The C2A domain of synaptotagmin alters the kinetics of voltage-gated Ca2+ channels Ca(v)1.2 (Lc-type) and Ca(v)2.3 (R-type).

Biochemical and genetic studies implicate synaptotagmin (Syt 1) as a Ca2+ sensor for neuronal and neuroendocrine neurosecretion. Calcium binding to Syt 1 occurs through two cytoplasmic repeats termed the C2A and C2B domains. In addition, the C2A domain of Syt 1 has calcium-independent properties required for neurotransmitter release. For example, mutation of a polylysine motif (residues 189-192) reverses the inhibitory effect of injected recombinant Syt 1 C2A fragment on neurotransmitter release from PC12 cells. Here we examined the requirement of the C2A polylysine motif for Syt 1 interaction with the cardiac Cav1.2 (L-type) and the neuronal Cav2.3 (R-type) voltage-gated Ca2+ channels, two channels required for neurotransmission. We find that the C2A polylysine motif presents a critical interaction surface with Cav1.2 and Cav2.3 since truncated Syt 1 containing a mutated motif (Syt 1*1-264) was ineffective at modifying the channel kinetics. Mutating the polylysine motif also abolished C2A binding to Lc753-893, the cytosolic interacting domain of Syt 1 at Cav1.2 1 subunit. Syt 1 and Syt 1* harboring the mutation at the KKKK motif modified channel activation, while Syt 1* only partially reversed the syntaxin 1A effects on channel activity. This mutation would interfere with the assembly of Syt 1/channel/syntaxin into an exocytotic unit. The functional interaction of the C2A polylysine domain with Cav1.2 and Cav2.3 is consistent with tethering of the secretory vesicle to the Ca2+ channel. It indicates that calcium-independent properties of Syt 1 regulate voltage-gated Ca2+ channels and contribute to the molecular events underlying transmitter release.

The synaptic vesicle protein Synaptotagmin I (Syt 1), 1 is proposed to function as a Ca 2ϩ sensor for neurotransmitter release (1,2). Consistent with its proposed role as a calcium sensor protein, Syt 1 binds calcium via two repeating structures termed C2A and C2B domains (3).
Furthermore, Ca 2ϩ binding to the C2A domain promotes its insertion into membranes via an interaction with the acidic phospholipids (8,9,13,14) consistent with the Ca 2ϩ requirements of neurosecretion. Microinjection of recombinant C2A domains and antibodies specific for this region impair neurotransmitter release from neuroendorine PC12 cells (15) and giant squid axons (16). Interestingly, the inhibitory effect of recombinant C2A fragments in PC12 cells occurs independently of its calcium binding properties and is mediated through a novel polybasic motif (17). Thus, the Syt 1 C2A domain contains calcium-dependent and -independent activities, which mediate Syt 1 function during neurotransmitter release. Furthermore, Syt 1 and Syt 4 were recently shown to promote transmitter release independently of Ca 2ϩ binding to the C2A domain (18).
Here we studied the functional interaction of Syt 1 with Ca v 1.2 and Ca v 2.3 by examining the relative contribution of the C2A polylysine motif on channel activity and binding to the cytosolic domain Lc 753-893 of the ␣ 1 1.2 of Ca v 1.2. Our data indicate that the C2A domain of Syt 1 modulates the activation kinetics of Ca v 1.2 and Ca v 2.3. Mutation of the C2A polylysine motif abolished the binding to the cytosolic interaction domains of the channel. Moreover, this mutation altered the modulatory effect of Syt 1 on Ca v 1.2 and Ca v 2.3 activity, impairing the ability of Syt 1 to reverse the syntaxin 1A inhibition of channel activity.
Electrophysiological Assays-Whole cell currents were recorded at room temperature (20 -24°C) by applying a standard two-microelectrode voltage clamp using a Dagan 8500 amplifier. Voltage and current agar-cushioned electrodes (0.3-0.6 M⍀) were filled with 3 M KCl (32). Current-voltage relationships were determined by voltage steps as indicated in the legend to figures, in Ba 2ϩ solution (mM): 5 Ba(OH) 2 /50 N-methyl-D-glucamine/1 KOH/40 tetraethylammonium/5 HEPES, pH 7.5 and titrated to pH 7.5 with (CH 3 ) 2 SO 4 . The activation kinetics was determined from leak-subtracted current traces by a mono-exponential fit of the pClamp8 software (Axon Inst.). The activation time constants were determined by fitting the raw current data with the equation: indicates the amplitude of current at time t, I max is the maximum amplitude, and act is the time constant for activation. Each trace was fitted separately according to Boltzmann, and the averaged values were plotted. Ca v 1.2 activation was fitted to single exponential function, while a two exponential function nicely described the data of Ca v 2.3 time course. Data presentation was done using Origin 6 software (Microcal). All quantitative results are given as the mean Ϯ S.E. (n ϭ 6 -10) Protein Expression-Protein expression in oocytes was tested for by Western analysis 5-7 days after cRNA injection. Oocytes were homogenized in buffer containing (in mM): 1 EDTA/250 sucrose/10 Tris-HCl, pH 7.0, and addition of a mixture of protease:aprotinin, phenylmethylsulfonyl fluoride, iodoacetamide, pepstatin A, and leupeptin at 4°C. Homogenates were centrifuged (12,000 ϫ g, 10 min); the pellet was discarded and supernatant was collected. Protein was determined by a micro-Bradford assay in enzyme-linked immunosorbent assay-reader plate using bovine serum albumin as standard (38). Protein samples (30 g) mixed with 100 mM dithiothreitol and 2% SDS, boiled 3 min, applied to 10% SDS-PAGE, transferred to nitrocellulose, and probed using affinity-purified monoclonal anti-syntaxin 1A (Sigma) followed by a horseradish peroxidase-conjugated anti-mouse antibody. Syntaxin 1A expression was detected by enhanced chemiluminescence (ECL system).
Affinity Determination Using the Surface Plasmon Resonance Spectroscopy and GST Binding Assays-The affinity of Lc 753-893 , the II-III loop that links domains II-III of the Ca V 1.2 ␣1 subunit (32) and GST-C2A wt (31) or GST-C2A* mutant (17) was determined using (i) Biacore 3000 system (Biacore AB) based on surface plasmon resonance methodology and (ii) glutathione-S-transferase (GST) binding assay using GST-agarose beads.
Purified His 6 -tagged Lc 753-893 was immobilized on a research-grade CM5 sensor chip in a flow cell coated with carboxyl-methyl dextran as the surface matrix using activated carboxyl groups and EDC coupling in HBS-EP buffer (150 mM NaCl, 3.4 mM EDTA, and 0.005% (v/v) 10 mM Hepes, pH 7.4, and surfactant P20) at a flow rate of 10 l/min. The surface was activated for 7 min with a mixture of N-hydroxysuccinimide (0.05 M) and EDC (0.02 M). His 6 -tagged Lc 753-893 was injected at a concentration of 20 g/ml in 10 mM sodium acetate, pH 3.5, until the desired level of binding was achieved. Ethanolamine (1 M, pH 8.5) was injected for 7 min to block the remaining activated groups. Control flow-cell surface was prepared by activating and then deactivating (blocking) the carboxyl groups as mentioned above. His 6 Lc 753-893 binding studies to wild type and mutant C2A domains were initiated by passing the recombinant fusion proteins GST-C2A wild type, GST C2A* mutant, and GST alone at increasing concentrations as indicated through the flow cells at a rate of 20 l/min in HBS-EP running buffer. Surface regeneration was carried out after each binding assay by a 10-l pulse of 1 M NaCl in 10 mM NaOH. The data were analyzed using the Kinetics Wizard of the Biacore control software with automatic corrections for nonspecific binding by subtraction of the responses obtained for the control surface from the data obtained. The kinetics of binding and affinity constants were calculated using the Biaevaluation software.
Binding of the cytoplasmic domain of Lc 753-893 (100 nmol) to GST fusion proteins, Syt 1, C2A, C2B, C2A*, and GST alone (100 pmol) using glutathione-agarose 4B beads (25 l) was performed as described (31,32). Immunoblots were probed using affinity-purified anti-Lc 753-893 antibody and visualized by enhanced chemiluminescence (ECL system). Oocytes injected with cRNA of ␣ 1 1.2 (5 ng/oocyte), ␣2␦1 (5 ng/oocyte), ␤2a (10 ng/oocyte), and a day later with Syt 1 (5 ng/oocyte) or Syt 1* (5 ng/oocyte) are shown. Inward Ba 2ϩ currents were elicited from a holding potential of Ϫ80 mV in response to a 160-ms pulse by voltage steps to potentials between Ϫ30 and ϩ45 mV in 5-mV increments. A, leaksubtracted peak current-voltage relationship: collected data from oocytes expressing the three channel subunits (E) together with Syt 1 () or Syt 1* (ƒ). The data points correspond to the mean Ϯ S.E. of current amplitude (n ϭ 8). B, the activation component of a typical current produced at each test pulse was fitted with a single exponential function between the lines marked by asterisks. C, the averaged time constants of activation ( act , mean Ϯ S.E., n ϭ 6) are plotted against test pulses in the absence (E) and in the presence of Syt 1 () and Syt 1* (ƒ). Two sample Student's t tests were applied, and p values Ͻ0.05 were obtained from the two-tailed tests. D, protein expression of Syt 1 and Syt 1* was determined at day 5 following cRNA (5 ng/oocyte) injection into Xenopus oocytes. The proteins were separated on 10% SDS-PAGE, transferred to nitrocellulose membrane, and subjected to Western analysis using anti-Syt 1 antibody and ECL detection.

RESULTS
The C2A Domain of Syt 1 Is Required for Functional Interactions with the Voltage-gated Ca 2ϩ Channels, Ca v 1.3-Functional interactions of voltage-gated Ca 2ϩ channels with the full-length Syt 1 have been previously described using the Xenopus oocytes expression system. To assess the role of the C2A polylysine motif (amino acids 189 -192) on Syt 1 interactions with the Ca 2ϩ channel, the C2A polylysine motif was substituted with alanine residues in full-length Syt 1 (Syt 1*) or a truncated form of Syt 1 (Syt 1 1-264 ) lacking the C2B domain ( Fig. 1).
Ca v 1.2 currents were elicited in oocytes co-expressing the three-channel subunits ␣ 1 1.2/␤2a/␣2␦1 with Syt 1 or Syt 1* from a holding potential of Ϫ80 mV to test potentials between Ϫ30 and ϩ45 mV in response to 160-ms test pulse (Fig. 2). Peak current amplitudes were not affected by Syt 1 (35) or Syt 1, as FIG. 3. Syt 1 and Syt 1* interact with Ca v 1.2 in the presence of syntaxin 1A. A, superimposition of macroscopic ␣ 1 1.2, ␣2␦1, and ␤2a currents evoked from a holding potential Ϫ80 mV by a single voltage step (160 ms) to ϩ20 mV in oocytes expressing the three-channel subunits in various combinations as indicated. B, peak-current amplitudes (data not shown) normalized to maximum current (I/I max ) plotted against test potentials were fitted according to Boltzmann; channel subunits (E) with syntaxin 1A (q), syntaxin 1A and Syt 1 (), or syntaxin 1A and Syt 1* (ƒ). The mid-point of activation (V 1/2 ) and Boltzmann slope (k) of ␣ 1 1.2/␣2␦1/␤2a were V 1/2 ϭ Ϫ21.6 Ϯ 1.2 mV, k ϭ 2.9 Ϯ 1.9; with syntaxin 1A, V 1/2 ϭ Ϫ7.5 Ϯ 1.75 mV, k ϭ 7.2 Ϯ 1.44; with syntaxin 1A and Syt 1 V 1/2 ϭ Ϫ18.6 Ϯ 2.1 mV, k ϭ 2.4 Ϯ 0.8; and with syntaxin 1A and Syt 1* , V 1/2 ϭ Ϫ19.0 Ϯ 2.2 mV and k ϭ 3.9 Ϯ 2.1. C, the activation time constants ( act , mean Ϯ S.E., n ϭ 8) are plotted against test pulses between Ϫ10 and ϩ30 mV; the channel alone (E), with syntaxin 1A (q), with syntaxin 1A and Syt 1 () and D, with syntaxin 1A and Syt 1* (ƒ). Two sample Student's t tests were applied, and p values Ͻ0.05 were obtained from the two-tailed tests. See Fig. 2 for cRNA/oocyte of channel subunits, syntaxin 1A (2 ng/oocyte). Inward Ba 2ϩ currents were elicited in oocytes co-expressing ␣ 1 1.2, ␣2␦1, ␤2a, syntaxin 1A, and syntaxin 1A with Syt 1 1-264 and syntaxin 1A with Syt 1* 1-264 from a holding potential of Ϫ80 mV by voltage steps of 160 ms applied in 5-mV increments at potentials between Ϫ30 and ϩ45 mV. A, superposition of macroscopic ␣ 1 1.2, ␣2␦1, and ␤2a currents activated from a holding potential Ϫ80 by a single voltage step of 160 ms to a test potential of 0 mV in various combinations as indicated. B, leak-subtracted peak current-voltage relationship: collected data from oocytes expressing the three-channel subunits (E) with syntaxin 1A (q), syntaxin 1A and Syt 1 1-264 (f), or C, syntaxin 1A and Syt 1* 1-264 (Ⅺ). The data points correspond to the mean Ϯ S.E. of current amplitude (n ϭ 7). D, peak current amplitudes normalized to maximum current (I/I max ) are plotted against test potentials (data from B and C) and were fitted according to Boltzmann equation. The mid-point of activation (V 1/2 ) and the Boltzmann slope (k) of demonstrated by current-voltage relationship ( Fig. 2A). The activation component of Ca v 1.2 current was measured at each test pulse and was fitted with a single exponential function between the lines marked by asterisks (Fig. 2B). Under these experimental conditions both Syt 1 and Syt 1* slightly reduced activation rate at voltage range of Ϫ15 to Ϫ5 mV, while at more positive potentials act approached control values ( Fig. 2C; Table I). Lysates of oocytes co-injected with Syt 1, Syt 1*, and Ca v 1.2 were prepared and analyzed for Syt 1/Syt *1 expression by Western analysis using anti-Syt 1 antibody (see "Experimental Procedures"). As shown in Fig. 2D (28,32,37). Since the inhibitory effect of syntaxin 1A on these channels is reversed by Syt 1 (31, 32, 33) we next examined the Syt 1* mutant for reversal of syntaxin 1A inhibitory effects on channel activity. Fig. 3 shows the results of co-expressing Ca v 1.2 and syntaxin 1A with Syt 1 and Syt*1 in Xenopus oocytes. Superimposed traces of macroscopic whole cell Ba 2ϩ currents showed an 80% inhibition of current amplitude by syntaxin 1A, which was fully reversed in the presence of Syt 1 and partially by Syt 1* ( Fig. 3A; Table I). Furthermore, peak current amplitudes normalized to maximum current (I/I max ) showed a large voltage shift in the half- This voltage shift was reverted to V 1/2 ϭ Ϫ18.6 Ϯ 2 mV by Syt 1 and Ϫ19 Ϯ 2.2 mV by Syt 1* (Fig. 3B). Similarly, a complete reversal of the syntaxin 1A effect on Ca v 1.2 activation was observed with co-expression of Syt 1 (Fig. 3C), and only partial reversal by Syt 1* (Fig. 3D). The mutation at the polylysine motif impaired Syt 1* capacity to reverse the inhibitory effects of syntaxin 1A on Ca v 1.2 current amplitude and activation kinetics.
Ca v 1.2 Interacts with Syntaxin 1A and the Truncated Syt 1  and Syt 1* 1-264 Mutants-The partial reversion of the syntaxin 1A effect on Ca v 1.2 activation by Syt 1* compared with Syt 1 suggests that the polylysine C2A motif couples Syt 1 to channel activation (Fig. 3, C and D). To isolate the contribution of C2A domain we co-expressed truncated Syt 1 lacking the C2B domain (Syt 1 1-264 ) with Ca v 1.2 and syntaxin 1A. Ca v 1.2 whole cell currents were activated from a holding potential of Ϫ80 to 0 mV test pulse (Fig. 4A). Both the superimposed traces as well as the current-voltage relationships (Fig.  4, A-C) showed diminished current amplitudes by syntaxin 1A that were only partially reversed by Syt 1 1-264 and Syt 1* 1-264 . Syt 1* 1-264 was significantly less effective than Syt 1  . Furthermore the large shift in the half-maximal voltage induced by syntaxin 1A, (see above) was shifted back to V 1/2 ϭ Ϫ20.6 Ϯ 2.3 mV by Syt 1 1-264 and only to Ϫ12.9 Ϯ 3 mV by Syt 1* 1-264 (Fig. 4D).
A more striking difference between Syt 1 1-264 and Syt 1* 1-264 was observed on channel activation (Fig. 4, E and F). The marked slowing effect of activation kinetics by syntaxin 1A was fully reversed by Syt 1 1-264 , (Fig. 4C; Table I). In contrast, Syt 1* 1-264 was completely ineffective (Fig. 4E). Together, these results suggest the involvement of the C2A polylysine motif in the interaction with the channel. Oocytes were injected with cRNA of ␣ 1 2.3 (5 ng/oocyte), ␣2␦1 (5 ng/oocyte), ␤2a (10 ng/oocyte), and a day later, with Syt 1 (5 ng/oocyte) or Syt 1* (5 ng/oocyte). Inward Ba 2ϩ currents were elicited from a holding potential of Ϫ80 mV in response to an 80-ms pulse to various test potentials between Ϫ30 and ϩ45 mV in 5-mV increments. A, leak-subtracted peak current-voltage relationship: collected data from oocytes expressing the three channel subunits (E) together with Syt 1 () or Syt 1* (ƒ). The data points correspond to the mean Ϯ S.E. of current (n ϭ 8). B, the activation component of a typical current produced by a test pulse was fitted with a single exponential function between the lines marked by asterisks and was applied to determine the time constant of activation ( act ). C, activation time constants ( act , mean Ϯ S.E., n ϭ 6) are plotted against potentials between Ϫ20 and ϩ30 mV in the absence (E) and in the presence of Syt 1 () or D, Syt 1* (ƒ). Two sample Student's t test were applied, and p values Ͻ0.05 were obtained from the two tailed tests domain is exposed on the surface of the ␤-sandwich of Syt 1 where they are accessible for interacting with potential effector molecules (7). To determine whether the loss of functional interaction with the channel is related to impaired binding to Lc 743-893 , the II-III linker of the Ca v 1.2 ␣1 subunit (32) f. Two types of binding studies of C2A and mutant C2A* domains  , and ␤2a current traces evoked in response to an 80-ms pulse from a holding potential of Ϫ80 mV by a single voltage step to a 0-mV test pulse in oocytes co-expressing the three-channel subunits alone and together with either Syt 1 or Syt 1*. B, leak-subtracted peak current-voltage relationship: collected data from oocytes expressing the three-channel subunits (E) with syntaxin 1A (q), syntaxin 1A and Syt 1 (-), and C, syntaxin 1A and Syt 1* (ƒ). The data points correspond to the mean Ϯ S.E. of current (n ϭ 8). D, Peak current amplitudes normalized to maximum current (I/I max ) (data from B and C) are plotted against test potentials displayed with a Boltzmann fit. The mid-point of activation (V 1/2 ) and the Boltzmann slope (k) of ␣ 1 2.3/␣2␦1/␤2a were V 1/2 ϭ Ϫ8.8 Ϯ 0.1 mV, k ϭ 2.3 Ϯ 0.1; with Syt 1 V 1/2 ϭ Ϫ2.8 Ϯ 0.8 mV, k ϭ 3.9 Ϯ 0.25, and with Syt 1* 1-264 V 1/2 ϭ Ϫ2.7 Ϯ 1.38 mV, k ϭ 4.2 Ϯ 0.4. E, the activation time constants ( act , mean Ϯ S.E., n ϭ 6) are plotted against test potentials between Ϫ20 and ϩ25 mV: the channel alone (E), with syntaxin 1A (q), syntaxin 1A and Syt 1 (), or F, syntaxin 1A and Syt 1* (ƒ). Two sample Student's t tests were applied, and p values Ͻ0.05 were obtained from the two-tailed tests. cRNA/oocyte, see Fig. 3; Syntaxin 1A (2 ng/oocyte). were preformed. (i) Recombinant GST-C2A, GST-C2A*, GST-Syt 1, and GST proteins were immobilized to GSH-agarose beads and incubated with equimolar concentrations of recombinant His 6 Lc 753-893 (0.5 M; 2.5 g). As shown by Western analysis using anti-Lc 753-893 antibody (32), C2A, C2B, and Syt 1 bind Lc 753-893 , but no binding of C2A* was observed (Fig. 5A).
(ii) The affinity of C2A and C2A* to Lc 753-893 was tested using the Biacore technology (Biacore; see "Experimental Procedures"). His 6 Lc 753-893 was immobilized on a sensor chip surface. Recombinant samples of GST-C2A, GST-C2A* at the indicated concentrations were injected into the flow cell of the system (Biacore), and changes in resonance units were recorded as a function of time to yield sensorgrams as shown in Fig. 5B. The C2A binding to Lc 753-893 is manifested as large amplitude of the surface plasmon resonance signal while no resonance signal was obtained by C2A*. GST alone showed no binding (data not shown). The calculated affinity of C2A was O.213 M ( 2 ϭ 4.06; Fig. 5B). Hence the C2A mutant does not bind to the intracellular domain of the channel that comprises the Syt 1 interaction (31,32,35,40,47).

The C2A Domain of Syt 1 Is Required for Functional Interactions with Ca v 2.3 (R-channel)-Ca v 2.3 currents were elicited
in oocytes co-expressing ␣ 1 2.3/␤2a/␣2␦1 subunits (41) and Syt 1 or Syt 1* (Fig. 6). Syt 1 or Syt 1* modified neither Ca v 2.3 current-voltage relationship nor peak-current amplitude (Fig.  6A). Conversely, Syt 1 strongly accelerated Ca v 2.3 activation in the range of Ϫ20 to ϩ5 mV, converging at more depolarized values (Ͼ5 mV) ( Fig. 6B; Table I (Fig. 6C). At negative potentials between Ϫ20 and Ϫ10 mV, the rate was accelerated by Syt 1* similar to Syt 1, but between Ϫ5 and 0 mV an abrupt decrease in the rate was observed, which was slower than the channel (Fig. 6D). At more positive potentials, in the range of ϩ5 to ϩ30 mV, act approached control values (Fig. 6, C and D).
Activation of Ca v 2.3 Requires the C2A Polylysine Motif-We next examined the requirement of the Syt 1 C2A polylysine motif on channel activation using the truncated Syt 1 mutants. Ca v 2.3 currents were elicited in oocytes co-expressing the three channel subunits along with Syt 1 1-264 and Syt 1* 1-264 . In both mutants the C2B domain is missing, and in Syt 1* 1-264 the polylysine motif was substituted with alanine residues. The effects of the truncated mutants on channel activity are shown in Fig. 7, A and B. Superimposed traces of whole cell current were activated from a holding potential of Ϫ80 mV by a single voltage step to 0-mV test pulse (Fig. 7A). Syt 1 1-264 appeared to inhibit current amplitude by 60% at 0 mV, while Syt 1* 1-264 displayed no effect on current amplitude but significantly slowed channel inactivation (inactivation kinetics were not explored in the present study). Current-voltage relationships were significantly shifted in the presence of Syt 1 1-264 but not Syt 1* 1-264 (Fig. 7, B and  C). Peak current amplitudes normalized to maximum current (I/I max ) show that the half-maximal voltage activation (V 1/2 ) was significantly displaced by Syt 1 1-264 from Ϫ8.8 Ϯ 0.1 mV to 0.5 Ϯ 0.1 mV and only marginally to Ϫ6.2 Ϯ 0.1 mV by Syt 1* 1-264 , (Fig. 7D). This voltage shift can account for the apparent reduction in current amplitude. The slope factors were directly comparable between control conditions and those expressing Syt 1 1-264 or Syt 1* 1-264 (Fig. 6D). Syt 1  was also efficient at accelerating Ca v 2.3 activation in the Ϫ20 to Ϫ5 mV range similar to full-length Syt 1 (Fig. 7E; Fig. 6C). In contrast, the acceleration of activation by Syt 1* 1-264 was smaller and was detected only in Ϫ20 to Ϫ10 mV range (Fig. 6F). The effects of Syt 1 1-264 and Syt 1* 1-264 on Ca v 2.3 kinetics were specific for this channel as co-expression of these proteins result in no effect on Ca v 1.2 activation kinetics (see Table I). Together, these data suggest that Ca v 2.3 activation involves the C2A domain of Syt 1. Moreover, mutation of the polylysine motif modifies the interaction of the C2A domain with the channel.
A Cross-interaction of Syntaxin 1A with Syt 1 Mutants and Ca v 2.3-Superimposed traces of macroscopic whole cell Ca v 2.3 current elicited from a holding potential of Ϫ80 mV by a single voltage step to a 0-mV test pulse showed a partial reversal by Syt 1 (from 50% to 18%) of the syntaxin 1A-mediated current inhibition but not by Syt 1* (Fig. 8A). Current-voltage relationships obtained in the presence of syntaxin 1A or syntaxin 1A with either one of the Syt 1 mutants indicated a shift toward more positive potentials by Syt 1* (Fig. 8, B and C) that could account for the reduction in current amplitude at 0 mV (Fig.  7A). Peak current amplitudes normalized to maximum current (I/I max ) showed no shift in the half-maximal voltage of Ca v 2.3 (V 1/2 ϭ Ϫ8.8 Ϯ 0.1 mV) by syntaxin 1A (V 1/2 ϭ Ϫ8.0 Ϯ 2.1 mV) (Fig. 8D), unlike the large shift induced by syntaxin 1A in Ca v 1.2 (Fig. 3B). In the presence of syntaxin 1A, V 1/2 was shifted toward more positive potentials to Ϫ4.3 Ϯ 1.2 mV by Syt 1 and to Ϫ0.4 Ϯ 0.7 mV by Syt 1*, (Fig. 8D). Hence, Syt 1 and Syt 1* differently modify the syntaxin-associated channel.
Ca v 2.3 activation was accelerated in cells expressing syntaxin 1A and was not modified further by Syt 1 (Fig. 8E). In contrast, Syt 1* slowed the activation kinetics in the presence of syntaxin 1A, in particular in potentials between 0 -15 mV (Fig. 8F). Thus, mutation of the polylysine motif increased the current voltage shift of Ca v 2.3 and was less effective than Syt 1 at reversing the syntaxin 1A inhibition. In addition, the mutation appeared to affect the interaction of the channel with syntaxin 1A, slowing the activation kinetics. Together, these data suggest that the Syt 1 C2A polylysine motif participates in the syntaxin 1A modulation of Ca v 2.3 activation.
Since Syt 1 and Syt 1 mutants affect syntaxin 1A modulation of the channel, their effect on syntaxin 1A expression in oocytes was tested (Fig. 9). Oocytes were injected with syntaxin 1A cRNA (5 ng/oocyte) and cRNA encoding the various Syt 1 mutants (5 ng/oocyte) as indicated. At day five after injection, oocytes were lysed and proteins were separated on SDS-PAGE and analyzed using monoclonal anti-syntaxin 1A antibody (Fig. 9). As shown by the Western blot analysis there were no significant changes in syntaxin 1A expression in the presence of either one of the four Syt 1 mutants (Fig. 9).
cated Syt 1 mutants lacking the C2B domain. Currents were evoked from a holding potential of Ϫ80 mV to 0 mV test pulse (Fig. 10A). As shown current amplitude was reduced by syntaxin 1A and was partially reversed by the two mutants. Current-voltage relationships showed that Syt 1 1-264 and Syt 1* 1-264 were equally effective at reverting syntaxin 1A inhibition of Ca v 2.3 current amplitude (Fig. 10B). The halfmaximal voltage (V 1/2 ) of the channel was not affected by syntaxin 1A (see Fig. 8D), but a small shift toward more positive potentials was observed by Syt 1 1-264 to 2.7 Ϯ 1.4 mV and to Ϫ2.8 Ϯ 0.8 mV by Syt 1* 1-264 (Fig. 10C). Syntaxin 1A accelerated Ca v 2.3 activation (Fig. 10D), which was further increased in the presence of Syt 1 1-264 (Fig. 10E). In contrast, Syt 1* 1-264 lost the ability to accelerate Ca v 2.3/syntaxin 1A activation (Fig.  10F). Together, these data show that the mutant failed to modify Ca v 2.3 interaction with syntaxin and suggest that the C2A polylysine motif participates in the syntaxin 1A cross-talk with Ca v 2.3.
We addressed the possibility that the voltage-gated Ca 2ϩ channel (Ca v 1.2, Ca v 2.1, Ca v 2.2, and Ca v 2.3), an established effector for Syt 1 (30 -32, 34, 40), may be functionally coupled through the polylysine C2A domain. Using the Xenopus oocytes expression system we examined the functional consequences of mutating the C2A polylysine motif on Ca v 1.2 (Lc-type) the channel that supports evoked secretion in PC12 cells and the neuronal Ca v 2.3 (R-type) channel. The changes induced in the activation kinetics and current amplitude of voltage-sensitive Ca 2ϩ channels demonstrate that the C2A polylysine motif participates in the interaction of Syt 1 with both Ca v 1.2 and Ca v 2.3.
Modulation of Ca v 1.2-Since the Syt 1/syntaxin 1A interaction occurs independently of the C2A polylysine motif (17), the observed differences in syntaxin 1A modulation of channel activity in the presence of Syt 1 may result from either a direct interaction with the channel or with a new site formed by the association of syntaxin 1A with the channel. The full-length Syt 1 reversed syntaxin 1A inhibition of Ca v 1.2 activity, while Syt 1* was significantly less effective. The marked slowing of activation kinetics of Ca v 1.2 by syntaxin 1A was reversed by Syt 1 1-264 but not by Syt 1* 1-264 . The mutation of the polylysine motif in the Syt 1 1-264 protein lacking a C2B domain (Syt 1* 1-264 ) results in a complete loss of function. These results were further substantiated when no binding of His 6 Lc 753-893 to the isolated C2A domain were observed. Lc 753-893 , the intracellular domain of Ca v 1.2 ␣1 subunit, was previously shown to be the site of interaction of Syt 1, C2A, and C2B (31,32). The four mutated KKKK residues abolished GST-C2A* binding to His 6 Lc 753-893 in two methods, GSH-agarose beads and plasmon resonance spectroscopy. Therefore, these results, in part, could provide an explanation to why unlike the intact C2A, polylysine-mutated C2A peptide when injected into PC12 was unable to interfere with transmitter release (17).
Interestingly, the effect of a mutated C2A polylysine motif in full-length Syt 1*, appeared to be partially attenuated by C2B domain, consistent with a functional relationship between the two C2 domains of Syt 1 and the calcium channel (24). Interactions through the Syt 1 C2B domain are also functionally important for neurosecretion (19,20,22,24). More recent studies using a genetic rescue approach in Drosophila reveals a role for the polylysine motif of the C2B domain in evoked release (25). Moreover, the C2B domain promotes the Ca 2ϩ -dependent binding of syntaxin 1A to C2A, suggesting a level of functional synergy between the two C2 domains of Syt 1 (44). Interestingly, secretion from PC12 cells carried out using the cracked cell method showed that dense-core vesicle exocytosis does not require vesicular synaptotagmin 1 but may use instead the plasma membrane synaptotagmins 3 and 7 as Ca 2ϩ sensors (45).
Modulation of Ca v 2.3-Previously, induction of faster activation by Syt 1 was observed for the neuronal Ca 2ϩ channels, Ca v 2.2 and Ca v 2.3, in contrast to no effect on Ca v 1.2 (28 -30).
Here we show that truncated Syt 1 (Syt 1 1-264 ) accelerated Ca v 2.3 activation suggesting that C2A and not C2B domain is responsible for the observed effects. Mutation of the polylysine motif in C2A abolished the stimulatory effect on Ca v 2.3 activation, indicating the role of this motif in the interaction of the vesicular protein with the channel as well as with Ca v 2.3 associated with syntaxin 1A.
The truncated mutant Syt 1* 1-264 partially restored (ϳ75%) current amplitude but did not reverse the syntaxin 1A effect on activation. Thus modulation of the syntaxin/Ca v 2.3 kinetics was affected by Syt 1 1-264 but was lost in Syt 1 1-264 *. In contrast, Syt 1 and Syt 1* effectively reversed the syntaxin 1A inhibition of Ca v 2.3 current amplitude. These results propose that the C2B domain partially compensates for the mutation in the C2A polylysine motif.
Together, the data indicate that the C2A polylysine motif affects the activation of the channel and modulates the kinetics of syntaxin 1A-associated channel. In Fig. 11 we showed a schematic model of putative interactions of the channel, syntaxin, and Syt 1.
In summary, our studies provide compelling evidence that the Syt 1 C2A domain is involved in a functional coupling of the vesicle with the voltage-gated Ca 2ϩ -channels, Ca v 1.2 and Ca v 2.3. The C2A polylysine motif appears to participate in this interaction and likely functions independently of the Syt 1 Ca 2ϩ -mediated interactions with phospholipids or syntaxin 1A (17). The effects of C2A polylysine motif on transmitter release in PC 12 cells as previously reported, may result from a direct modification of the activation kinetics of the Ca 2ϩ channel or function indirectly by competing with endogenous Syt 1 for interactions with the channel. The ability of Syt 1, syntaxin 1A, and the Ca 2ϩ channel to interact is consistent with the formation of a functional exocytotic unit, the excitosome (32). The excitosome complex composed of the Ca 2ϩ channel, syntaxin 1A, SNAP-25, and Syt 1 displays distinct kinetic properties required for calcium-triggered secretion (28,29,32,33,39,46). Therefore, inhibition of neurotransmitter release by C2A domain might occur by interfering with generating the excitosome complex and the ensuing propagation of the signal from the channel to the fusion/docking machinery rather than Ca 2ϩ binding to Syt 1 (30). The physiological relevance and the consequences of the different modulation of neuroendocrine (Ca v 1.2) and neuronal (Ca v 2.3) Ca 2ϩ channels by Syt 1 during the steps leading to transmitter release will require further studies.