Disruption of pancreatic beta-cell lipid rafts modifies Kv2.1 channel gating and insulin exocytosis.

In pancreatic beta-cells, the predominant voltage-gated Ca(2+) channel (Ca(V)1.2) and K(+) channel (K(V)2.1) are directly coupled to SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein (SNAP) receptor) proteins. These SNARE proteins modulate channel expression and gating and closely associate these channels with the insulin secretory vesicles. We show that K(V)2.1 and Ca(V)1.2, but not K(V)1.4, SUR1, or Kir6.2, target to specialized cholesterol-rich lipid raft domains on beta-cell plasma membranes. Similarly, the SNARE proteins syntaxin 1A, SNAP-25, and VAMP-2, but not Munc-13-1 or n-Sec1, are associated with lipid rafts. Disruption of the lipid rafts by depleting membrane cholesterol with methyl-beta-cyclodextrin shunts K(V)2.1, Ca(V)1.2, and SNARE proteins out of lipid rafts. Furthermore, methyl-beta-cyclodextrin inhibits K(V)2.1 but not Ca(V)1.2 channel activity and enhances single-cell exocytic events and insulin secretion. Membrane compartmentalization of ion channels and SNARE proteins in lipid rafts may be critical for the temporal and spatial coordination of insulin release, forming what has been described as the excitosome complex.

In the pancreatic islets of Langerhans, glucose uptake by ␤-cells initiates a cascade of cellular events resulting in insulin secretion. A key response leading to insulin release is the change in transmembrane potential associated with the opening and closing of ion channels. Glucose uptake and metabolism increases the ratio of ATP/ADP, leading to the blockade of ATP-sensitive potassium (K ϩ -ATP) channels. Inhibition of these channels results in cell membrane depolarization and subsequent activation of voltage-gated Ca 2ϩ (Ca V ) 1 channels. Influx of extracellular Ca 2ϩ through Ca V channels causes os-cillatory elevations in [Ca 2ϩ ] i , fusion of insulin-containing vesicles with the cell membrane, and insulin release (reviewed in Ref. 1). This entire process is suppressed or terminated by the opening of voltage-gated K ϩ (K V ) channels (2). The integrated process of channel gating is critical for the coordination of insulin release and thus the consequent maintenance of proper plasma glucose levels.
Pancreatic ␤-cells and clonal insulinoma cells express four different families of K V channels (K V 1, K V 2, K V 3, K V 4) in variable levels (2)(3)(4). K V 2.1 is the most abundant K V channel isoform expressed in both isolated islet ␤-cells and insulinoma cells. To support this notion, the dominant-negative knockout of K V 2.1 channel or pharmacological blockade with a selective K V 2.1 antagonist reduces steady-state outward K V currents by ϳ60 -70% (2,5). In addition to K V 2.1, other K V channel ␣ subunits are expressed in pancreatic ␤-cells to a lesser extent, including K V 1.4 and K V 1.6, which account for less than 25% of outward K ϩ currents measured in these cells (2).
The central role of Ca V channels in insulin secretion is well recognized (1). The predominant Ca V channel in ␤-cells is the L-type channel (long-lasting; Ca V 1.2/␣ 1C-a and Ca V 1.3/␣ 1D ) (6,7). The L-type Ca V channel antagonists nifedipine and verapamil inhibit glucose-induced insulin release (6,7), whereas L-type Ca V channel agonists (BAY-K 8644 and CGP28392) augment the release of insulin (8,9). Recent evidence from Hofmann and co-workers (7) has demonstrated a greater role of Ca V 1.2 in regulating insulin secretion in comparison with Ca V 1.3 in mouse ␤-cells. Specifically, tissue-directed knockout of Ca V 1.2 inhibited the first phase of insulin secretion (7), whereas in contrast, Ca V 1.3 knockout either had no effect on ␤-cell function (10) or caused some developmental alterations in ␤-cell growth and proliferation (6).
Lipid rafts are specialized membrane microdomains highly enriched with glycosphingolipids and cholesterol unlike other membrane regions, which are comprised mainly of phospholipids (reviewed in Refs. 11 and 12). The lipid composition of these raft domains and the tight packing of the acyl lipid chain make these membrane fractions resistant to Triton X-100 or Na 2 CO 3 solubilization. This characterization and their buoyancy on sucrose gradients have been utilized to purify lipid rafts. Caveolin, a key structural protein of caveolae, has been used as a marker protein to isolate caveolae (13). Interest in caveolae has emerged due to the numerous associated membrane and cell signaling proteins within this complex, including ion channels, G-protein coupled receptors, insulin receptor, kinases, adenylate cyclase, structural proteins, and SNARE proteins, which are collectively important for exocytosis of neurotransmitters and hormones (13).
In the present study, we examined the localization of pancreatic ␤-cell ion channels and SNARE proteins in cholesterolrich lipid raft domains. We demonstrate that K V 2.1 and Ca V 1.2, but not Kv1.4, Kir6.2, or SUR1, are enriched in lipid rafts. Furthermore, the SNARE proteins syntaxin-1A, SNAP-25, and VAMP-2 target to lipid rafts but not n-Sec1 or Munc-13-1. Depletion of membrane cholesterol results in redistribution of these proteins out of lipid rafts and consequently alter K V 2.1 but not Ca V 1.2 channel amplitude and gating. Moreover, there is a marked enhancement of glucose-induced insulin secretion and single-cell exocytic events following cholesterol depletion. These results suggest that the targeting of ion channels and SNARE proteins to cholesterol-rich lipid rafts in ␤-cells is important for coordinating cellular excitability and insulin exocytosis.
Rat Islet and ␤-Cell Isolations-Isolation of rat islets by collagenase digestion was performed as described previously (15). Dispersion of islets with 0.015% trypsin in Ca 2ϩ -and Mg 2ϩ -free phosphate-buffered saline was used to isolate single ␤-cells. Both islets and ␤-cells were cultured in low glucose Roswell Park Memorial Institute (RMPI) 1640 medium supplemented with 10% fetal bovine serum, 0.25% HEPES, and 100 units/ml penicillin-100 g/ml streptomycin. ␤-cells were cultured for up to 4 days.
Cell Culture-Hamster HIT-T15 cells were provided to us by Dr. R. P. Robertson (Pacific Northwest Research Institute, Seattle, WA). Rat INS-1 cells were obtained from the American Tissue Culture Collection, and mouse MIN6 cells were a gift from Dr. Susumu Seino (Chiba University, Chiba, Japan). HIT-T15 and INS-1 cells were cultured in RPMI 1640 medium, and MIN6 cells were grown in Dulbecco's modified Eagle's medium. The media were supplemented with 10% fetal bovine serum and 100 units/ml penicillin-100 g/ml streptomycin. For INS-1 and MIN6 insulinoma cells, the media were supplemented with 2 l/500 ml ␤-mercaptoethanol. Membrane cholesterol depletion of HIT-T15 cells was performed by incubating 10 mM methyl-␤-cyclodextrin for 30 min at 37°C.
Confocal Immunofluorescence Microscopy-Laser confocal immunofluorescence microscopy was performed as described previously (16). Rat islets and ␤-cells were fixed in 2% formaldehyde for 0.5 h at room temperature, blocked with 5% normal goat serum and 0.1% saponin for 0.5 h at room temperature, and then immunolabeled with mouse monoclonal anti-caveolin-1 or -2 (1:20 dilution) and guinea pig anti-insulin (1:200 dilution) overnight at 4°C. The coverslips were rinsed with 0.1% saponin and phosphate-buffered saline and then incubated with the appropriate fluorescent-labeled secondary antibodies (either rhodamine red or fluorescein isothiocyanate) for 1 h at room temperature and mounted on slides in a fading retarder (0.1% p-phenylenediamine in glycerol). Images were obtained using a Zeiss LSM-410 laser scanning confocal imaging system (Carl Zeiss, Oberkochen, Germany).
Lipid Raft Isolation-HIT-T15 cells were harvested and lysed in a sodium carbonate solution (500 mM Na 2 CO 3 , pH 11, supplemented with protease inhibitors) using a sonicator. Lysed cells were centrifuged at 2000 rpm for 15 min at 4°C. The supernatant was diluted with an equal volume of a 90% sucrose solution in MES-buffered saline (25 mM MES, 150 mM NaCl, pH 6.5, supplemented with protease inhibitors) and placed into an Ultracentrifuge tube. A 5-30% discontinuous sucrose gradient was formed above the 45% sucrose-HIT-T15 lysate, and the sample was centrifuged at 39,000 rpm for 16 -20 h at 4°C. Gradient fractions (600 l) were collected from the top, and 30 l of each fraction was loaded onto an SDS-PAGE gel. The protein was transferred to polyvinylidene difluoride-plus membranes (Fisher) and immunoblotted with the desired primary antibodies. Primary antibodies were labeled with peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology). Proteins were detected by chemiluminescence (ECL-Plus, Amersham Biosciences), and membranes were exposed to x-ray film (Eastman Kodak Co.).
Insulin Secretion Assay-HIT-T15 cells were plated in 12-well culture dishes at 5 ϫ 10 5 cells/well. The following day, cells were washed twice and preincubated for 30 min in Krebs-Ringer bicarbonate buffer containing (in mM): 115 NaCl, 5 KCl, 24 NaHCO 3 , 2.5 CaCl 2 , 1 MgCl 2 , 10 HEPES, and 0.1% bovine serum albumin with or without 10 mM methyl-␤-cyclodextrin. After the preincubation period, cells were washed thoroughly with Krebs-Ringer bicarbonate buffer and incubated in Krebs-Ringer bicarbonate buffer supplemented with 10 mM glucose for 1 h at 37°C. Aliquots (200 l) were taken at the end of the incubation period and stored at Ϫ20°C until assayed for insulin. Radioimmunoassays were performed using the rat insulin radioimmunoassay kit (Linco Research, Inc., St. Charles, MO).
Single-cell Capacitance Measurement-Patch electrodes were coated with orthodontic wax (Butler, Guelph, ON, Canada) close to the tips and fire-polished. Pipette resistance ranged from 2 to 5 megaohms when pipettes were filled with the intracellular pipette solutions. The pipette solution contained (in mM): 125 K-glutamate, 10 KCl, 10 NaCl, 1 MgCl 2 , 5 HEPES, 0.05 EGTA, 0.1 cAMP, and 4 MgATP, pH to 7.1. The extracellular solutions consisted of (in mM): 138 NaCl, 5.6 KCl, 1.2 MgCl 2 , 2.6 CaCl 2 , 5 HEPES, and 2.5 D-glucose, pH to 7.4. Some rat islet ␤-cells were preincubated with the extracellular solutions containing 10 mM methyl-␤-cyclodextran for 20 -50 min at 32°C before recordings. Cell capacitance was estimated by the Lindau-Neher technique (17), implementing the "Sine ϩ DC" feature of the Lock-in module (40 mV peakto-peak and a frequency of 500 Hz) in the standard whole cell configuration. Recordings were conducted using an EPC9 patch clamp amplifier and Pulse software. Exocytic events were elicited by a train of eight 500-ms depolarizing pulses (1-Hz stimulation frequency) from Ϫ70 to 0 mV. All recordings were performed at 32°C.
Statistical Analysis-Data points represent mean Ϯ S.E. An unpaired Student's t test was used to compare control values from methyl-␤-cyclodextrin-treated groups. A p Ͻ 0.05 was considered statistically significant.

RESULTS
The presence of lipid rafts and caveolin-1 has been described in the neuroendocrine pheochromocytoma PC12 cell line and pancreatic exocrine acinar cells and shown to be important in exocytosis (18 -20). The thorough identification and characterization of lipid rafts in ␤-cells has yet to be reported. Therefore, we initially examined the expression of caveolin, a protein localized to specialized lipid raft domains (caveolae), in pancreatic islets and insulinoma cells. Western blot analyses revealed abundant expression of both caveolin-1 and -2 in isolated rat whole islets and considerably lower expression of both caveolin proteins in HIT-T15, MIN6, and INS-1 cells (Fig. 1A). No expression of muscle-specific caveolin-3 was detected in either islets or insulinoma cells (data not shown). Pancreatic islets are comprised of insulin-secreting ␤-cells (Ͼ60%), as well as non-␤-cells (i.e. ␣and ␦-cells). Therefore, we performed confocal immunofluorescence microscopy to determine the cellular distribution and localization of caveolin in the islet cells. Immunolabeling of both caveolin-1 (Fig. 1B) and caveolin-2 (Fig.  1D) demonstrated their abundant expression in rat islets. Coimmunolabeling of the islets with insulin to identify the ␤-cells revealed that both caveolin-1 and caveolin-2 were present in ␤-cells. Caveolin-2 was also present in non-insulin-containing cells. Caveolin is important for the formation of caveolae, which are uncoated flask-shaped invaginations on the plasma membrane, and for trafficking of cholesterol to the cell surface membrane (11). Since caveolae were not observed in electron micrographs of islet ␤-cells (21), we therefore examined the cellular localization of caveolin-1 in enzymatically dispersed islet ␤-cells. To our surprise, caveolin-1 only showed intracellular labeling and very little expression on the plasma membrane. Moreover, caveolin-1 appeared to be colocalized with insulin (Fig. 1C).
Ion channels and SNARE proteins have been demonstrated to be enriched in non-caveolar lipid rafts (19,20,22). Proteins localized in lipid rafts cannot be removed by sodium carbonate or high salt, are resistant to Triton X-100 solubilization, and can be separated on discontinuous sucrose gradients (18). As HIT-T15 cells had relatively more abundant caveolin expression as compared with the other insulinoma cell lines (Fig. 1A), this is an appropriate model to examine for the presence of lipid rafts and its content proteins. We probed HIT-T15 cells membranes, which were treated with sodium carbonate (pH 11) and separated on a sucrose gradient, for the presence of ion channels and SNARE proteins in the lipid raft fractions. Western blot analyses detected that K V 2.1 and Ca V 1.2 channels comigrated with caveolin-1 and -2 to the 5-30% sucrose interface, suggesting that these proteins target to lipid rafts. In contrast, neither K V 1.4 nor the K ϩ -ATP channel subunits, Kir6.2 and SUR1, migrated to the lipid raft fractions. We and other investigators have demonstrated both K V 2.1 and Ca V 1.2 channels to be functionally coupled with the SNARE proteins, syntaxin-1A and SNAP-25 (15,(23)(24)(25)(26). We found syntaxin-1A, SNAP-25, and VAMP-2 (syntaptobrevin-2), but not Munc-13-1 or n-Sec1, to localize to the lipid raft rich fractions.
Effects of Cholesterol Depletion on Channel Gating-To determine that K V 2.1 and Ca V 1.2 channels and the SNARE proteins were localized to cholesterol-rich lipid rafts, we depleted membrane cholesterol with methyl-␤-cyclodextrin. Incubation of HIT-T15 cells with 10 mM methyl-␤-cyclodextrin resulted in the redistribution of K V 2.1 and Ca V 1.2 out of the lipid raft fraction (Fig. 2B), as well as the SNARE proteins, syntaxin-1A, SNAP-25, and VAMP-2 (Fig. 2B), indicating that these proteins are indeed localized to lipid raft domains. However, caveolin-1 (Fig. 2B) and caveolin-2 (data not shown) remained in the buoyant fractions after cyclodextrin pretreatment. This is not unexpected given that these proteins appear to be predominantly targeted intracellularly (Fig. 1, B and C), and methyl- ␤-cyclodextrin can only deplete surface membrane cholesterol since it is membrane-impermeable.
Membrane cholesterol depletion by methyl-␤-cyclodextrin has been shown to alter the gating but not current amplitude of heterologous expressed K V 2.1 in LtkϪ cells (22). We examined the effects of cholesterol depletion on K V and Ca V channel gating in HIT-T15 cells. There was a marked reduction in peak K V current amplitude in HIT-T15 cells after pretreatment with methyl-␤-cyclodextrin (Fig. 3, A and B). Peak outward K V currents were reduced from 276 Ϯ 15 pA/pF (control) to 187 Ϯ 31 pA/pF in methyl-␤-cyclodextrin-treated cells (p Ͻ 0.05; n ϭ 4). The remaining outward K V currents displayed a prominent inactivating component, similar to the currents we observed following inhibition of K V 2.1 expression or blockade of these channels in ␤-cells (2,5). We have identified these currents to be K V 1.4 channels. The reduction in K V currents is not the result of a large hyperpolarizing shift in the steady-state current inactivation as observed by Martens et al. (22), which could lead to reduction in measured K V currents. There was a small, yet significant, leftward shift in the steady-state inactivation curve with a V1 ⁄2 of Ϫ27 Ϯ 2 mV in control to Ϫ36 Ϯ 1 mV after treating the cells with methyl-␤-cyclodextrin (p Ͻ 0.05; n ϭ 4 for each group) (Fig. 3C). We observed no changes in the voltage dependence of channel activation. In contrast, there were no significant changes to either Ca V current amplitude (Fig. 4, A and B) or gating (data not shown).
Cholesterol Depletion Enhances Insulin Secretion-Cholesterol depletion in neuroendocrine PC-12 cells reduces the number of SNARE protein clusters on the membrane and resulted in a reduction in exocytosis and the rate of dopamine secretion (19,20). Therefore, we determined whether cholesterol depletion might have similar effects on ␤-cells (Fig. 5A). Surprisingly, in contrast to PC-12 cells, methyl-␤-cyclodextrin enhanced glucose-stimulated insulin secretion from HIT-T15 cells from 164 Ϯ 12 pg of insulin/h to 261 Ϯ 18 pg of insulin/h (n ϭ 3; p Ͻ 0.01). We then examined the effects of methyl-␤-cyclodextrin on insulin granule exocytosis per se by performing patch clamp capacitance measurements (⌬Cm) on single islet ␤-cells (16,17). We elicited exocytosis by eight depolarizing pulses from Ϫ70 mV to 0 mV to measure both the size of the readily releasable pool of insulin granules and the extent of refilling of this pool from the reserve pool (1). A single depolarizing pulse increased cell capacitance by 115 Ϯ 34 fF (n ϭ 6) in control ␤-cells. This value is similar to a previous study suggesting that this represents the readily releasable pool of insulin vesicles docked at the plasma membrane (27) (Fig. 5B). Preincubation of the cells with 10 mM methyl-␤-cyclodextrin increased ⌬Cm by 213 Ϯ 34 fF (n ϭ 6); however, this did not achieve statistical significance from control (p ϭ 0.07) (Fig. 5). A more marked effect was observed following seven additional depolarizing trains whereby control ⌬Cm increased by 266 Ϯ 47 fF from baseline, whereas cholesterol depletion resulted in a rise of 698 Ϯ 74 fF (n ϭ 6; p Ͻ 0.001), indicating a greater capacity to fill up the readily releasable pool and amplify exocytosis. Furthermore, although the ⌬Cm increases leveled off by the 5th-6th depolarizing pulse in control cells, we observed a further incremental increase in ⌬Cm even up to the 8th pulse in the methyl-␤-cyclodextrin-treated cells. Taken together, these results suggest that altering the distribution of channels and SNARE proteins out of the lipid rafts can enhance insulin secretion.

DISCUSSION
Lipid rafts have garnered much interest due to the targeting of numerous membrane proteins to these domains and the association of caveolin in the normal cell biology and pathobiology of a variety of diseases including diabetes, cancer, atherosclerosis, Alzheimer's disease, and muscular dystrophy (11,12). The existence and role of lipid rafts in pancreatic ␤-cells has not been fully elucidated. We have detected the presence of caveolin-1 and caveolin-2 in pancreatic islets and lower expression in insulinoma cells. Caveolin-1 was expressed in primary ␤-cells, whereas caveolin-2 appeared to be expressed in both ␤and non-␤-cells. The lower abundance or absence of caveolin expression in insulinoma cells is not too surprising. Caveolin-1 is abundantly expressed in differentiated cells but is downregulated in transformed cells, and thus, caveolin has been suggested to act as a tumor suppressor (28,29). Neuroblastoma PC-12 cells express low amounts of caveolin-1 (19,30), which can be up-regulated during differentiation (30). Surprisingly, expression of caveolin-1 in primary ␤-cells was concentrated in intracellular compartments in a similar manner to insulin and not at the cell surface. Indeed, caveolin-1 appeared to be colocalized with insulin. Both endocrine and exocrine secretory granular membranes are rich in cholesterol and caveolin-1 (18,31). However, due to the limited resolution of our immunofluorescent images, we cannot rule out the possibility that caveolin-1 is localized to separate and distinct cholesterol transport vesicles in the endoplasmic reticulum/Golgi network (32).
their study resulted in a marked voltage shift in the steadystate inactivation curve without any effect on K V 2.1 current amplitude. This is in stark contrast to the present study, in which we observed a dramatic decrease in K V 2.1 currents fol-lowing treatment with methyl-␤-cyclodextrin. The reasons for these differences on K V 2.1 current magnitude are not currently known but may result from the effects of lipid raft disruption on accessory K V channel subunits or differences in the lipid FIG. 3. Effects of methyl-␤-cyclodextrin (10 mM) pretreatment on ␤-cell K V current amplitude, channel activation, and steady-state inactivation. A, whole cell currents were recorded from HIT-T15 cells. A marked reduction in peak outward K ϩ currents and acceleration of current inactivation are observed following methyl-␤-cyclodextrin pretreatment. B, a significant reduction in peak K V current amplitude is observed following treatment with methyl-␤-cyclodextrin (left panel), but there is no significant shift in the voltage dependence of channel activation (right panel). C, steady-state inactivation was measured using a standard two-pulse protocol. Methyl-␤-cyclodextrin significantly elicited a hyperpolarizing shift in the steady-state inactivation curve from Ϫ27 Ϯ 1.6 mV (Control) to Ϫ36 Ϯ 1 mV (p Ͻ 0.05) with no change in the slope factor (9.8 Ϯ 1.3 mV versus 9.3 Ϯ 0.7 mV) (n ϭ 4). microenvirnoment between insulin-secreting cells and LtkϪ cells. In contrast to the marked effects on K V 2.1 channels, we observed no effect of cholesterol depletion on L-type Ca V currents, although Ca V 1.2 channels were no longer associated with lipid rafts. These results are similar to the lack of changes observed on L-type Ca 2ϩ channel gating and amplitude in cardiac and smooth muscle cells after methyl-␤-cyclodextrin treatment (33) and suggest that these channels are more resistant to changes in the surrounding lipid milieu. Lastly, Kir6.2 and SUR1 are situated more abundantly on insulin granules than on the plasma membrane as shown by immunohistochemical and electron microscopy studies (34). We have found caveolin-1 to be colocalized with insulin, but neither Kir6.2 nor SUR1 migrated to lipid raft fractions on sucrose gradients. Given that lipid rafts are less than 50 nm in diameter (12), which is much smaller than insulin granules (300 nm in diameter) (1), it is conceivable that Kir6.2/SUR1 are located in non-lipid raft domains on these granules.
It is not clear why only K V 2.1 and Ca V 1.2 channels are localized to lipid rafts. However, we and other investigators have shown that both of these channels directly interact with the SNARE proteins, syntaxin-1A and SNAP-25 (15,23,24,25). These channels may cluster at the limited and discrete active zones on the plasma membranes, where insulin granules are docked (35), forming what has been referred to as the excitosome complex (26). It is conceivable that the trafficking of cholesterol, via caveolin on the insulin granules, to the active zones incorporates cholesterol into the surrounding membrane around these channels after granule fusion. Although L-type Ca 2ϩ current densities were not affected by methyl-␤-cyclodextrin, the reduction in K V 2.1 currents could account in part for the increase in insulin secretion. Knock-down of K V 2.1 channel expression or selective pharmacological blockade of these channels in ␤-cells markedly enhanced glucose-induced insulin secretion (2,5).
As in neuroendocrine PC12 cells, we observed the association of the SNARE proteins syntaxin-1A, SNAP-25, and VAMP-2 in lipid rafts, but not Munc-13-1 or n-Sec1. SNARE proteins play an important role in the structural and spatial organization of secretory vesicles for exocytosis. A striking difference between ␤-cells and PC12 cells is that the depletion of membrane cholesterol in ␤-cells resulted in enhanced insulin release and exocytic events, whereas cholesterol depletion in PC12 cells inhibited exocytosis and dopamine release (19,20). It is not known what is the basis for the differences between these results. The dispersion of the SNARE proteins in PC12 cells has been suggested to account for the inhibition of exocytosis. We observed an increase, albeit insignificant (p ϭ 0.07), in the readily releasable pool of

FIG. 4. Effects of methyl-␤-cyclodextrin on L-type Ca V channels in HIT-T15 insulinoma cells.
A, representative whole cell recordings of L-type Ca V currents recorded using 20 mM BaCl 2 in the bath solution and 20 mM tetraethylammonium inside and outside to block K V currents. Currents were elicited from a holding potential of Ϫ80 mV with step depolarizations from Ϫ60 to ϩ80 mV. A 30-min preincubation with 10 mM methyl-␤-cyclodextrin had no significant effect on peak I Ba currents. B, current-voltage relationship of L-type Ca V channels in control (f) and following preincubation with methyl-␤-cyclodextrin (E) (n ϭ 5). Cholesterol depletion had no effect on the gating or amplitude of L-type Ca V channels.
FIG. 5. Cholesterol depletion enhances single-cell exocytic events. Changes in cell capacitance (⌬Cm) were measured from single ␤-cells using a train of eight depolarizing pulses (500 ms in duration) from Ϫ70 mV to 0 mV. A, representative recordings of ⌬Cm from a control ␤-cell and a separate ␤-cell pretreated with 10 mM methyl-␤cyclodextrin. B, summarized changes in ⌬Cm from control and methyl-␤-cyclodextrin pretreated ␤-cells. Points represent the mean Ϯ S.E. from six cells. *, p Ͻ 0.05; †, p Ͻ 0.01; ‡, p Ͻ 0.001 from control. insulin granules and a major increase in the refilling of this pool from the reserve pool, suggesting that membrane cholesterol may profoundly influence the sequential steps of insulin exocytosis, including vesicular mobilization, docking, or priming of insulin granules with the plasma membrane. Alternatively, membrane cholesterol might influence the lipid mixing during the exocytic membrane fusion step and perhaps the stability of the fusion pore opening. Future studies are warranted to determine the cellular mechanism of the augmented insulin secretion following cholesterol depletion.
In summary, our data demonstrate the presence of caveolin in pancreatic ␤-cells and the selective association of K V 2.1, Ca V 1.2, and the SNARE proteins syntaxin-1A, SNAP-25, and VAMP-2 in cholesterol-rich lipid rafts. Unlike what has been reported in most cells, caveolin is not clustered on the cell surface membrane in caveolar domains but is localized on the insulin granules. A similar finding has been shown in pancreatic acinar cells, where caveolin is believed to play a role in lipid transport (18). The cellular role of caveolin in ␤-cells remains to be determined. The localization of K V 2.1 in lipid rafts is important for channel gating given that cholesterol depletion results in a large decrease in K V 2.1 current amplitude. The integrity of lipid rafts also regulates glucose-induced insulin secretion. These results suggest that the localization and targeting of ion channels and SNARE proteins to cholesterol-rich lipid rafts is important for the sequential regulation of cellular excitability and insulin exocytosis in ␤-cells. Noninsulin-dependent diabetes mellitus or type II diabetes is the most common form of the disease; however, its etiology is not fully known. The primary characteristic of this disease is the impaired secretion of insulin in response to elevations in plasma glucose concentrations, in addition to peripheral insulin resistance and lipotoxicity due to increases in plasma cholesterol, free fatty acids, and triglycerides (36). Abnormal cholesterol metabolism may result in changes to the composition of the lipid raft domains that may contribute to disruption in the interaction of the different components of the insulin secretory machinery. Changes in the surrounding lipid environment may also contribute to the alterations in ion channel function observed in prediabetic and diabetic animal models (37)(38)(39), which may, in turn, contribute to the progression and development of this disease.