Uncoupling of Ca_v1.2 From Ca²⁺-Induced Ca²⁺ Release and SK Channel Regulation in Pancreatic β-Cells

Abstract

We investigated the role of Ca_v1.2 in pancreatic β-cell function by expressing a Ca_v1.2 II-III loop/green fluorescent protein fusion in INS-1 cells (Ca_v1.2/II-III cells) to disrupt channel-protein interactions. Neither block of K_ATP channels nor stimulation of membrane depolarization by tolbutamide was different in INS-1 cells compared with Ca_v1.2/II-III cells, but whole-cell Ca_v current density was significantly increased in Ca_v1.2/II-III cells. Tolbutamide (200 μM) stimulated insulin secretion and Ca²⁺ transients in INS-1 cells, and Ca_v1.2/II-III cells were completely blocked by nicardipine (2 μM), but thapsigargin (1 μM) blocked tolbutamide-stimulated secretion and Ca²⁺ transients only in INS-1 cells. Tolbutamide-stimulated endoplasmic reticulum [Ca²⁺] decrease was reduced in Ca_v1.2/II-III cells compared with INS-1 cells. However, Ca²⁺ transients in both INS-1 cells and Ca_v1.2/II-III cells were significantly potentiated by 8-pCPT-2′-O-Me-cAMP (5 μM), FPL-64176 (0.5 μM), or replacement of extracellular Ca²⁺ with Sr²⁺. Glucose (10 mM) + glucagon-like peptide-1 (10 nM) stimulated discrete spikes in [Ca²⁺]_i in the presence of verapamil at a higher frequency in INS-1 cells than in Ca_v1.2/II-II cells. Glucose (18 mM) stimulated more frequent action potentials in Ca_v1.2/II-III cells and primary rat β-cells expressing the Ca_v1.2/II-II loop than in control cells. Further, apamin (1 μM) increased glucose-stimulated action potential frequency in INS-1 cells, but not Ca_v1.2/II-III cells, suggesting that SK channels were not activated under these conditions in Ca_v1.2/II-III loop-expressing cells. We propose the II-III loop of Ca_v1.2 as a key molecular determinant that couples the channel to Ca²⁺-induced Ca²⁺ release and activation of SK channels in pancreatic β-cells.

Type 2 diabetes is typified by hyperglycemia resulting from insufficient insulin secretion from pancreatic β-cells, diminished sensitivity of target tissues to the effects of insulin, or both (1). Glucose-stimulated increases in intracellular Ca²⁺ concentration are the primary drivers of insulin secretion from β-cells (2), and derangements in Ca²⁺ signaling are observed in β-cells from diabetic rodents (3, 4). Uptake of glucose into pancreatic β-cells and subsequent metabolism of glucose increases the ATP:ADP ratio within β-cells (5). ATP binds to and blocks β-cell K_ATP channels (6), which are composed of the sulfonylurea receptor 1 (SUR1) and the inwardly rectifying K⁺ channel K_ir6.2 (7), resulting in membrane depolarization (8). If of sufficient magnitude, membrane depolarization activates voltage-gated Ca²⁺ (Ca_v) channels (9). Ca²⁺ influx via the L-type channels present in β-cells, Ca_v1.2 and 1.3 (10), is thought to be particularly important for driving insulin secretion in response to elevated glucose or sulfonylurea stimulation (11). Sulfonylurea drugs have long been used to stimulate insulin secretion from pancreatic β-cells in persons with type 2 diabetes (12). Similar to ATP, binding of sulfonylureas to SUR1/K_ir6.2 decreases K⁺ efflux through these channels and triggers membrane depolarization (13). One role of Ca²⁺ influx via L-type Ca²⁺ channels into pancreatic β-cells is to stimulate further release of Ca²⁺ from the endoplasmic reticulum (ER), via Ca²⁺-induced Ca²⁺ release (CICR). This process is activated in β-cells depolarized with elevated concentrations of extracellular KCl (14) but apparently does not contribute to [Ca²⁺]_i oscillations observed in β-cells stimulated with high glucose concentrations in the absence of other cosecretagogues (15, 16).

The SUR1/K_ir6.2 channel is known to interact directly or indirectly with several proteins that reside at the plasma membrane and play a role in insulin exocytosis. The scaffolding proteins Rab 3 interacting molecule (RIM)2 (17) and Piccolo (18) also interact with the intracellular II-III loop domain of Ca_v1.2 (19). We have previously reported that endogenously expressed RIM2 is present in lipid rafts and is immunoprecipitated by a Ca_v1.2/II-III-enhanced green fluorescent protein loop fusion protein, but not by a Ca_v1.3/II-III loop-enhanced green fluorescent protein fusion protein, in INS-1 cells (20). Further, we reported that overexpression of the Ca_v1.2/II-III loop causes accumulation of Ca_v1.2, but not Ca_v1.3, outside of lipid rafts in the rat insulinoma cell line, INS-1. Conversely, overexpression of the Ca_v1.3/II-III loop selectively excluded Ca_v1.3, but not Ca_v1.2, from lipid rafts (20). The guanine nucleotide exchange factor exchange protein directly activated by cAMP 2 (EPAC2) interacts directly with both SUR1 and Piccolo (21) whereas Piccolo forms a dimer with RIM2, in a Ca²⁺-dependent manner (21). Thus, a network of protein-protein interactions connecting the main components of the pathway that leads to insulin secretion stimulated by sulfonylureas implicates Ca_v1.2 in this process. This signaling complex of scaffolding proteins (RIM2, Piccolo), cAMP effector (EPAC2), and ion channel subunits (SUR1, Ca_v1.2) is of special significance because membrane depolarization-dependent calcium influx via Ca_v1.2 channels has been implicated in triggering Ca²⁺-induced Ca²⁺ release from the ER in pancreatic β-cells (22), a process that is amplified by cAMP, at least in part, through EPAC2 (22, 23). However, it is not clear whether interaction of Ca_v1.2 with other proteins via the II-III loop has functional significance in sulfonylurea-stimulated insulin secretion or [Ca²⁺]_i transients in pancreatic β-cells. We show here that expression of the Ca_v1.2 intracellular II-III loop in the pancreatic β-cell line INS-1 doesn't disrupt SUR1 or EPAC2 localization to lipid rafts but does disrupt CICR stimulated by the sulfonylurea tolbutamide, as assessed by several different experimental approaches.

In addition to modulation of insulin secretion, increases in intracellular Ca²⁺ concentration in pancreatic β-cells may also modulate electrical activity stimulated by glucose. One potential mechanism for fine tuning of glucose-stimulated action potentials is via activation of the small conductance Ca²⁺-activated K⁺ channel, SK (24). SK subtypes 1–3 were detected in the rat insulinoma cell line INS-1 (25). Activation of SK channels occurs as intracellular Ca²⁺ accumulates at the plasma membrane, and K⁺ efflux via SK channels limits action potential frequency (26). Indeed, block of SK channels with specific pharmacologic agents such as apamin (27) or UCL 1684 (28) can markedly increase glucose-stimulated action potential frequency in pancreatic β-cells, but the major source of the Ca²⁺ that activates SK channels in pancreatic β-cells is not clear. Here we report that modulation of glucose-stimulated action potential frequency by SK channels is disrupted in both INS-1 cells stably expressing the Ca_v1.2/II-III loop and primary rat β-cells expressing the Ca_v1.2/II-III loop via adenoviral transduction.

Materials and Methods

Chemicals

Antibodies were acquired from the following sources: EPAC2 and eukaryotic translation initiation factor 3 subunit e (eIF3e), Santa Cruz Biotechnology; SUR1, Abcam; K_ir6.2, Millipore Corp; green fluorescent protein (GFP) and horseradish peroxidase-conjugated secondary antibodies, Bio-Rad Laboratories. 8-pCPT-2′-O-Me-cAMP-AM was from Biolog. Apamin was from Tocris Cookson. All other chemicals were from Sigma-Aldrich.

Cell culture

INS-1 cells (29) and Ca_v1.2/II-III cells (30) were cultured in RPMI medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (HyClone Laboratories), 11 mg/mL sodium pyruvate, 10 mM HEPES, 100 U/mL penicillin, 100 μg/mL streptomycin, and 50 μM β-mercaptoethanol at 37°C, 5% CO₂ (INS-1 RPMI). Ca_v1.2/II-III cells were cultured in 200 μg/mL G418 to maintain expression of the Ca_v1.2 channel II-III loop. INS-1 cells used for current clamp recordings were cultured in low glucose (2.5 mM) for 18–24 hours prior to experiments.

Sucrose density gradient fractionation

INS-1 cells and Ca_v1.2/II-III cells were lysed in the presence of protease inhibitors (800 nM aprotinin, 50 μM leupeptin, 1 μg/mL pepstatin, 1 mM benzimidine, 1 mM AEBSF, 10 μg/mL calpain I and II inhibitors; all from Sigma) and fractionated over a discontinuous 5%/30%/40% sucrose gradient as described previously (20). Fractions (1 mL) were removed from the top of each gradient, for a total of 11 fractions. Like fractions from multiple tubes were pooled and concentrated using Amicon Ultra-4 concentrator tubes (Millipore Corp). Protein concentrations were determined using the BCA assay (Pierce Chemical Co) before they were analyzed via Western blot.

Immunoprecipitation assays

Experiments in which the Ca_v1.2/II-III or Ca_v1.3/II-III loops fused to GFP were immunoprecipitated from INS-1 cell lysates with antibodies against GFP were performed as previously described (20). Immunoprecipitation of eIF3e was performed as previously described except that cell lysates were incubated with antibodies against eIF3e overnight at 4°C, and then with agarose-immobilized Protein A Plus (Thermo Scientific) overnight at 4°C before washing. Protein was eluted by incubating at 80°C in Laemmli buffer.

Western blotting

Protein (20 μg) from each fraction of sucrose gradients was separated by SDS-PAGE using 4% acrylamide (29:1 acrylamide-bisacrylamide) stacking gels, and 8% resolving gels for all blots except for K_ir6.2, for which 10% acrylamide resolving gels were used. A constant volume of fraction 1 (50 μl) was loaded on each gel because it consistently lacked any protein as detected by the BCA assay. Proteins were resolved on polyacrylamide gels in 25 mM Tris, 192 mM glycine, 0.1% sodium dodecyl sulfate, and then transferred to polyvinylidene difluoride membranes using a Bio-Rad Transblot Cell in 10 mM CAPS, 10% methanol, at 70 V for 35 minutes to 2 hours, depending on the size of the protein to be detected. Polyvinylidene difluoride membranes were blocked by soaking in 5% nonfat milk in PBS with 0.1% Tween 20 (PBST) for 2 hours at room temperature. Blocked membranes were incubated with primary antibodies overnight at 4°C at the following dilutions (EPAC2, 1:100; SUR1, 1:500; K_ir6.2, 1:200; eIF3e 1:100, GFP 1:1000) and then washed 3 times with PBST and incubated with secondary antibodies conjugated to horseradish peroxidase (1:1000) for 1 hour at room temperature. Excess secondary antibodies were removed by washing membranes 3 times in PBST, and proteins were detected using enhanced chemiluminescence or chemifluorescence (Amshersham Biosciences).

Electrophysiological assays

Electrophysiological measurements in INS-1 cells and Ca_v1.2/II-III cells were recorded at room temperature using an Axopatch 200B amplifier (Molecular Devices) and filtered at 1 kHz (6-pole Bessel filter, −3 dB). Patch pipettes were pulled from borosilicate glass (VWR) and fire polished to resistances of 2–5 MΩ. INS-1 cells used for current clamp recordings were cultured in low glucose (2.5 mM) for 18–24 hours prior to experiments. For current clamp experiments and voltage clamp experiments to measure K_ATP channel current, the intracellular solution contained (in mM concentration) 90 K₂SO₄, 10 NaCl, 1 MgCl₂, 1.1 EGTA, 0.1 CaCl₂, 5 HEPES, 0.3 ATP, 0.2 GTP. The extracellular solution used for measuring K_ATP currents contained (in mM concentration) 138 NaCl, 5.6 KCl, 11.1 glucose, 10 HEPES, 1.2 MgCl₂, 2.6 CaCl₂, and the same solution with 2.5 mM glucose instead of 11.1 mM glucose was used in current-clamp experiments. The pH of solutions was adjusted to 7.4 with NaOH, and the osmolality was adjusted to 290–300 mOsm. Whole-cell K_ATP currents were elicited by 1.3-second steps of ± 20 mV from a holding potential of −70 mV. Data were acquired at a sampling frequency of 1 kHz. The membrane potential of INS-1 cells and Ca_v1.2/II-III cells was measured using gap-free recording at a sampling frequency of 1 kHz in I = 0 current clamp mode. The K_ATP channel opener, diazoxide (300 μM), was transiently applied to maximally open K_ATP channels, before application of tolbutamide. Tolbutamide solutions were prepared from stocks dissolved in 0.1 M NaOH, made fresh daily. Diazoxide solutions were prepared from stocks dissolved in dimethylsulfoxide. For recordings of voltage-gated Ca²⁺ channel currents, the bath solution contained (in mM concentration) 150 Tris, 10 BaCl₂, 4 MgCl₂. The intracellular solution contained (in mM concentration) 130 N-methyl-d-glucamine, 10 EGTA, 60 HEPES, 2 ATP, and 1 MgCl₂. The pH of both solutions was adjusted to 7.3 with methanesulfonic acid, and the osmolality was corrected to 290–300 mOsM. Current-voltage relationship data were collected by applying 100-millisecond test depolarizations from −50 to +60 mV in 10-mV increments, from a holding potential of −70 mV. Data were acquired at a sampling frequency of 10 kHz and filtered at 1 kHz. For current clamp recordings of membrane depolarization with glucose, the perforated patch technique was used to keep the intracellular signaling environment intact. The bath solution contained (in mM concentration) 138 NaCl, 5.6 KCl, 2.5 glucose, 10 HEPES, 1.2 MgCl₂, 2.6 CaCl₂. The intracellular solution contained (in mM concentration) 10 KCl, 10 NaCl, 70 K₂SO₄, 2 MgCl₂, 10 HEPES, 100–120 μg/mL amphotericin B. The pH of both solutions was adjusted to 7.4 with NaOH and the osmolality was adjusted to 290–300 mOsm. A 30 mg/mL amphotericin stock was prepared in dimethylsulfoxide from powder prior to all experiments. The stock solution and the intracellular solution containing amphotericin were sonicated for at least 15 minutes prior to use, were used within 1 hour, and were protected from light. Recording pipettes were front-filled with intracellular solution without amphotericin and were back-filled with intracellular solution containing amphotericin. After formation of the cell-attached patch, the access resistance (R_a) was monitored. After 10–15 minutes, when R_a was less than 50 MΩ, the patch was considered to be perforated enough to begin current clamp recordings. Upon adequate access, the membrane potential was measured and 300 μM diazoxide was briefly applied to fully hyperpolarize the membrane potential. Once a stable baseline membrane potential was achieved, 18 mM glucose was perfused into the recording chamber to stimulate membrane depolarization and action potential spiking. Whole-cell SK channel current recording was performed using a holding potential of −70 mV with steps to −50 mV. The intracellular and extracellular solutions for SK channel whole-cell recordings were made as described previously (25). The intracellular solution consisted of (in mM concentration): 110 potassium-gluconate, 10 KCl, 10 NaCl, 1MgCl₂, 3 Mg-ATP, 5 HEPES, 10 EGTA, and 9.6 CaCl₂ to achieve 2 μM free [Ca²⁺] (MaxChelator http://www.maxchelator.stanford.edu); pH and the osmolality were adjusted to 7.2 and 290–310, respectively. Extracellular solution contained (in mM concentration): 140 NaCl, 4 KCl, 2 NaHCO₃, 1 NaH₂PO₄, 1 MgSO₄, 5 HEPES, 2.5 CaCl₂, with pH and osmolality adjusted to 7.4 and 290–310, respectively.

Insulin secretion assay

INS-1 cells or Ca_v1.2/II-III cells were plated in 24-well tissue culture plates at 50%–70% confluency and incubated overnight in RPMI medium with 2.5 mM glucose at 37°C, 5% CO₂. Immediately before the assay, cells were washed twice in PBS (137 mM NaCl, 10 mM phosphate, 2.7 mM KCl [pH 7.4]), and preincubated with a modified Krebs-Ringer buffer (KRBH: 134 mM NaCl, 3.5 mM KCl, 1.2 mM KH₂PO₄, 0.5 mM MgSO₄, 1.5 mM CaCl₂, 5 mM NaHCO₃, 10 mM HEPES [pH 7.4]) supplemented with 0.05% fatty acid free BSA (KRBH buffer) alone, or containing indicated concentrations of inhibitors, for 30 minutes at 37°C, 5% CO₂. The buffer was decanted and replaced with fresh KRBH buffer alone (basal condition) or KRBH containing the indicated concentrations of tolbutamide or glucose with or without inhibitors for 1 hour at 37°C, 5% CO₂. Secreted insulin was assayed using an ELISA for rat insulin (High-Range EIA kit; ALPCO Diagnostics). Cells were lysed in 20 mM Na₂HPO₄, 150 mM NaCl, 0.1% Triton X-100, 800 nM aprotinin, 50 μM leupeptin, 1 μg/ml pepstatin, 1 mM benzamidine, 1 mM 4-(2-Aminoethyl) benzenesulfonylfluoride, 10 μg/mL calpain inhibitor I, 10 μg/mL calpain inhibitor II (pH 7.4), and cellular protein in each well was determined using the BCA assay (Thermo Scientific).

Intracellular Ca²⁺ assays in 96-well plates

INS-1 cells or Ca_v1.2/II-III cells were plated at 100% confluency in black-walled 96-well plates (Corning Life Sciences) in RPMI supplemented as described above and incubated overnight at 37°C, 5% CO₂. Cells were washed twice with PBS and incubated with 5 μM Fura-2AM (Molecular Probes) diluted in KRBH for 1 hour at 37°C, 5% CO₂. The KRBH containing Fura-2AM was then removed, and the cells were washed twice with KRBH, and equilibrated for 30 minutes at 37°C, 5% CO₂, in the KRBH alone, or with indicated concentrations of inhibitors or 8-pCPT-2′-O-Me-cAMP-AM. Cells were stimulated by injection of the indicated concentration of sulfonylurea or glucose (or buffer control), and changes in intracellular Ca²⁺ concentrations were measured by recording the ratio of fluorescence intensities at 508/20 nm resulting from excitation of Fura-2 at 340/11 nm or 380/20 nm (center/bandpass) using a Synergy 4 multimode microplate reader (BioTek). Ratios were acquired every 0.7 second for 15 seconds before injection and 2 minutes after injection of tolbutamide. In experiments with 18 mM glucose as the stimulus, ratios were acquired every minute for 1 hour. Data were corrected for any injection artifact by subtracting the change in fluorescence ratio measured in cells injected with KRBH alone.

Single-cell intracellular Ca²⁺ assays

Cells were seeded on a poly-d-lysine coated round glass coverslip and incubated overnight in INS-1 RPMI medium. Cells were washed with PBS before loading with 3 μM cell-permeant Ca²⁺-sensitive dye Fura-2 (Life Technologies) at room temperature for 1 hour. The loading solution was removed, and cells were washed with KRBH followed by incubation in 0.5% BSA KRBH at room temperature for 30 minutes. The coverslip was mounted on a perfusion chamber attached to the stage of an Olympus, IX50 inverted microscope (Olympus) equipped with a PlanApo 40× objective lens (0.95 na). Solution was perfused to the chamber through an in-line solution heater (Warner Instrument) at a constant flow rate (1 mL/ min) and constant temperature (37 ± 1°C). Cells were first perfused with KRBH for 30 seconds followed by 30 mM KCl to depolarize the cells for 5 minutes. Verapamil (50 μM) was added during the final 4 minutes with 30 mM KCl to bring the [Ca²⁺]_i to basal. A combination of 10 mM glucose, 10 nM glucagon-like peptide 1 (GLP-1), and 50 μM verapamil was applied for the rest of the experiment to elicit CICR. Cells were alternatively excited at 340 nm and 380 nm using a filter wheel (Sutter Instrument), and the images (emission signal at 510 nm) were recorded in time lapse (one ratio every 2 seconds) using a Clara CCD camera (Andor Technology) at the beginning of KRBH application. Single and isolated cells were selected as regions of interest (ROI) and the averaged intensities (340:380 nm ratio) for each ROI at different time points were measured using Metamorph (Molecular Devices). Basal [Ca²⁺]_i level of a cell was obtained by averaging the intensities of that cell within the first 30 seconds when only KRBH was present, and the basal was used to normalize the intensities at all time points. A single cell Ca²⁺ transient trace was obtained by plotting all the normalized intensities against lapsed time points. The slope of a peak was used to define a CICR peak. A spike in a trace is considered as a CICR event if the slope of the peak is equal or greater than that of the KCl-induced spike.

Measurement of ER Ca²⁺ dynamics

INS-1 cells or Cav1.2/II-III cells were transiently transfected with the pcDNA3-D1ER plasmid (Addgene) encoding an ER-targeted fluorescence resonance energy transfer (FRET)-based Ca²⁺ sensor in which Ca²⁺ binding enables FRET between cyan fluorescent protein (CFP) and citrine (31). Cells were split 24–48 hours after transfection into 4 chambered 35-mm glass bottom dishes (In Vitro Scientific) and imaged 48–72 hours after transfection. Cell culture media were replaced with Krebs-Ringer buffer 30 minutes prior to imaging. The basal FRET signal was collected for 1 minute prior to treatment and 10 minutes following the addition of the positive control carbachol (500 μM) or tolbutamide (200 μM), which were diluted in Krebs-Ringer buffer. To determine the minimum FRET value, or R_min, cells were treated with thapsigargin (1 μM) for 30 minutes prior to imaging and subsequently stimulated with carbachol to exhaust ER calcium stores. The D1ER FRET signal was measured with a ×20 objective lens using a Nikon A1 Confocal System with a Perfect Focus Ti-E Inverted Microscope. The CFP fluorophore of the D1ER FRET sensor was excited at 457 nm with an Argon laser. CFP and citrine emission were collected at 482 nm and 525 nm, respectively, using 482/35nm and 525/50 filter cubes. The citrine signal intensity, or FRET signal, was normalized to the intensity of the CFP emission, which served as an expression of ER calcium levels. The time course of the FRET signal was analyzed using NIS Elements software. Only those cells that displayed a high FRET signal were chosen for analysis, and these values were averaged among the chosen cells for each experiment at all time points. R_min was averaged among several separate experiments and represents the point at which the D1ER sensor no longer detects a decrease in ER calcium. This value was subtracted from each FRET time point in all experiments in order to normalize for the difference in R_min between INS-1 and Ca_v1.2/II-III cells. To compare ER calcium changes among experiments for each cell type, the FRET/CFP values over the time course were then normalized to the average FRET-CFP ratio over the first minute of no treatment, yielding a baseline value over the first 60 seconds of 100% or R_max. The peak change in the FRET-CFP ratio, or decrease in ER calcium levels, was calculated based on the maximum difference from baseline over a variable time span following stimulation. Change in FRET was calculated as a percentage of the dynamic ER calcium range, or difference between the R_max and R_min, in either INS-1 or Ca_v1.2/II-III cells.

Isolation and adenoviral transduction of rat pancreatic islets

Pancreatic islets were isolated from 2-month-old, male Wistar rats (∼225 g) via collagenase digestion by the Islet Core Facility of the Indiana Diabetes Research Center, Indiana University School of Medicine. Islets were maintained in islet RPMI medium (RPMI-1640 medium containing 8 mM glucose, 10 mM HEPES, 10% fetal bovine serum (Hyclone), 11 mg/mL sodium pyruvate, 100 U/mL penicillin, and 100 μg/mL streptomycin) and kept at 37°C and 5% CO₂ for 24 hours. Groups of 150 islets were handpicked into fresh islet RPMI medium for adenoviral transduction. Adenovirus encoding GFP alone or GFP fused to the C-terminal end of the Ca_v1.2/II-III loop domain (30) was obtained from ViraQuest. Islets were transduced at an MOI of 50 (assuming 1000 cells/islet). After 24 hours with the virus, the islets were washed with PBS and transferred to fresh islet RPMI medium. After another 24 hours, the islets were dissociated into single cells.

Islet dissociation

Islets were digested in 1 mL of 0.25% Trypsin-EDTA for every 100 islets and triturated repeatedly for 3 minutes. Trypsin was neutralized with the addition of 3 mL of RPMI medium, and cells were centrifuged at 800 rpm for 3 minutes, resuspended in islet RPMI medium containing 2.5 mM glucose, and plated into poly-d-lysine coated 35-mm tissue culture dishes. Dissociated cells were incubated at 37°C, 5% CO₂ until they were used for electrophysiological experiments.

Data analysis

Electrophysiological data were analyzed using Clampfit 8 (Molecular Devices). Area-under-the-curve (AUC) analysis and dose-response curve fits were performed using SigmaPlot 11 (Systat Software). Single-cell calcium imaging data were analyzed with Metamorph software (Molecular Devices). Analyzed data are shown as mean values ± SE. Statistical significance was determined using one-way ANOVA and the Holm-Sidak post hoc test unless otherwise indicated. Differences with P values < .05 were considered significant.

Results

Sucrose-density gradient fractionation of proteins involved in depolarization-induced insulin secretion in INS-1 cells and Ca_v1.2/II-III cells

The K_ATP channel, composed of K_ir6.2 and SUR1 subunits, plays a central role in the insulin secretion stimulated by sulfonylureas and glucose. We examined the localization of K_ir6.2, SUR1, and EPAC2 in lipid rafts by fractionating the Triton X100-insoluble portion of INS-1 and Ca_v1.2/II-III cell lysates on discontinuous sucrose gradients. EPAC2 is reported to interact directly with both Piccolo (21) and SUR1 (19), and we found that both EPAC2 and SUR1 are highly concentrated in lipid raft fractions of sucrose gradients (the interface of 5% and 30% sucrose) in both INS-1 cells and Ca_v1.2/II-III cells (Figure 1). We also assessed the localization of the K_ATP channel subunit K_ir6.2 and found that although it is present at the 5%/30% sucrose interface, it was also distributed throughout the 40% sucrose fractions in both INS-1 cells and Ca_v1.2/II-III cells (Figure 1). The lipid raft-resident protein caveolin 1 was detected at the 5%/30% sucrose interface but also distributed throughout the sucrose gradient in samples from both INS-1 and Ca_v1.2/II-III cells. This distribution of caveolin 1 is similar to that observed in a previous study using the pancreatic β-cell line HIT-T15 (32). Thus, the K_ATP channel subunits SUR1 and K_ir6.2, along with the interacting protein EPAC2, are present in lipid rafts in INS-1 cells, and their distribution on discontinuous sucrose gradients is not perturbed by expression of the Ca_v1.2 intracellular II-III loop.

Figure 1.

K_ATP channel subunits and the cAMP effector EPAC2 are present in lipid rafts in both INS-1 cells and Ca_v1.2/II-III cells. Western blots detecting the indicated proteins are shown for each fraction of the sucrose-density gradients for cell lysates from INS-1 cells or Ca_v1.2/II-III cells. Fractions 2 and 3, at the interface of 5% and 30% sucrose, contained the Triton X-100 insoluble, low-density material (lipid rafts). K_ir6.2, the pore-forming subunit of the K_ATP channel, is detected in, but not restricted to, lipid raft fractions. This distribution is not affected by expression of the Ca_v1.2/II-III loop. The dashed line across the K_ir6.2 blots indicates the boundary between 2 separate polyacrylamide gels containing fractions 1–9 (top) and fractions 10 and 11 (bottom) that were simultaneously transferred to a single polyvinylidene difluoride membrane before western blotting. EPAC2 and SUR1 are highly localized to the lipid raft fractions, and this localization was not affected by expression of the Ca_v1.2/II-III loop. Caveolin 1 is present in lipid raft fractions but also distributed into the 40% sucrose fractions in both INS-1 cells and Ca_v1.2/II-III cells. Fractions 10 and 11 are not shown for the SUR1, EPAC2, and caveolin 1 blots because they contained little or none of the indicated proteins. Molecular weight standards are shown. Each blot is representative of at least 3 independent experiments.

Electrophysiological characterization of Ca_v1.2/II-III cells

Ca_v1.2 is reported to exist in a complex with proteins essential for stimulation of pancreatic β-cells by sulfonylureas; therefore, we compared the modulation of electrical activity in INS-1 cells and Ca_v1.2/II-III cells by tolbutamide. Figure 2A shows a whole-cell voltage-clamp experiment with a Ca_v1.2/II-III cell held at −70 mV, with alternating steps to −50 and −90 mV. Application of tolbutamide via external perfusion blocked both the inward and outward K⁺ current in a dose-dependent manner. Plots of the percent current blocked by tolbutamide concentrations between 100 nM and 500 μM are shown in Figure 2A. Fits to these plots yielded EC₅₀ values for tolbutamide of 2.6 ± 0.7 μM and 3.8 ± 0.2 μM for INS-1 cells and Ca_v1.2/II-III cells, respectively. Because block of K_ATP channels by tolbutamide leads to membrane depolarization in pancreatic β-cells, we performed current clamp experiments to compare the potency of tolbutamide depolarizing the membrane potential in INS-1 cells and in Ca_v1.2/II-III cells. As shown in Figure 2B, 200 μM tolbutamide elicited strong membrane depolarization, leading to initiation of action potentials in both INS-1 cells (left panel) and Ca_v1.2/II-III cells (right panel). Neither the resting membrane potential, nor the membrane depolarization elicited by 10, 50, 200, or 500 μM tolbutamide were significantly different in INS-1 cells or Ca_v1.2/II-III cells (Figure 2C). Finally, we compared the density of Ba²⁺ current conducted by voltage-gated Ca²⁺ channels and the density of tolbutamide-sensitive K⁺ current in INS-1 cells and Ca_v1.2/II-III cells. As we reported previously, the I_Ba density is markedly increased in Ca_v1.2/II-III cells compared with INS-1 cells (20); however, the K_ATP channel current density is not significantly different (Figure 2D). Thus, our data demonstrate that expression of the Ca_v1.2 II-III loop increases voltage-gated Ca²⁺ channel current density but does not affect either K_ATP channel current density, the sensitivity of K_ATP channels to tolbutamide block, or the ability of tolbutamide to depolarize the membrane potential in INS-1 cells in a dose-dependent manner.

Figure 2.

Tolbutamide induces similar K_ATP block and membrane depolarization in INS-1 and Ca_v1.2/II-III cells. A (left), representative trace of the dose-dependent tolbutamide block of K_ATP current measured in Ca_v1.2/II-III cells using whole-cell voltage clamp. Whole-cell current was elicited by a ± 20 mV step from a holding potential of −70 mV. A (right), percentage of the whole-cell K_ATP current blocked by tolbutamide in INS-1 cells (open circles, n = 4–11) and Ca_v1.2/II-III cells (closed circles, n = 5–7). IC₅₀ values were determined by fitting the data using the equation % Current Blocked = min + (max − min)/(1 + 10̂(logIC₅₀ − [tolbutamide])). B, Representative traces of the change in membrane potential elicited by 200 μM tolbutamide in INS-1 cells and Ca_v1.2/II-III cells. C, dose-dependent effect of tolbutamide on membrane depolarization in INS-1 cells (open bars, n = 8–18) and Ca_v1.2/II-III cells (closed bars, n = 7–13). EC₅₀ values were determined by fitting the data using the equation membrane potential = 1/(1 + ([tolbutamide]/EC₅₀)ⁿ). D, Whole-cell current density comparison of voltage-gated calcium channels (Ca_v) and K_ATP channels in INS-1 cells (open bars) and Ca_v1.2/II-III cells (closed bars). I_Ba density (pA/pF) was measured at 0 mV from a holding potential of −70 mV. The difference in total whole-cell I_Ba density for Ca_v1.2/II-III cells compared with INS-1 cells is statistically significant (***, P < .001; n = 13–15). Current density of K_ATP channels was determined by measuring whole-cell current elicited by a ± 20 mV step from a holding potential of −70 mV in the presence of 300 μM diazoxide to maximally open K_ATP channels. There is no significant difference in K_ATP channel current density between INS-1 cells and Ca_v1.2/II-III cells (n = 20). s, seconds; Tolb, tolbutamide.

Insulin secretion and intracellular Ca²⁺ transients stimulated by tolbutamide in Ca_v1.2/II-III cells don't require intact intracellular stores of Ca²⁺

Given the central role of the L-type channel Ca_v1.2 in the stimulation of insulin secretion, we examined tolbutamide-stimulated insulin secretion in INS-1 cells and Ca_v1.2/II-III cells using the static incubation method (Figure 3). As expected from the electrophysiological characterization of tolbutamide in each cell line, 200 μM tolbutamide stimulated insulin secretion that was significantly different from basal secretion (Figure 3A). The L-type-selective Ca²⁺ channel blocker, nicardipine (2 μM), completely inhibited secretion stimulated by tolbutamide in each cell line (Figure 3A). To assess the role of internal stores of Ca²⁺ in tolbutamide-stimulated insulin secretion, the ability of 1 μM thapsigargin, an inhibitor of the sarcoplasmic/ER Ca²⁺-ATPase, to inhibit insulin secretion stimulated by tolbutamide was determined. In INS-1 cells, we found that thapsigargin significantly, but incompletely, blocked insulin secretion stimulated by 200 μM tolbutamide (Figure 3A), suggesting that both influx of Ca²⁺ via L-type Ca²⁺ channels and release of internal stores of Ca²⁺ are required for maximal insulin secretion stimulated by tolbutamide in INS-1 cells. In sharp contrast, thapsigargin treatment does not inhibit tolbutamide-stimulated secretion in Ca_v1.2/II-III cells (Figure 3A). Therefore, our results suggest that release of internal Ca²⁺ stores is not required for maximal insulin secretion in response to tolbutamide in Ca_v1.2/II-III cells.

Figure 3.

Insulin secretion and Ca²⁺ transients stimulated by tolbutamide are sensitive to thapsigargin in INS-1 cells but not Ca_v1.2/II-III cells. A, Insulin secretion stimulated by 200 μM tolbutamide is completely blocked by 2 μM nicardipine (black bars) in both INS-1 cells and Ca_v1.2/II-III cells. (***, P < .001 compared with basal; ###, P < .001 compared with nicardipine (n = 9 in 3 separate experiments). Insulin secretion stimulated by 200 μM tolbutamide is partially inhibited by pretreatment with 1 μM thapsigargin (white bars) in INS-1 cells, but not in Ca_v1.2/II-III cells (***, P < .001; **, P < .01; *, P < .05 compared with basal; ##, P < .01 compared with tolbutamide alone (n = 9 in 3 separate experiments)). B, Ca²⁺ transients stimulated by 200 μM tolbutamide are mediated by L-type Ca²⁺ channels in both INS-1 cells and Ca_v1.2/II-III cells. Cells pretreated with 500 nM FPL 64176 (black circles) exhibit markedly greater increases in intracellular Ca²⁺ concentration in response to tolbutamide stimulation than cells pretreated with KRBH buffer only (white circles). Pretreatment of INS-1 cells or Ca_v1.2/II-III cells with 2 μM nicardipine (white diamonds) completely inhibits the Ca²⁺ transients elicited by 200 μM tolbutamide + 500 nM FPL 64176. Data shown are mean ± SE from experiments representative of 3 performed in quadruplicate. C, The rapid, initial phase of Ca²⁺ transients stimulated by 200 μM tolbutamide is inhibited by pretreatment with 1 μM thapsigargin in INS-1 cells but not in Ca_v1.2/II-III cells. White circles represent the thapsigargin-sensitive portion of the tolbutamide response. Black circles represent the response to tolbutamide in the absence of thapsigargin. In contrast, the rapid peak in [Ca²⁺]_i stimulated by tolbutamide is greatly amplified in both cell lines by replacing extracellular Ca²⁺ with 2 mM Sr²⁺ (open diamonds). Data shown are mean ± SE from experiments representative of 3 performed in quadruplicate. Tolb, tolbutamide; Thaps, thapsigargin.

To compare changes in intracellular Ca²⁺ concentration induced by tolbutamide, INS-1 cells and Ca_v1.2/II-III cells in 96-well plates were loaded with the Ca²⁺ indicator Fura-2AM, and changes in the ratio of fluorescence intensity at 510 nm after excitation at 340 nm and 380 nm (340-nm to 380-nm ratio) were measured upon injection of tolbutamide (final concentration of 200 μM). The changes in 340-nm to 380-nm ratio upon injection of tolbutamide were corrected by subtracting the changes in 340-nm to 380-nm ratio measured in replicate wells of cells upon injection of buffer only. The net change in Fura-fluorescence ratio, reflecting the net change in intracellular Ca²⁺ concentration stimulated by 200 μM tolbutamide, is shown in Figure 3. Tolbutamide induced a biphasic rise in intracellular Ca²⁺ concentration in INS-1 cells and Ca_v1.2/II-III cells. In INS-1 cells, a rapid peak in intracellular Ca²⁺ concentration is followed by decay to an elevated plateau level that persisted for the duration of the 2-minute measurement (white circles: tolbutamide, Figure 3B). Tolbutamide induced Ca²⁺ transients in Ca_v1.2/II-III cells that rose to a similar peak as in INS-1 cells, but Cav1.2/II-III transients appeared to decay more slowly (white circles: tolbutamide, Figure 3B). Ca²⁺ transients induced by tolbutamide in INS-1 cells and Ca_v1.2/II-III cells are mediated by L-type Ca²⁺ channels because pretreatment with 500 nM FPL-64176, an L-type-specific Ca²⁺ channel agonist, greatly enhanced the level of intracellular Ca²⁺ in response to tolbutamide stimulation (black circles: tolbutamide + FPL-64176, Figure 3B). When cells were pretreated with 2 μM nicardipine, the Ca²⁺ transient stimulated by tolbutamide and FPL-64176 was greatly reduced (white diamonds: tolbutamide + FPL-64176 + nicardipine, Figure 3B). Thus, in both INS-1 cells and Ca_v1.2/II-III cells, the increase in intracellular Ca²⁺ concentration stimulated by tolbutamide is controlled by influx of Ca²⁺ via L-type Ca²⁺ channels.

Because insulin secretion stimulated by tolbutamide was differentially sensitive to thapsigargin in INS-1 and Ca_v1.2/II-III cells, we examined the effect of thapsigargin on changes in intracellular Ca²⁺ concentration stimulated by tolbutamide in these 2 cell lines (Figure 3C). Pretreatment of INS-1 cells with 1 μM thapsigargin before stimulation with tolbutamide selectively inhibited the rapid peak in intracellular Ca²⁺ concentration but had a minimal effect on the sustained plateau phase of the transient. In the Ca_v1.2/II-III cells, pretreatment with thapsigargin was much less effective in inhibiting the peak in Ca²⁺ concentration. The portion of the Ca²⁺ response to tolbutamide that was thapsigargin sensitive [tolbutamide response − (tolbutamide + thapsigargin response)] is shown as white circles in Figure 3C for INS-1 and Ca_v1.2/II-III cells. The insensitivity of the Ca_v1.2/II-III cells to pretreatment with thapsigargin could indicate either an uncoupling of Ca²⁺ influx via L-type Ca²⁺ channels from CICR or a decreased sensitivity of the Ca²⁺ release channel RYR2 to gating by Ca²⁺. To test these possibilities, we measured the tolbutamide-stimulated rise in intracellular Ca²⁺ concentration with extracellular Ca²⁺ replaced with 2 mM Sr²⁺. Sr²⁺ binds to Fura-2 with a higher K_D value than Ca²⁺ (33), is able to stimulate release of Ca²⁺ from internal stores via activation of RYR2 in β-cells (14), and is not buffered as strongly in the cytoplasm as Ca²⁺ (34). We found that in the presence of 2 mM extracellular Sr²⁺, 200 μM tolbutamide stimulated marked transient elevations of Fura-2 340-nm to 380-nm fluorescence ratio in both INS-1 cells and Ca_v1.2/II-III cells (Figure 3C, open diamonds). Taken together, the results presented in Figure 3 suggest that influx of Ca²⁺ via L-type channels in Ca_v1.2/II-III cells does not couple efficiently to release of Ca²⁺ from internal stores, but that Sr²⁺ influx strongly stimulates release of Ca²⁺ from internal stores in both INS-1 cells and Ca_v1.2/II-III cells.

Potentiation of tolbutamide-stimulated intracellular Ca²⁺ transients and insulin secretion by 8-pCPT-2′-O-Me-cAMP-AM

Because tolbutamide-stimulated Ca²⁺ transients were not sensitive to thapsigargin in Ca_v1.2/II-III cells, we asked whether Ca²⁺ transients stimulated by tolbutamide would be differentially potentiated by the EPAC2-selective cAMP analog 8-pCPT-2′-O-Me-cAMP-AM in INS-1 and Ca_v1.2/II-III cells. We tested the ability of 1 μM and 5 μM 8-pCPT-2′-O-Me-cAMP-AM to potentiate tolbutamide-stimulated Ca²⁺ transients and insulin secretion in INS-1 and Ca_v1.2/II-III cells. Cells were loaded with Fura-2AM and were pretreated for 30 minutes with 0, 1, or 5 μM 8-pCPT-2′-O-Me-cAMP-AM prior to injection of either buffer only or 200 μM tolbutamide. Figure 4A shows the net change in Fura-2 fluorescence ratio (tolbutamide response − buffer-only response) for representative experiments in INS-1 cells and Ca_v1.2/II-III cells. Pretreatment with 5 μM 8-pCPT-2′-O-Me-cAMP-AM increased the early peak of the Ca²⁺ transient over pretreatment with buffer alone or 1 μM 8-pCPT-2′-O-Me-cAMP-AM in both INS-1 cells and Ca_v1.2/II-III cells. AUC analysis for the entire postinjection time course of 3 independent experiments in INS-1 cells revealed that 5 μM 8-pCPT-2′-O-Me-cAMP-AM, but not 1 μM, significantly increased the Ca²⁺ transient induced by tolbutamide (Figure 4B). However, AUC analysis in Ca_v1.2/II-III cells revealed that 1 μM and 5 μM 8-pCPT-2′-O-Me-cAMP-AM significantly increased the Ca²⁺ transient induced by tolbutamide (Figure 4B). Because tolbutamide-stimulated Ca²⁺ transients in INS-1 cells and Ca_v1.2/II-III cells are potentiated by 5 μM 8-pCPT-2′-O-Me-cAMP-AM and 1 and 5 μM 8-pCPT-2′-O-Me-cAMP-AM, respectively, we next asked whether tolbutamide-stimulated insulin secretion would be similarly potentiated by the same concentrations of 8-pCPT-2′-O-Me-cAMP-AM. Insulin secretion was stimulated with 200 μM tolbutamide in the absence or presence of 0, 1, or 5 μM 8-pCPT-2′-O-Me-cAMP-AM. Figure 4C shows that 5 μM 8-pCPT-2′-O-Me-cAMP-AM significantly potentiated tolbutamide-stimulated insulin secretion in INS-1 and Ca_v1.2/II-III cells.

Figure 4.

The EPAC-selective cAMP analog 8-pCPT-2′-O-Me-cAMP (ESCA) potentiates insulin secretion and Ca²⁺ transients stimulated by tolbutamide in both INS-1 cells and Ca_v1.2/II-III cells. A, Ca²⁺ transients stimulated by 200 μM tolbutamide in both INS-1 cells and Ca_v1.2/II-III cells are potentiated by pretreatment of cells with ESCA. Data shown are mean ± SE from representative experiments performed in quadruplicate. B, Area under the curve (AUC: 2 minutes post injection) analysis of 9 (INS-1 cells) or 3 (Ca_v1,2.II-III cells) independent experiments in which cells were pretreated with 0, 1, or 5 μM ESCA before stimulation with 200 μM tolbutamide. ***, P < .001; **, P < .01; *, P < .05 compared with tolbutamide alone; ###, P < .001 compared with tolbutamide + 1 μM ESCA. C, Tolbutamide-stimulated insulin secretion is potentiated by 5 μM ESCA in INS-1 and Ca_v1.2/II-III cells (***, P < .001; **, P < .01 compared with basal; ###, P < .001 compared with 1 μM ESCA + tolbutamide). Tolb, tolbutamide.

Tolbutamide stimulation reduces ER Ca²⁺ concentration to a greater extent in INS-1 cells than in Ca_v1.2/II-III cells

To further substantiate that CICR is disrupted in Ca_v1.2/II-III cells, we directly examined the change in ER [Ca²⁺] in response to tolbutamide using the ER-targeted calmodulin-based Ca²⁺ sensor D1ER (31). D1ER comprises the fluorescent proteins CFP and citrine connected by calmodulin and a mutant calmodulin-binding peptide in such a way that Ca²⁺ binding permits FRET between CFP and citrine. Thus, in the high [Ca²⁺] environment of the ER, FRET is detected, and stimuli that reduce ER Ca²⁺ concentration will reduce D1ER FRET. INS-1 cells or Ca_v1.2/II-III cells were transiently transfected with the pcDNA₃-D1ER plasmid, and FRET between CFP and citrine was clearly detected. The basal FRET value was taken to be the FRET ratio measured under basal conditions (R_max), and the minimum FRET value was taken as the FRET ratio measured in cells pretreated for 30 minutes with thapsigargin and further stimulated with 500 μM carbachol (R_min). The FRET window for each cell line was taken as R_max − R_min. The FRET window for Ca_v1.2/II-III cells was significantly lower than that measured for INS-1 cells (Figure 5A). To normalize for this difference, changes in FRET ratios in subsequent experiments were expressed as a percentage of the FRET window for that cell line. In both INS-1 cells and Ca_v1.2/II-III cells, application of 500 μM carbachol triggered a rapid and persistent reduction in FRET, indicating a decrease in ER [Ca²⁺] (Figure 5B). In contrast, stimulation of both INS-1 cells and Ca_v1.2/II-III cells with 200 μM tolbutamide resulted in a rapid but transient change in FRET ratio that returned to basal levels approximately 1 minute after stimulation (Figure 5C). Comparison of the average maximum decrease in FRET ratio upon stimulation over 7–8 experiments (246–271 cells for each condition) revealed that the decrease in ER [Ca²⁺] in response to carbachol was not different between INS-1 cells and Ca_v1.2/II-III cells, but that the decrease in ER [Ca²⁺] in response to tolbutamide was significantly reduced in Ca_v1.2/II-III cells compared with INS-1 cells (Figure 5D). Thus, the data in Figure 5C show a rapid, transient decrease in ER [Ca²⁺] corresponding with the rapid, transient thapsigargin-sensitive rise of [Ca²⁺]_in stimulated by tolbutamide shown in Figure 3C. Further, the tolbutamide-stimulated transient drop in ER [Ca²⁺] and the corresponding rise of cytosolic [Ca²⁺] detected with Fura-2AM in INS-1 cells are both markedly reduced in Ca_v1.2/II-III cells.

Figure 5.

The drop in ER [Ca²⁺] in response to tolbutamide stimulation is diminished in Ca_v1.2/II-III cells compared with INS-1 cells. A, Quantification of the D1ER FRET window (difference between the R_max and R_min) in INS-1 and Ca_v1.2/II-III cells. INS-1 cells have a significantly greater dynamic range of ER calcium than Ca_v1.2/II-III cells. The R_max was calculated prior to any type of stimulation and represents the average of 26 experiments (total of 799 cells) in INS-1 cells and 21 experiments (total of 727 cells) in Ca_v1.2/II-III in cells (INS-1 cells: 2.81 ± 0.04; Ca_v1.2/II-III: 2.33 ± 0.04). The R_min was calculated through depletion of ER calcium stores using 30-minute treatment with thapsigargin followed by stimulation with carbachol. The minimal FRET ratio represents the average of 6 experiments (total of 259 cells) in INS-1 cells and 4 experiments (total of 139 cells) in Ca_v1.2/II-III in cells (INS-1 cells: 2.11 ± 0.07; Ca_v1.2/II-III: 1.99 ± 0.03). The D1ER window was calculated by subtracting the minimal FRET value from the basal FRET value (***, P < .001; Student's unpaired t test). B, Stimulation of INS-1 and Ca_v1.2/II-III cells with carbachol (500 μM; added at the 60-seconds time point) causes a rapid and prolonged decrease in the D1ER FRET signal. The INS-1 trace is a representative experiment that includes 49 cells. The Ca_v1.2/II-III trace is a representative experiment, which includes 83 cells. C, Stimulation of INS-1 and Ca_v1.2/II-III cells with tolbutamide (200 μM; added at the 60-seconds time point) causes a transient decrease in the D1ER FRET signal followed by a rapid recovery to baseline. The peak reduction in ER calcium levels is greater in INS-1 cells than Ca_v1.2/II-III cells. Data shown are representative experiments including 78 INS-1 cells, and 61 Ca_v1.2/II-III cells. D, Quantification of the decrease in D1ER FRET signal in INS-1 and Cav1.2/II-III cells following stimulation with carbachol or tolbutamide. Stimulation of ER calcium release by tolbutamide but not carbachol is significantly greater in INS-1 cells than in Ca_v1.2/II-III cells. Percent decrease in basal FRET ratio was calculated by dividing the change in FRET ratio by (R_max − R_min), and this was averaged among all experiments for both cell lines (carbachol: INS-1 cells 22.29 ± 1.79; Ca_v1.2/II-III cells 23.18 ± 2.39; Tolbutamide: INS-1 cells 8.43 ± 1.06; Ca_v1.2/II-III cells: 4.01 ± 1.76) (*, P < .05, INS-1 cells compared with Ca_v1.2/II-III cells, Student's unpaired t test). Carb, carbachol; Tolb, tolbutamide.

Glucose and GLP-1 in the presence of verapamil stimulate Ca²⁺-induced Ca²⁺ release events at a lower frequency in Ca_v1.2/II-III cells than in INS-1 cells

To further examine the ability of Ca²⁺ influx across the membrane to activate Ca²⁺ release from internal stores in INS-1 cells and Ca_v1.2/II-III cells, we measured changes in intracellular Ca²⁺ concentration in single cells stimulated with 10 mM glucose and 10 nM GLP-1 in the presence of 50 μM verapamil (2). The rationale for this protocol is that glucose and GLP-1 potentiate CICR, whereas verapamil decreases the frequency of L-type channel openings on the plasma membrane, allowing observation of discrete CICR events (35). Figure 6A shows 5 examples of INS-1 cells in which changes in intracellular Ca²⁺ concentration were measured using Fura-2. Cells were initially depolarized with application of 30 mM KCl in KRBH for 5 minutes in the absence of glucose, resulting in a large, transient rise in intracellular Ca²⁺ concentration. One minute into the 30 mM KCl depolarization, perfusion of 50 μM verapamil was initiated and maintained for the duration of the experiments. Four minutes after verapamil perfusion was initiated, 30 mM KCl perfusion was terminated, and perfusion with 10 mM glucose and 10 nM GLP-1 was initiated and maintained for the duration of the experiments. Perfusion with glucose and GLP-1 stimulated discrete, transient spikes in intracellular Ca²⁺ concentration that arose from, and resolved to, baseline levels in INS-1 cells. Figure 6B shows 5 examples of the same single-cell experiment performed with Ca_v1.2/II-III cells. In both INS-1 cells and Ca_v1.2/II-III cells, some cells displayed a gradual increase in baseline Ca²⁺ concentration, whereas other cells displayed little, if any, persistent increase in intracellular Ca²⁺ concentration. However, intermittent Ca²⁺ spikes were observed in both backgrounds. These intermittent Ca²⁺ spikes stimulated by glucose and GLP-1 in the presence of verapamil were significantly more frequent in INS-1 cells than in Ca_v1.2/II-III cells (Figure 6C; P < .05). In addition, the average spike amplitude and slope were both significantly reduced in Ca_v1.2/II-III cells compared with INS-1 cells (Figure 6, D and E; P < .001 and P < .05, respectively). Taken together, the data in Figure 6 suggest that influx of Ca²⁺ in response to glucose and GLP-1 stimulation in Ca_v1.2/II-III cells is less efficiently coupled to release of Ca²⁺ from internal stores than in INS-1 cells.

Figure 6.

Discrete spikes in intracellular Ca²⁺ concentration in response to glucose and GLP-1 in the presence of verapamil are less frequent in Ca_v1.2/II-III cells than in INS-1 cells. Representative single cell Ca²⁺ traces of INS-1 cells (A) or Ca_v1.2/II-III cells (B). Cells were loaded with 3 μM Fura-2AM, and perfused with KRBH for 30 seconds (basal) followed by 30 mM KCl for 1 minute. Coapplication of 50 μM verapamil for 4 minutes was used to bring the intracellular Ca²⁺ level ([Ca²⁺]_i) to basal; 10 mM glucose (Glu) +10 nM GLP-1+ 50 μM verapamil (verap) was perfused for the rest of the experiment to elicit CICR. C, The CICR peak numbers per cell for INS-1 and Ca_v1.2/II-III (2.261 ± 0.509 (n = 46) and 0.925 ± 0.290 (n = 40), respectively; *, P < .05, one-way ANOVA with the Tukey post hoc test). D, The average amplitude of CICR peaks (Δ340/380) is significantly greater in INS-1 cells than in Ca_v1.2/II-III cells (0.065 ± 0.003 [n = 108] and 0.03 ± 0.003 [n = 40], respectively; ***, P < .001, one-way ANOVA with the Tukey post hoc test). E, The initial slopes of CICR peaks in INS-1 cells are significantly greater than those of Ca_v1.2/II-III cells. The CICR peak slope is calculated as dividing the peak amplitude (Δ340/380) by the time (sec) from the lowest point to the highest point. The CICR peak slopes in INS-1 and Ca_v1.2/II-III cells are 0.02 ± 0.002 (n = 108) and 0.012 ± 0.001 (n = 40), respectively (*, P < .05, one-way ANOVA with the Tukey post hoc test). Glu, glucose; Verap, verapamil.

Glucose stimulates action potentials at a greater frequency in INS-1 cells and primary rat β-cells expressing the Ca_v1.2/II-III loop

Because glucose uptake and metabolism stimulate membrane depolarization and action potential firing in pancreatic β-cells, we measured changes in membrane potential in intact INS-1 cells and Ca_v1.2/II-II cells using the perforated patch configuration of the current clamp. Figure 7A shows representative traces of membrane depolarization stimulated by application of 18 mM glucose (arrows indicate time at which glucose was applied) to an INS-1 cell and a Ca_v1.2/II-III cell. The higher frequency of action potentials stimulated by glucose in the Ca_v1.2/II-III cells is clearly visible in these traces. Indeed, analysis of ≥ 20 cells of each type revealed that glucose stimulated action potentials at a significantly higher frequency in Ca_v1.2/II-III cells than in INS-1 cells (P < .01) (Figure 7D). This increase in action potential frequency was not the result of a stronger depolarization in response to glucose because neither the resting membrane potential (2.5 mM glucose) nor the average baseline membrane potential after application of 18 mM glucose (ie, the membrane potential from which the action potentials arose) were different between INS-1 cells and Ca_v1.2/II-III cells (Figure 7C). Activation of the small conductance, calcium-activated potassium channel (SK channel), is known to regulate action potential frequency in pancreatic β-cells (27). Therefore, we examined whether or not dysregulation of SK channel activity might play a role in the increased action potential frequency observed in Ca_v1.2/II-III cells. In the presence of the SK channel blocker apamin (1 μM), the frequency of 18 mM glucose-stimulated action potentials in INS-1 cells is significantly increased (Figure 7, B and D) but is not different from the frequency measured in Ca_v1.2/II-III cells in the absence of apamin (Figure 7D), suggesting that Ca²⁺ activation of SK channels in Ca_v1.2/II-III cells is aberrant. Indeed, in Ca_v1.2/II-III cells, application of 1 μM apamin doesn't significantly alter the frequency of action potentials stimulated by 18 mM glucose (Figure 7, B and D). To further understand the mechanism leading to increased glucose-stimulated action potential frequency in Ca_v1.2/II-III cells, we compared the magnitude of the action potential after hyperpolarizations (AHPs) measured in INS-1 cells and Ca_v1.2/II-III cells during trains of action potentials stimulated with 18 mM glucose. Figure 7E shows example traces of action potentials recorded from an INS-1 cell or a Ca_v1.2/II-III cell. Even though the initial action potentials both arise from a membrane potential of approximately −40 mV, the AHP in the Ca_v1.2/II-III cell is of lower amplitude (ie, less negative) than the INS-1 cell. Consequently, within the time frame of the traces, another action potential is generated in the Ca_v1.2/II-III cells, but not in the INS-1 cell. The mean AHP amplitude was significantly reduced in Ca_v1.2/II-III cells (−11.4 mV ± 0.1 mV; n = 16 cells, 1564 AHPs) compared with INS-1 cells (−13.5 mV ± 0.1 mV; n = 18 cells, 1087 AHPs; P < .001). Application of the SK channel blocker apamin (1 μM) significantly reduced the AHP amplitude of glucose-stimulated action potentials in INS-1 cells (−10.8 ± 0.1 mV; n = 8 cells, 430 AHPs; P < .001 compared with control INS-1 cells), but didn't significantly affect the AHP amplitude in Ca_v1.2/II-III cells (−11.1 ± 0.1 mV; n = 7 cells, 476 AHPs). Interestingly, the average AHP amplitude in Ca_v1.2/II-III cells in the absence of apamin was not significantly different from that observed in INS-1 cells in the presence of apamin. Taken together, the data in Figure 7 suggest that normal regulation of action potential frequency is disrupted in Ca_v1.2/II-III cells and that loss of SK channel activation could account for the significant increase in glucose-stimulated action potential frequency in these cells.

Figure 7.

SK channel modulation of glucose-stimulated action potentials is disrupted in Ca_v1.2/II-III cells. A, Representative current-clamp traces from an INS-1 cell and a Ca_v1.2/II-III cell showing the development of membrane depolarization and action potentials in response to perfusion with18 mM glucose (arrow). B, Representative current-clamp traces from an INS-1 cell and a Ca_v1.2/II-III cell showing the effect of bath application of 1 μM apamin (arrow) on a train of action potentials stimulated with 18 mM glucose. C, Membrane depolarization induced by 18 mM glucose is similar in INS-1 cells (white bar; n = 40) and Ca_v1.2/II-III cells (black bar; n = 23). D, Action potentials induced by 18 mM glucose in Ca_v1.2/II-III cells (black bar) have a frequency significantly greater than the frequency measured in INS-1 cells (white bar) (**, P < .01; n = 20–22). The SK channel blocker apamin (1 μM) increases the frequency of action potentials in INS-1 cells (*, P < .05; n = 8–22). The frequency of glucose-stimulated action potentials in Ca_v1.2/II-III cells is not significantly different in the presence or absence of apamin (n = 7–20). E, AHPs of glucose-stimulated action potentials are reduced in Ca_v1.2/II-III cells compared with INS-1 cells. Example action potentials stimulated by glucose in an INS-1 cell (black trace) or a Ca_v1.2/II-III cell (gray trace) are shown. Analysis of 18 INS-1 cells (1087 AHPs) and 16 Ca_v1.2/II-III cells (1564 AHPs) revealed a significant decrease in the amplitude of the AHP of glucose-stimulated action potentials in Ca_v1.2/II-III cells compared with INS-1 cells (***, P < .001). Apamin (1 μM) significantly reduces AHP amplitude in INS-1 cells (8 cells, 430 AHPs; ***, P < .001), but not in Ca_v1.2/II-III cells (7 cells, 470 AHPs).

Given the marked effects of Ca_v1.2/II-III loop expression on glucose-stimulated electrical activity observed in INS-1 cells, we asked whether this could be recapitulated in primary rat β-cells. To this end, we constructed an adenovirus encoding the same Ca_v1.2/II-III loop-GFP fusion expressed in Ca_v1.2/II-III cells and used it to transduce primary rat β-cells. Figure 8A shows representative current-clamp traces recorded from primary rat β-cells transduced with either a control adenovirus encoding GFP only (left panel) or with the adenovirus encoding the Ca_v1.2/II-IIII loop-GFP fusion (right panel). The arrows indicate the time at which the perfusate was switched from 2.5 mM glucose to 18 mM glucose. Neither the resting membrane potential (measured at 2.5 mM glucose) nor the baseline membrane potential measured at 18 mM glucose in cells transduced with the Ca_v1.2/II-III-GFP encoding virus were different from those measured in cells transduced with the control virus (Figure 8B). However, cells transduced with the Ca_v1.2/II-III-GFP virus fired action potentials in response to 18 mM glucose at a significantly greater frequency than cells transduced with the control virus encoding GFP alone (Figure 8C)(P < .01). Analysis of AHP amplitude in GFP and Ca_v1.2/II-III-GFP-transduced β-cells (Figure 8D) revealed that the AHP amplitude of glucose-stimulated action potentials was significantly reduced in cells expressing Ca_v1.2/II-III-GFP (−8.6 ± 0.1 mV; n = 9 cells, 432 AHPs) compared with controls expressing GFP only (−9.8 ± 0.2 mV; n = 6 cells, 71 AHPs) (P < .001, Student's unpaired t test). Thus, consistent with what was observed in the INS-1 and Ca_v1.2/II-III cell lines, acute expression of the Ca_v1.2/II-III loop in primary rat β-cells results in an increase in the frequency of glucose-stimulated action potentials.

Figure 8.

Glucose-induced action potential frequency is greater in rat β-cells expressing Ca_v1.2/II-III compared with controls expressing GFP alone. A, Representative current clamp traces from a rat primary β-cell transduced with adenovirus encoding GFP only (left panel) and a rat primary β-cell transduced with adenovirus encoding Ca_v1.2/II-III-GFP (right panel) in response to perfusion with 18 mM glucose (arrow), which triggers membrane depolarization and action potentials. B, Resting membrane potential and membrane depolarization induced by 18 mM glucose are not different in β-cells transduced with the GFP adenovirus (white bar; n = 26) or the Ca_v1.2/II-III-GFP adenovirus (black bar; n = 20). C, Glucose stimulates action potentials at a greater frequency in β-cells transduced with the Ca_v1.2/II-III-GFP adenovirus (black bar; n = 9) than in the control β-cells transduced with the GFP adenovirus (white bar; n = 6) (**, P < .01). D, The amplitude of AHPs measured during trains of glucose-stimulated action potentials were significantly reduced in primary rat β-cells transduced with the Ca_v1.2/II-III-GFP adenovirus compared with cells transduced with adenovirus encoding GFP alone (***, P < .001, Student's t test).

SK-channel current is detected in both INS-1 cells and Ca_v1.2/II-III cells

Given the apparent lack of SK channel regulation of glucose-stimulated action potentials in β-cells expressing the Ca_v1.2/II-III loop, we asked whether or not SK channel activity could be detected in Ca_v1.2/II-III cells. Whole-cell K⁺ currents were measured in INS-1 cells and Ca_v1.2/II-III cells using steps to −50 mV from a holding potential of −70 mV, with a calculated free Ca²⁺ concentration of 2 μM in the recording pipette. Under these conditions, the major K⁺ channels available to conduct current are the K_ATP channel and SK channels, because voltage-dependent channels are not activated. This protocol elicited an outward current at −50 mV in both INS-1 and Ca_v1.2/II-III cells (Figure 9A). Application of 100 nM apamin did not block a significant fraction of this current in either cell line if the pipette solution did not contain free Ca²⁺ (Figure 9B). In contrast, if 2 μM free Ca²⁺ was included in the pipette solution, 100 nM apamin blocked a significant portion of the K⁺ currents measured using this protocol (Figure 9B). The density of apamin-sensitive outward K⁺ current measured under these conditions was not different between INS-1 and Ca_v1.2/II-III cells (Figure 9C). Thus, the lack of modulation of glucose-stimulated AP frequency by apamin in Ca_v1.2/II-III cells shown in Figures 7 and 8 is not the result of the absence of functional SK channels. We next asked whether block of SK channels with apamin could modulate glucose-stimulated insulin secretion in either INS-1 cells or Ca_v1.2/II-III cells, reasoning that it may also be differentially regulated by apamin in INS-1 cells and Ca_v1.2/II-III cells. However, we found that glucose-stimulated insulin secretion was significantly potentiated by apamin in both cell types (Figure 9D). To understand this difference in apamin modulation of glucose-dependent events, we examined the time course of the rise in intracellular [Ca²⁺] stimulated by 18 mM glucose in INS-1 cells and Ca_v1.2/II-III cells using Fura-2. As shown in Figure 9E, the rise in intracellular [Ca²⁺] was delayed over the first several minutes of glucose stimulation in Ca_v1.2/II-III cells compared with INS-1 cells. Taken together, the data in Figure 9 suggest that apamin does not modulate glucose-stimulated action potentials in Ca_v1.2/II-III cells because the glucose-dependent rise in cytoplasmic [Ca²⁺] is not sufficient to activate SK channels within the time frame of current-clamp experiments (ie, several minutes after addition of glucose). However, within the longer time frame of the glucose-stimulated insulin secretion assay (ie, 60 minutes of glucose stimulation) SK channels are eventually activated in Ca_v1.2/II-III cells and exert the expected inhibitory effect on secretion, which is relieved by apamin blockade of the channels.

Figure 9.

SK channels are activated by intracellular Ca²⁺ in both INS-1 and Ca_v1.2/II-III cells. A, Representative whole-cell voltage clamp traces of INS-1 (right panel) and 1.2/II-III (left panel) cells demonstrating the Ca²⁺-dependent K⁺ current blocked by apamin. Cells were depolarized to −50 mV for 1 second from a holding potential of −70 mV. When the pipette solution contained 0 μM free Ca²⁺, no current was blocked by 100 nM apamin. Tolbutamide block of K_ATP channel current was detected both in the absence and presence of free Ca²⁺ in the pipette, and was readily reversed upon washout. B, Comparison of fraction current remaining after 100 nM apamin between INS-1 (n = 8 at 2 μM [Ca²⁺]_in and n = 5 at 0 [Ca²⁺]_in) and 1.2/II-III (n = 8 at 2 μM [Ca²⁺]_in and n = 6 at 0 [Ca²⁺]_in) cells. Data are shown as mean values ± SEM. (0.65 ± 0.08 (n = 8) and 1.05 ± 0.08 (n = 5) at 2 μM [Ca²⁺]_in and 0 μM [Ca²⁺]_in, respectively) and in 1.2/II-III cells (0.69 ± 0.05 (n = 8) and 0.920 ± 0.0651 (n = 6) at 2 μM [Ca²⁺]_in and 0 μM [Ca²⁺]_in, respectively). The fraction of current remaining in the presence of 100 nM apamin was significantly lower with 2 μM free Ca²⁺ in the pipette than in the absence of Ca²⁺ (*, P < .05, one-way ANOVA with Tukey post -hoc test). C, Comparison of SK current density between INS-1 (n = 8) and 1.2/II-III (n = 8) cells. Bar graphs represent the mean current density of apamin-sensitive current with 2 μM Ca²⁺ in the pipette solution expressed as pA/pF ± SEM (INS-1 cells: 1.280 ± 0.314, n = 8; 1.2/II-III cells: 0.993 ± 0.381, n = 8)). The SK current density in Ca_v1.2/II-III cells is not different from that measured in INS-1 cells (P = .234, Mann-Whitney Rank Sum Test). D, Glucose-stimulated insulin secretion is potentiated by 100 nM apamin in both INS-1 and Ca_v1.2/II-III cells. Cells were stimulated with 10 mM glucose for 1 hour, and insulin secretion was assayed as described in Materials and Methods. The addition of apamin significantly increased insulin secretion compared with glucose alone (P < 0.05; Student's unpaired t test). E, The rise in intracellular Ca²⁺ in response to 18 mM glucose is delayed in Ca_v1.2/II-III cells compared with INS-1 cells. Cells were loaded with Fura-2, and stimulated by injection of 18 mM glucose. The 340 nm to 380 nm ratio was measured every minute for 1 hour. Control experiments were performed with injection of KRBH containing no glucose. The data shown are corrected by subtraction of the slow, linear increase in 340:380 ratio observed in the absence of glucose. Data shown are the means ± SEM of an experiment done in triplicate and are representative of 3 independent experiments. ms, milliseconds; tolb, tolbutamide; wash, washout.

IQGAP1 and eIF3e coimmunoprecipitate with the Ca_v1.2/II-III loop

Given the marked effects of the expression of the Ca_v1.2/II-III loop in INS-1 cells and primary β-cells on Ca²⁺ signaling, we examined the ability of the Ca_v1.2/II-III loop to interact with proteins in INS-1 cells using coimmunoprecipitation experiments. The Ca_v1.2/II-III loop was previously reported to bind directly to the C2 domains of RIM2 (36) and Piccolo (19) and to coimmunoprecipitate RIM2 from INS-1 cell lysates (20). The GTP-binding protein CDC42 is reported to play a key role in regulating actin polymerization during glucose-stimulated insulin secretion (37) (38). IQGAP1 is a scaffold protein that binds CDC42 in a manner that is regulated by Ca²⁺-calmodulin (39). Further, the C-terminal regions of IQGAP1and -2 were recently found to form a pseudo-C2 fold, capable of binding phosphatidylinositol 3,4,5,-trisphosphate (40). Therefore, we examined whether the Ca_v1.2/II-III loop could interact with IQGAP1 by immunoprecipitating the loop with antibodies against GFP and blotting for IQGAP1. As shown in Figure 10A, IQGAP1 coimmunoprecipitated with neither GFP transiently expressed in INS-1 cells, nor with the GFP-tagged Ca_v1.3/II-III loop stably expressed in INS-1 cells. In contrast, immunoprecipitation of the GFP-tagged Ca_v1.2/II-III loop from Ca_v1.2/II-III cell lysates coimmunoprecipitated IQGAP1.

Figure 10.

The Ca_v1.2/II-III loop interacts with IQGAP1 and eIF3e in INS-1 cells. A, IQGAP1 coimmunoprecipitates with the Ca_v1.2/II-III loop/GFP fusion peptide, but not the Ca_v1.3/II-III loop fusion peptide, from INS-1 cell lysates. Cell lysates were incubated with anti-GFP antibodies coupled to agarose, washed, and eluted by incubation at 80°C in Laemmli buffer. Ly, 100 μg of cell lysate; Ub, unbound proteins removed after incubation with agarose-immobilized antibodies; W, protein in the last wash fraction before elution; E, protein eluted by incubation at 80°C in Laemmli buffer. Samples were blotted with antibodies against IQGAP1 (upper panel) or GFP (lower panel). Blots shown are representative of 3 separate experiments. B, The Ca_v1.2/II-III loop coimmunoprecipitates with eIF3e. Ca_v1.2/II-III lysates were incubated with antibodies to eIF3e and protein A-sepharose, washed, and eluted with Laemmli buffer. Lanes are labeled as described above. Samples were blotted with antibodies against GFP to detect the Ca_v1.2/II-III loop-GFP fusion peptide (upper panel) or with antibodies against eIF3e (lower panel). Blots shown are representative of 4 separate experiments. IB, immunoblotting; IP, immunoprecipitation.

The translation initiation factor eIF3e was recently reported to mediate Ca²⁺-dependent internalization of Ca_v1.2 in INS-1 cells (41). eIF3e was previously reported to bind directly to a 3-amino acid sequence toward the C-terminal end of the II-III loop of Ca_v1.1, 1.2, 1.3, 2.2, and 2.3 (42). Disruption of a mechanism leading to internalization of Ca_v channels could account for the marked increase in Ca_v current density that we observed in Ca_v1.2/II-III cells (Figure 2D). We therefore asked whether the Ca_v1.2/II-III loop was interacting with eIF3e using immunoprecipitation of eIF3e and Western blotting with antibodies to GFP to detect the Ca_v1.2/II-III loop fusion. As shown in Figure 10B, immunoprecipitation of eIF3e did coimmunoprecipitate the Ca_v1.2/II-III loop. We were unable to detect coimmunoprecipitation of the Ca_v1.2/II-III loop and eIF3e using antibodies against GFP (data not shown). It is possible that eIF3e binding to the Ca_v1.2/II-III loop/GFP fusion hinders binding of antibodies to GFP because the eIF3e binding motif and the GFP fusion site are both on the C-terminal end of the loop. These data suggest that eIF3e and the Ca_v1.2/II-III loop interact in Ca_v1.2/II-III cells in a manner that may result in accumulation of Ca_v channels in the plasma membrane of these cells.

Discussion

Key proteins involved in triggering insulin secretion are present in lipid rafts in INS-1 cells

Lipid raft localization of Ca_v1.2 has been reported in insulin-secreting cells (20, 32), but the functional significance of this localization is largely unclear. However, the scaffolding proteins RIM2 and Piccolo were also detected in lipid rafts prepared using sucrose gradient centrifugation (20) and are reported to bind directly to the Ca_v1.2 II-III loop via their C2 domains (19). Moreover, RIM2 is coimmunoprecipitated from INS-1 cell lysates by the II-III loop of Ca_v1.2 but not that of Ca_v1.3 (20). Further, the K_ATP channel subunit SUR1 binds to the cAMP effector EPAC2 (19). Activation of EPAC2 by cAMP in pancreatic β-cells is linked to amplification of CICR via activation of PLC-ϵ (43) and a 2-APB-sensitive Ca²⁺ influx (44). Here we show that SUR1 and EPAC2 are highly enriched in lipid rafts prepared from INS-1 cell lysates (Figure 1), suggesting that these proteins, along with Ca_v1.2, may indeed colocalize to lipid rafts on the membrane of INS-1 cells. In contrast, K_ir6.2, although present in lipid rafts, is more widely distributed throughout sucrose density gradients, consistent with reports that it also resides within secretory granules in β-cells (45, 46), and that K_ir6.2 is present in secretory granules in β-cells from Sur1^−/− mice (47). Interestingly, expression of the Ca_v1.2/II-III loop did not appear to affect the distribution of Piccolo and RIM2 (20), or SUR1, EPAC2, and K_ir6.2 (this study) on sucrose density fractions, in contrast to the marked shift of Ca_v1.2 to higher density fractions in Ca_v1.2/II-III cells (20) reported previously. Our finding that SUR1 and K_ir6.2 are found in lipid rafts in INS-1 cells contrasts with a previous study, using HIT-T15 cells, which found them to be excluded from lipid rafts, whereas Ca_v1.2 was found to be highly localized to lipid rafts (32). The reason for this discrepancy is not clear; however, the colocalization of Ca_v1.2, SUR1, K_ir6.2, Piccolo, RIM2, and EPAC2 in lipid rafts is consistent with reports that these proteins interact physically and functionally (19, 21).

Expression of the Ca_v1.2/II-III loop does not disrupt K_ATP channel activity but increases Ca_v channel current density

We compared the electrophysiological properties of the Ca_v1.2/II-III cells with those of INS-1 cells (Figure 2). We found no difference between Ca_v1.2/II-III cells and INS-1 cells in regard to either whole-cell K_ATP channel current density and sensitivity to block by the sulfonylurea tolbutamide, or to the potency of membrane depolarization and action potential frequency (data not shown) stimulated by tolbutamide in current clamp experiments conducted using the whole-cell configuration. However, as reported previously (20), we did find that whole-cell Ca_v current density (measured using Ba²⁺ as the permeant ion) was significantly increased in Ca_v1.2/II-III cells compared with INS-1 cells. This increase in Ca_v current density was not confined to L-type channels because the portion of current blocked by 10 μM nifedipine was not different in INS-1 cells and Ca_v1.2/II-III cells (20). The mechanism for this increase in Ca_v current density is suggested by the finding that the translation initiation factor eIF3e binds to the II-III loop of multiple Ca_v channels in a Ca²⁺-dependent manner to mediate internalization from the plasma membrane (42). Very recently, this mechanism has been reported to regulate the density of Ca_v1.2 channel on the plasma membrane of INS-1 cells (41). The 3-amino acid motif in the Ca_v1.2 II-III found to be critical for interaction with eIF3e (33) is present in the Ca_v1.2/II-III loop-GFP fusion used in this study. Indeed, we found that antibodies to eIF3e could coimmunoprecipitate the Ca_v1.2/II-III loop. Thus, the marked increase in Ca_v current density is likely driven, at least in part, by Ca_v1.2/II-III/GFP fusion interference with retrieval of Ca_v channels from the plasma membrane via the eIF3e-dependent mechanism. However, even with the increase in Ca_v current density, CICR is diminished in Ca_v1.2/II-III cells. It is tempting to speculate that interruption of the Ca_v1.2 channel-piccolo and or -RIM2 interactions by expression of the Ca_v1.2/II-III loop excludes Ca_v1.2 from a signaling complex critical for CICR, but does not otherwise interfere with insertion of Ca_v1.2 into the plasma membrane. However, this study cannot exclude the possibility that loss of eIF3e modulation of Ca_v1.2 plasma membrane expression, or interruption of a Ca_v1.2/IQGAP1 interaction, is contributing to the deficits in CICR and SK channel activation observed in Ca_v1.2/II-III cells. Interestingly, IQGAP1 is an essential component of the Wnt-receptor-actin-myosin-polarity structure, which recruits cortical ER to the plasma membrane during directional cell movement (48). It will be of interest to determine the relative roles of Ca_v1.2 interactions with scaffolding proteins and/or eIF3e in CICR and SK channel activation in pancreatic β-cells.

Ca²⁺-induced Ca²⁺ release is diminished in Ca_v1.2/II-III cells compared with INS-1 cells

Given that we detected no deficits in K_ATP channel function or sensitivity to tolbutamide in Ca_v1.2/II-III cells compared with INS-1 cells, we examined both insulin secretion and Ca²⁺ transients simulated by tolbutamide in these cell lines (Figure 3). Secretion stimulated by tolbutamide in INS-1 cells is completely inhibited by nicardipine and partially inhibited by pretreatment with thapsigargin. This pharmacologic profile is consistent with contributions by both Ca²⁺ flux across the membrane through L-type channels (thapsigargin-resistant portion) as well as from Ca²⁺-induced release of Ca²⁺ from internal stores (thapsigargin-sensitive portion) to tolbutamide-stimulated insulin secretion. In contrast, insulin secretion stimulated by tolbutamide in Ca_v1.2/II-III cells was completely blocked by nicardipine, but completely resistant to pretreatment with thapsigargin, suggesting that release of Ca²⁺ from internal stores does not contribute to tolbutamide-stimulated insulin secretion in these cells. Indeed, the strong inhibition of the rapid tolbutamide-stimulated peak in intracellular Ca²⁺ concentration by thapsigargin in INS-1 cells was markedly reduced in Ca_v1.2/II-III cells, supporting the conclusion that Ca²⁺-induced Ca²⁺ release from internal stores is disrupted in these cells. This conclusion was strongly supported by measurements of ER [Ca²⁺] changes with the D1ER Ca²⁺ indicator (Figure 5). Indeed, the kinetics of the decrease in ER [Ca²⁺] upon tolbutamide stimulation closely reflected the corresponding increase in cytosolic Ca²⁺ measured with Fura-2 in INS-1 cells, and both responses were similarly inhibited in Ca_v1.2/II-III cells. Interestingly, the observation that the D1ER FRET window was significantly lower in Ca_v1.2/II-III cells compared with INS-1 cells suggests that localization of Ca_v1.2 via the II-III loop may contribute to maintenance of ER Ca²⁺ stores.

To determine whether the above observations were unique to sulfonylurea stimulation, we compared CICR in INS-1 and Ca_v1.2/II-III cells stimulated with GLP-1 and glucose in the presence of 50 μM verapamil (Figure 6). The rationale behind these experiments is to enhance the sensitivity of RYR2-mediated release of Ca²⁺ from stores via the actions of GLP-l (49, 50), but minimize the Ca²⁺ influx via L-type Ca²⁺ channels, which is also enhanced under these conditions (51), using the Ca²⁺ channel blocker verapamil to reduce the probability of channel opening (52). This protocol is reported to generate discrete CICR events in pancreatic β-cells (2), and indeed, we observed, discrete, intermittent Ca²⁺ spikes using this protocol with INS-1 cells. When this protocol was applied to Ca_v1.2/II-III cells, these discrete Ca²⁺ spikes were detected less frequently and were of a significantly lower amplitude and slope compared with those detected in INS-1 cells. These data further support the conclusion that Ca²⁺ influx via Ca_v1.2 channels is not efficiently coupled to release of Ca²⁺ from internal stores in Ca_v1.2/II-III cells during glucose + GLP-1 stimulation.

Given the increased Ca_v current density in Ca_v1.2/II-III cells, we propose that the observed disruption of CICR is the result of a spatial uncoupling of Ca_v1.2 and RYR2 in Ca_v1.2/II-III cells. An alternative explanation might be a deficit in the RYR2 response to increases in intracellular Ca²⁺ concentration in Ca_v1.2/II-III cells. However, our findings that replacement of extracellular Ca²⁺ with Sr²⁺ restored the thapsigargin-sensitive, tolbutamide-stimulated early phase of the Ca²⁺ transient in Ca_v1.2/II-III cells (Figure 3), and that the EPAC2-selective cAMP analog 8-pCPT-2′-O-Me-cAMP-AM strongly potentiated both tolbutamide-stimulated insulin secretion and Ca²⁺ transients in Ca_v1.2/II-III cells (Figure 4) argue against this interpretation.

Ca_v1.2/II-III cells exhibit an increased frequency of action potentials in response to glucose

If coupling of Ca²⁺ influx via Ca_v1.2 to intracellular signaling has, in fact, been disrupted in Ca_v1.2/II-III cells, then electrical excitability in response to glucose could be enhanced in Ca_v1.2/II-III cells compared with INS-1 cells, if it contributes to activation of small-conductance Ca²⁺-activated K⁺ (SK) channels in pancreatic β-cells. SK channels are expressed in mouse β-cells (53), INS-1 cells (25), and human β-cells (27). Pancreatic β-cells from SK-4 channel knockout mice exhibited an increase in glucose-stimulated action potential frequency (54), and block of SK channels in human β-cells with the toxin apamin increases glucose-stimulated action potential frequency, reflecting the role of SK channels in mediating the AHP (25). Our finding that glucose-stimulated action potentials in both Ca_v1.2/II-III cells and in primary rat β-cells transduced with an adenovirus encoding the Ca_v1.2/II-III loop occurred at a frequency approximately twice that of their appropriate controls (Figures 7 and 8), supports the conclusion that interaction of Ca_v1.2 with other proteins via the II-III loop is required for normal regulation of SK channels. Loss of SK channel activation as the mechanism for this increase in glucose-stimulated excitability is strongly supported by the insensitivity of glucose-stimulated action potentials in Ca_v1.2/II-III cells to apamin and the highly significant decrease in the magnitude of the AHP during trains of glucose-stimulated action potentials in Ca_v1.2/II-III cells and primary β-cells transduced with the Ca_v1.2/II-III-GFP virus. Further, the data in Figure 9 demonstrate that SK channels are activated in Ca_v1.2/II-III cells if 2 μM Ca²⁺ is provided in the pipette solution. Thus, the primary deficit leading to dysregulation of SK channels in Ca_v1.2/II-III cells is likely the prolonged delay in the rise of [Ca²⁺]_in in response to glucose.

In conclusion, this study provides evidence that the Ca_v1.2 II-III loop is critical for efficient CICR and regulation of SK channel activity in pancreatic β-cells. Specific interactions between Ca_v1.2 and RIM2, Piccolo, IQGAP1, and eIF3e may contribute to maintenance of CICR and/or SK channel regulation, but the role that each of these individual proteins plays remains to be determined. In type 2 diabetes, defects in β-cell Ca²⁺ responses to glucose are commonly observed (55). A recent study reported that transgenic mice expressing a phosphomimic, gain-of-function mutant of RYR2 (S2814D) had diminished Ca²⁺ responses to glucose and KCl and developed glucose intolerance (56). Thus, it is conceivable that derangements in CICR could contribute to type 2 diabetes. Interestingly, exposure of INS-1 cells to 28 mM glucose for 72 hours disrupted lipid rafts and inhibited insulin secretion in response to both KCl and glucose (57), an effect that was attributed to glucose inhibition of cholesterol synthesis. In addition, exposure of pancreatic islets to palmitate inhibited glucose-stimulated insulin secretion and dispersed microdomains of elevated intracellular Ca²⁺ concentration in response to brief episodes of membrane depolarization in pancreatic β-cells (58). It will be of interest to examine whether disruption of Ca_v1.2 localization, and the loss of CICR efficiency and regulation of glucose-stimulated electrical excitability described herein, plays a role in pancreatic β-cell dysfunction induced by glucotoxicity or lipotoxicity.