter construct only. Ionomycin (1µM) was dissolved in the external solution indicated above and applied using a fast perfusion system (ALA VC 3 -8SP; ALA Scientific Instruments, Inc., NY). Analysis was performed using Igor Pro software version 6.0 (Wavemetrics).

Co-immunoprecipitation
experiments-Frozen rat brain tissue was homogenized in (10% w/v) of lysis buffer (in millimolar: 150 NaCl, 50 Tris, pH 7.5 and 1% NP-40 including a protease inhibitor cocktail (Roche)). The homogenate was centrifuged at ~16 000 g for 10 min at 4°C in a micro centrifuge and supernatant collected. For immunoprecepitation, 200 µL of aliquots of supernatant pre-cleared by addition of 40 µL of a 50% slurry (v/v) of rehydrated protein A-Sepharose beads (Amersham Pharmacia), followed by rotation at 4°C for 1 hr. Pre-cleared supernatants were then separately incubated with 5 µg of either anti-Ca v 3.1, anti-Ca v 3.2 or anti-Ca v 3.3 antibodies for 16 hrs at 4°C. To collect immuneprecipitated complex, the samples were further incubated for 2 hrs with 40 µL protein A-Sepharose beads (50% slurry). The beads with immunoprecipitated complex were washed four times with ice cold lysis buffer and resuspended in 40 µL of 2X SDS PAGE sample buffer. Heated at 70°C for 15 min and the soluble proteins were then resolved by SDS-PAGE and analyzed by Western blotting. Affinity-purified, rabbit polyclonal antibodies recognizing Ca v 3.1, Ca v 3.2 and Ca v 3.3 4 were purchased from Alomone Lab and mouse monoclonal syntaxin 1A antibody was from Chemicon International (Temecula, CA, USA). Co-immunoprecipitation experiments from tsA-201 cells were performed from cells expressing Stx1A-myc or SNAP-25 along with the CD4-EGFP-Ca v 3.2 Cter fusion protein were solubilized with 400 L of 50mM Tris/HCl pH 7.5, 0.05 mM EDTA and 10 mM Chaps Buffer supplemented with proteases inhibitor (1X, Roche) and lysed by sonication. Lysates were cleared by centrifugation at ~8000 g for 10 min. All steps were carried out on ice or at 4°C. Lysates were incubated for 3 hrs with 15 L biotinylated anti-Myc antibody (Santa Cruz) or for 16 hrs with anti-GFP antibody (1/1000, Abcam) and then for 45 min with streptavidine beads or protein G beads (Invitrogen) at 4°C and were washed twice with PBS (1X) Tween 0.1% and twice with PBS (1X). The beads were finally resuspended in PBS (1X). Samples were resolved on Tris-glycine SDSpolyacrylamide gels and transferred to nitrocellulose membranes using standard methods. Membranes were blocked in 10% (w/v) milk in PBS (0.1% (v/v) Tween-20) and then incubated with either anti-GFP or anti-SNAP-25 primary antibodies for 2 hrs at room temperature. After being washing once for 5 min, membranes were incubated with horseradish peroxidase conjugated protein A (Abcam). The washed membranes were developed using enhanced chemiluminescence.
nRT culture, immunostaining and confocal imaging-The Nucleus reticularis area was isolated from lateral thalamus of P0-P1 rat pups, cut into small pieces and then digested in containing papain (Worthington, LS003126) culture media. After digestion, the tissue was washed and triturated for neuron dissociation. nRT neurons were seeded at low density onto coverslips pretreated with poly-D-lysine (Sigma, P7280) followed by Laminin (Sigma, L2020) in 24-well plate. Culture medium consisted of BME (Invitrogen, 21010) supplemented with 1% B-27 supplement (Invitrogen, 17504-044), 2mM L-glutamine (Invitrogen, 25030-049), 1mM sodium pyruvate (Invitrogen, 11360-070), 10 mM HEPES(Sigma, H7523), 0.6% D-glucose (Sigma, G7528), 1% pen strep (Invitrogen, 15140), and 5% fetal bovine serum (Invitrogen, 26140). The next day after plating, the existing media was replaced by fresh culture media, after that every 4-5 days half of the media was removed and replaced by fresh culture media. For immunostaining, cells were fixed with 4% paraformaldehyde in PBS for 20 min at room temperature, washed and permeabilized in PBST (PBS containing 0.02% Triton X-100) for 3 min, and blocked with blocking buffer (3% BSA in PBS) for 30 min. Cells were then incubated overnight at 4°C with primary polyclonal rabbit anti-Ca v 3.2 (Alomone) and mouse monoclonal anti-syntaxin-1A (Sigma) antibodies diluted at 1:400 and 1:500, respectively, in blocking buffer. Cells were then washed, incubated 1h at room temperature with anti-rabbit Alexa488-conjugated and anti-mouse Alexa555-conjugated secondary antibodies, washed and mounted on glass slide. Confocal images were acquired with a Zeiss LSM510 Meta microscope and analysed using ImageJ software. The intensity correlation analysis was performed using the ImageJ plugin previously described (32).

RESULTS
Syntaxin-1A associates with Ca v 3.2 channels in central neurons -The existence of a relationship between T-type Ca 2+ channels and neurotransmitter release (23-26) prompted us to examine the possible colocalization of these channels with SNARE proteins in central neurons. Isolated rat reticular thalamic neurons (nRT) that are known to express predominantly the Ca v 3.2 isoform of T-type channels (33) were immunolabeled, and T-type channel and syntaxin-1A distribution analyzed by confocal microscopy. The anti-Ca v 3.2 antibody labeling (green) revealed a somatodendritic pattern consistent with previous electrophysiological studies (34)(35)(36). Interestingly, intense syntaxin-1A labeling was also similarly observed along the somatodendritic axis (red) and visual inspection of the overlaid images and colocalized pixels suggests colocalization of the proteins (Fig. 1A). We used the intensity correlation analysis (ICA) method to test for a 5 staining relation between Ca v 3.2 channel and syntaxin-1A as previously described (32). The ICA plots of Ca v 3.2 and syntaxin-1A performed both on the soma and on dendrite regions were strongly skewed toward positive values (Fig. 1B), consistent with a highly dependent staining pattern. Furthermore, the calculated intensity correlation quotient values (ICQ) were positive and highly significant (+0.30 ± 0.01 and +0.31 ± 0.01 for the soma and dendrite respectively (n = 12; p < 0.001). These data thus strongly suggest that a pool of syntaxin-1A and Ca v 3.2 channels are present at the same subcellular level. We then investigated by coimmunoprecipitations whether T-type calcium channels and syntaxin-1A associate with each other at the protein level. As shown in Figure 1B, specific anti-Ca v 3.1, Ca v 3.2 and Ca v 3.3 antibodies precipitated syntaxin-1A from rat brain homogenate, thus suggesting the existence of Ttype Ca 2+ channel/syntaxin-1A complexes (Fig.  1C).  Table S1). In syntaxin-1A expressing cells, the half-inactivation potential of Ca v 3.2 channels was shifted to more hyperpolarized potentials by -16.3 ± 1.2 mV (p < 0.001) (to -86.9 ± 1.2 mV, n = 28) compared to control cells (-70.6 ± 0.5 mV, n = 27). The syntaxin-1A-mediated modulation of Ca v 3.2 inactivation was abolished upon cleavage of syntaxin-1A by co-expression of the botulinium neurotoxin C1 (BoNT/C1) ( Fig. 2C and Table S1), indicating that the membrane insertion domain of syntaxin-1A was necessary for the observed functional effects. Consistent with this observation, co-expression of a syntaxin-1A construct lacking the transmembrane domain was unable to modulate Ca v 3.2 channel inactivation ( Fig. 2C and Table S1). Similar results were obtained with Ca v 3.1 and Ca v 3.3 channels, indicating that syntaxin-1A regulation is conserved across the entire T-type channel family (Fig. S1).

Syntaxin-1A modulates
It is well established that syntaxin-1A exists in both "open" and "closed" conformations (37), and that syntaxin-1A state is an important determinant of N-type channel modulation (11). Similar to previous observations in N-type channels, co-expression of a mutant syntaxin-1A construct that is locked permanently in the "open" conformation failed to induce a hyperpolarizing shift of the steady-state in half-inactivation potential (Fig. 2C), indicating that syntaxin-1A mediated modulation of Ca v 3.2 channel is syntaxin state-dependent.

Syntaxin-1A interacts within the carboxyterminal domain of the Ca v 3.2 subunit -Previous
studies have shown that functional interaction of N-and P/Q-type Ca 2+ channels can be disrupted by a peptide containing the synprint site (12,13). In order to investigate the Ca v 3.2 molecular determinants of syntaxin-1A regulation, we examined the capability of the different intracellular regions of the Ca v 3.2 subunit to prevent syntaxin-1A regulation. T-type Ca v 3.2 channels were coexpressed in tsA-201 cells along 6 with syntaxin-1A +/-a construct in which the major intracellular regions of the channels were fused to a CD4 epitope as well as the green fluorescent protein (EGFP) (Fig. 3A). Barium currents were recorded at -20 mV from a holding potential of -85 mV (where 56% of the channels are tonically inactivated in the presence of syntaxin-1A versus 11% in the absence of syntaxin-1A) before and after a 5 sec hyperpolarizing prepulse to -130 mV. A 5 sec hyperpolarizing prepulse to -130 mV produced a 2.1-fold increase in current amplitude in cells expressing syntaxin-1A (versus 1.1-fold increase in the absence of syntaxin-1A) due to the recovery of channels from inactivation (Fig. 3B). This prepulse-induced current facilitation was found significantly decreased upon co-expression of the CD4-EGFP-Ca v 3.2 Cter fusion protein (from 2.1 ± 0.2-fold (n = 20) to 1.4 ± 0.1 (n = 17), p < 0.001), whereas expression of the CD4-EGFP-Ca v 3.2 Cter construct by itself did not affect Ba 2+ currents in the absence of syntaxin-1A (indicating that the CD4-EGFP-Ca v 3.2 Cter construct by itself does not alter the voltage-dependence of inactivation of Ca v 3.2 channels) (Fig. 3B). In contrast, expression of the other intracellular regions of the channel failed to prevent syntaxin-1A regulation (Fig. 3B), suggesting a potential interaction of syntaxin-1A within the carboxy-terminus of the channel. Consistent with a possible specific binding of syntaxin-1A within the C-terminal domain of Ca v 3.2, co-expression of syntaxin-1A in COS cells promoted translocation of a EGFP-Ca v 3.2 Cter fusion protein to the plasma membrane whereas the EGFP-Ca v 3.2 II-III linker fusion protein remained diffusely expressed inside the cell (Fig. 3C). The trafficking study was performed on COS cells because they present the advantage to have a smaller nucleus than tsA-201 cells and more diffuse cytoplasm, allowing a better analysis of protein trafficking and plasma membrane staining. Finally, co-immunoprecipitation experiments from tsA-201 cells using the CD4-EGFP-Ca v 3.2 Cter fusion protein revealed that syntaxin-1A specifically interacts with the carboxy-terminus domain of the Ca v 3.2 subunit (Fig. 3D), which is in stark contrast with the reported syntaxin-1A interaction site on HVA channels (II-III loop). Altogether, these findings demonstrate that syntaxin-1A modulates Ca v 3.2 channel activity by interacting within the C-terminal domain of the channel and that co-expression of the putative interaction site of the channel competitively inhibits syntaxin-1A-induced channel modulation.
Like syntaxin-1A, SNAP-25 was also found to associate with the Ca v 3.2 C-terminus (Fig.  S3). While SNAP-25 by itself produced only slight effects on T-type channel activity, its coexpression reversed the effects of syntaxin-1A on channel inactivation properties (Fig. S3 and Table  S1), in analogy with previous observations with Ntype channels (15,16).
Ca v 3.2-SNARE complexes contribute to low-threshold exocytosis -Besides conferring a regulation of channel availability, the interaction between syntaxin-1A and T-type channels may also serve as a means for optimizing the coupling between T-type channel-mediated Ca 2+ entry and secretion. To test for this possibility, Ca v 3.2 channels were over-expressed in MPC 9/3L-AH cells, a chromaffin cell line already fully endowed with the vesicle release machinery, but devoid of voltage-gated Ca 2+ channels (6). Voltagedependent exocytosis was evaluated by monitoring membrane capacitance changes (indicative of vesicle fusion) in response to a 100 ms depolarizing step to -20 mV from a holding potential of -90 mV. Voltage-dependent activation of over-expressed Ca v 3.2 channels produced a robust increase in exocytosis (estimated by normalizing capacitance changes by Ca 2+ charge density) which was reduced by 96% (p < 0.001) when SNARE proteins were uncoupled from the channel by co-expression of the CD4-EGFP-Ca v 3.2 Cter fusion protein (Fig. 4C-D), even though Ca 2+ entry per se was not affected (1.9 ± 0.7 pC/pF, (n = 10) versus 1.8 ± 0.4 pC/pF, (n = 15)) ( Fig. 4A-B). In contrast, co-expression of the CD4-EGFP-Ca v 3.2 II-III linker fusion protein, which does not interact with syntaxin-1A, had no effect on Ca v 3.2-induced voltage-dependent exocytosis. The observation that Ca 2+ entry was not affected upon co-expression of the CD4-EGFP-Ca v 3.2 Cter fusion protein is consistent with the notion that syntaxin-1A modulation is abolished when Ca v 3.2 channels are functionally coupled to a full SNARE complex.
To confirm that the decrease in exocytosis observed upon expression of the CD4-EGFP-Ca v 3.2 Cter fusion protein is due to the uncoupling of 7 Ca v 3.2 channels from endogenous SNARE proteins and not to a direct distortion of the vesicle release machinery, exocytosis was evaluated in EGFP-and CD4-EGFP-Ca v 3.2 Cter -expressing MPC 9/3L-AH cells (in the absence of Ca v 3.2 channel) in response to a Ca 2+ ionophore. Exposition of EGFP-expressing cells to ionomycin (1 µM) elicited an increase in membrane capacitance that was significantly reduced by 94% (p < 0.001) after removal of extracellular Ca 2+ (273 ± 95 fF, (n = 13) versus 17 ± 11 fF, (n = 9)) ( Fig. 5A-B). However, neither the magnitude (273 ± 95 fF, (n = 13) versus 288 ± 81 fF, (n = 11), p = 0.37), nor the kinetic (tau) of the exocytosis response (9.0 ± 2.1 sec versus 8.9 ± 1.9 sec, p = 0.98) was affected by overexpression of the CD4-GFP-Ca v 3.2 Cter construct ( Fig. 5A-C). These results indicate that the negative impact of the CD4-EGFP-Ca v 3.2 Cter fusion protein on the voltage-dependent exocytosis from MPC 9/3L-AH cells expressing Ca v 3.2 channels is not due to a direct distortion of the exocytotic machinery by the fusion protein but rather to the uncoupling of the channel from endogenous SNARE proteins.
Altogether, these findings indicate that Ca v 3.2 channels have the propensity to support low-threshold exocytosis, and that this process requires a close proximity between the vesicle release machinery and the channel.

DISCUSSION
In the present study, we provide compelling evidence for the existence of a syntaxin-1A/Ca v 3.2 T-type Ca 2+ channel signaling complex. We demonstrate that this interaction not only potently modulates Ca v 3.2 channel gating but also regulates T-type channel mediated exocytosis.
It was previously shown that Ca v 2.1 and Ca v 2.2 Ca 2+ channels colocalize with syntaxin-1A at presynaptic nerve terminals (38)(39)(40) and that these channels can be isolated as a complex with SNARE proteins (41)(42)(43). We show that similarly to Ca v 2.1 and Ca v 2.2 channels, Ca v 3.2 channels also colocalize with syntaxin-1A along the somatodendritic axis of nRT neurons, and that these two proteins can be co-immunoprecipitated, thus suggesting the existence of a native syntaxin-1A/Ca v 3.2 channel signaling complex. Moreover, in vitro binding experiments indicate that syntaxin-1A interacts within the carboxy-terminal domain of the Ca v 3.2 subunit, which indicates that T-type channels use different structural determinants compared with the synprint site utilized by members of the HVA channel family. At this point it is unclear whether distinct syntaxin-1A structural determinants are involved in the interactions with Ca v 3 and Ca v 2 channels.
Electrophysiological analyses of Ca v 3.2 channels in tsA-201 cells revealed that syntaxin-1A potently decreases channel availability by shifting the steady-state inactivation toward more hyperpolarized potentials. This is qualitatively similar to what has been reported for N-and P/Qtype Ca 2+ channels. Our observation that coexpression of the carboxy-terminal region of the Ca v 3.2 subunit prevents syntaxin-1A-mediated modulation of Ca v 3.2 channel strongly suggest that syntaxin-1A modulates channel activity by virtue of its direct binding to the carboxy-terminal of the Ca v 3.2 subunit, rather than the mere presence of a syntaxin-1A molecule in the cell. The notion that syntaxin-1A-mediated modulation of Ca v 3.2 channel inactivation is abolished by botulinium neurotoxin C1 or upon deletion of the syntaxin-1A transmembrane domain indicates that membrane anchoring of syntaxin-1A is critical for channel modulation, and is analogous with previous observations on HVA channels (44). The inability of the "open" form of syntaxin-1A to alter T-type channel gating also parallels previous observations in Ca v 2.2 (11). Given that syntaxin-1A undergoes a conformational switch from a "closed" state to an "open" during the vesicle release cycle (37,45,46), this suggests the possibility that syntaxin-1A may be able to dynamically regulate Ca v 3.2 channel availability during various stages of exocytosis. The observation that SNAP-25 coexpression could reverse the syntaxin-1A mediated effect on channel gating is reminiscent of findings with Ca v 2.2 (but not Ca v 2.1) channels (14)(15)(16). At this point, we do not know whether this is due to a direct competition between SNAP-25 and syntaxin-1A for the same binding site within the C-terminus region of the channel, or whether both proteins can bind to the channel simultaneously. Irrespective of these structural information, the syntaxin-1A mediated alteration in the voltage-dependence of inactivation and the ability of SNAP-25 to regulate this process may represent an important regulatory pathway for T-type Ca 2+ channels. This is further 8 underscored when considering that these channels are ~50% inactivated near neuronal resting membrane potentials such that even small changes in the voltage dependence of inactivation can bring about dramatic changes in Ca 2+ entry (47).
Besides modulating T-type Ca 2+ channel gating, the physical association between Ca v 3.2 channel and SNARE proteins may contribute significantly to optimize Ca v 3.2-mediated exocytosis. Consistent with previous studies showing that Ca v 2.1, Ca v 2.2 or Ca v 3.1 channels can trigger voltage-dependent exocytosis when heterologously expressed in MPC-9/3L cells (6,31,48), we show that expression of Ca v 3.2 channel in these cells induced robust voltagedependent exocytosis that was drastically reduced when SNARE proteins were competitively uncoupled from the channel by co-expression of the carboxy-terminal domain of the Ca v 3.2 subunit. This result is consistent with previous studies showing that disruption of N-and P/Q-type Ca 2+ channels-SNAREs interaction by a peptide containing the synprint site prevents neurotransmitter release (12,13). It was proposed that uncoupling of Ca 2+ channels from SNARE proteins may not only lead to a spatial dissociation of vesicles and channels, but also to a loss of feedback regulation that collectively alter exocytosis efficiency (49,50). In this context, our data fit with the idea that T-type Ca 2+ channels support low-threshold exocytosis by virtue of their close association with the SNARE protein machinery. Such tight coupling may be particularly important for T-type channels due to their small unitary conductance and low open probability.
In conclusion, although HVA and T-type channels utilize completely different channel structural determinants for their interactions with the SNARE machinery, they are functionally regulated by syntaxin-1A in a strikingly similar manner. This may perhaps underscore the fundamental evolutionary importance for providing tight control over (and localizing the exocytosis machinery close to the source of) Ca 2+ entry. Overall, our findings reveal unrecognized new insights into the regulation of T-type Ca 2+ channels by SNARE proteins and provide the first molecular mechanism by which these channels contribute to low-threshold exocytosis.