Modulation of L-Type Ca²⁺ Channels by Distinct Domains Within SNAP-25

Islets of Langerhans were isolated from mice and dispersed into single cells that were then grown in RPMI-1640 medium (Gibco, Middlesex, UK).

HIT-T15 cells were grown at 37°C in 5% CO₂/95% air in RPMI-1640 medium supplemented with 20 mmol/l glutamine, 10% FCS (Gibco, Gaithersburg, MD), penicillin (100 units/ml), and streptomycin (100 mg/ml). When the cells were confluent, they were passaged and transiently transfected with 2 mg empty plasmid vector (control) or plasmid DNA encoding for SNAP-25_1–206, BoNT/A light chain, or SNAP-25_1–197 (16). The transfection of plasmid DNA was facilitated using Pfx-1 lipid (Invitrogen, Carlsbad, CA) at a ratio of 1:6 (wt/wt). After a 4-h incubation at 37°C in 5% CO₂/95% air, normal RPMI-1640 medium was replaced. Transfection efficiency was determined by using pcDNA3/His/Lac Z (Invitrogen) as a control plasmid for transfections, followed by determination of β-galactosidase expression. In a previous study (16), HT-T15 cells were found to uniformly intake the plasmids. Therefore, successfully transfected cells were confirmed by coexpression and visualization of the GFPs (Clontech, Palo, CA).

Whole-cell Ca²⁺ currents were recorded in pancreatic β-cells from adult male and female mice. Pipettes were pulled from borosilicate glass capillaries (Hilgenberg, Malsfeld, Germany) on a horizontal programmable puller (DMZ Universal Puller; Zeitz-Instrumente, Augsburg, Germany). Typical electrode resistance was 3–5 MΩ. Electrodes were filled with a standard internal solution containing glutathione S-transferase (GST) (10⁻⁹ mol/l), GST SNAP-25_1–206 (10⁻⁹ mol/l), Lc_753–893 (10⁻⁶ mol/l), or Lc_753–893 (10⁻⁶ mol/l) combined with GST SNAP-25_1–206 (10⁻⁹ mol/l). The standard internal solution contained the following (in mmol/l): 150.0 N-methyl-D-glucamine, 10.0 EGTA, 1.0 MgCl₂, 2.0 CaCl₂, 5.0 HEPES, and 3.0 Mg ATP (pH 7.2 adjusted with HCl). The cells were bathed in a solution containing the following (in mmol/l): 138.0 NaCl, 5.6 KCl, 1.2 MgCl₂, 10.0 CaCl₂, 5.0 HEPES, and 10.0 tetraethylamonium (TEA) (pH 7.4 adjusted with NaOH). After obtaining a seal, the holding potential was set at −70 mV during the experiment, and depolarizing voltage pulses (70 mV, 100 ms, 0.05 Hz) were applied. The resulting currents were recorded with an Axopatch 200 amplifier (Axon Instruments, Foster City, CA). All recordings were made at room temperature (∼22°C). The amplitude of whole-cell Ca²⁺ currents was normalized by the cell capacitance. Acquisition and analysis of data were done using the software program pCLAMP (Axon Instruments).

On the third day after the transient transfection, the HIT-T15 cells were trypsinized and dispersed into single cells for patch-clamp study. Drops of medium containing these cells were placed in a 1-ml chamber mounted on the stage of an inverted phase-contrast microscope and allowed to stick to the bottom for 10 min. The RPMI-1640 medium was replaced by external solution composed of (in mmol/l): 20.0 BaCl₂, 90.0 NaCl, 5.0 CsCl, 1.0 MgCl₂, 20.0 TEA-Cl, and 10.0 HEPES (pH 7.4). Whole-cell configuration of the patch-clamp technique was applied to record membrane currents in single HIT-T15 cells, which showed GFP brightness under ultraviolet light. The pipette solution contained the following components (in mmol/l): 70.0 cesium aspartate, 1.0 MgCl₂, 5.0 EGTA, 4.0 Mg-ATP, and 20.0 HEPES (pH 7.2). Recordings were performed using an Axopatch 1D amplifier (Axon Instruments). Whole-cell voltage-clamp protocols were generated by pCLAMP6 software (Axon Instruments). All signals were filtered at 1 kHz by an on-board eight-pole Bessel filter before digitization with a Digiata 1200 analog-to-digital converter (Axon Instruments). Cell capacitance was determined by integration of the capacity transient. Data were presented as means ± SE and compared by unpaired Student’s t test for single comparisons and by ANOVA for multiple comparisons. P < 0.05 was considered statistically significant. All experiments were performed at room temperature.

HIT-T15 cells with different transfectants were cultured on coverslips. On the third day after transfection, the coverslips were placed in a 1-ml chamber and loaded with low Ca²⁺-HEPES buffer solution containing 4 μmol/l Fura 2-AM (Molecular Probes, Eugene, OR) at room temperature for 45 min. Cells were then rinsed three times with a standard extracellular solution containing (in mmol/l) 125.0 NaCl, 5.0 KCl, 1.8 CaCl₂, 2.0 MgCl₂, 0.5 NaH₂PO₄, 5.0 NaHCO₃, 10.0 HEPES, and 10.0 glucose (pH 7.2). Single-cell Ca²⁺ imaging was performed at room temperature using a Merlin imaging system (Life Sciences Resources, Cambridge, UK). Fura 2-AM fluorescence was calibrated to Ca²⁺ concentration with 10 μmol/l ionomycin in the presence of 5 mmol/l CaCl₂ for maximal fluorescence ratio or with no added Ca²⁺ and 10 mmol/l EGTA for minimal fluorescence ratio (17). Cells with successful transfection were determined by concomitant transfection of GFP, which was illuminated with light at 470 nm. Two measures were taken to minimize the possible influence of GFP on fluorescence density from Fura 2. First, control cells were also transfected with GFP, as we did in electrophysiologic studies. Second, we only examined the cells with just-visible GFP fluorescence and eliminated those with maximal GFP fluorescent brightness at 470 nm to minimize optical interference.

To test for possible modulation of voltage-gated Ca²⁺ currents by SNAP-25, pancreatic β-cells were dialyzed with GST (10⁻⁹ mol/l, control), GST SNAP-25_1–206 (10⁻⁹ mol/l), Lc_753–893 (10⁻⁶ mol/l) alone, or Lc_753–893 (10⁻⁶ mol/l) together with GST SNAP-25_1–206 (10⁻⁹ mol/l) through the recording pipette. Whole-cell voltage-gated Ca²⁺ currents from cells subjected to different treatments were evoked by step- depolarizing voltage pulses from a holding potential of −70 to 0 mV at 0.05 Hz. As shown in Fig. 1, whole-cell Ca²⁺ currents recorded from cells treated with GST SNAP-25_1–206 gradually and markedly decreased during recording in comparison with those from control or Lc_753–893-treated cells. Interestingly, the gradual reduction in whole-cell Ca²⁺ currents induced by the intracellular application of GST SNAP-25_1–206 was counteracted by coapplication of Lc_753–893. Furthermore, there was no significant difference in Ca²⁺ current density between control cells and cells treated with Lc_753–893 (10⁻⁶ mol/l) alone or Lc_753–893 (10⁻⁶ mol/l) together with GST SNAP-25_1–206.

To record L_Ca currents efficiently, Cs⁺ was included in the pipette solution, and Ba²⁺ and TEA were added to the bath solution to block voltage-gated outward K⁺ currents. Ba²⁺ also increased the L_Ca conductance. Among 140 cells (15.6 ± 0.1 pF) tested, 119 displayed inward currents in response to depolarizing pulses from a holding potential of −70 mV (Fig. 2). The peak current-voltage relationship from a control cell shows that the inward current appeared at a high voltage of −30 mV, reached the maximum at 10 mV, and reversed at ∼50 mV (Fig. 2B). These electophysiological properties were consistent with the L_Ca current reported in insulin-secreting β-cells by other authors (18,19). Furthermore, the inward current was greatly enhanced by 10⁻⁶ mol/l Bay K8644 and blocked considerably by 10⁻⁶ mol/l nifedipine, which are selective L_Ca agonists and antagonists, respectively (Fig. 2). Bay K8644 (10⁻⁶ mol/l) increased the peak amplitude by 185.4 ± 19.2% (n = 6) as well as shifted the current-voltage relationship to the left by 10 mV (Fig. 2B). In the presence of nifedipine (10⁻⁶ mol/l), the peak inward current amplitude was reduced to 13.7 ± 2.0% of control (n = 6). Thus, the predominant inward current (>85%) in HIT-T15 cells results from the activation of L_ca.

To examine the effect of wild-type SNAP-25 protein on L_Ca currents, we introduced the full-length recombinant GST SNAP-25 protein into HIT-T15 cells via the patch pipette. This exogenously applied recombinant SNAP-25 also allows us to examine the time course of action of SNAP-25 on L_Ca. It has been previously shown that a protein with a molecular weight of 50 kDa can diffuse through a pipette (access resistance of 10 MΩ) into the cytosol with a time constant of 5 min (20). As shown in the lower panel of Fig. 3A, the inward current began to decrease from 2 min after membrane rupture in a HIT-T15 cell treated with 10⁻⁹ mol/l GST SNAP-25, whereas such a decline was not seen with 10⁻⁹ mol/l GST alone in the patch pipette (Fig. 3A, upper panel). The inhibition of the inward current became significant around 4 min. At 5 min, the peak inward current amplitude in GST SNAP-25–treated cells (n = 8) was reduced by 32% compared with that in GST-treated cells (n = 8, P < 0.05) (Fig. 3B). In a preliminary study, we found that GST alone (10⁻⁹ mol/l, n = 4) did not alter the time course of the inward current seen in control HIT-T15 cells (n = 4). Thus, cells treated with GST were used as vehicle control cells. This result indicates that SNAP-25 protein can inhibit L_Ca currents in HIT-T15 cells, and this is consistent with the previous finding that coexpression of SNAP-25 and L_Ca suppressed the Ca²⁺ currents (4). We then examined whether the inhibitory effect caused by SNAP-25 was related to its direct interaction with the cytoplasmic domain separating repeats II and III l-loop of the α1C subunit of the L_Ca (8) by adding both synthetic II-II l-loop (1 μmol/l Lc_753–893, a generous gift from Dr. D. Atlas, Jerusalem, Israel) and GST SNAP-25 (10⁻⁹ mol/l) to the pipette solution. As shown in Fig. 3C, SNAP-25 failed to suppress the inward current in the concomitant presence of Lc_753–893. Lc_753–893 itself had no significant effect on the inward current (Fig. 3C), consistent with the previous report (8). Our results indicate that Lc_753–893 blocks the binding site at SNAP-25 for the L_Ca and subsequently prevents the SNAP-25-induced inhibition. This was surprising for us because we previously thought that Lc_753–893 would merely extend the distance between Ca²⁺ entry through L_Ca and the Ca²⁺ sensing site on the insulin granule, thereby disrupting exocytosis, but would not directly affect L_Ca activity (8). These results suggest that interaction between SNAP-25 and the II-III l-loop of L_Ca is necessary for SNAP-25 to negatively modify L_Ca activity in HIT-T15 cells.

We then explored which domain within SNAP-25 protein is responsible for its modulation of L_Ca activities. Acute treatment with BoNT/A cleaves SNAP-25 at Gln¹⁹⁷-Arg¹⁹⁸ to generate a membrane-bound NH₂-terminal SNAP-25_1–197 fragment and a cytosolic COOH-terminal consisting of nine amino acids, SNAP-25_198–206 (11,21). These cleavage products could compete with each other and with the remaining intact SNAP-25, making it difficult to distinguish their independent actions. We therefore used the strategy of overexpressing SNAP-25 proteins and BoNT/A light chain. The overexpressed SNAP-25 proteins will be targeted to the plasma membrane to effectively compete with endogenous SNAP-25, and the overexpressed BoNT/A will have access to all cellular SNAP-25 proteins, including those in complexes that would have cycled into uncomplexed forms. Furthermore, sufficient time for cytosolic proteolysis would deplete the cleavage products.

We first compared the effects of the overexpression of wild-type SNAP-25 (SNAP-25_1–206) and the mutant NH₂-terminal SNAP-25_1–197. Mutant SNAP-25_1–197, like SNAP-25_1–206, is capable of binding membrane syntaxin 1a (22,23), and the regions within this portion of the truncated SNAP-25 protein have been shown to directly bind to neuronal N-type Ca²⁺ (24) and neuroendocrine L_Ca (8). Our previous report demonstrated that transfection with wild-type SNAP-25 and mutant SNAP-25_1–197 resulted in targeting of these proteins specifically to the plasma membrane and that these SNAP-25 protein–transfected cells were identified by the coexpressed GFP (16). In a control study, the peak inward current-voltage relationship in HIT-T15 cells (n = 4) cotransfected with empty vectors and vectors with cDNA encoding for GFP did not show significant difference from that of untransfected cells (n = 4, data not shown). Therefore, those cells cotransfected with empty vectors and GFP vectors were used as control cells in all sets of experiments involving various transfectants, if not indicated otherwise. Figure 4A displayed the current traces evoked by voltage steps from −60 to 50 mV from four individual cells with similar membrane capacitance to minimize the cell size variation. Transfection of wild-type SNAP-25 reduced inward current amplitudes (Fig. 4A). Peak current-voltage relationships demonstrated that the inward currents were decreased in HIT-T15 cells expressing SNAP-25 protein in a voltage-dependent manner (Fig. 4B). The inhibition became significant at 0 mV, reached maximum at 10 mV, and disappeared at potential positive to 20 mV. At 10 mV, cell transfection with SNAP-25_1–206 attenuated the peak current amplitude by 27%, similar to the 31% inhibition observed in cells in which GST SNAP-25_1–206 was introduced to the cell interior via the patch pipette. The current-voltage curve, however, did not shift in any direction (Fig. 4B). These data further confirm that SNAP-25_1–206 protein can inhibit L_Ca activity in HIT-T15 cells.

However, in contrast to the wild-type SNAP-25, expression of the mutant SNAP-25_1–197 did not decrease but instead paradoxically increased the inward currents (Fig. 4A). Figure 4B shows that inward current amplitudes were significantly augmented in cells expressing SNAP-25_1–197 at membrane potentials from −20 to 10 mV. At 10 mV, overexpression of SNAP-25_1–197 increased the peak current amplitude by 19%. We next investigated whether the excitatory effect of SNAP-25_1–197 was associated with the Lc loop of the α1C subunit. We used another set of experiments in which 1 μmol/l Lc_753–893 peptide was injected via the patch pipette into HIT-T15 cells already expressing mutant SNAP-25_1–197. The data were collected 5 min after the membrane breakthrough to allow Lc_753–893 peptides to diffuse into the cell. As shown in Fig. 5A, injection of the Lc_753–893 peptide blocked the stimulatory effect of SNAP-25_1–197 at potentials from −20 to 10 mV. For example, the inward current density at 10 mV was −8.7 ± 0.5 pA/pF in SNAP-25_1–197-transfected cells (n = 8, P < 0.05 vs. control) and reversed to −7.3 ± 0.5 pA/pF after injecting Lc_753–893 into SNAP-25_1–197-transfected cells (n = 7), which was close to −7.0 ± 0.3 pA/pF in control cells (n = 8) (Fig. 5B). Taken together, these results indicate that SNAP-25_1–197 has an opposite effect on L_Ca currents, as compared with SNAP-25_1–206, and that its excitatory action is also associated with direct interaction with the Lc loop of the L_Ca.

We then examined the effects of depletion of endogenous SNAP-25 protein by transfection with BoNT/A light chain. We predicted that because wild-type SNAP-25 inhibits L_Ca, its depletion by BoNT/A should relieve this inhibition. Surprisingly, Fig. 4A shows that transfection with BoNT/A paradoxically decreased the L_Ca currents to an even greater extent than wild-type SNAP-25_1–206. The current-voltage relationship demonstrated that the BoNT/A-induced inhibition was also voltage dependent, with statistical significance between 0 and 20 mV (Fig. 4B). At a membrane potential of 10 mV, BoNT/A inhibited the peak inward current amplitude by 41%, whereas wild-type SNAP-25_1–206 inhibited it by only 27% (Fig. 4B).

The greater inhibition by BoNT/A expression is likely due to absence of a tonic-positive regulatory effect of the NH₂-terminal SNAP-25_1–197 domain on L_Ca currents. Wild-type SNAP-25–mediated inhibition could therefore not be ascribed to this product. COOH-terminal SNAP-25_198–206 peptide, another BoNT/A cleavage product, was therefore tested for its ability to mimic the inhibitory effects of wild-type SNAP-25. As shown in Fig. 6, injecting synthetic SNAP-25_198–206 (10⁻⁹ mol/l) into HIT-T15 cells (n = 6) began to suppress the inward current from 5 min after membrane rupture and reached the maximum inhibition of 47% at 10 min, as compared with control cells (n = 6). Furthermore, introduction of synthetic SNAP-25_198–206 (10⁻⁹ mol/l) into HIT-T cells transfected with SNAP-25_1–197 also resulted in the reduction of the inward current, with a maximum inhibition of 34% occurring at 10 min after membrane rupture, as compared with control cells (Fig. 6). These data suggest that SNAP-25_198–206 is likely the putative domain within SNAP-25 that inhibits L_Ca currents and exerts a dominant effect over the NH₂-terminal SNAP-25_1–197 domain.

Because L_Ca current amplitudes were affected by different transfections, we further explored how these transfections could affect L_Ca kinetics. Therefore, both voltage- and time-dependent activation and inactivation were examined in the following studies. Voltage-dependent activation curves were obtained by plotting normalized peak tail currents against test pulses. In this set of experiments, only expression of the truncated SNAP-25_1–197 in HIT-T15 cells shifted the activation curve 10 mV toward the left (Fig. 7), indicating that the L_Ca is easier to activate at lower voltages in the presence of SNAP-25_1–197, consistent with its stimulatory effect on this channel (Fig. 4B). However, transfection with SNAP-25_1–206 or BoNT/A did not affect activation kinetics (Table 1). Parameter values for voltage-dependent activation kinetics are summarized in Table 1.

Voltage-dependent inactivation of the L_Ca was assessed using a standard two-step protocol in which long duration prepulses (15 s) of varying test potentials were followed by a short step to 10 mV. The steady-state inward current, after the 10-mV test pulse from each prepulse, was measured as a fraction of the maximal inward current. The voltage-dependent inactivation curve was obtained by plotting these fractional values against prepulse voltages and best-fitted by the Boltzmann equation (Fig. 7). No transfections significantly altered the inactivation curve. Only control and SNAP-25_1–197 data for voltage-dependent inactivation are shown in Fig. 7. The values for voltage-dependent inactivation parameters are also summarized in Table 1.

Time constants during activation were obtained by fitting depolarization-evoked inward current traces with a mono-exponential function. Figure 8A shows that the inward current was activated faster in the cell transfected with SNAP-25_1–197 but almost unaffected in the cell transfected with SNAP-25_1–206 or BoNT/A. By calculation, the activation time constant was reduced from 2.8 ± 0.2 ms in control cells (n = 11) to 2.0 ± 0.1 ms in SNAP-25_1–197 (n = 11, P < 0.05), suggesting that SNAP-25_1–197 accelerates the activation rate of the L_Ca, again consistent with its stimulatory effect on this channel. The values for activation time constants were given in Table 2.

Time-dependent inactivation of the L_Ca was studied using a long (9 s) depolarizing pulse to 10 mV (Fig. 8B). Time constants (τ₁ and τ₂) of the inward current decay were calculated by fitting current traces with a biexponential equation. As shown in Table 2, SNAP-25_1–206 accelerated the inactivation rate of the L_Ca, with τ₁ and τ₂ increased by 1.5- and 1.6-fold, respectively, which was in agreement with the previous study performed in Xenopus oocyte (4). The absolute values for τ₁ were larger in our experiment. The reason for this was probably the replacement of the bath solution Ca²⁺ with Ba²⁺, which could slow down the inactivation rate of the L_Ca. Transfection with BoNT/A also increased the inactivation rate, but transfection with SNAP-25_1–197 significantly decelerated it (Fig. 8B and Table 2). These data may therefore partially explain changes in inward current amplitudes, as seen in Fig. 4. That is, faster inactivation caused by SNAP-25_1–206 and BoNT/A results in smaller inward current amplitudes, whereas slower inactivation induced by SNAP-25_1–197 leads to larger inward current amplitudes.

In pancreatic β-cells, Ca²⁺ entry through l-type channels triggers the exocytosis of insulin granules (25). Therefore, modulation of L_Ca channel activity by SNAP-25 proteins could consequently affect Ca²⁺ influx and cytoplasmic-free Ca²⁺ concentration ([Ca²⁺]_i). To test this, we measured [Ca²⁺]_i in single HIT-T15 cells using Fura 2 as Ca²⁺ indicator. The HIT-T15 cells studied were transfected with the empty vector (control), SNAP-25_1–206, BoNT/A, or SNAP-25_1–197. Transfection was again determined by coexpression of GFP in each experiment. HIT-T15 cells were placed in a 1-ml chamber with Ca²⁺-containing physiological solution (see research design and methods). After the basal [Ca²⁺]_i level was recorded, the cells were treated with 10 mmol/l KCl, which depolarizes the cell membrane and subsequently activates L_Ca. The difference between the peak transient rise in [Ca²⁺]_i and the basal [Ca²⁺]_i was considered as net Ca²⁺ influx through L_Ca. To minimize the possible inter- and intra-assay variations, [Ca²⁺]_i values from cells (usually 3–4) with efficient transfection in each field were averaged, and only one field was counted from each chamber. Figure 9A shows that expression of SNAP-25_1–206 and BoNT/A in HIT-T15 cells decreased the peak transient rise in [Ca²⁺]_i, whereas SNAP-25_1–197 had an opposite effect. Data for the peak transient rise in [Ca²⁺]_i were averaged in Fig. 9B. In control cells, the peak transient rise in [Ca²⁺]_i was 199.1 ± 10.4 nmol/l (n = 6). SNAP-25_1–206 and BoNT/A reduced it to 150.6 ± 5.4 nmol/l (n = 6, P < 0.05) and 134.1 ± 7.8 nmol/l (n = 6, P < 0.05), respectively, whereas SNAP-25_1–197 raised it to 248.1 ± 18.1 nmol/l (n = 6, P < 0.05). These data, therefore, are in agreement with our L_Ca current recordings (Fig. 4) and provide further evidence that domains within SNAP-25 protein have distinct effects on the L_Ca.

In this study, we demonstrate for the first time that SNAP-25 specifically modulates L_C subtype Ca²⁺ channel activity in insulin-secreting cells. This indicates that SNAP-25 not only functions as a key component in a core complex to trigger insulin exocytosis but also interacts with the voltage-gated L_Ca to control Ca²⁺ influx, which plays an important role in the formation of the core complex (4,8). The degree of inhibition of full-length SNAP-25 on whole-cell Ca²⁺ currents in the primary pancreatic β-cell was smaller than that observed in the HIT-T15 cell. This is probably due to the relative small contribution of the L_C subtype Ca²⁺ channel to the total L-type Ca²⁺ currents in normal β-cells. The predominant subtype of the L_Ca in islet β-cells is the L_D subtype (26–28), whereas HIT-T15 cells predominantly contain the L_C subtype. These results also suggest that although the L_C and L_D subtypes of L_Ca share a 70% homology (29), the L_C subtype appears to be more sensitive to SNAP-25. In contrast, syntaxin 1A appears to similarly inhibit both the L_C and L_D subtypes of β-cell L_Ca (4,8,28).

We also present the first evidence that domains within the SNAP-25 protein are capable of causing distinct changes in L_Ca kinetics. Wild-type SNAP-25 reduced the inward current amplitude and accelerated its inactivation rate, indicating its inhibitory regulatory role on the L_Ca, which is consistent with the previous report (4). It is likely that SNAP-25_1–206 acts on the inactivation state of the channel and results in its faster decay. However, BoNT/A expression, which depleted all cellular SNAP-25 proteins, resulted in an even greater inhibition of L_Ca currents. Specifically, BoNT/A expression decreased the current amplitude, accelerated the inactivation rate, and reduced the Ca²⁺ influx. BoNT/A expression has the advantage over the acute application of BoNT/A. The latter could acutely generate cleavage products that compete with each other and with the remaining intact SNAP-25, therefore preventing assessment of the independent effect of the functional domains within SNAP-25. The greater inhibition observed with BoNT/A expression suggests that SNAP-25 contains not only negative but also positive regulatory domains. Indeed, we found the COOH-terminal SNAP-25_198–206 peptide to confer the inhibitory effect of SNAP-25 and of BoNT/A, whereas the membrane-bound NH₂-terminal SNAP-25_1–197 conferred a paradoxical stimulatory effect on the L_Ca. The fact that COOH-terminal SNAP-25_198–206 peptide is able to override the positive regulatory effect of SNAP-25_1–197 indicates that the negative regulatory role of SNAP-25 on the L_Ca probably predominates in the islet β-cell. Nonetheless, this dominant-negative regulatory effect of SNAP-25 is balanced by the positive regulatory domain, as evidenced by its complete absence effected by BoNT/A transfection, which caused an even greater inhibition of the Ca²⁺ channel than the SNAP-25 overexpression. It therefore seems likely that both negative and positive regulatory domains of the SNAP-25 protein are operative in modulating the Ca²⁺ channel in the islet β-cell.

SNAP-25_1–197 increased the current amplitude (Fig. 4), activated L_Ca at lower membrane potentials (Fig. 7), and decelerated its inactivation rate (Fig. 8B). These results indicate that SNAP-25_1–197 may interact with more than one site at the L_Ca. Injecting the Lc_753–893 peptide into HIT cells transfected with SNAP-25_1–197 or wild-type SNAP-25_1–206 restored the current-voltage curve to control levels, indicating that both SNAP-25 proteins modulated the L_Ca by direct interaction with the channel. This is further supported by the previous study showing a direct binding between SNAP-25 and L_Ca (8). Nonetheless, wild-type SNAP-25 and mutant SNAP-25_1–197 must be bound to the II-III L-loop of L_Ca in distinct conformations to assert these apparent opposite effects on L_Ca and the resulting Ca²⁺ influx (Fig. 9). In our previous report, where data collected were from a field of both transfected and untransfected cells (transfection efficiency ∼45%), we did not observe an obvious difference in KCl-evoked Ca²⁺ influx between SNAP-25_1–197– and wild-type SNAP-25-transfected cells (16). However, in the current study, data were collected from cells with the highest protein expression, as determined by GFP coexpression, therefore permitting the effects to be seen. Despite its positive modification of the L_Ca, overexpression of SNAP-25_1–197 caused a net inhibition of insulin secretion in HIT-T15 cells (16), likely due to the formation of nonfunctional exocytotic SNARE complexes (22,23).

Injection of the COOH-terminal SNAP-25_198–206 inhibited L_Ca currents to a greater extent than wild-type SNAP-25, and interestingly, these inhibitory effects dominated over the stimulatory effects of SNAP-25_1–197 expression. The greater inhibition by overexpression of BoNT/A over the wild-type SNAP-25 further supports our hypothesis that these two domains of SNAP-25, NH₂-terminal SNAP-25_1–197, and COOH-terminal SNAP-25_198–206 exert a “yin yang” effect on the L_Ca. This might in part explain why the presence of higher levels of extracellular Ca²⁺ or higher membrane depolarization could abrogate the inhibitory effects of the acutely applied BoNT/A (13,15), wherein the COOH-terminal SNAP-25_198–206, which initially blocks the L_Ca, binds insulin granule VAMP-2 to form a nonfunctioning fusion complex (16). However, its subsequent accelerated proteolysis in the cytosol would leave the relatively intact membrane-bound SNAP-25_1–197 unopposed to act on the L_Ca to increase Ca²⁺ influx. Whereas SNAP-25_1–197 also forms a nonfunctional docking complex (16), the increased Ca²⁺ influx might overcome this inhibition. In support, Coorssen et al. (30) demonstrated that increased Ca²⁺ levels could effect exocytosis independent of the SNARE complex and that the latter may serve to modulate Ca²⁺ sensitivity in driving fusion. Furthermore, in contrast to a neuron that has a higher proportion of secretory granules in a readily releasable pool of docked vesicles (∼10%), <4% of insulin secretion granules in β-cells are in this pool (25,31). The vast majority of insulin granules are located more distantly from the membrane, within a reserve pool. The initial phase of [Ca²⁺]_i rise is important for exocytosis of the readily releasable pool (32,33), whereas the subsequent sustained elevation of [Ca²⁺]_i acts to prime and mobilize the reserve pool to the readily releasable pool (32,33). The SNAP-25_1–197 stimulatory effects on the L_Ca may compensate for the reduced insulin exocytosis caused by the nonfunctional docking complex by mobilizing more secretion granules from the reserve pool to the readily releasable pool.

Taken together, we have now provided evidence that domains within SNAP-25 protein have distinct effects on the L_Ca. These results have the potential to elucidate how SNARE proteins may not only regulate physiological insulin exocytosis but also be of pathophysiological significance in the dysregulation of Ca²⁺ influx–mediated islet β-cell injury in diabetes (1–3). In fact, SNARE protein expression levels were diminished in islets of type 2 diabetic models, including Zucker fa/fa and GK rats (34,35), which exhibit abnormal regulation of Ca²⁺ influx and insulin exocytosis (36). Reconstitution of these SNARE proteins, particularly SNAP-25 and syntaxin, partially restored insulin secretion to normal levels (35), likely by restoring the diabetic dysregulation of the distal components of the insulin exocytotic machinery, in particular the L_C excitosome. Finally, the positive regulatory domains within these SNARE proteins, particularly within the NH₂-terminal SNAP-25_1–197 protein, could serve as targets for novel drug design to treat diabetes.

This work was supported by grants to H.Y.G. from the Canadian Diabetes Association and the Canadian Institutes for Health Research as well as grants from the National Institutes of Health (DK55160) and Juvenile Diabetes Research Foundation. Support for this work was also obtained from grants from the Swedish Medical Research Council, the Swedish Diabetes Association, the Nordic Insulin Foundation Committee, the Fredrik and Ingrid Thurings Foundation, Funds of Karolinska Institutet, the Berth von Kantzows Foundation, the Novo Nordisk Foundation, the Swedish Society of Medicine, and the Swedish National Board of Health and Welfare.

We thank Daphne Atlas, Robert Tsushima, Eva Pasyk, and Anne Marie Salapatek for their critical review of the manuscript and technical advice.