A Novel KCNJ11 Mutation Associated with Congenital Hyperinsulinism Reduces the Intrinsic Open Probability of β-Cell ATP-sensitive Potassium Channels*

The β-cell ATP-sensitive potassium (KATP) channel controls insulin secretion by linking glucose metabolism to membrane excitability. Loss of KATP channel function due to mutations in ABCC8 or KCNJ11, genes that encode the sulfonylurea receptor 1 or the inward rectifier Kir6.2 subunit of the channel, is a major cause of congenital hyperinsulinism. Here, we report identification of a novel KCNJ11 mutation associated with the disease that renders a missense mutation, F55L, in the Kir6.2 protein. Mutant channels reconstituted in COS cells exhibited a wild-type-like surface expression level and normal sensitivity to ATP, MgADP, and diazoxide. However, the intrinsic open probability of the mutant channel was greatly reduced, by ∼10-fold. This low open probability defect could be reversed by application of phosphatidylinositol 4,5-bisphosphates or oleoyl-CoA to the cytoplasmic face of the channel, indicating that reduced channel response to membrane phospholipids and/or long chain acyl-CoAs underlies the low intrinsic open probability in the mutant. Our findings reveal a novel molecular mechanism for loss of KATP channel function and congenital hyperinsulinism and support the importance of phospholipids and/or long chain acyl-CoAs in setting the physiological activity of β-cell KATP channels. The F55L mutation is located in the slide helix of Kir6.2. Several permanent neonatal diabetes-associated mutations found in the same structure have the opposite effect of increasing intrinsic channel open probability. Our results also highlight the critical role of the Kir6.2 slide helix in determining the intrinsic open probability of KATP channels.

Pancreatic ATP-sensitive potassium (K ATP ) channels play a key role in glucose-stimulated insulin secretion by coupling glucose metabolism to ␤-cell excitability (1)(2)(3). Each K ATP channel consists of four poreforming Kir6.2 subunits encoded by KCNJ11 and four regulatory sulfonylurea receptor 1 (SUR1) receptors encoded by ABCC8 (4 -6). The activity of K ATP channels is subject to regulation by intracellular nucleotides ATP and ADP (1,7,8). ATP inhibits channel activity by binding to the Kir6.2 subunit, whereas ADP in complex with Mg 2ϩ stimulates channel activity by interacting with SUR1. Changes in ATP and ADP concentrations during glucose metabolism are, thus, linked to K ATP channel activity, which in turn controls ␤-cell membrane potential and insulin secretion. As such, genetic mutations in Kir6.2 or SUR1 that alter channel sensitivity to ATP or MgADP impair the ability of K ATP channels to convert metabolic signals to electrical signal, resulting in insulin secretion diseases. In addition to intracellular nucleotides, both long chain acyl-CoAs and membrane phosphoinositides have been shown to stimulate channel activity and reduce channel sensitivity to ATP inhibition by interacting with Kir6.2 (9 -12, 14 -17). However, compared with intracellular nucleotides, the role of membrane phosphoinositides or long chain acyl-CoAs in controlling K ATP channel activity in physiological and pathological conditions is less clear and still under investigation (8).
Congenital hyperinsulinism is a disease characterized by persistent insulin secretion despite life-threatening hypoglycemia (3, 18 -21). Dominant or recessive mutations in ABCC8 or KCNJ11 that reduce or abolish channel activity are the major genetic cause of the disease (3, 18 -21). Most mutations identified to date reside in the SUR1 subunit (1,3,22,23). Functional studies of a considerable number of missense SUR1 mutations have established loss of channel response to MgADP and impaired expression of functional channels at the cell surface as two major defects (24 -27). Relatively few mutations have been reported in the Kir6.2 subunit (1, 28); these mutations either introduce a premature stop codon that results in truncated nonfunctional Kir6.2 protein or introduce missense mutations that cause rapid degradation of the protein (28,29). Here, we report the identification and functional characterization of a dominant congenital hyperinsulinism-associated Kir6.2 missense mutation F55L. We show that this mutation greatly reduces the open probability of K ATP channels in intact cells without affecting channel expression. The low channel activity is likely due to reduced channel response to membrane phosphoinositides and/or long chain acyl-CoAs, as application of exogenous PIP 2 3 or oleoyl-CoA restores channel activity to that seen in wild-type (WT) channels. Our finding identifies impaired K ATP channel response to phospholipids and/or long chain acyl-CoAs as a novel mechanism underlying congenital hyperinsulinism.

MATERIALS AND METHODS
Clinical and Genetic Analyses-Tests of acute insulin responses to calcium (2 mg/kg), leucine (15 mg/kg), glucose (0.5 gm/kg), and tolbutamide (25 mg/kg) sequentially infused intravenously at intervals of 1 h were carried out as previously described (27). Peripheral blood was obtained for isolation of genomic DNA from the proband, a male infant, and his father. Direct sequencing of DNA from the patient and family members was done as previously described (27).
Molecular Biology-Rat Kir6.2 cDNA is in pCDNAI/Amp vector (a generous gift from Dr. Carol A. Vandenberg) and SUR1 in pECE. Sitedirected mutagenesis was carried out using the QuikChange site-directed mutagenesis kit (Stratagene), and the mutation was confirmed by sequencing. Mutant clones from two independent PCR reactions were analyzed to avoid false results caused by undesired mutations introduced by PCR.
86 Rb ϩ Efflux Assay-COSm6 cells were plated onto 35-mm culture dishes and transfected with wild-type SUR1 and control or mutant rat Kir6.2 cDNA using FuGENE. Cells were incubated for 24 h in culture medium containing 86 RbCl (1 Ci/ml) 2 days after transfection. Before measurement of 86 Rb ϩ efflux, cells were incubated for 30 min at room temperature in Krebs-Ringer solution with metabolic inhibitors (2.5 g/ml oligomycin and 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, cells were lysed. The 86 Rb ϩ in the aspirated solution and the cell lysate was counted. The percentage 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 (25).
Western Blotting and Chemiluminescence Assay-Cell surface expression level of the mutant channel was assessed by Western blot and by a quantitative chemiluminescence assay using a SUR1 that was tagged with a FLAG epitope (DYKDDDDK) at the N terminus (fSUR1), as described previously (30). COSm6 cells grown in 35-mm dishes were transfected with 0.4 g of rat Kir6.2 and 0.6 g of fSUR1 and lysed 48 -72 h post-transfection in 20 mM HEPES, pH 7.0, 5 mM EDTA, 150 mM NaCl, 1% Nonidet P-40 (IGAPEL) with Complete TR protease inhibitors (Roche Applied Science). Proteins in the cell lysate were separated by SDS-PAGE (7.5% for SUR1 and 12% for Kir6.2), transferred to nitrocellulose membranes, analyzed by incubation with appropriate primary antibodies followed by horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences), and visualized by enhanced chemiluminescence (Super Signal West Femto; Pierce). The primary antibodies used are: M2 mouse monoclonal anti-FLAG antibody for fSUR1 (Sigma) and rabbit polyclonal anti-Kir6.2 for Kir6.2 (from Santa Cruz Biotechnology, Santa Cruz, CA). For chemiluminescence assay, cells were fixed with 2% paraformaldehyde for 30 min at 4°C. Fixed cells were preblocked in phosphate-buffered saline (PBS) plus 0.1% bovine serum albumin (BSA) for 30 min, incubated in M2 anti-FLAG antibody (10 g/ml) for 1 h, washed 4 ϫ 30 min in PBS plus 0.1% BSA, incubated in horseradish peroxidase-conjugated anti-mouse (Jackson Immuno Research, Inc., 1:1000 dilution) for 20 min, and washed again 4 ϫ 30 min in PBS plus 0.1% BSA, all at room temperature. Chemiluminescence of each dish was quantified in a TD-20/20 luminometer (Turner Designs) after 15 s of incubation in Power Signal enzyme-linked immunosorbent assay luminol solution (Pierce). All steps after fixation were carried out at room temperature.
Electrophysiology-Patch clamp recordings were performed in the inside-out configuration as previously described (30). Briefly, COSm6 cells were transfected with cDNA encoding WT or mutant channel proteins as well as cDNA for the green fluorescent protein to help identify positively transfected cells. Patch clamp recordings were made 36 -72 h post-transfection. Micropipettes were pulled from non-heparinized Kimble glass (Fisher) with resistance typically ϳ1-2 megaohms. The 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 and ADP were added as the potassium salt. For measuring ATP sensitivity, 1 mM EDTA was included in K-INT to prevent channel rundown (31). All currents were measured at a membrane potential of Ϫ50 mV (pipette voltage ϭ ϩ50 mV), and inward currents are shown as upward deflections. Data were analyzed using pCLAMP software (Axon Instrument). Off-line analysis was performed using Microsoft Excel programs. The MgADP or diazoxide response was calculated as the current in a K-INT solution containing either 0.1 mM ATP and 0.5 mM ADP or 0.1 mM ATP and 0.3 mM diazoxide (both with1 mM free Mg 2ϩ ) relative to that in plain K-INT solution.
Data Analysis-ATP dose response curve fitting was performed with Origin 6.1. Channel intrinsic P o was estimated from stationary fluctuation analysis of short recordings (ϳ1 s) of macroscopic currents in K-INT/EDTA solution or K-INT/EDTA plus 5 mM ATP (32)(33)(34). Currents were filtered at 1 kHz. Mean current (I) and variance ( 2 ) in the absence of ATP were obtained by subtraction of the mean current and variance in 5 mM ATP. Single channel current (i) was assumed to be Ϫ3.6 pA at Ϫ50 mV (corresponding to single channel conductance of 72 picosiemens). P o was then calculated using the following equation: P o ϭ 1 Ϫ ( 2 /(i ϫ I)). Statistical analysis was performed using independent two-population two-tailed Student's t test.

Patient History and Identification of the F55L Missense Mutation in
Kir6.2-The proband was a male infant diagnosed with persistent hypoglycemia and hyperinsulinism at 12 h of age (14 microunits/ml with a blood glucose level of 31 mg/dl; 13 microunits/ml with a blood glucose of 24 mg/dl; normal, Ͻ3 microunits/ml). Tests of acute insulin responses to secretogogues were performed in the proband and his father, and results compared with hyperinsulinism patients with recessive ABCC8 null mutations (3992-9 G3 A and ⌬F1388) and glutamate dehydrogenase mutations and normal adult controls (35). The proband showed a pattern of responses similar to those seen in patients with recessive ABCC8 mutations, an abnormal positive insulin response to calcium stimulation, as well as possibly diminished responses to glucose and tolbutamide ( Table 1). The father of the proband had mildly abnormal acute insulin responses, including abnormally increased insulin responses to calcium and leucine and a lower response to tolbutamide than to glucose stimulation. This pattern of insulin responses resembles that seen in children with hyperinsulinism due to recessive ABCC8 and KCNJ11 mutations that retain partial channel activity (35). Subsequent

Calcium Leucine Glucose Tolbutamide
Proband ( FEBRUARY 3, 2006 • VOLUME 281 • NUMBER 5 genetic testing revealed that the proband and the father shared a 165C3 A mutation in the KCNJ11, resulting in a missense mutation, F55L, in the Kir6.2 subunit of the K ATP channel. Of note, although the father showed mildly abnormal acute insulin responses as well as evidence of protein-sensitive hypoglycemia (blood glucose nadir of 63 mg/dl at 60 min after ingestion of oral protein; normal, Ͼ70 mg/dl), he demonstrated normal fasting blood glucose (84 mg/dl after a 12-h fast). The more severe hypoglycemia seen in the proband, at least in the newborn period, is likely attributed in part to perinatal stress, as has been reported previously (36,37). Treatment with diazoxide at 15 mg/kg/day rapidly improved the hypoglycemia in the proband. Within 4 days, he was able to fast 12 h without hypoglycemia. Periodic reassessments up to 4 years of age showed that, whereas on treatment with a low dose of diazoxide at 4 mg/kg/day, the proband was able to fast successfully for 12-15 h without hypoglycemia and did not develop hypoglycemia in response to an oral protein tolerance test (1 g/kg).

KCNJ11 Mutation, K ATP Channel Gating, and Hyperinsulinism
The F55L Mutation Reduces Open Probability of K ATP Channels-To investigate how the F55L mutation in Kir6.2 affects channel function, we reconstituted the mutant channels in COS cells by co-expression of F55L-Kir6.2 and SUR1. To first confirm reduction or loss of channel activity in intact cells during glucose deprivation, as one would predict based on the disease phenotype, we performed 86 Rb ϩ efflux assays to assess channel activity in response to metabolic poisoning. As shown in Fig. 1A, homomeric mutant channels are much less active upon metabolic inhibition compared with WT channels. Because the F55L mutation is heterozygous in the patient, we simulated the heterozygous state by co-expressing WT and mutant Kir6.2 at a 1:1 cDNA molar ratio with SUR1 (referred to as hetF55L). The hetF55L channels, although are much more active than homomeric F55L mutant channels (referred to as homF55L), are still less active than WT channels. These results demonstrate the causal role of the Kir6.2 F55L mutation in congenital hyperinsulinism.
The reduced activity of the mutant channel upon metabolic inhibition could be due to reduced surface expression or altered gating properties (3). To distinguish between these possibilities, we compared the level of surface expression of mutant channels to that of WT channels. Western blots showed that the steady-state mutant Kir6.2 protein level is similar to WT Kir6.2 (Fig. 1B, left). Moreover, fSUR1 co-expressed with the F55L mutant Kir6.2 exhibits core-glycosylated as well as complex-glycosylated forms comparable with that observed in fSUR1 coexpressed with WT Kir6.2 (Fig. 1B, left), indicating the mutation does not affect surface expression of the channel. Quantification of surface expression by chemiluminescence assays further corroborated the conclusion that the F55L mutation does not affect channel expression (Fig. 1B, right).
The above results suggest that the reduced activity seen in the mutant is likely due to altered gating regulation. We, therefore, carried out detailed electrophysiological analysis using the inside-out patch clamp recording technique. Upon patch excision into ATP-free solution, WT channels gave averaged currents of ϳ2 nA at Ϫ50 mV with symmetrical potassium concentrations on both sides of the membrane. By contrast, the averaged current amplitude from the F55L mutant channels was only ϳ0.2 nA. To test whether the reduction in current amplitude results from reduced single channel conductance or channel open probability, we performed single channel recording of the mutant. The F55L mutant exhibited single channel conductance (70.4 Ϯ 1.6 picosiemens) comparable with that of WT channel (68.0 Ϯ 3.1 picosiemens). However, the channel open probability appears greatly reduced (Fig. 1C). These results show that the mutation reduces the ATP-independent open probability (referred to as the intrinsic open probability in this study) of K ATP channels.
Effects of PIP 2 and Oleoyl-CoA on the F55L Mutant Channel-Membrane phosphoinositides, in particular PIP 2 , are known to affect the open probability of K ATP channels (15,17,38). Recently, we reported that expression of Kir6.2 constructs with deliberate mutations that abolish channel sensitivity to PIP 2 (R176A, R177A, R206A) in the insulinsecreting cell INS-1 results in phenotypes mimicking those seen in congenital hyperinsulinism. These include reduced K ATP channel activity, depolarized membrane potential, and excessive insulin secretion at basal glucose concentration. We, therefore, hypothesize that the reduced channel open probability observed in the F55L mutant could be due to attenuated channel response to PIP 2 . Supporting this hypothesis, we found that in inside-out patch clamp recordings application of exogenous PIP 2 (5 M) to the cytoplasmic face of the membrane markedly increased the current amplitude of the F55L mutant channels (Fig. 2A). The -fold increase in current amplitude upon PIP 2 stimulation for WT, hetF55L, and homF55L channels is 1.27 Ϯ 0.08, 2.27 Ϯ 0.34, and 10.02 Ϯ 2.39, respectively (Table 2). Based on the -fold increase in current amplitude, we expect the P o of WT, hetF55L, and homF55L channels to be ϳ0.78, 0.44, and 0.1, respectively, assuming the maximal P o after PIP 2 stimulation is near 1 for all channels. Using noise analysis (33, 34), we estimated the P o of WT, hetF55L, and homF55L channels before and after PIP 2 stimulation. As shown in Table 1, the P o for WT, hetF55L, and homF55L channels before PIP 2 application was 0.81 Ϯ 0.02, 0.66 Ϯ 0.04, and 0.11 Ϯ 0.01, respectively. After PIP 2 stimulation, the P o for WT, hetF55L, and homF55L channels increased to 0.94 Ϯ 0.01, 0.89 Ϯ 0.02, and 0.72 Ϯ 0.04. These numbers are in general agreement with the trend expected based on macroscopic current amplitude increase. However, note that the P o estimated by the noise analysis is likely higher because underestimation of the noise due to filtering (32)(33)(34)39). Also, the maximal P o estimated by noise analysis after PIP 2 stimulation for homF55L mutant channels is 0.72, significantly lower than 1 (p Ͻ 0.01). Therefore, the intrinsic P o of the homF55L mutant channel is likely below 0.1 (based on a ϳ10-fold increase in macroscopic currents after PIP 2 stimulation and the estimated maximal P o of 0.72).
Long chain acyl-CoAs, like PIP 2 , stimulate K ATP channel activity by interacting with Kir6.2 and have been proposed to play a role in modulating channel activity in ␤-cells (9, 10, 14, 40). At 5 M, oleoyl-CoA stimulated F55L channel activity to a similar extent (13.12 Ϯ 2.18-fold increase) as 5 M PIP 2 (Fig. 2B, Table 1). However, in a given patch the time course of stimulation is faster with oleoyl-CoA than with PIP 2 (Fig.  2C). In addition, unlike the effect of PIP 2 , the effect of oleoyl-CoA is readily reversible (Fig. 2C). Taken together, these results are consistent with the notion that the F55L mutation renders the channel less responsive to endogenous membrane phosphoinositides and/or long chain acyl-CoAs, leading to lower channel open probability, thereby persistent membrane depolarization and hyperinsulinism.
Reduced channel sensitivity to phosphoinositides/long chain acyl-CoAs could result from reduced binding to these ligands or impaired gating movements after ligand binding. To gain further insight into the mechanism, we substituted Phe-55 with additional amino acids with distinct biophysical properties. We reasoned that if the residue is directly involved in lipid binding, mutation to a positively charged residue should increase channel activity, whereas mutation to a negatively charged residue should decrease channel activity, as shown previously for other PIP 2 /CoA binding residues (15, 40 -42). Mutation of Phe-55 to either positively charged lysine or negatively charged glutamate yielded channels with very low activity, which increased markedly upon exposure to oleoyl-CoA (Fig. 3) or PIP 2 (not shown). The current increase after 5 M oleoyl-CoA stimulation was Ͼ20-fold for both F55E and F55K (n ϭ 3; Fig. 3). In comparison, mutation of Phe-55 to hydrophobic methionine yielded channels whose intrinsic open probabilities are lower than WT but higher than F55L, with the -fold increase in FIGURE 2. Reduced sensitivity of F55L mutant channels to PIP 2 and oleoyl-CoA. A, representative inside-out patch clamp recordings from cells expressing WT (n ϭ 6), homF55L (F55L; n ϭ 10), or hetF55L (WT ϩ F55L; n ϭ 3). Note the very low activity of the homomeric F55L mutant channel upon patch excision (arrowhead) and the large increase in current amplitude upon exposure to PIP 2 . B, same as A except that channels were exposed to 5 M oleoyl-CoA (n ϭ 7, 15, and 5 for WT, WTϩF55L, and F55L, respectively). C, direct comparison of F55L mutant channel response to oleoyl-CoA and PIP 2 in the same patch (an example from a total of four such recordings). Dashed line indicates zero current. Note that the effect of oleoyl-CoA could be easily washed out, whereas the effect of PIP 2 was irreversible. Also, some patches showed obvious reduction in sensitivity to inhibition by 5 mM ATP, as previously reported (15,17). All recordings were made at ϩ50 mV pipette potential with symmetrical K-INT/EDTA solution on both sides of the membrane patch. Inward currents are shown as upward deflections. FEBRUARY 3, 2006 • VOLUME 281 • NUMBER 5 macroscopic current upon stimulation by 5 M oleoyl-CoA for F55M being 3.11 Ϯ 0.18 (n ϭ 5). By contrast, introduction of aromatic amino acids tryptophan or tyrosine at this position had very mild or no effect on channel activity. The -fold increase in macroscopic current after oleoyl-CoA stimulation for F55W and F55Y is 1.65 Ϯ 0.01 (n ϭ 3) and 1.14 Ϯ 0.06 (n ϭ 4), respectively. These results suggest that the aromatic side chain rather than charge at this amino acid position is important for the slide helix structure and function of the channel (see "Discussion").

KCNJ11 Mutation, K ATP Channel Gating, and Hyperinsulinism
Nucleotide Sensitivity and Diazoxide Response of F55L Mutant Channels-To test whether the F55L mutation also affects channel sensitivity to intracellular nucleotides, we examined mutant channel response to ATP and MgADP. ATP dose response curves were constructed by exposing channels to various concentrations of ATP in K-INT plus 1 mM EDTA solution (see "Materials and Methods"); EDTA was used because it effectively prevents channel rundown, allowing for more accurate measurement of ATP sensitivity (31). The half-maximal inhibition concentration (IC 50 ) for WT, hetF55L, and homF55L chan-nels is 23. 4 Ϯ 1.4, 17.4 Ϯ 0.3, and 14.5 Ϯ 0.5 M, respectively (Fig. 4A); the IC 50 of homF55L channels is slightly, although significantly lower than that of WT channels (p Ͻ 0.01). Next, the response of the F55L mutant channel to MgADP stimulation was examined. In these experiments, we compared the current in 0.1 mM ATP with that in 0.1 mM ATP and 0.5 mM ADP (free Mg 2ϩ concentration ϳ1 mM). In WT channels, 0.5 mM ADP in the presence of Mg 2ϩ antagonized the inhibitory effect of 0.1 mM ATP and stimulated channel activity (Fig. 4B). This stimulatory effect was fully retained in the F55L mutant channel (Fig.  4B), indicating the mutation does not disrupt the coupling between SUR1 and Kir6.2. As described above, the patient carrying the F55L mutation responded to treatment by the potassium channel opener diazoxide. Because the mutation is heterozygous, the diazoxide response could simply be conferred by the WT allele. We, therefore, examined whether the F55L mutation in Kir6.2 affects diazoxide sensitivity. Under the same experimental paradigm, the response of homF55L mutant channel to diazoxide stimulation was comparable with that seen in WT channels (Fig. 4C). Thus, the F55L mutation does not affect the channel ability to respond to diazoxide. It is worth noting however, that although both MgADP and diazoxide stimulated mutant channel activity, the current amplitude at maximal stimulation even after prolonged exposure (Ͼ10 min; not shown) did not exceed the initial current observed in K-INT solution (Fig. 4, B and C). These results indicate the nucleotide stimulatory effect of SUR1 cannot overcome the low intrinsic open probability defect caused by the F55L mutation in Kir6.2.

DISCUSSION
Because the linkage between K ATP channel mutations and congenital hyperinsulinism was established a decade ago, much effort has been devoted to understanding the molecular and cellular mechanisms by which mutations alter channel function, as this information is necessary for developing pharmacological strategies toward disease treatment (1,3,19,43). Studies thus far, mostly from SUR1 mutations, have shown a loss of surface channel expression due to defects in channel biogenesis or trafficking and loss of channel sensitivity to MgADP as two major mechanisms leading to channel dysfunction (3). In this work we have identified a KCNJ11 mutation 165C3 A that results in a missense mutation, F55L, in Kir6.2. Functional analysis of the mutant channel revealed a novel defect. Unlike previously reported disease mutations, the missense F55L mutation in Kir6.2 does not affect channel expression nor channel response to MgADP. Instead, it renders the channel much less active by decreasing the intrinsic channel open probability by at least 10-fold. The reduced channel open probability is consistent with the mutation playing a causal role in the disease.
Mechanisms Underlying the Reduced Open Probability of F55L Mutant Channels-In the absence of nucleotides, WT K ATP channels have reported open probability of ϳ0.5-0.8 (38, 44 -46). The level of phosphoinositides such as PIP 2 in the plasma membrane has been positively correlated with the intrinsic open probability of K ATP channels  . Recording conditions were as described in Fig. 2. Response of the various mutant channels to PIP 2 was similar to their response to oleoyl-CoA (not shown). (15,17,38). The fact that the F55L mutant channel activity is markedly increased by application of exogenous PIP 2 supports the notion that reduced channel sensitivity to PIP 2 accounts for the reduced open probability seen in the mutant. Currently available experimental approaches do not allow for definitive tests of whether the mutation reduces channel sensitivity to PIP 2 by directly affecting binding, indirectly affecting the gating steps after PIP 2 binding, or both. Several positively charged residues in the cytoplasmic domain of Kir6.2 have been proposed as PIP 2 binding residues, including Arg-54, which is immediately adjacent to Phe-55 (15,39,41,42). Deliberate mutation of Arg-54 to neutral alanine or glutamine or to negatively charged glutamate results in channels with reduced PIP 2 response and low open probability similar to the F55L mutant (39,41), whereas mutation to positively charged lysine results in a channel with WT-like phenotype (41). These observations have led to the proposal that Arg-54 may be directly involved in PIP 2 binding (41). Phe-55 is the first amino acid in the predicted amphipathic "slide helix" that runs parallel to the lipid bilayer; the slide helix has been proposed to move laterally along the membrane during channel gating (47). Given our results that substitution of Phe-55 by either lysine or glutamate greatly reduced channel open probability, whereas mutation of Phe-55 to tryptophan or tyrosine, two other amino acids also containing an aromatic side chain, gave rise to channels resembling WT, Phe-55 is unlikely a PIP 2 /CoA binding residue. Rather, our data suggest that the reduced sensitivity to PIP 2 /CoA seen in the F55L mutant may be due to impaired transduction steps after PIP 2 binding. The leucine mutation may disrupt the amphipathic helical structure and interfere with the ability of the helix to move along the membrane during gating. Alternatively, F55L may interfere with PIP 2 /CoA binding indirectly by preventing other residues from interacting with PIP 2 /CoA. In addition to reduced PIP 2 sensitivity and reduced open probability, a slightly increased ATP sensitivity was also observed in the F55L mutant. We do not believe this increased ATP sensitivity is a result of increased ATP binding affinity but is a consequence of allosteric effect. This is consistent with previous studies showing that channel open probability is negatively correlated with ATP sensitivity (38). The increased ATP sensitivity is not likely to contribute to the disease since at physiological concentrations of ATP (mM range) most WT channels are already closed (48). Some mutations in Kir6.2 that reduce channel sensitivity to PIP 2 , such as R176C and R177C, have been reported to cause uncoupling between SUR1 and Kir6.2 and abolish channel response to MgADP (49). This is not the case for F55L, as the channel exhibits normal response to both MgADP and diazoxide. Because both R176C and R177C are believed to directly reduce PIP 2 binding affinity, the distinction between these two mutants and F55L is consistent with our above suggestion that F55L reduces channel sensitivity to PIP 2 via a different mechanism.
The reduced intrinsic channel open probability caused by the F55L mutation in Kir6.2 is in direct contrast to the increased intrinsic channel open probability reported in a group of neonatal diabetes-causing Kir6.2 mutations including V59G, which is also located in the slide helix of Kir6.2 (50,51). The V59G mutation results in channels with very high intrinsic open probability that cannot be reversed by prolonged incubation with the PIP 2 binding polycation neomycin, suggesting that the mutant channel is locked in an open state (50). These studies show that disruption of the slide helix structure by mutations can lead to either gain or loss of channel open probability, highlighting the critical role of the slide helix in K ATP channel gating.
Membrane Phosphoinositides and Long Chain CoAs in K ATP Channel Regulation-The gating effects of membrane phosphoinositides and long chain CoAs on K ATP channels in isolated membranes have been well documented (9 -17). Structure-functional studies indicate that modulation of channel activity by the two classes of lipid molecules involve the same set of Kir6.2 residues (9 -11, 40). Our results that both PIP 2 and oleoyl-CoA rescue the P o defect of the F55L mutant channel are consistent with this notion. However, because phosphoinositides and long chain CoAs appear to share the same gating mechanism, it has been difficult to study the relative role of the two classes of lipids in determining channel activity in physiological or pathological conditions (8,52). Recently, we reported that manipulations of channel-PIP 2 interactions in the insulin-secreting cell line INS-1 have profound effects on channel activity and the coupling between glucose and insulin secretion (53). A Kir6.2 polymorphism, E23K/I337V, that has been implicated in type II diabetes was recently shown to increase channel sensitivity to long chain CoAs (12,16,54). Our study presented here demonstrates that a naturally occurring mutation in human that reduces channel response to PIP 2 /CoA results in congenital hyperinsulinism. Together, these studies are consistent with both classes of lipid molecules playing a role in controlling channel activity. They raise the possibility that membrane phosphoinositides and long chain CoAs could serve as potential pharmacological targets for treating insulin secretion disorders caused by K ATP channel mutations.