Sulfonylureas correct trafficking defects of ATP-sensitive potassium channels caused by mutations in the sulfonylurea receptor.

The pancreatic ATP-sensitive potassium (K(ATP)) channel, a complex of four sulfonylurea receptor 1 (SUR1) and four potassium channel Kir6.2 subunits, regulates insulin secretion by linking metabolic changes to beta-cell membrane potential. Sulfonylureas inhibit K(ATP) channel activities by binding to SUR1 and are widely used to treat type II diabetes. We report here that sulfonylureas also function as chemical chaperones to rescue K(ATP) channel trafficking defects caused by two SUR1 mutations, A116P and V187D, identified in patients with congenital hyperinsulinism. Sulfonylureas markedly increased cell surface expression of the A116P and V187D mutants by stabilizing the mutant SUR1 proteins and promoting their maturation. By contrast, diazoxide, a potassium channel opener that also binds SUR1, had no effect on surface expression of either mutant. Importantly, both mutant channels rescued to the cell surface have normal ATP, MgADP, and diazoxide sensitivities, demonstrating that SUR1 harboring either the A116P or the V187D mutation is capable of associating with Kir6.2 to form functional K(ATP) channels. Thus, sulfonylureas may be used to treat congenital hyperinsulinism caused by certain K(ATP) channel trafficking mutations.

The pancreatic ATP-sensitive potassium (K ATP ) channel, a complex of four sulfonylurea receptor 1 (SUR1) and four potassium channel Kir6.2 subunits, regulates insulin secretion by linking metabolic changes to ␤-cell membrane potential. Sulfonylureas inhibit K ATP channel activities by binding to SUR1 and are widely used to treat type II diabetes. We report here that sulfonylureas also function as chemical chaperones to rescue K ATP channel trafficking defects caused by two SUR1 mutations, A116P and V187D, identified in patients with congenital hyperinsulinism. Sulfonylureas markedly increased cell surface expression of the A116P and V187D mutants by stabilizing the mutant SUR1 proteins and promoting their maturation. By contrast, diazoxide, a potassium channel opener that also binds SUR1, had no effect on surface expression of either mutant. Importantly, both mutant channels rescued to the cell surface have normal ATP, MgADP, and diazoxide sensitivities, demonstrating that SUR1 harboring either the A116P or the V187D mutation is capable of associating with Kir6.2 to form functional K ATP channels. Thus, sulfonylureas may be used to treat congenital hyperinsulinism caused by certain K ATP channel trafficking mutations.
ATP-sensitive potassium (K ATP ) 1 channels present in the plasma membrane of pancreatic ␤-cells play a central role in mediating glucose-induced insulin secretion (1)(2)(3)(4). The activity of K ATP channels, which regulates ␤-cell membrane potential, is determined by the relative concentrations of intracellular ATP and ADP. When the blood glucose level rises, the increased intracellular [ATP/ADP] ratio favors K ATP channel closure, resulting in membrane depolarization, Ca 2ϩ influx, and insulin secretion. When the blood glucose level falls, the above molecular events reverse, and insulin release is stopped. In the event where K ATP channels fail to open during glucose starvation, ␤-cell membrane potential remains depolarized, and in-sulin secretion persists, leading to severe hypoglycemia. These symptoms are found in patients suffering from congenital hyperinsulinism (5), also known as persistent hyperinsulinemia hypoglycemia of infancy (PHHI) (6). Indeed, mutations in the K ATP channel genes, sulfonylurea receptor 1 (SUR1) and the inward rectifier potassium channel Kir6.2, that lead to a loss of channel function have been shown to be major causes of PHHI (4,6).
The pancreatic K ATP channel complex consists of four poreforming Kir6.2 subunits and four regulatory SUR1 subunits (7)(8)(9)(10). Gating of K ATP channels occurs as a result of the interplay between both channel subunits and intracellular ATP and ADP. Binding of ATP to the Kir6.2 subunit inhibits channel activity, whereas binding of Mg 2ϩ -complexed ATP or ADP to the SUR1 subunit stimulates channel activity (11)(12)(13)(14). SUR1 is a member of the ATP-binding cassette transporter family; it has three transmembrane domains: TM0, TM1, and TM2, and two large cytoplasmic nucleotide binding domains: NBD1 and NBD2 (15,16). Structure-function studies suggest that the chemical energy derived from nucleotide binding and hydrolysis at the NBDs is relayed through TM0 to the pore subunit Kir6.2 to induce channel opening (17)(18)(19)(20)(21). When intracellular [ATP/ADP] ratio is high, ATP inhibition predominates, and channel activity decreases. Conversely, when intracellular [ATP/ADP] ratio is low, MgADP stimulation predominates, and channel activity increases. Importantly, many PHHI-associated SUR1 mutations specifically reduce or abolish the ability of the channel to be stimulated by MgADP (13,22,23), indicating that the primary mechanism for channel activation during glucose starvation is a rise in ADP. In addition to conferring MgADP stimulation, SUR1 also mediates channel regulation by sulfonylureas and the potassium channel opener diazoxide (2,3,24). Sulfonylureas, such as glibenclamide and tolbutamide, inhibit channel activity by binding to SUR1; they are clinically effective in the treatment of type II diabetes (25,26). Diazoxide, by contrast, stimulates channel activity when bound to SUR1 and has been successfully used to treat some cases of PHHI (4,27).
Aside from functional regulation, SUR1 and Kir6.2 subunits are also both required for cell surface expression of the channel. Assembly of the K ATP channel complex occurs in the endoplasmic reticulum (ER) (7). To ensure that only correctly assembled functional channels are exported to the plasma membrane, the cell has developed a quality control mechanism involving a RKR tripeptide motif present in both SUR1 and Kir6.2 (28). It has been proposed that in preassembled and partially assembled channel proteins, the RKR motif is exposed, providing a signal for ER retention. Upon complete assembly of the channel complex, the RKR motifs become shielded to allow the channel to pass the ER quality control checkpoint and traffic forward to the plasma membrane. Inactivation of the motif by deletion or mutation to AAA leads to unregulated surface expression of individual subunits and partially assembled channel complexes (14,28). Although trafficking signals other than the RKR sequence motif have been reported to regulate the abundance of K ATP channels in the plasma membrane, their mechanisms are not as well understood (21,29,30).
Defective protein trafficking caused by genetic mutations underlies many human diseases. Examples include the cystic fibrosis transmembrane conductance regulator, the HERG (human ether-a-go-go-related gene) potassium channel and the gap junction protein connexin (31)(32)(33). Recent studies show that such a mechanism also explains how some SUR1 mutations lead to loss of channel function and consequently the disease PHHI (30, 34 -36). So far, the SUR1 mutations that have been reported to cause channel trafficking defects are located within or downstream of the second nucleotide-binding domain of the protein. Here, we report that two PHHI-associated SUR1 mutations, A116P and V187D (2,21,29,37), located in the first transmembrane domain (TM0), prevent trafficking of K ATP channels from the ER to the plasma membrane. Unlike the previously reported trafficking mutants, trafficking defects caused by these two mutations can be corrected by sulfonylureas, and the rescued channels are fully functional.

EXPERIMENTAL PROCEDURES
Molecular Biology-FLAG epitope (DYKDDDDK) was inserted at the N terminus of the hamster SUR1 cDNA by sequential overlap extension PCR. Point mutations of SUR1 were introduced into hamster SUR1 cDNA in the pECE plasmid using a QuikChange site-directed mutagenesis kit (Stratagene). The FLAG epitope tag and mutations were confirmed by DNA sequencing. All SUR1-Kir6.2 fusion constructs were also in pECE vector (22). Rat Kir6.2 cDNA is in pCDNA3 vector. Mutant clones from multiple PCRs were analyzed in all experiments to avoid false results caused by undesired mutations introduced by PCR. 86 Rb ϩ Efflux Assay-The Cells were incubated for 24 h in culture medium containing 86 RbCl (1 Ci/ml) 2-3 days after transfection with SUR1 and Kir6.2. Before measurement of Rb efflux, the cells were incubated for 30 min at 25°C in Krebs-Ringer solution with metabolic inhibitors (2.5 g/ml oligomycin plus 1 mM 2-deoxy-D-glucose). At selected time points, the solution was aspirated from the cells and replaced with fresh solution. At the end of a 40-min period, the cells were lysed in 2% SDS-Ringer solutions. The 86 Rb ϩ in the aspirated solution and the cell lysate were counted. The percentage of efflux at each time point was calculated as the cumulative counts in the aspirated solution divided by the total counts from the solutions and the cell lysates.
Immunofluorescence Staining-COSm6 cells were plated in 6-well tissue culture plates, transfected with 0.6 g of SUR1 and 0.4 g of Kir6.2/well using FuGENE 6 (Roche Applied Science) according to the manufacturer's directions. The cells were analyzed 48 -72 h post-transfection. For surface staining, the cells were incubated with anti-FLAG M2 mouse monoclonal antibody (Sigma; diluted to 10 g/ml in Opti-MEM containing 0.1% BSA) for 1 h at 4°C, washed with ice-cold PBS, and then incubated with Cy-3-conjugated donkey anti-mouse secondary antibodies (Jackson) for 30 min at 4°C. After three 5-min washes in ice-cold PBS, the cells were fixed with 4% paraformaldehyde and viewed using an Olympus Fluoview confocal microscope. For total cellular staining of FLAG-tagged SUR1, the cells were fixed with cold (Ϫ20°C) methanol for 5 min. The fixed cells were incubated with the anti-FLAG M2 monoclonal antibody (10 g/ml PBS containing 1% BSA) at room temperature for 1 h, washed in PBS, incubated with Cy3conjugated donkey anti-mouse secondary antibodies for 30 min at room temperature, and washed again in PBS before imaging.
Immunoblotting-Immunoblotting analyses were performed using COS-1 cells instead of COSm6 cells to reduce background (22). Transfection was carried out as described above. The cells were lysed 48 -72 h post-transfection in a buffer (referred to as the lysis buffer) containing 20 mM HEPES, pH 7.0, 5 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, and Complete TR protease inhibitors (Roche Applied Science). The proteins in cell lysates were separated by SDS/PAGE (8%), transferred to nitrocellulose membrane, analyzed by M2 anti-FLAG antibody followed by horseradish peroxidase-conjugated anti-mouse secondary antibodies (Amersham Biosciences), and visualized by chemiluminescence (Super Signal West Femto; Pierce).
Chemiluminescence Assay-COSm6 cells were plated in 35-mm dishes and transfected with K ATP channel subunits using FuGENE 6. Drug treatment was initiated 32-40 h post-transfection and lasted for 4 -24 h. The cells were then processed for chemiluminescence assays as described previously (36,38). Briefly, the cells were fixed with 2% paraformaldehyde for 30 min at 4°C, preblocked in PBS with 0.1% BSA for 30 min, incubated in M2 anti-FLAG antibody (10 g/ml) for an hour, washed four times for 30 min in PBS with 0.1% BSA, incubated in horseradish peroxidase-conjugated anti-mouse (Jackson, 1:1000 dilution) for 20 min, and washed again four times for 30 min in PBS with 0.1% BSA. Chemiluminescence of each dish was quantified in a TD-20/20 luminometer (Turner Designs) following 5 s of incubation in Power Signal ELISA 32 Femto luminol solution (Pierce). All steps after fixation were carried out at room temperature.
Metabolic Labeling and Immunoprecipitation-COS1 cells grown in 35-mm dishes were transfected with fSUR1 and Kir6.2 for 24 h. The cells were starved in methionine/cysteine-free Dulbecco's modified Eagle's medium supplemented with 5% dialyzed fetal bovine serum for 30 min prior to labeling with L-[ 35 S]methionine (ICN Tran 35 S-Label, 150 -250 Ci/ml) for 30 min. Labeled cultures were chased for various times in regular medium supplemented with 10 mM methionine at 37°C. At the end of the chase, the cells were lysed in 500 l of the lysis buffer (see above). For immunoprecipitation, 500 l of cell lysate was incubated with 10 g of anti-FLAG M2 antibodies for 1 h at 4°C and then with protein A-Sepharose 4B (Bio-Rad) for 2 h at 4°C. The precipitate was washed three times in the lysis buffer, and the proteins were eluted using the Laemmli sample buffer. The eluted proteins were separated by 8% SDS-PAGE, and the gel was subjected to fluorography. The dried gels were analyzed using a Bio-Rad PhosphorImager.
Patch Clamp Recordings-COSm6 cells were transfected using Fu-GENE 6 and plated onto coverslips. The cDNA for the green fluorescent protein was co-transfected with SUR1 and Kir6.2 to facilitate identification of positively transfected cells. Patch clamp recordings were made 36 -72 h post-transfection. All of the experiments were performed at room temperature as previously described. Micropipettes were pulled from nonheparinized Kimble glass (Fisher) on a horizontal puller (Sutter Instrument, Co., Novato, CA). Electrode resistance was typically 1-2 M⍀ when filled with K-INT solution (below). Inside-out patches were voltage-clamped with an Axopatch 1-D amplifier (Axon Inc., Foster City, CA). The standard bath (intracellular) and pipette (extracellular) solution (K-INT) had the following composition: 140 mM KCl, 10 mM K-HEPES, 1 mM K-EGTA, pH 7.3. ATP was added as the potassium salt. All currents were measured at a membrane potential of Ϫ50 mV (pipette voltage ϭ ϩ50 mV). The data were analyzed using pCLAMP8 software (Axon Instrument). Off-line analysis was performed using Microsoft Excel programs. The data were presented as the means Ϯ S.E.

Both the A116P and V187D Mutations in SUR1 Prevent
Normal Cell Surface Expression of K ATP Channels-Several recent studies have shown that defective K ATP channel trafficking is an underlying mechanism of congenital hyperinsulinism. The trafficking mutations reported to date are all located in the NBD2 and the distal C-terminal region of the molecule, including ⌬F1388, R1394H, and L1544P (34 -36). To examine whether mutations located in other parts of the molecule also affect channel trafficking, we focused our attention to two mutations, A116P and V187D, that are located in the first transmembrane domain, or TM0, of SUR1 ( Fig. 1), and that have previously been reported to not form functional channels when co-expressed with Kir6.2 (6,21,37). To investigate how the A116P and V187D mutations lead to loss of functional K ATP channels, we first performed Western blot analysis. Both mutations were engineered into a SUR1 construct that has been tagged with a FLAG epitope at its extracellular N terminus (referred to as fSUR1 from here on) to facilitate detection (34). The N-terminal FLAG tag has been tested in many SUR1 constructs in previous studies and has been shown not to affect channel phenotype (34,36,38,39). The SUR1 protein has two N-linked glycosylation sites. As the core-glycosylated immature form (the lower band) traffics through the Golgi, its N-linked oligosaccharides become further modified to yield the complex glycosylated mature form (the upper band). Fig. 2A shows that, although both the immature and mature forms were seen with cells co-expressing WT-fSUR1 and Kir6.2, only the immature form was evident in cells co-expressing Kir6.2 and the A116P-or the V187D-fSUR1 mutants. These results demonstrate that the mutant protein is synthesized and suggest that the mutant channel may fail to traffic to the cell surface. The lack of surface expression is confirmed by immunofluorescent staining experiments. In contrast to the abundant surface staining observed in cells transfected with Kir6.2 and WT-fSUR1, surface staining in cells transfected with Kir6.2 and A116P-fSUR1 or V187D-fSUR1 was barely detectable (Fig.  2B, top panels). However, upon fixation and membrane permeabilization, both mutant proteins were found inside the cell, with fluorescent signals concentrated in the perinuclear region suggesting ER retention. These results led us to conclude that the A116P and V187D mutations cause loss of functional K ATP channels by preventing channels from trafficking to the cell surface.
The Trafficking Defects of the A116P and V187D Mutants Are Intrinsic to SUR1-One potential explanation for the trafficking defects seen with the A116P-and V187D-fSUR1 mutations is that the SUR1 mutants are unable to associate with Kir6.2. Because a prerequisite for either SUR1 or Kir6.2 to exit the ER is proper assembly of the two subunits into the octameric K ATP channel complex that shields the RKR ER retention/retrieval signals, failure of SUR1 to associate with Kir6.2 would result in the absence of SUR1 in the plasma membrane, as we have observed. To address this possibility, we used a heterotandem dimer construct in which the C terminus of the mutant fSUR1 has been fused to the N terminus of Kir6.2 (referred to as A116P-or V187D-fSUR1/Kir6.2 fusion) to achieve obligatory physical association between the two subunits; similar SUR1/ Kir6.2 fusion constructs have been used previously by a number of groups for structure-function and trafficking studies (7,9,10,28,34). Surface expression of the fusion protein was quantified using chemiluminescence assays as described previously (36,38). Although WT fSUR1/Kir6.2 fusion protein was expressed at a level comparable with that observed in cells transfected with WT-fSUR1 and Kir6.2 as individual subunits, fusion proteins carrying the A116P-or V187D-SUR1 mutation had poor surface expression, ϳ10 and 20% that of WT fSUR1/ Kir6.2 fusion, respectively (Fig. 3A). The fact that obligatory association between SUR1 and Kir6.2 does not overcome the trafficking defects caused by either mutation indicates that the defects are unlikely because of the inability of the mutant SUR1 to associate with Kir6.2. Moreover, we found that inac-tivation of the RKR ER retention/retrieval motif in SUR1 by mutation to AAA (referred to as fSUR1 AAA ), which should allow the protein to bypass the ER surveillance system and traffic to the cell surface in the absence of Kir6.2 (28, 34, 36), did not improve the surface expression of the mutant SUR1 when In contrast, only the immature band is observed in cells expressing Kir6.2 and A116P-or V187D-fSUR1. The total steady-state protein level of A116P-and V187D-fSUR1 also appears less than that of WT-fSUR1. Molecular mass markers (kDa) are indicated on the right. B, top panels, surface staining of COSm6 cells transiently transfected with Kir6.2 and either WT-, A116P-, or V187D-fSUR1 using the M2 anti-FLAG mouse monoclonal antibodies followed by Cy-3-conjugated anti-mouse secondary antibody. Staining was performed in living cells at 4°C, and the cells were then fixed with 4% paraformaldehyde and viewed by confocal microscopy (Olympus Fluoview 300). Whereas cells expressing WT-fSUR1 channels have abundant surface staining, those expressing A116P-or V187D-fSUR1 channels have barely detectable staining. Bottom panels, total cellular expression of WT or mutant fSUR1. Cells co-transfected with Kir6.2 and the various fSUR1 constructs were fixed and permeabilized with methanol and stained for the FLAG epitope using the M2 anti-FLAG mouse monoclonal antibodies followed by Cy-3-conjugated anti-mouse secondary antibody. Both A116P-and V187D-fSUR1 were detected inside the cell, with a perinuclear staining pattern. fWT, WT-fSUR1; fA116P, A116P-fSUR1; fV187D, V187D-fSUR1. expressed in the absence of Kir6.2. The surface expression levels of A116P-fSUR1 AAA and V187D-fSUR1 AAA are 10 and 20% that of WT-fSUR1 AAA , respectively (Fig. 3B). These findings are consistent with the idea that the trafficking defects are intrinsic to the SUR1 molecule. In addition, although the sur-face expression levels of the A116P and V187D mutant channels were very low (7 and 19% of WT; see Figs. 3C and 4A), we occasionally found membrane patches that contained enough channels to allow electrophysiological recordings. In these experiments, both mutant channels exhibited normal functional properties, including ATP, MgADP, and diazoxide sensitivities (data not shown, but see Fig. 7, B and C). These observations provide strong evidence that the two SUR1 mutants are capable of associating with the pore forming Kir6.2 subunits, and that the mutations do not interfere with the functional coupling between SUR1 and Kir6.2. Taken together, the results led us to propose that the A116P and V187D mutations cause trafficking defects in SUR1, possibly by promoting protein misfolding, but that the mutant proteins retain the ability to associate with Kir6.2 to form functional channels.
Another potential mechanism for the deficient surface expression of K ATP channels is that the A116P and V187D mutations interfere with proper shielding of the RKR signals in the channel complex. This mechanism has been proposed to underlie the trafficking defect caused by the L1544P-SUR1 mutation, which could be partially overcome by removal of the RKR motif in either SUR1 or Kir6.2 and completely overcome by simultaneous removal of the RKR motif in both channel subunits (36). We found that inactivation of the RKR signal in SUR1 slightly increased surface expression of the A116P mutant channels (from 6 to 18% of normal expression level) but not the V187D mutant, and removal of the RKR signal in Kir6.2 (Kir6.2⌬C25) also had very little effect on the surface expression of either mutant (Fig. 3C). These results suggest that the two mutations under study do not cause defective channel trafficking by preventing the RKR motifs in SUR1 or Kir6.2 from being properly concealed.
Glibenclamide Corrects the Channel Trafficking Defects Caused by the A116P and V187D Mutations in SUR1-The data presented so far suggest that the two mutations likely cause defective channel trafficking by promoting misfolding of SUR1. The trafficking defects of some membrane proteins caused by protein misfolding, including the cystic fibrosis transmembrane conductance regulator, human P-glycoprotein, HERG channels, and ␤-glucosidase, can be partially overcome by treating cells with chemical chaperones (40,41). We tested two types of potential chemical chaperones: the general chemical chaperone glycerol, and pharmacological agents, sulfonylureas and diazoxide, that bind specifically to SUR1. Glycerol and other polyols are known to stabilize protein conformation, increase the rate of in vitro protein refolding, and increase the kinetics of oligomeric assembly. Glycerol has been shown to partially rescue the trafficking defect of the cystic fibrosis transmembrane conductance regulator carrying the ⌬F508 mutation (42) and aquaporin-2 associated with nephrogenic diabetes insipidus (43,44), among others (40). Agonists or antagonists specific to a protein have also been used successfully as pharmacological chaperones to correct protein mislocalization caused by mutations (40). Examples include ␣-galactosidase A, ␤-glucosidase, human P-glycoprotein, HERG potassium channel, and vasopressin receptors (41,(45)(46)(47)(48). We found that treating cells with 5% glycerol for 24 h slightly improved surface expression of both the A116P and V187D mutants as well as the WT channel (Fig. 4A). Treating cells with 5 M glibenclamide, however, dramatically increased surface expression of A116P-fSUR1, from 5 to 55%, and of V187D-fSUR1, from 19% to 70% of normal WT channel expression level (Fig. 4A). A small increase in surface expression of WT-fSUR1 was also consistently observed in cells treated with 5 M glibenclamide (Fig. 4A), as reported previously (36).
The effect of glibenclamide on the A116P and V187D mu- The trafficking defects of the A116P-and V187D-fSUR1 mutants are intrinsic to SUR1. A, obligatory association between SUR1 and Kir6.2 does not overcome trafficking defects caused by A116P or V187D. Fusion fSUR1/Kir6.2 constructs containing either A116P or V187D mutation still exhibit poor surface expression compared with the WT fusion construct. In this and subsequent figures, cell surface expression was quantified using chemiluminescence assays. Each data point represents the mean Ϯ S.E. of 3-10 experiments (dishes in duplicate or triplicate for each experiment). For each experiment, RLUs from mock transfected cells were subtracted from the RLUs of each dish to obtain the net RLUs. Surface expression was then normalized to that of cells expressing Kir6.2 and WT-fSUR1 (black bar) by dividing the net RLUs from each dish to the averaged net RLUs of dishes expressing WT-fSUR1 channels. B, the A116P and V187D mutations cause trafficking defects in SUR1. In WT-fSUR1, mutation of the RKR signal to AAA (fWT AAA ) enables the protein to traffic to cell surface in the absence of Kir6.2. However, introducing A116P or V187D to fSUR1 AAA (fA116P AAA and fV187D AAA ) abolishes the ability of the proteins to traffic to the cell surface. C, trafficking defects caused by A116P and V187D do not involve improper shielding of RKR signals in the channel complex. Inactivation of the RKR signal in SUR1 (SUR1 RKR/AAA ) only slightly improved surface expression of A116P but not V187D, whereas removal of RKR in Kir6.2 (Kir6.2⌬C25) slightly improved surface expression of V187D but not A116P. fWT, WT-fSUR1; fA116P, A116P-fSUR1; fV187D, V187D-fSUR1. tants was also clearly seen in immunostaining and in immunoblotting experiments. Glibenclamide treatment led to the appearance of the complex glycosylated mature form (Fig. 4B) and a marked increase in cell surface staining (Fig. 4C) for both mutants. The response of the A116P-and the V187D-fSUR1 mutants to glibenclamide was specific; another SUR1 ligand, FIG. 4. Glibenclamide rescues surface expression of the A116P and V187D mutant K ATP channels. A, cells co-expressing Kir6.2 and WT-, A116P-, or V187D-fSUR1 were subjected to the different drug treatments indicated for 24 h, and surface expression of fSUR1 was quantified using the chemiluminescence assay as described in Fig. 3. Without any drug treatment, expression levels of the A116P and V187D mutants are 6.4 Ϯ 1.2 and 19.1 Ϯ 4.8% that of WT, respectively. Glycerol at 5% slightly improved the surface expression of WT and both mutants to 106.42 Ϯ 1.2, 11.91 Ϯ 3.0, and 35.05 Ϯ 0.65% of normal WT, respectively. Glibenclamide treatment at 5 M for 24 h dramatically improved surface expression of both A116P-and V187D-fSUR1 (to 55.4 Ϯ 4.8 and 70.4 Ϯ 14.5% of WT, respectively) but only slightly increased WT expression (by 6.6 Ϯ 2.8%). In contrast, diazoxide had no or negative effects on surface expression of all channels tested. B, Western blots showing that in cells expressing Kir6.2 and the A116P or V187D mutant fSUR1, treatment with 5 M glibenclamide for 24 h led to appearance of the mature band, which was not detected in untreated cells ( Fig. 2A). C, immunostaining experiments demonstrating the effect of glibenclamide on surface expression of mutant channels. Top panels, surface staining of COSm6 cells transfected with Kir6.2 and either WT-, A116P-, or V187D-fSUR1, as described for Fig. 2B. The cells received 5 M glibenclamide treatment for 24 h prior to staining. In contrast to the data shown in Fig. 2B, cells expressing A116P-or V187D-fSUR1 mutant channels had strong surface staining that is nearly comparable with cells expressing WT-fSUR1 channels. Bottom panels, same as the top panel except that cells were fixed and permeabilized in methanol to reveal total cellular staining of f-SUR1. The staining in glibenclamide-treated cells appears more associated with the plasma membrane (compare with the bottom panels of Fig. 2B), closer to what was observed in cells expressing WT-fSUR1 channels. fWT, WT-fSUR1; fA116P, A116P-fSUR1; fV187D, V187D-fSUR1; Glib, glibenclamide. diazoxide, was without effect at concentrations between 1 nM and 300 M (only 100 M is shown in Fig. 4A). In fact, diazoxide slightly decreased the surface expression of both A116P and V187D mutants.
To titrate the concentration of glibenclamide needed to correct the trafficking deficiency of the mutants, we measured channel surface expression in cells treated with glibenclamide ranging from 1 nM to 10 M. Fig. 5A shows that the effect of glibenclamide is detected at concentrations as low as 1 nM for both mutants and saturates at ϳ1-5 M. We also examined the time course of the glibenclamide effect. The effect was seen as early as 2 h, climbing rapidly to ϳ60% of the maximal effect at 8 h and slowly increasing to the maximum at 24 h after treatment (Fig. 5B).

Glibenclamide Rescues Surface Expression of A116P Mutant Channels by Slowing the Degradation of the Mutant SUR1
Protein-We presume that glibenclamide acts as a chemical chaperone to facilitate folding of the mutant fSUR1 in the ER, thereby increasing maturation and cell surface expression of the channel complex. To test this, we examined the effect of glibenclamide on metabolically labeled A116P-fSUR1 in cells co-expressing Kir6.2. For these experiments, the cells were pulse-labeled for 30 min and chased in the presence or absence of 5 M glibenclamide. In the absence of glibenclamide, A116P-fSUR1 was detected as a core glycosylated immature band following a 30-min pulse-labeling period and remained as the immature form throughout the chase period of up to 24 h (Fig.  6A). In the presence of glibenclamide, however, the immature band was slowly converted to the mature form between 4 and 6 h of chase, and the mature band remained relatively stable for up to 24 h of chase (Fig. 6A). We quantified the degradation rate of A116P-fSUR1 co-expressed with Kir6.2 in control or in glibenclamide-treated cells (the sum of both immature and mature forms) and compared it with that of WT-fSUR1 (co-expressed with Kir6.2) in control cells (the sum of both immature and mature forms). As shown in Fig. 6B, whereas the overall degradation rate of the A116P-fSUR1 mutant protein in glibenclamide-treated cells is similar to that of WT-fSUR1 in control cells, it is markedly slower than that of A116P-fSUR1 in control cells. To further determine whether glibenclamide stabilizes the A116P-fSUR1 mutant protein by stabilizing the mutant SUR1 itself, we measured the degradation rate of A116P-fSUR1 in the absence of Kir6.2. Cells expressing A116P-fSUR1 alone were pulse-labeled for 30 min and chased for up to 18 h in the presence or absence of 5 M glibenclamide. We found that glibenclamide indeed slowed the degradation of A116P-fSUR1, although the rate is still faster than that of WT-fSUR1 in control cells (Fig. 6C). These results are consistent with the notion that glibenclamide increases cell surface expression of the mutant channel by facilitating folding of the SUR1 proteins and preventing them from being rapidly degraded.
Tolbutamide Also Rescues SUR1 A116P Mutant Channels to the Cell Surface, and the Expressed Channels Are Fully Functional after Tolbutamide Washout-Following our observation that glibenclamide significantly improves cell surface expression of the A116P and V187D mutants, the question arises as to whether the rescued channels are still glibenclamide-bound and whether they are physiologically functional. Glibenclamide inhibits K ATP channel activity when bound to SUR1 by abolishing channel response to MgADP (49). Without being able to respond to MgADP stimulation, the glibenclamide-rescued cell surface channels would not be able to sense the rise in ADP concentration upon glucose deprivation and hence would be physiologically nonfunctional. We found that in glibenclamidetreated cells, although the K ATP current density increased dramatically, the channels failed to respond to MgADP, even after 48 h of washout (data not shown). This suggests that glibenclamide is relatively stable and remains bound to SUR1 in the

FIG. 5. Concentration and time dependence of the effect of glibenclamide on surface expression of A116P-and V187D-fSUR1 mutant K ATP channels. A, cells transfected with
Kir6.2 and WT-, A116P-, or V187D-fSUR1 were treated with different concentrations of glibenclamide for 24 h, and the surface expression of fSUR1 was quantified by the chemiluminescence assay. Improvement on surface expression of the mutant channel was evident even at 1 nM. Glibenclamide also increased surface expression of WT-fSUR1 channels slightly (5-15%) but consistently. Each data point represents the mean Ϯ S.E. of 3-10 experiments. Calculation of relative surface expression is as described in the legend to Fig. 3. B, time course of the effect of glibenclamide. The cells expressing A116P-or V187D-fSUR1 mutant channels were treated with 5 M for different periods of time as indicated, and surface expression of the mutant channel was quantified by chemiluminescence assays. The effect of glibenclamide at each time point was calculated by subtracting the net RLUs obtained from cells not receiving any glibenclamide and expressed as a percentage of the maximal effect observed at 24 h after treatment. Each data point represents the average from 2-3 experiments, and the error bar is the deviation from the average or the S.E. fWT, WT-fSUR1; fA116P, A116P-fSUR1; fV187D, V187D-fSUR1. time course we examined. We therefore turned to another sulfonylurea drug, tolbutamide. Tolbutamide, like glibenclamide, binds to SUR1 and inhibits channel activity; however, it binds with lower affinity and can be washed out more easily. Fig. 7A shows that in chemiluminescence assays, tolbutamide also increases surface expression of both A116P-and V187D-fSUR1 mutant channels, at concentrations of 100 and 300 M that we tested (only 300 M is shown). Consistently, in insideout patch clamp recording experiments, tolbutamide treatment led to parallel increases in the current size of both mutant channels; the average patch current amplitudes in K-INT for A116P-and V187D-fSUR1 channels are 3.24 Ϯ 0.75 nA (n ϭ 12) and 5.40 Ϯ 1.19 nA (n ϭ 13), respectively, compared with 7.05 Ϯ 1.02 nA (n ϭ 17) for control WT-fSUR1 channels. Both mutant channels also exhibited single channel conductance (ϳ80 pS at ϩ50 mV, n ϭ 2) and current-voltage relations similar to those of WT channels (Ϫ100 mV to ϩ100 mV, n ϭ 5-6 in each case; data not shown). Most importantly, the rescued mutant channels showed normal responses to MgADP within 2 h of tolbutamide removal (Fig. 7, B and C). The normal MgADP response indicates that the rescued mutant channels are physiologically competent in responding to metabolic signals. In addition, both mutant channels responded to the potassium channel opener diazoxide, like WT channels (Fig. 7, B  and D). The fact that both mutant channels have normal gating properties again demonstrates that these mutations do not affect the functional coupling between SUR1 and Kir6.2.
The ability of the rescued mutant channels to respond to metabolic changes was further examined by 86 Rb ϩ efflux experiments using the A116P mutant as an example. Without tolbutamide treatment, the cells transfected with A116P exhibited very low K ATP channel activities upon metabolic inhibition (11% efflux in 40 min; compare with 9% in untransfected cells) in contrast to cells transfected with WT channels (81% efflux). Following tolbutamide treatment (300 M for 24 h) and subsequent washout of tolbutamide (for 12 h), cells transfected with Kir6.2 and A116P exhibited a substantial increase in channel activities upon metabolic inhibition (ϳ40% efflux; compare with ϳ80% in cells transfected with Kir6.2 and WT-fSUR1 and 10% in untransfected cells). Thus, tolbutamide can be used as a pharmacological chaperone to recruit A116P mutant channels to the cell surface. The recruited channels are physiologically functional, and their presence in the plasma membrane persists even after 12 h of tolbutamide removal.

DISCUSSION
Pancreatic ␤-cell function relies on proper metabolic regulation of K ATP channel activity, which is in turn dependent on proper plasma membrane localization of the channel. Failure of K ATP channels to express on the cell surface results in the loss of channel function, persistent membrane depolarization, and excessive insulin secretion. There has been an increase in the number of disease-associated mutations in SUR1 reported to cause defective trafficking and mislocalization of K ATP channels (30, 34 -36, 50). In this study, we show that two SUR1 point mutations, A116P and V187D, identified in patients with congenital hyperinsulinism (2, 37) cause defective trafficking and a lack of cell surface expression of K ATP channels. The trafficking defects can be corrected by the sulfonylurea drugs glibenclamide and tolbutamide, which act as pharmacological chaperones to facilitate folding and maturation of the mutant proteins.
Mechanisms of Trafficking Defects Caused by the A116P and V187D Mutations-Multiple steps are involved in the proper expression of K ATP channels on the cell surface. Both channel subunits, SUR1 and Kir6.2, have to be correctly translated, folded, and co-assembled into an octameric channel complex. In addition, channels that traffic to the plasma membrane have to maintain normal metabolic stability. In our pulse-chase labeling experiments, the mutant A116P-fSUR1 never became complex-glycosylated, arguing that the lack of surface expression was a result of ER retention rather than increased degradation of surface channels. Our results also suggest that the ER retention is unlikely to be attributed to the inability of the mutant protein to associate with Kir6.2 or improper shielding of the RKR signals but instead is due to lowered folding efficiency of the mutant SUR1. Such an interpretation is strongly sup- ported by our metabolic pulse-chase labeling experiments, which showed that glibenclamide slowed the degradation rate of the A116P mutant SUR1 and, in the presence of Kir6.2, promoted maturation of the mutant protein. The ability of the mutant SUR1 to form functional channels with Kir6.2 either before or after sulfonylurea rescue further demonstrates that both SUR1 mutants are capable of associating with Kir6.2. This conclusion differs somewhat from that reached by Chan et al. (29), who proposed that the A116P and V187D mutations cause PHHI by preventing association between SUR1 and FIG. 7. Tolbutamide rescues functional A116P and V187D mutant channels to the cell surface. A, cells transfected with Kir6.2 and A116P-or V187D-fSUR1 were treated with 300 M tolbutamide for 24 h, and surface expression of channels was measured by chemiluminescence assays. At 300 M, tolbutamide was nearly as effective as 5 M glibenclamide and restored surface expression of A116P-and V187D-fSUR1 channels from 6.4 Ϯ 1.2 to 49.9 Ϯ 7.6% and from 19.1 Ϯ 4.8 to 48.9 Ϯ 6.0% of normal levels, respectively. Each data point represents the mean Ϯ S.E. of 3-9 experiments. Calculation of relative surface expression is as described in the legend to Fig. 3. B, surface mutant channels rescued by tolbutamide displayed normal response to MgADP and diazoxide after tolbutamide washout. The MgADP or diazoxide response was quantified as described previously (38). First, the response was calculated as the current in a K-INT solution plus 0.1 mM ATP, 0.5 mM ADP or 0.25 mM diazoxide, and 1 mM free Mg 2ϩ , relative to that in plain K-INT solution (see C and D below). The response in mutant channels was then normalized to the response obtained in WT channels, which was derived from the average of 10 -12 patches. The average normalized response to MgADP is 106.58 Ϯ 5.43% (n ϭ 11) and 101.68 Ϯ 5.16% (n ϭ 6) for fA116P and fV187D channels, respectively. The average normalized response to diazoxide is 93.65 Ϯ 12.37% (n ϭ 6) and 108.74 Ϯ 6.63% (n ϭ 6) for fA116P and fV187D channels, respectively. C, representative K ATP current traces recorded from inside-out membrane patches containing WT-fSUR1, A116P-fSUR1, or V187D-fSUR1 channels 2 h after tolbutamide removal. The patches were exposed to ATP and ADP at concentrations indicated above each current trace. Free [Mg 2ϩ ] in all nucleotide-containing solutions was 1 mM. The holding potential was Ϫ50 mV, and the inward currents were shown as upward deflection. D, same as C except that the channels were tested for their response to 250 M diazoxide. fWT, WT-fSUR1; fA116P, A116P-fSUR1; fV187D, V187D-fSUR1.
Kir6.2, based on the observation that an SUR1 TM0 domain containing either mutation failed to co-immunoprecipitate with Kir6.2. It is possible that the discrepancy arises because only the TM0 domain of SUR1, rather than full-length SUR1, was analyzed in their experiments. However, a more likely explanation may be that there are two populations of mutant SUR1 in the ER: a large population that exists in an unstable folding state and a small population that exists in a stable folding state. The population that has reached a stable folding state can associate with Kir6.2 to form functional channels and traffic to the cell surface, whereas the population that is not properly folded cannot and is rapidly degraded. Sulfonylureas increase the folding efficiency of the SUR1 mutants and shift the equilibrium toward the stably folded population that can associate with Kir6.2. The existence of a pool of molecules in an unstable folding state may also hold true for the WT-fSUR1 protein. We have noticed that even in the absence of Kir6.2, the degradation curve of WT-fSUR1 cannot be fit well by a single exponential and appears to have a fast and a slow component (Fig. 6C). We speculate that the quickly degrading component may reflect the population of molecules that never reached the correctly folded state.
The predicted location of residue Ala 116 in SUR1 is in the third transmembrane segment of TM0, and Val 187 is in the fifth transmembrane segment of TM0 (15,51). It is expected that mutation of alanine 116 to a proline and valine 187 to a charged aspartate would be disruptive to the transmembrane ␣-helix structure and affect the normal protein folding process. Several regions of SUR1 have been implicated in glibenclamide binding, including cytoplasmic loops 3 and 8 and possibly TM2 (52,53). However, the first transmembrane domain (TM0) where the A116P and V187D mutations are located has not been implicated in glibenclamide binding. Interestingly, in the cystic fibrosis transmembrane conductance regulator, some mutations in the first transmembrane domain prevent proper folding of the full-length protein not by disrupting folding of the first transmembrane itself but by affecting how this domain interacts with downstream domains (54). It is possible that the two TM0 mutations also destabilize the folding of the downstream domains, which can be stabilized by binding of sulfonylureas. Alternatively, sulfonylureas could directly interact with TM0. Future systematic deletion mutation analysis will help determine the mechanism. Remarkably, despite their adverse effects on protein folding and trafficking, the two TM0 mutations have no apparent effects on the functional coupling between SUR1 and Kir6.2; both mutant channels have normal responses to ATP, MgADP, and diazoxide. It is interesting that the mutant proteins exhibit selectivity in their response to pharmacological agents. Tolbutamide, another sulfonylurea, also rescues the two mutant channels to the cell surface. By contrast, diazoxide, which too binds SUR1 but is structurally quite different from sulfonylureas and results in channel stimulation, does not correct the trafficking defect of either A116Por V187D-fSUR1.
Comparison with Other Trafficking Mutants-A number of missense or point deletion mutations in SUR1 have been reported to reduce or prevent cell surface expression of K ATP channels, including ⌬F1388, R1394H, L1544P, A1457T, V1550D, and L1551V (34 -36, 50). These mutations are all clustered in the NBD2 and C-terminal tail of the protein. Among them, ⌬F1388, R1394H, and L1544P have been characterized in detail. Although like A116P and V187D they all result in a lack of surface channel expression phenotype, the mechanisms leading to this phenotype differ, as revealed by their responses to the different rescuing strategies. For instance, the L1544P mutant channels are retained in the ER likely because the mutation affects the shielding of the RKR motifs in the channel complex; elimination of the RKR signals completely restores surface expression of the mutant (36). By contrast, elimination of the RKR signals only improved surface expression of the other trafficking mutants slightly, probably by allowing a small fraction of the mutant protein to escape the ER quality control mechanism. Another notable difference is the response of the mutants to potential pharmacological chaperones. Although sulfonylureas rescue the surface expression of mutant channels bearing the A116P or the V187D mutation, they do not rescue surface expression of either ⌬F1388 or L1544P mutant channels (36). It is conceivable that protein misfolding occurring late in the molecule, such as might be the case for ⌬F1388, is beyond the influence of chemical chaperones like sulfonylureas. As more trafficking mutations in SUR1 are uncovered, it would be interesting to determine whether the ability of sulfonylureas to act as chemical chaperones is correlated with the location of the mutations. Finally, Partridge et al. (35) reported that the R1394H-SUR1 mutation causes K ATP channels to be trapped in the Golgi apparatus, a defect that can be corrected by diazoxide treatment. However, we did not observe such trafficking defect when expressing either the FLAG-tagged or untagged R1394H-SUR1 with Kir6.2 in COS cells, as assayed by immunoblotting, chemiluminescence, and patch clamp recording experiments. 2 It is not clear whether this discrepancy arises from the different cell lines used for expression. In this regard, it is interesting to note that when they expressed R1394H-SUR1 mutant channels in Xenopus oocytes, the channels exhibited normal trafficking and cell surface expression.
Therapeutic Implications-Our study shows that sulfonylurea drugs, in addition to their therapeutic value in the treatment of type II diabetes, may also be used to correct defective channel trafficking caused by certain PHHI mutations. Although genetic and clinical data on the A116P mutation have not been published, the V187D mutation has been shown to account for the majority of PHHI cases in Finland (37). All patients carrying the mutation, either in one or both alleles, had severe disease and required subtotal pancreatectomy. The results presented in this study show that sulfonylureas have rapid, potent, and long lasting (for at least 12 h after drug removal) effects on rescuing the A116P and V187D mutant channels to the cell surface. Most importantly, both A116P-and V187D-SUR1 mutant channels rescued to the cell surface by tolbutamide are fully functional upon drug removal and respond to MgADP and diazoxide stimulation like WT channels. Thus, mutant channels rescued to the surface will be able to respond to metabolic signals and to diazoxide treatment. It should be noted that sulfonylureas, in the time course of our experiments, also cause a small (10 -15%) but consistent increase in surface expression of WT K ATP channels. Such ligand-induced improvement in protein folding has been documented for the ␦-opioid receptor (55). Curiously, Kawaki et al. (56) reported that long term treatment of the pancreatic ␤-cell line MIN6 cells with glibenclamide (at 10 M for 14 days) reduces surface expression of K ATP channels but increases the total SUR1 protein level, suggesting an effect on the trafficking of K ATP channels. It would be important to determine in the future the mechanisms by which sulfonylureas exert their short term and long term effects on K ATP channel trafficking. Many sulfonylurea derivatives have been developed that have different potencies and specificities in interacting with ␤-cell K ATP channels (26). Studies examining the ability of these various compounds to chaperone SUR1 trafficking mutants will provide insight into how sulfonylureas facilitate protein folding and may help develop new drugs tailored for specific clinical uses in the treatment of type II diabetes and congenital hyperinsulinism.