Constitutive Endocytic Recycling and Protein Kinase C-mediated Lysosomal Degradation Control KATP Channel Surface Density*

Pancreatic ATP-sensitive potassium (KATP) channels control insulin secretion by coupling the excitability of the pancreatic β-cell to glucose metabolism. Little is currently known about how the plasma membrane density of these channels is regulated. We therefore set out to examine in detail the endocytosis and recycling of these channels and how these processes are regulated. To achieve this goal, we expressed KATP channels bearing an extracellular hemagglutinin epitope in human embryonic kidney cells and followed their fate along the endocytic pathway. Our results show that KATP channels undergo multiple rounds of endocytosis and recycling. Further, activation of protein kinase C (PKC) with phorbol 12-myristate 13-acetate significantly decreases KATP channel surface density by reducing channel recycling and diverting the channel to lysosomal degradation. These findings were recapitulated in the model pancreatic β-cell line INS1e, where activation of PKC leads to a decrease in the surface density of native KATP channels. Because sorting of internalized channels between lysosomal and recycling pathways could have opposite effects on the excitability of pancreatic β-cells, we propose that PKC-regulated KATP channel trafficking may play a role in the regulation of insulin secretion.

The ATP-sensitive potassium (K ATP ) channel plays a key role in the regulation of glucose-induced insulin secretion by the pancreatic ␤-cells (1)(2)(3). Central to this role is its unique ability to couple changes in the metabolism of glucose to changes in membrane potential. A rise in blood glucose levels increases the uptake and metabolism of glucose, resulting in an increase in the intracellular [ATP]/[ADP] ratio and inhibition of K ATP channels. This leads to depolarization of the ␤-cell membrane (caused by inhibition of K ϩ efflux), stimulating opening of voltage-activated calcium channels, Ca 2ϩ influx, and insulin release (4). Regulation of K ATP channel function by products of metabolism (e.g. nucleotides) as well as other cellular signals (e.g. protein kinases, lipids) has been extensively studied (1,5,6). By comparison, little is known about how the number of channels at the plasma membrane of the cell is controlled, although there is growing evidence that changes in the membrane density of the channel underlie disease states (7,8).
Structurally, K ATP channels exist as octamers formed from four subunits of the inwardly rectifying potassium channel Kir6.1 or Kir6.2, together with four sulfonylurea receptor (SUR1, SUR2A, or SUR2B) subunits (5, 9 -11). The pancreatic K ATP channel comprises Kir6.2 and SUR1 subunits, which are encoded by the genes KCNJ11 and ABCC8, respectively (5,6,10,11). Mutations in both of these genes are associated with disorders of insulin secretion including congenital hyperinsulinism and neonatal diabetes (4,7). Studies have shown that although some mutations affect the nucleotide regulation of the channel (12)(13)(14), others alter the density of the channels at the cell surface by affecting trafficking (15)(16)(17)(18)(19). Recent large scale genome-wide studies have established a strong link between the KCNJ11 gene and type 2 diabetes (20); however, the underlying mechanisms are unknown.
The genetic and cell biological evidence that changes in cell surface density of K ATP channels can have profound effects on insulin secretion raises the possibility that changes in the surface density could play a role in the regulation of insulin secretion in normal ␤-cells (7,17,18,21). Although there are no data for ␤-cells, studies of cardiac and neuronal cells have demonstrated that activation of PKC 3 down-regulates K ATP channels (22). Given the evidence that PKC enzymes in the ␤-cell are activated by glucose stimulation (23,24) and that activation of PKC augments insulin secretion (23,25,26), it is reasonable to speculate that PKC could down-regulate the number of K ATP channels in ␤-cells, thereby enhancing ␤-cell excitability and insulin secretion. Intriguingly, a recent report (27) has suggested that AMP-activated protein kinase-mediated increases in surface density of K ATP channels at least partially underlie the inhibition of insulin secretion under low glucose conditions. These findings further highlight the potential importance of regulation of K ATP channel trafficking in the normal control of insulin secretion.
The cell surface density of membrane proteins is often determined by a balance between forward trafficking from the endoplasmic reticulum, endocytosis, and recycling (28 -30). Studies * This work was supported by the Medical Research Council, United Kingdom. have shown that assembly and forward trafficking of K ATP channels is controlled by endoplasmic reticulum localization (31) and exit (19) signals; this process, however, appears to be too slow (32) to produce rapid changes in cell surface density. On the other hand, both endocytosis and recycling could occur rapidly to produce prompt changes in the cell surface density of the channel. It is therefore not surprising that endocytic mechanisms are often targeted by cellular signals, such as protein kinases, to regulate the density of membrane proteins at the plasma membrane (29,30,33). Hu et al. (22) reported that activation of PKC stimulates endocytosis of K ATP channels and that little or no channel internalization occurs in the absence of PKC stimulation. However, a later study demonstrated that K ATP channels can undergo constitutive endocytosis in the absence of PKC stimulation using a tyrosine-based endocytic signal located on the Kir6.2 subunit (17). The fate of channels following endocytosis and the contribution of post-endocytic mechanisms to the regulation of cell surface density of K ATP channels are not known.
Here, we set out to examine the fate of endocytosed K ATP channels. Our results show that K ATP channels undergo rapid constitutive endocytosis and subsequent recycling to the plasma membrane. Additionally, we demonstrate that the PKC-induced decrease in the surface density of K ATP channels is brought about via a reduction in their recycling with endocytosis being unaffected. Finally, PKC activation ultimately causes channels to accumulate in a Lamp1-positive late endosomal or lysosomal compartment with a concomitant increase in the rate of channel degradation. We would suggest that the PKCmediated decrease in K ATP channel surface density will likely prove to have important consequences for ␤-cell excitability and insulin secretion.

EXPERIMENTAL PROCEDURES
Cell Culture and cDNA Expression-HEK293 cells were obtained from ATCC (CRL-1573), The HEK293-MSRII cell line was provided by Glaxo-Smith Kline (Stevenage, UK). All of the cell lines were grown in a humidified atmosphere containing 5% CO 2 . HEK cell lines were maintained in Dulbecco's modified Eagle's medium/F-12 medium (Invitrogen) supplemented with 10% fetal calf serum and penicillin and streptomycin (100 g ml Ϫ1 ). For HEK293-MSRII cells, this was supplemented with 500 g ml Ϫ1 G418 (Invitrogen). INS1e cells were a kind gift of Dr. C. B. Wollheim (University Medical Center, Geneva, Switzerland) and were cultured as described previously (17). cDNA constructs encoding mouse Kir6.2 with an extracellular HA epitope and an additional 11-amino acid linker (pcDNA3-HA-Kir6.2) and hamster SUR1 (pcDNA6-SUR1) are as described previously (17). The cell lines were transfected using FuGENE 6 (Roche Applied Science) 48 h prior to experimentation. A HEK293 cell line stably co-expressing HA-Kir6.2 and SUR1 (hereafter referred to as HA-tagged K ATP channels) was generated and propagated in the presence of G418 (500 g ml Ϫ1 ) and blasticidin (10 g ml Ϫ1 ) (Invitrogen) as described previously (18).
Drug Treatments-Phorbol 12-myristate 13-acetate (PMA), 4␣-phorbol 12,13-didecanoate (4␣-PDD), bisindolylmaleimide I, and chelerythrine chloride were purchased from Calbiochem. Stock solutions of PMA and chelerythrine were stored in DMSO at Ϫ20°C and diluted in either Dulbecco's modified Eagle's medium/F-12 medium ϩ 1% ovalbumin (Sigma) for immunocytochemistry and chemiluminescence experiments or extracellular solution for patch clamp experiments. Unless otherwise stated, the cells were pretreated with drugs for 30 min prior to and throughout experimentation. The drug concentrations used throughout were: PMA, 100 nM; and chelerythrine, 10 M. Control cells were treated with vehicle alone (DMSO) at 0.1%.
General Methods of Immunocytochemistry-The cells transfected with HA-Kir6.2 and SUR1 on coverslips were subjected to internalization assay as described previously using the rat monoclonal anti-HA antibodies (clone 3F10; Roche Applied Science) (17). Where required, a third clone encoding a green fluorescent protein-tagged marker protein was included in the transfections. The samples were imaged on a Zeiss 510 META laser scanning confocal microscope under a 63ϫ oil immersion lens (NA 1.40). Green fluorescent protein, fluorescein isothiocyanate, and AlexaFluor 488 (494-nm excitation and 519-nm emission) were excited using an argon laser fitted with 488-nm filters, and Cy3 (550-nm excitation and 570-nm emission) was excited using a helium/ neon laser fitted with 543-nm filters.
Recycling Double Stain-To assess recycling of internalized K ATP channels, the cells were incubated with rat anti-HA antibody (0.2 g ml Ϫ1 ) for 2 h at 37°C to allow antibody labeled channels to undergo internalization. The cells were then washed with PBS, and surface-bound antibody was removed by washing twice with ice-cold acidic buffer (0.5 M NaCl, 0.5% acetic acid, pH 2.4). The cells were then incubated at 37°C to allow recycling of the antibody-labeled internalized channels to occur for the required length of time, before fixing with 2% paraformaldehyde for 10 min. Surface (recycled) channels are then labeled with AlexaFluor 488 -conjugated donkey anti-rat antibody (1:500; Molecular Probes). The cells were then permeabilized with methanol:acetone (50:50) for 5 min and labeled with Cy3-conjugated goat anti-rat antibodies (1:500; Jackson Immunoresearch) to stain the nonrecycled channels. In experiments where the dynamics of recycling and reinternalization were followed (see Fig. 2) surface-bound rat anti-HA antibodies were blocked with unconjugated goat anti-rat secondary antibodies (Sigma; 2 g ml Ϫ1 ) for 1 h at 4°C. Following three washes with ice-cold PBS to remove unbound antibody, the cells were incubated at 37°C in the continuous presence of AlexaFluor 488 -conjugated anti-rat secondary antibody to label the recycled channels. The cells were chilled at the end of the desired time interval, washed, fixed, and permeabilized, before staining with Cy3-conjugated goat anti-rat antibodies to label the nonrecycled channels. We refer to the latter method in the text as the secondary antibody capture assay.
Co-localization Quantification-Co-localization analysis was carried out using Image J software (National Institutes of Health). Pearson correlation coefficients were calculated for red and green channels excluding zero pixels.
Chemiluminescence Assays-For quantification of steadystate surface levels of K ATP channels a modification of a method described previously was employed (22) as described by Mankouri et al. (17). Briefly, the cells were fixed in 2% paraformaldehyde for 10 min, and the cell surface channels were labeled with rat anti-HA antibodies for 1 h as above. Following extensive washes with PBS, the channels were labeled with HRP-conjugated anti-rat secondary antibodies (1:500 Sigma) for 1 h and washed, and the cells were lysed in 0.5% sodium deoxycholate in 50 mM Tris-HCl, pH 7.4. The amount of secondary antibody present in cell lysates was measured as the increase in luminescence 2 min following the addition of HRP reporter substrate PS ATTO (Lumigen), using a POLARstar OPTIMA luminometer.
For quantification of channel endocytosis, cell surface K ATP channels were labeled with rat anti-HA antibody for 1 h at 4°C followed by washing with chilled PBS. The cells were then incubated at 37°C for the desired time period to permit channel internalization followed by fixation in 2% paraformaldehyde for 10 min. K ATP channels remaining at the cell surface were then labeled with HRP-conjugated anti-rat antibodies. The cells were subsequently washed thoroughly and lysed before measuring luminescence. Channel internalization was taken as the loss of luminescence relative to the levels at time 0.
To quantify channel recycling, the cells were incubated for 2 h at 37°C in the presence of rat anti-HA antibodies to label the endocytic pool of K ATP channels. The excess antibodies were removed by washing with chilled PBS, and the cells were subsequently incubated at 37°C in the presence of HRP-conjugated anti-rat antibodies to label the channels returning to the cell surface. At desired time points, recycling was halted by transferring cells to 4°C. The cells were then washed thoroughly and lysed, and the chemiluminescence associated with the HRP bound to recycling channels was measured. Channel recycling was presented as the increase in the luminescent signal relative to time 0.
The data were obtained from a number of independent experiments (stated as n in the figure legends), each carried out in duplicate. In all cases the luminescent signal was normalized to lysate protein content estimated by bicinchoninic acid assay. Background chemiluminescence measured from untransfected controls was subtracted from all of the test values. The data are expressed as the means Ϯ S.E. of means. Statistical significance was determined using Student's t test; p Ͻ 0.05 was considered significant.
Electrophysiology-24 h prior to recording, the cells were detached from culture vessels using 0.05% trypsin/EDTA (Invitrogen) and seeded at low density onto 35-mm culture dishes. Borosilicate glass microelectrodes were pulled to a resistance of 2.5-4 M⍀ when filled with recording solution. Extracellular recording solution contained 138 mM NaCl: 5.6 mM KOH, 2.6 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, titrated to pH 7.4 with HCl. Intracellular recording solution contained 100 mM KCl, 40 mM KOH, 1.7 mM MgCl 2 , 1 mM CaCl 2 , 10 mM EGTA, 10 mM HEPES, titrated to pH 7.2 with HCl. The currents were recorded in 300-ms 10-mV pulse steps between Ϫ100 and ϩ50 mV from a holding potential of Ϫ40 mV. The recordings were made in standard whole cell patch clamp configuration using an EPC10 amplifier and Patchmaster software (HEKA). The currents were filtered at 3 kHz and sampled at 10 kHz and analyzed offline using Fitmaster software (HEKA). The corrections were made for liquid-liquid junction potential, and 80% series resistance compensation was routinely applied. All of the recordings were made at room temperature. Current (I)-voltage (V) relationships were determined first in the absence and then in the presence of tolbutamide (200 M). Conductance was measured from the slopes of the I-V curves in the linear region between Ϫ100 and Ϫ50 mV. K ATP conductance was determined by subtracting tolbutamide-insensitive K ϩ conductance from total conductance and normalized to cell capacitance.
Biochemistry-To assess the stability of the surface pool of K ATP channels, we employed a modification of the cell surface immunoprecipitation method described previously (34). Briefly, HEK 293 cells stably expressing HA-tagged K ATP channels were surface-labeled at 4°C with biotinylated anti-HA Fab fragments (1:500 dilution) (3F10 clone; Roche Applied Science). Excess unbound antibody was removed by multiple washes in chilled PBS. The cells were then treated with either vehicle or PMA as indicated and incubated in Dulbecco's modified Eagle's medium/F-12 medium supplemented with 1% ovalbumin for the desired time at 37°C to allow channel endocytosis and degradation. The cells were subsequently collected and lysed overnight at 4°C in 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 1ϫ protease inhibitors (complete mini-EDTA free; Roche Applied Science). The protein concentration of each sample was assessed by BCA assay and 500 g of protein was used. Channel antibody complexes were then pulled down using neutravidin-agarose beads (Thermo scientific), eluted with 1ϫ SDS buffer, and immunoblotted with rabbit anti-HA antibodies (1:2000 dilution; Immune Systems Ltd.). For quantification of band intensity, densitometry was performed using Image J software (National Institutes of Health).
Statistical Analysis-Statistical significance was assessed at p Ͻ 0.05 using Student's t test or analysis of variance with Tukey post hoc means comparison as appropriate. FEBRUARY 19, 2010 • VOLUME 285 • NUMBER 8

JOURNAL OF BIOLOGICAL CHEMISTRY 5965
Reagents-Unless otherwise stated, all of the reagents were purchased from Sigma.

Endocytosed K ATP Channels Are Recycled to the Plasma
Membrane-We have previously demonstrated that plasma membrane K ATP channels undergo rapid and constitutive endocytosis (17). The high rate of channel endocytosis together with the relatively long half-life of K ATP channels reported earlier (32) led us to believe that channel recycling could play an important role in the regulation of cell surface density. We therefore first set out to examine whether the endocytosed K ATP channels recycle back to the cell surface. HEK293 cells co-expressing SUR1 and Kir6.2 with an extracellular HA tag were used in immunofluorescence experiments to look for any recycling of endocytosed channels. Transfected cells were incubated with rat anti-HA antibodies for 2 h at 37°C to allow the antibodies to bind to the exposed HA epitope of the channels and undergo co-internalization with the channels. Antibodies bound to channels remaining at the cell surface following this incubation were removed by washing with an ice-cold acid strip buffer, pH 2.5. Subsequently the cells were transferred to 37°C for 1 h to permit recycling of internalized channels.
Recycled channels were detected by staining unpermeabilized cells with AlexaFluor 488 -labeled anti-rat secondary antibodies (green). Nonrecycled channels were detected by permeabilizing the cells and staining with Cy3-labeled anti-rat secondary antibodies (red). Fig. 1A shows that the acid strip step is effective in removing surface antibody labeling (compare left panel with the middle panel) and that internalized channels are recycled to the cell surface (right panel). Recycling could be blocked by treatments known to inhibit recycling; thus lowering the temperature to 19°C, where endocytosis can occur, but recycling is known to be inhibited, fully blocked recycling (Fig.  1B). Likewise, pretreatment of the cells with monensin (35) or primaquine (36), agents that block recycling by preventing acidification of endosomes, also prevented recycling (Fig. 1B).
To eliminate the possibility that the anti-HA antibodies undergo internalization and recycling independent of K ATP channels, we have examined the ability of a mutant channel (Y330A mutation in Kir6.2) to recycle. This mutation destroys the tyrosine-based internalization motif in Kir6.2 and prevents endocytosis (17), thereby leaving no channels in the endosomal pool to recycle. As expected, we found robust surface expression (left panel) but no recycling (right panel) of the mutant channel (Fig. 1C). Control experiments demonstrate the absence of antibody uptake in untransfected cells (supplemental Fig. S1). Together, these data demonstrate that internalized K ATP channels are subsequently recycled back to the cell surface.
Channels Undergo Multiple Rounds of Endocytosis and Recycling-The appearance of a large pool of apparently nonrecycled channels (Fig. 1A, right panel) led us to ask whether these were in fact channels that had undergone recycling but subsequently had been reinternalized. To address this question we employed a secondary antibody capture assay, as outlined in Fig. 2A.
Cells expressing the HA-tagged K ATP channels were incubated at 37°C for 2 h in the presence of rat anti-HA antibodies to allow accumulation of channels labeled with the antibody within the cell via endocytosis. Following incubation with unconjugated anti-rat secondary antibody to block anti-HA antibody remaining at the cell surface, the cells were switched to 37°C to allow recycling to occur. Recycling was performed in the presence of Alexafluor 488 -conjugated anti-rat secondary antibody so that any channel that returns to the cell surface is labeled with the fluorescent antibody. The cells were then fixed, permeabilized, and stained with Cy3-conjugated anti-rat antibodies to label channels that failed to recycle. The results (Fig.  2B, top panels) show the appearance of surface green stain within 5 min of initiation of recycling, indicating that the endocytosed channels were able to recycle back rapidly. In addition to the surface stain, we also begin to see green stain inside the cells, indicating that during this period the channels were not only able to recycle back to the cell surface but had undergone reinternalization. Red stain (Fig. 2B, lower panels) was present throughout the time course of the experiment, indicating that a proportion of the channels had failed to recycle during the period tested.
We next adapted this secondary antibody capture assay to measure the recycling of channels by quantitative chemiluminescence (Fig. 2C). The cells were first incubated with the primary anti-HA antibody at 37°C to label surface channels undergoing endocytosis. After washing at 4°C to remove excess antibody, the cells were returned to 37°C to initiate recycling. Recycling was performed in the presence of HRP-conjugated anti-rat secondary antibody to allow continuous labeling of recycled channels with HRP. After washing, chemiluminescence caused by bound HRP was measured from cell lysates. Chemiluminescence at time 0 of recycling (representing the number of channels at cell surface) was normalized to one. The increase in chemiluminescence as a result of recycling was expressed as the fold increase relative to the zero time value. The data (Fig. 2C) demonstrate that the rate of recycling is fairly rapid, with an equivalent to around ϳ80% of cell surface channels undergoing recycling within 10 -15 min (Fig. 2C). Taken together with the previous findings (17), it is evident that both internalization and recycling of K ATP channels occur on a rapid time scale.
PKC Activation Reduces the Surface Levels of K ATP Channels-As outlined in the introduction, PKC appears to have an important role in the regulation of the surface density of K ATP channels. We therefore next examined the effects of PKC activation upon K ATP channel surface density and trafficking using the PKC activator PMA. The results (Fig. 3A) show that activation of protein kinase C by PMA, but not treatment with 4␣-PDD, an inactive analogue of PMA, leads to a decrease in the numbers of K ATP channels at the cell surface, as judged by the loss of staining at the cell periphery and an increase in the accumulation of channels inside the cell. Inhibition of PKC by chelerythrine blocked the appearance of K ATP channels in perinuclear compartments and led to a largely peripheral distribution of channels. We next applied a quantitative chemiluminescence assay to a cell line stably expressing both Kir6.2-HA and SUR1. 1 h of PMA treatment led to a significant decrease (62 Ϯ 3.6%) in the levels of K ATP channels at the cell surface; again, this effect was absent from control cells treated with 4␣-PDD (Fig. 3B). In agreement with the previous report (22), however, treatment of cells with chelerythrine did not result in any significant (as one would predict) increase in the cell surface density of channels. This suggested that PKC inhibition might redistribute the channel to peripheral endosomes that lie close to the cytoplasmic face of the membrane rather FIGURE 2. K ATP channels undergo multiple rounds of endocytosis and recycling. A, schematic of secondary antibody capture assay (see "Experimental Procedures"). HEK293 cells stably expressing Kir6.2-HA and SUR1 were incubated with rat anti-HA antibodies for 2 h at 37°C to allow labeling and subsequent endocytosis of cell surface channels. Following blocking of surface-bound antibody with unlabeled secondary antibody, a subsequent incubation at 37°C was performed in the presence of extracellular Alexafluor 488 -conjugated anti-rat antibodies (green) to label channels returning to the cell surface. The cells were then fixed and permeabilized, and nonrecycled channels were stained with Cy3-conjugated anti-rat antibodies (red). B, representative images of time course of channel recycling assayed as outlined by the schematic in A. The scale bar equals 10 m. C, time course of channel recycling determined by quantitative chemiluminescence assay (see "Experimental Procedures"). The data points represent the means, and the error bars indicate S.E. (n ϭ 3). The increase in recycling was expressed relative to surface channels present at time 0. than recycling into the membrane. Consistent with this suggestion, acid strip buffer failed to remove the anti-HA antibody from the majority of labeled channels (Fig. 3C), indicating their intracellular (subplasma membrane) location. To further confirm the involvement of PKC in the PMA-mediated reduction in channel surface density, the cells were treated with inhibitors of PKC prior to and throughout PMA application. The effects of PMA were blocked by a 30-min treatment of cells with the PKC inhibitors chelerythrine or bisindolylmaleimide I prior to PMA exposure (Fig.  3D).
The large decrease seen in K ATP channel surface density is expected to reduce the whole cell K ATP currents following PKC activation. To investigate this, we subjected HEK293 cells stably expressing SUR1 and Kir6.2-HA to the whole cell patch clamp technique and measured K ATP currents elicited by dialysis of intracellular ATP (Fig. 3, E-G). After the establishment of the whole cell configuration with a pipette solution lacking ATP, K ATP currents increased rapidly, reaching a plateau between 4 and 8 min. Pretreatment with PMA for 1 h prior to establishing whole cell configuration led to a significant decrease in the tolbutamide-sensitive conductance (12.2 Ϯ 2.3 nS pF Ϫ1 in PMAtreated cells versus 22.9 Ϯ 4.1 nS pF Ϫ1 for vehicle-treated cells; p Ͻ 0.05, n ϭ 9) (Fig. 3G). The K ATP specificity of this effect was confirmed by the observation that tolbutamide-insensitive currents remained unchanged following PKC activation (Fig. 3E).
PKC Activation Reduces the Surface Levels of K ATP Channels by Decreasing Their Recycling-A previous study reported that down-regulation of K ATP channels during PKC activation is brought about by an increase in channel endocytosis (22). However, this report did not assess the possible involvement of channel recycling in the regulation K ATP surface density. PKC activation with PMA almost completely prevented recycling of pancreatic K ATP channels compared with vehicle controls (Fig. 4A). To assess the relative effects of PKC activation on endocytosis and recycling, and hence their contribution to the regulation of cell surface density, we have examined the effect of PMA on the rate of endocytosis. Fig. 4B shows that activation of PKC by PMA had no effect upon channel internalization (76.4 Ϯ 7.4% in control cells versus 75.5 Ϯ 7.9% for PMA-treated cells after 10 min). Thus reduced recycling rather than enhanced endocytosis appears to account for the PKC-induced down-regulation of K ATP channel surface density.
Because the above data suggested that PKC activation has little effect on the rate of endocytosis, we sought to obtain further evidence to support this. First, we tested the effect of chelerythrine on endocytosis in the presence of the drug monensin, a known blocker of recycling. The results show that chelerythrine was unable to prevent endocytosis as indicated by the reduced surface density and significant accumulation of K ATP channels in an intracellular location (Fig. 4C). Second, we would predict that if channel internalization was blocked by chelerythrine, then channel recycling should also be attenuated, as was the case for the endocytosis defective Y330A mutant (Fig. 1C). The recycling experiment (Fig. 4D), however, shows that chelerythrine pretreatment failed to prevent intracellular accumulation of K ATP channels as indicated by their ability to recycle and bind to fluorescein isothiocyanate-conjugated secondary antibody in the bathing medium. Together, these data suggest that although intracellular accumulation of channels is inhibited by chelerythrine, there appears to be no block of channel endocytosis.
PKC Activation Diverts Endocytosed K ATP Channels to a Lamp1-positive Compartment and Increases Their Rate of Degradation-To get an insight into the mechanism underlying PKC activation on K ATP channel down-regulation, we carried out a screen for co-localization of endocytosed channels with markers for specific endocytic organelles. The results indicated that endocytosed K ATP channels reside primarily in early and late endosomal compartments (supplemental Fig. S2). Based on these observations, we asked whether PKC activation leads to a redistribution of the internalized channel between the early and late endosomes and lysosomes. Commensurate with the decrease in the surface pool of channels, we saw an increase in channel co-localization with all of the markers examined. The increase in co-localization with the early endosome marker EEA1 was modest (Fig. 5, A and B); however, the increase in co-localization with both the late endosome marker, CI-M6PR (Fig. 5, C and D), and the lysosomal marker Lamp1 (Fig. 5, E and  F) was comparatively greater. These results suggest that PKC activation predominantly targets endocytosed channels to late endosomes or lysosomes.
The increase in late endosomal/lysosomal co-localization suggested that PKC might divert the endocytosed channels to lysosomal degradation. To test this, we first labeled the surface pool of channels with a biotin-conjugated anti-HA antibody at 4°C and then incubated at 37°C in the presence and absence of PMA for different time periods. We then isolated the labeled channels using neutravidin-resin pulldown and measured the levels of the channel protein in the precipitate by Western blotting. The results (Fig. 6A) show that following PKC activation, degradation of the surface pool of channels is markedly increased. Furthermore, pretreatment of cells with lysosomal inhibitors, NH 4 Cl, chloroquine, and leupeptin prevented the PKC-induced degradation of K ATP channels (Fig. 6, B and C). Taken together, we conclude that PKC diverts endocytic channels to lysosomal degradation.

PKC Activation Decreases K ATP Channel Surface Density and Currents in INS1e
Cells-All of the data presented so far were obtained using the HEK293 cell line because of the advantages that recombinant systems offer. We therefore wanted to ascertain the relevance of our findings in a closer approximation of the pancreatic ␤-cell, INS1e. In INS1e cells, HA-tagged K ATP channels are internalized constitutively as reported previously (17). The distribution of endocytosed HA-tagged K ATP channels in INS1e cells is qualitatively similar to that observed in HEK293 cells, with the majority of the internalized channels co-localized with markers for early (EEA1-positive) and late (CI-M6PR-positive) endosomes (Fig. 7A); there was little or no co-localization with the trans-Golgi network marker TGN38. Further, PKC activation with PMA caused a substantial decrease in the surface levels of the channels, with a concomitant increase in the intracellular compartments, an effect that was blocked when PKC was inhibited with chelerythrine (Fig. 7B).
Having demonstrated that the intracellular distribution and the effect of PKC on the endocytosis of HA-tagged K ATP channels in INS1e cells is similar to that observed in the recombinant HEK cell line, we next examined the effect of PKC activation on the current density of native K ATP channels in this model cell line by whole cell patch clamp (Fig. 7, C and D). In line with observations in HEK293 cells, 2 h of PMA treatment signifi-cantly decreased the K ATP channel density (2.7 Ϯ 1.3 nS pF Ϫ1 for the PMA-treated group versus 7.6 Ϯ 1.1 nS pF Ϫ1 for control cells; p Ͻ 0.05, n ϭ 3). The effect is specific because PMA failed to reduce the current density of outwardly rectifying whole cell currents recorded from INS1e cells; PMA rather caused a small but significant increase in the current density when K ATP was inhibited with high (5 mM) intracellular ATP (supplemental Fig. S3). These data suggest that the observed decrease in K ATP channel density may represent an important and specific ␤-cell response to elevated PKC activity.

DISCUSSION
Membrane proteins are removed from the cell surface by endocytosis and are then either recycled back to the membrane or subjected to degradation (28,29,33). These sorting mechanisms provide the cell with the ability to adjust the density of a given protein at the plasma membrane. Despite the increasing evidence that changes in the K ATP channel density could result in diseases of inappropriate insulin secretion (7,8), mechanisms that contribute to the regulation of K ATP channel density at the plasma membrane are poorly understood. Here, we demonstrate that (i) plasma membrane K ATP channels undergo endocytosis constitutively; (ii) once internalized, a proportion of the channels are recycled back to the plasma membrane, and the remainder is targeted to lysosomes for degradation; and (iii) PKC shifts the balance in favor of lysosomal degradation. Because stimulants of insulin secretion, such as glucose and acetylcholine, activate PKC, we suggest that regulated endocytic trafficking may play a role in the regulation of insulin secretion. In further support of this hypothesis, a recent study has shown AMP-activated protein kinase to dynamically regulate the surface density of K ATP channels in response to the metabolic status of the ␤-cell (27). The authors of this study suggest that an AMP-activated protein kinase-mediated increase in K ATP channel surface density is a key factor in the inhibition of insulin secretion under low glucose conditions. FIGURE 5. PMA treatment increases the co-localization of endocytosed K ATP channels with the lysosomal marker Lamp1. A-F, effect of PKC activation on the intracellular distribution of endocytosed K ATP channels. HEK293 cells expressing HA-Kir6.2 and SUR1 (A-D) plus Lamp1-green fluorescent protein (E and F) were allowed to internalize anti-HA antibodies at 37°C for 1 h in the presence of either vehicle (DMSO, 0.1%) or PMA (100 nM). Following fixation, the cells were co-stained for HA-Kir6.2 (red) and the indicated marker proteins (green). EEA1 was for early endosomes, and the cation-independent mannose-6-phosphate receptor (M6PR) was for late endosomes. A, C, and E are representative of at least three independent experiments; the larger images in each panel represent a merger of the small images stained for HA-Kir6.2 (red) and marker proteins (green) at the bottom. The scale bars equal 10 m. In B, D, and F, co-localization was quantified by particle analysis using the Image J software to generate Pearson correlation coefficients from at least 10 cells and from each of three independent experiments. The data shown are means Ϯ S.E. * denotes significance at the p Ͻ 0.05 level assessed by Student's t test.
Recycling of K ATP Channels-Previous studies have shown that K ATP channels undergo rapid internalization with ϳ80% of the channels being removed from the plasma membrane within 10 min (17). Present results demonstrate that internalized channels can recycle rapidly back to the plasma membrane ( Figs. 1 and 2). Given the report that biosynthetic delivery is a slow process (32), rapid recycling may represent the major mechanism by which K ATP channels removed from the membrane by constitutive endocytosis are replenished. Furthermore, it appears that K ATP channels are capable of undergoing multiple rounds of internalization and recycling (Fig. 2) while slowly being removed to degradation (Fig. 6). These results suggest that any signal that affects either internalization or recycling could bring about rapid changes in the cell surface density of the channel. Although little is known about the ability of potassium channels to undergo multiple rounds of internalization and recycling, a number of transporters, receptors and ligand-gated ion channels utilize multiple cycles of internalization and recycling as a means to control the flow of material and information across the cell membrane. For example, glucose transporter 4, which mediates the uptake of glucose by muscle and adipose tissue, is recruited into the cell membrane in response to insulin stimulation and is internalized into storage vesicles when the stimulus is removed (37). Desensitization and resensitization of cells to signaling molecules such as adrenaline, angiotensin, and ␣-amino-3-hydoxy-5-methy-4-isoxazole-propionic acid is achieved by endocytosis and recycling of their respective receptors (29,30). Although the significance of rapid rates of internalization and recycling of K ATP channels is unclear at present, such rapid dynamics are capable of promptly adjusting the cell surface density of the channel in response to physiological cues.
PKC Activation Inhibits Recycling-Previous studies have reported that the cell surface density of K ATP channels expressed in neuronal, cardiac, and COS-7 cells is down-regulated when PKC activity was stimulated with PMA (22). Conversely, when PKC was inhibited with chelerythrine, very little channel was found within the cell; the majority of the channel density was found at the periphery of the cell (22). We have obtained similar results with K ATP channels expressed in HEK 293 (Fig. 3) and the model ␤-cell line, INS1e (Fig. 7). Based on their data, Hu et al. (22) concluded that down-regulation of K ATP channels is brought about by the PKC-induced increase in endocytosis. In light of our current finding that internalized K ATP channels can recycle back to the membrane, we asked whether PKC could inhibit channel recycling and thereby also contribute to down-regulation. Our results revealed that PKC activation almost completely prevents recycling; by contrast, there was no significant effect on the rate of endocytosis (Fig. 4, A and B). Because this result contrasts the previous conclusion by Hu et al. (22), we sought further evidence. We found that inhibition of PKC with chelerythrine had no effect on the internalization of the channel (Fig.  3C). Furthermore channels internalized in the presence of chelerythrine were able to recycle back when the drug was withdrawn (Fig. 4D). Taken together, our results argue that it is the reduced recycling, rather than increased endocytosis, that underlies the PKC stimulated down-regulation of K ATP channels. In this respect, down-regulation of Kir6.2/SUR1 is mechanistically different from other Kir channels, including Kir3.4 (38) and Kir6.1/SUR2B (39) where PKC activation appears to stimulate endocytosis rather than affecting recycling. There are also fundamental differences in terms of the mechanisms that different Kir channels use to undergo endocytosis. Kir6.2/SUR1 is internalized via clathrin-dependent endocytosis (17), whereas Kir3.4 and Kir6.1/SUR2B are internalized via non-clathrin-dependent pathways into macropinosomes and caveosomes, respectively. Thus the current study describes a previously unrecognized mechanism for PKC regulation of Kir channel density.
PKC Activation Diverts K ATP Channels to Lysosomal Degradation-PKC activation caused a significant diversion of the internalized channels to Lamp1-positive late endosomal/ lysosomal compartments (Fig. 5). Furthermore, PKC activation with PMA resulted in a marked increase in the rate of lysosomal degradation of internalized channels (Fig. 6). Taken together with the finding that internalization (17) and recycling of K ATP channels (Figs. 2-4) occurs rapidly, these results suggest that K ATP channels undergo many rounds of internalization and recycling before being subjected to degradation.
Although PKC appears to target the endosomal sorting machinery to regulate the cell surface density of K ATP channels, the precise mechanisms are yet to be understood. At least two alternative possibilities could be suggested. One is that PKC activation diverts internalized channels to lysosomes, thereby reducing the number of channels available for recycling. The other is that PKC inhibits recycling directly, resulting in an increase in the pool of channels available for diverted trafficking to lysosomes. Inhibition of recycling by PKC activation has been reported for other membrane proteins; examples include ␥-aminobutyric acid, type A receptors (40) and dopamine transporters (41). PKC can also influence lysosomal targeting of membrane proteins. For example, PKC enhances ubiquitination and, thereby, the lysosomal targeting of the dopamine transporter (41). Ubiquitination of K ATP channels and its role in the proteosomal degradation of the K ATP channel have been demonstrated (42), but there are no published data on lysosomal targeting.
Potential Physiological Implications-The potential role of PKCmediated down-regulation of K ATP channels in neuronal, cardiac, and vascular functions has been discussed by Hu et al. (22,43). The significance of down-regulation of K ATP channels in ␤-cell physiology, however, remains to be explored. Numerous studies have shown that ␤-cells express multiple PKC isoforms (23,24). PKC knock-out studies revealed that PKC plays a key role in the regulation of insulin secretion (44,45). However, because PKC appears to affect a variety of targets involved in regulation of insulin secretion (24), it makes it difficult to assign a definite role for the PKC-mediated downregulation of K ATP channels. But the findings that several isoforms of PKC are activated in response to key stimulants of insulin secretion, including glucose and acetylcholine (24) and the evidence that K ATP channels play a central role in glucose-stimulated insulin secretion (4) suggest a possible link between PKC-induced down-regulation of K ATP channels and the regulation of insulin secretion.
In summary, our results demonstrate that endocytosed pancreatic K ATP channels are rapidly recycled back to the plasma membrane and that they undergo multiple rounds of rapid internalization and recycling. In addition, a fraction of the internalized channels are slowly diverted to degradation in the lysosomes. Thus the balance between recycling and degradation appears to be a key determinant of the number of K ATP channels at the plasma membrane (Fig. 8). When PKC is activated, this balance is shifted toward degradation, leading to significant down-regulation of cell surface channels. Because several PKC isoforms are activated during glucose elevation, we propose that this will reduce the plasma membrane density of K ATP channels during periods of elevated blood glucose