Essential Role of Phosphoinositide 3-Kinase in Leptin-induced K ATP Channel Activation in the Rat CRI-G1 Insulinoma Cell Line*

The mechanism by which leptin increases ATP-sensi-tive K 1 ( K ATP ) channel activity was investigated using the insulin-secreting cell line, CRI-G1. Wortmannin and LY 294002, inhibitors of phosphoinositide 3-kinase (PI3-kinase), prevented activation of K ATP channels by leptin. The inositol phospholipids phosphatidylinositol bisphosphate and phosphatidylinositol trisphosphate (PtdIns(3,4,5)P 3 ) mimicked the effect of leptin by in- creasing K ATP channel activity in whole-cell and inside- out current recordings. LY 294002 prevented phosphatidylinositol bisphosphate, but not PtdIns(3,4,5)P 3 , from increasing K ATP channel activity, consistent with the latter lipid acting as a membrane-associated messenger linking leptin receptor activation and K ATP channels. Signaling cascades, activated downstream from PI 3-ki-nase, utilizing PtdIns(3,4,5)P 3 as a second messenger and commonly associated with insulin and cytokine action (MAPK, p70 ribosomal protein-S6 kinase, stress-ac-tivated protein kinase 2, p38 MAPK, and protein kinase B), do not appear to be involved in leptin-mediated activation of K ATP channels

The hormone leptin, secreted by adipocytes, has a major influence on body weight homeostasis (1,2). Although the hypothalamus is considered the main target for leptin, particularly with respect to body weight regulation, it is clear that this hormone has distinct actions on other peripheral, target organs. There have been several reports that leptin reduces insulin secretion from pancreatic beta cells (3)(4)(5)(6), although this view is not shared by all investigators (7). One mechanism proposed to explain the leptin-induced reduction in insulin secretion is via activation of ATP-sensitive K ϩ (K ATP ) channels (8,9). This increase in potassium current results in beta cell hyperpolarization, reduced calcium entry, and hence decreased insulin secretion. In addition, there are features common to both insulin-secreting cells and leptin-sensitive hypothalamic neurones (10,11), most notably glucose responsiveness and the presence of K ATP channels, which are activated by exposure of the cells to leptin. The apparent involvement of both leptin receptors and K ATP channel activation in key systems involved in metabolic homeostasis has led us to examine the likely signal transduction pathways underlying this effect.
The leptin receptor belongs to the class I cytokine receptor superfamily (1,2), members of which are thought to signal via janus-tyrosine kinases. Activated janus-tyrosine kinases can mediate signaling via insulin receptor substrate proteins (12)(13)(14), which following tyrosine phosphorylation become docking sites for Src homology 2-containing enzymes like phosphoinositide 3-kinase (PI 3-kinase). 1 Indeed, leptin is reported to stimulate glucose transport in C 2 C 12 myotubules via a PI 3kinase-dependent process (15), and leptin signaling in these cells employs janus-tyrosine kinase 2 and insulin receptor substrate 2 in this process (16). There is also growing evidence that PI 3-kinase plays a pivotal role in the signal transduction pathways linking insulin receptor activation and various cellular responses (14). Wortmannin and LY 294002 are inhibitors of PI 3-kinase (17,18) that are reported to inhibit insulininduced activation of protein kinase B (PKB) (19), mitogenactivated protein kinase (MAPK) (20,21), and p70 S6k (22). We have shown recently that leptin activation of K ATP channels in CRI-G1 cells is sensitive to both wortmannin and LY 294002, suggesting the possible involvement of a PI 3-kinase and that insulin occludes leptin activation of K ATP channels (23), indicating cross-talk between insulin and leptin-stimulated signaling pathways. Thus, the cell-signaling pathways initiated by leptin action in CRI-G1 cells may overlap with those of insulin. In support of this possibility, recent reports indicate that leptin-induced proliferation involves activation of MAPK in C3H10T1/2 cells (24) and the insulin-secreting MIN6 cell line (25). It has also been demonstrated (26) that in Chinese hamster ovary cells stably expressing leptin receptor isoforms, tyrosine phosphorylation of MAPK is enhanced. Most of the cellular roles of PI 3-kinase have been attributed to the lipid products of these enzymes, which bind to target molecules via specific lipid binding domains. There is mounting evidence that phosphatidylinositol trisphosphate (PtdIns (3,4,5)P 3 ) levels increase in cells following stimulation by a variety of agonists and that the increased PtdIns(3,4,5)P 3 levels precede activation of downstream signaling molecules (27).
The possibility that phosphoinositides are involved in signaling to potassium channels is indicated by numerous reports suggesting that phosphatidylinositol bisphosphate (PtdIns(4,5)P 2 ) can directly activate cloned inward rectifier K ϩ channels (28) and native (29,30) and cloned (consisting of the inward rectifier channel, Kir6.2, and the sulfonylurea receptor, SUR1 subunits) K ATP channels (31,32). In this report we compare the effect of leptin action on K ATP channel activity in CRI-G1 cells with that induced by PtdIns(4,5)P 2 or PtdIns(3,4,5)P 3 using whole-cell and inside-out current recordings.
Electrophysiological Recording and Analysis-Experiments were performed using whole-cell current clamp recordings to monitor membrane potential with excursions to voltage clamp mode to examine macroscopic currents and excised inside-out recordings to examine single channel responses, as described previously (9). During voltage clamp recordings, the membrane potential was held at Ϫ50 mV, and 10 mV steps of 100 ms duration were applied every 200 ms (range Ϫ120 to Ϫ30 mV). Current and voltage were measured using an Axopatch 200 B amplifier (Axon Instruments), and currents evoked in response to the voltage step protocol were analyzed using pCLAMP 6.0 software (Axon Instruments), whereas current clamp data were recorded onto digital audiotapes and replayed for illustration on a Gould TA 240 chart recorder. Single channel data were analyzed for current amplitude and channel activity (N f P o , where N f is the number of functional channels and P o is the open probability) as described previously (34). All data are expressed as mean Ϯ S.E., and statistical analyses were performed using Student's unpaired t test (unless otherwise stated). p Ͻ 0.05 was considered significant.
Recording electrodes were pulled from borosilicate glass and had resistances of 1-5 M⍀ for whole cell recordings and 8 -12 M⍀ for inside-out experiments when filled with electrolyte solution. The pipette solution for whole-cell recordings comprised 140 mM KCl, 0.6 mM MgCl 2 , 2.73 mM CaCl 2 , 5.0 mM ATP, 10 mM EGTA, 10 mM HEPES, pH 7.2 (free [Ca 2ϩ ] of 100 nM), whereas for single channel recordings the pipette solution contained: 140 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, pH 7.2. The bath solution for whole cell recordings comprised of normal saline: 135 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM HEPES, pH 7.4, whereas for inside-out excised patches the bath solution contained: 140 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM EGTA, 10 mM HEPES, pH 7.2 (free Ca 2ϩ of 30 nM). The free Ca 2ϩ concentrations were calculated using the "METLIG" program (P. England and R. Denton, University of Bristol, UK). All solution changes were achieved by superfusing the bath with a gravity feed system at a rate of 10 ml min Ϫ1 , which allowed complete bath exchange within 2 min. All experiments were performed at room temperature (22-25°C).
Determination of PKB Activity-CRI-G1 cells were deprived of serum overnight in Dulbecco's modified Eagle's medium and then stimulated for various times at 37°C with either insulin (100 nM) or leptin (10 nM) in the presence or absence of 100 nM wortmannin. Each 10-cm dish of cells was lysed in 1 ml of ice-cold lysis buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM EGTA, 1% v/v Triton X-100, 1 mM sodium orthovanadate, 10 mM sodium glycerophosphate, 50 mM NaF, 5 mM sodium pyrophosphate, 1 M microcystin-LR, 0.27 M sucrose, 1 mM benzamidine, 0.2 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 0.1% v/v 2-mercaptoethanol). The lysates were centrifuged at 4°C for 5 min at 13,000 ϫ g, and 100 g of each lysate was incubated for 60 min with 5 l of protein G-Sepharose coupled to 5 g of PKB isoform antibodies (35). The protein G-Sepharose-antibody-PKB complex was washed twice with 1 ml of lysis buffer containing 0.5 M NaCl and twice with 50 mM Tris-HCl, pH 7.5, 0.1 mM EGTA, and 0.1% (v/v) 2-mercaptoethanol. The immunoprecipitates were assayed for PKB activity using Cross-tide as a substrate (36). One unit of activity is defined as the amount, which catalyzes the incorporation of 1 nmol of phosphate into Cross-tide in 1 min. Each experimental assay was performed 3 times, in triplicate.
Phosphatidylinositol (3,4,5) Trisphosphate Measurement-PtdIns (3,4,5)P 3 levels were determined by radioligand displacement assay (37,38). Briefly, CRI-G1 cells were cultured in 80-cm 2 flasks as described previously. The cell monolayer was rinsed with normal saline prior to incubation with either insulin (100 nM) or leptin (10 nM) all prepared in normal saline containing 0.2% bovine serum albumin. After 30 min the medium was aspirated, and the cells were quenched in ice-cold 10% trichloroacetic acid. The trichloroacetic acid pellet was harvested by scraping and collected by centrifugation for 2-5 min at 13,000 rpm in a microfuge. The pellet was pre-extracted with neutral chloroform/methanol (1:2 v/v) to remove polar molecules and inositol phosphates, and then lipids were isolated by Bligh & Dyer extraction with chloroform/ methanol/HCl (40:80:1 v/v). Lipids were vacuum dried and stored at Ϫ80°C until assayed, after alkaline hydrolysis to release lipid headgroups, by radioligand displacement of inositol 1,3,4,5-tetrakisphosphate as described previously (37,38). The remaining protein was vacuum-dried and resuspended in 1 M NaOH, and the protein content was assayed using the DC protein assay kit (Bio-Rad). Results were expressed as pmol PtdIns (3,4,5)P 3 mg Ϫ1 protein, and each assay was carried out a minimum of nine times. Drugs-Tolbutamide, rapamycin, Mg-ATP, U71322, insulin, and wortmannin were obtained from Sigma. Recombinant human leptin, LY 294002, and PtdIns(4,5)P 2 were obtained from Calbiochem. PD 98059 was a gift from Professor John Hughes, Parke Davis Neuroscience Research Center (Cambridge UK). SB 203580 was a gift from Professor P. Cohen, University of Dundee. Tolbutamide was made up as a stock solution in 100 mM stock solution in Me 2 SO. Mg-ATP was made up as a 100 mM stock solution in 10 mM HEPES at pH 7.2 and kept at Ϫ4°C until required. Leptin was prepared as a 1 M stock solution, and insulin as a 100 M stock, in normal saline containing 0.2% (w/v) bovine serum albumin as a carrier. Wortmannin was stored as a 300 M stock solution in Me 2 SO, whereas LY 294002 was stored as 10 mM solution in 1% methanol. PtdIns(4,5)P 2 and PtdIns(3,4,5)P 3 were stored as 1 mM stock solutions in distilled water and sonicated before use. PD 98059 and SB 203580 were stored as a 10 mM stock solutions in Me 2 SO, whereas rapamycin was prepared as a 100 M solution in Me 2 SO. U73122 was stored as a 2 mM stock solution in 1% ethanol.

RESULTS
Leptin Activates K ATP Channels in CRI-G1 Cells-Under current clamp conditions with 5 mM ATP in the pipette solution to maintain K ATP channels in the closed state, the mean resting membrane potential of CRI-G1 cells was Ϫ37.0 Ϯ 1.3 mV (n ϭ 4). Application of leptin (10 nM) hyperpolarized the CRI-G1 cells to Ϫ64.8 Ϯ 3.2 mV (n ϭ 4; Fig. 1A). Examination of the voltage-clamped macroscopic currents indicates that prior to the addition of leptin the slope conductance of the cells was 0.65 Ϯ 0.09 nS, and following exposure to leptin (10 nM), this increased to 2.23 Ϯ 0.64 nS (n ϭ 4; Fig. 1B). The reversal potential (obtained from the point of intersection of the currentvoltage relationships) associated with the leptin-induced conductance increase was Ϫ77.8 Ϯ 2.1 mV (n ϭ 4), which is close to the calculated value for E K of Ϫ84 mV in this system, indicating an increase in K ϩ conductance. Application of tolbutamide (100 M) completely reversed the leptin-induced hyperpolarization and decreased conductance to pre-leptin levels (n ϭ 3), indicating that leptin activates K ATP channels in this cell line. Application of inhibitors of PI 3-kinase, wortmannin (10 nM; n ϭ 5; data not shown) and LY294002 (10 M; n ϭ 5; Fig. 1, C and D) had no significant affect on the resting membrane potential or slope conductance of CRI-G1 cells but prevented the leptin-induced cell hyperpolarization and activation of K ATP currents. These data are in agreement with our previous studies demonstrating that leptin activates K ATP channels in the rat CRI-G1 insulinoma cell line (9) presumably via activation of PI 3-kinase (23).
PtdIns(4,5)P 2 Mimics Leptin Activation of K ATP Channel Currents-There have been several studies demonstrating that phosphoinositides, most particularly, PtdIns(4,5)P 2 , can directly regulate the activity of various ion transporters and channels, including members of the inward rectifying family of K ϩ channels, to which the K ATP channel belongs (29 -32). Consequently, CRI-G1 cells were dialyzed with an electrode solu-tion containing 5 mM ATP and 50 M PtdIns(4,5)P 2 (Fig. 2, A and B). The initial mean resting membrane potential and slope conductance values obtained within 1-1.5 min of achieving the whole-cell configuration were Ϫ38.0 Ϯ 0.92 mV and 0.53 Ϯ 0.02 nS, respectively (n ϭ 5). However, after approximately 10 -15 min of dialysis the membrane potential hyperpolarized to Ϫ67.0 Ϯ 4.1 mV with an associated increase in slope conductance to 2.62 Ϯ 0.66 nS (n ϭ 5). This effect of PtdIns(4,5)P 2 was due to the activation of K ATP channels as tolbutamide (100 M) completely reversed the actions of PtdIns(4, 5)P 2 , resulting in a mean resting membrane potential and slope conductance of Ϫ38.2 Ϯ 0.43 mV and 0.59 Ϯ 0.07 nS, respectively, with an associated reversal potential of Ϫ78.1 Ϯ 0.55 mV (n ϭ 4). These data indicate that, in a manner similar to leptin, PtdIns(4,5)P 2 activates K ATP channels in CRI-G1 insulin-secreting cells. This was confirmed by examination of the actions of PtdIns(4,5)P 2 (1 M) on K ATP channel activity when applied directly to insideout membrane patches (Fig. 2E). PtdIns(4,5)P 2 induced activation of channel activity in the presence of ATP with values for N f P o of 1.09 Ϯ 0.28, 0.03 Ϯ 0.01, and 0.65 Ϯ 0.21 (n ϭ 3) in control, 0.1 mM MgATP, and MgATP ϩ 1 M PtdIns(4,5)P 2 , respectively.
PtdIns(4,5)P 2 is the precursor for the well studied phospholipase C (PLC)-linked second messengers, inositol 1,4,5trisphosphate and diacylgylcerol, and as such it is possible that leptin-induced K ATP channel activation may be attributed to a PLC-dependent process. Therefore, the effects of the PLC inhibitor, U73122, were investigated on leptin-induced K ATP current activation. Under current clamp conditions with 5 mM ATP and 2 M U73122 present in the electrode solution, the mean resting membrane potential and slope conductance of CRI-G1 cells were Ϫ35.4 Ϯ 0.85 mV and 0.61 Ϯ 0.09 nS, respectively (n ϭ 4). Subsequent application of leptin (10 nM) resulted in hyperpolarization and an increased slope conductance to Ϫ68.2 Ϯ 4.0 mV and 4.07 Ϯ 0.96 nS, respectively (n ϭ 4; data not shown) with an associated reversal potential of Ϫ79.0 Ϯ 1.1 mV (n ϭ 4) indicating the involvement of a K ϩ conductance. Tolbutamide (100 M) readily reversed the leptininduced hyperpolarization and increase in K ϩ conductance to pre-leptin levels (Ϫ38.2 Ϯ 4.2 mV and 0.52 Ϯ 0.02 nS (n ϭ 4)). Therefore it is unlikely that a phosphoinositide-specific PLC is involved in leptin activation of K ATP channels.
LY 294002 Prevents PtdIns(4,5)P 2 Activation of K ATP Channel Currents-Stimulation of cells with insulin or certain growth factors results in the activation of a Type I PI 3-kinase, which primarily phosphorylates PtdIns(4,5)P 2 to PtdIns(3,4,5)P 3 (14). Because PtdIns(4,5)P 2 is a key substrate for PI 3-kinase and inhibitors of this enzyme system prevent leptin activation of K ATP channels (23), we have examined the sensitivity of the PtdIns(4,5)P 2 -mediated increase in K ATP current to PI 3-kinase inhibitors. Cells were incu- bated with LY 294002 (10 M) for 25-30 min prior to obtaining the whole-cell configuration and subsequent dialysis with 5 mM ATP and 50 M PtdIns(4,5)P 2 . Immediately after achieving the whole cell configuration, the resting membrane potential of the cells was Ϫ40.6 Ϯ 0.95 mV (n ϭ 4) with a slope conductance under voltage clamp conditions of 0.52 Ϯ 0.05 nS (n ϭ 4). However, in the pres-ence of LY 294002, following 11-15 min dialysis, PtdIns(4,5)P 2 (50 M) had no effect on the resting membrane potential (Ϫ39.1 Ϯ 1.0 mV; n ϭ 4) or slope conductance (0.56 Ϯ 0.03 nS; n ϭ 4) of CRI-G1 cells (Fig. 2, C and D). Furthermore, the activation of K ATP by PtdIns(4,5)P 2 was completely inhibited in isolated patches pretreated with 10 M LY 294002 (Fig. 2F). Mean channel activity FIG. 2. PtdIns(4,5)P 2 activates K ATP channel currents via a PI 3-kinase-dependent process. A, representative current clamp record of a CRI-G1 cell dialyzed with 5 mM ATP and 50 M PtdIns(4, 5)P 2 . Dialysis with PtdIns(4,5)P 2 resulted in a slowly developing hyperpolarization of the cell membrane to Ϫ72 mV, an action readily reversed by the sulfonylurea, tolbutamide (100 M). B, plot of the current voltage relations for the currents obtained in A: q, control (before any change in resting membrane potential); Ⅺ, PtdIns(4, 5)P 2 ; and E, tolbutamide. PtdIns(4,5)P 2 increased the membrane conductance relative to control and tolbutamide reversed the PtdIns(4,5)P 2 -induced increase in conductance with a reversal potential of Ϫ83 mV. C, current clamp trace of a cell dialyzed with 5 mM ATP and 50 M PtdIns(4,5)P 2 . Incubation of cells with LY 294002 (10 M) for at least 20 min prior to obtaining the whole cell configuration prevented the PtdIns(4,5)P 2 -induced hyperpolarization. D, plot of the current voltage relations for the currents obtained in C: LY 294002 (q) and LY 294002 and PtdIns(4, 5)P 2 (E). E, inside-out recordings of single K ATP channel activity at a membrane potential of Ϫ40 mV. Application of 0.1 mM Mg-ATP markedly reduced channel activity. Addition of 1 M PtdIns(4,5)P 2 in the presence of Mg-ATP elicited a substantial activation of K ATP channel activity. F, identical experiment to E using a separate patch that was exposed to 10 M LY 294002. Note that the addition of 1 M PtdIns(4,5)P 2 in the presence of LY 294002 does not induce an increase in K ATP channel activity.
(N f P o ) was 1.00 Ϯ 0.39 in control, 0.05 Ϯ 0.01 in the presence of 10 M LY 294002 ϩ 0.1 mM Mg-ATP, and 0.03 Ϯ 0.01 (n ϭ 3) upon further addition of 1 M PtdIns(4,5)P 2 . These data demonstrate that blockade of PI 3-kinase prevents PtdIns(4,5)P 2 activation of K ATP channels in these cells, indicating that this molecule does not directly mediate leptin activation of K ATP channels.

Effects of Insulin and Leptin on PtdIns(3,4,5)P 3 Levels in CRI-G1 Cells-
To determine whether leptin mediates its effects on K ATP channel activity in CRI-G1 cells via enhanced synthesis of PtdIns(3,4,5)P 3 , the levels of this lipid were analyzed. As a positive control, we also examined the effects of insulin on PtdIns(3,4,5)P 3 levels in CRI-G1 cells as we have demonstrated that these cells respond to the presence of 100 nM insulin with a significant increase in PKB activity. Application of 100 nM insulin (30-min exposure) induced a significant increase (Fig. 4) in PtdIns(3,4,5)P 3 levels (p Ͻ 0.001). However, 10 nM leptin (cells stimulated for 30 min) had no significant effect on PtdIns(3,4,5)P 3 levels (Fig. 4). Consequently, although leptin-induced activation of K ATP channels is via a PI 3-kinasedependent process and PtdIns(3,4,5)P 3 can clearly activate K ATP channels directly, there is no direct correlation between this leptin response and global PtdIns(3,4,5)P 3 levels in this cell line. However, this result does not preclude the possibility that a small and/or localized, increase in PtdIns(3,4,5)P 3 may be elicited by leptin receptor stimulation. Such an increase may not be detected by the global assay but could activate K ATP channels directly or be sufficient to activate downstream effectors which in turn could activate K ATP channels.
Effects of Leptin on PKB and p70 S6k Activities-Activation of PI 3-kinase and the subsequent production of 3-phosphorylated inositol lipids is generally sufficient to trigger many insulinstimulated pathways (14), indicating that this enzyme is situated at the apex of divergent signal transduction cascades. Consequently, we have examined the likelihood that the downstream signaling molecules identified with insulin action are also associated with leptin activation of K ATP channels. Recently, evidence has accumulated linking the effects of insulin and other growth factors on cell function with protein kinase B (also known as Akt), which lies downstream of PI 3-kinase (19,27). It is generally considered that PKB is regulated through activation of PI 3-kinase by a dual mechanism. Direct binding of PI 3-kinase-derived lipids to the pleckstrin homology domain of PKB allows recruitment of the enzyme to the plasma membrane where it is phosphorylated (and therefore activated) by 3-phosphoinositide-dependent kinase 1, an enzyme that may also require the presence of PI 3-kinase products. We have determined whether leptin stimulation of K ATP channel activity in CRI-G1 cells is correlated with any change in PKB isoform (PKB␣, -␤, and -␥) activities compared with insulin as a positive control. The control activity of PKB isoforms, in the absence of hormones, was 0.87 Ϯ 0.2, 0.24 Ϯ 0.05, and 1.98 Ϯ 0.35 milliunit mg Ϫ1 protein (n ϭ 3) for PKB␣, -␤, and -␥, respectively. Fig. 5A shows that insulin (10-min exposure) induced a clear activation of PKB␣ and PKB␥ isoforms with little effect on PKB␤ activity, and the increase in PKB␣ and PKB␥ isoform activity was inhibited by a 10-min pre-exposure of the cells to 100 nM wortmannin (n ϭ 3). In contrast, 10 nM leptin (cells stimulated for between 5 and 60 min) produced no consistent alteration in PKB isoform activity (n ϭ 3; Fig. 5B). These results are consistent with the PtdIns(3,4,5)P 3 measurements following challenge of CRI-G1 cells with these hormones. Previously, we have shown (23) that insulin stimulation of CRI-G1 cells results in the occlusion of leptin-induced activation of K ATP channels indicating some cross-talk between signaling systems. However, application of 10 nM leptin (5-30 min) prior to insulin challenge (Fig. 5C) or 100 nM insulin (10 min) prior to leptin challenge (data not shown) had no significant effect on insulin-stimulated PKB␣ and PKB␥ isoform activity.
These data indicate that activation of PKB or p70 S6k is unlikely to underlie leptin activation of K ATP channels in this cell line. Although PtdIns(3,4,5)P 3 itself can activate K ATP channels, it is plausible that the PI 3-kinase-dependent actions of leptin are not mediated by the synthesis of this lipid, but by the protein serine/threonine kinase activity associated with type I PI 3-kinases. Indeed, Wymann and colleagues (41)  . E, representative single channel currents recorded from an inside-out membrane patch exposed to symmetrical 140 mM KCl at a membrane potential of Ϫ40 mV. Application of 0.1 mM Mg-ATP markedly reduced channel activity. Addition of 1 M PtdIns(3,4,5)P 3 in the presence of Mg-ATP resulted in an substantial increase in K ATP channel activity. F, identical experiment to E, using a separate patch that was exposed to 10 M LY 294002. Note that the addition of 1 M PtdIns(3,4,5)P 3 in the presence of LY 294002 does induce a large increase in K ATP channel activity. cently made use of a mutated form of p110␥, which lacks lipid kinase but retains protein kinase activity, to show that wortmannin-sensitive activation of MAPK occurs via the serine/ threonine kinase activity of PI 3-kinase.
MAPK Subfamilies Do Not Mediate Leptin Activation of K ATP Channels-Activation of MAPK has been implicated as a signaling intermediate for both insulin (42,43) and leptin (24 -26) in various cell types, actions that are sensitive to inhibition by PD 98059, a synthetic inhibitor of the activation of MAPK kinase (42,44). Under current clamp conditions with 5 mM ATP and 50 M PD 98059 in the pipette solution (Fig. 6C), the mean resting potential of CRI-G1 cells following dialysis for at least 10 -15 min was Ϫ40.0 Ϯ 3.5 mV (n ϭ 6) and examination of the macroscopic currents under voltage clamp conditions (Fig. 6D) gave a mean slope conductance of 0.57 Ϯ 0.07 nS (n ϭ 6), indicating that PD 98059 had no effect per se on K ATP currents. Application of leptin (10 nM) after a 10 -15-min dialysis hyperpolarized cells to Ϫ67.2 Ϯ 3.2 mV (n ϭ 6) and increased the slope conductance to 2.86 Ϯ 1.63 nS with an associated reversal potential of Ϫ79.0 Ϯ 0.86 mV (n ϭ 6). Tolbutamide (100 M) completely reversed these actions of leptin (Fig. 6, C and D), such that the membrane potential and slope conductance values returned to Ϫ40.2 Ϯ 1.8 mV and 0.53 Ϯ 0.07 nS, respectively (n ϭ 6), values not significantly (p Ͼ 0.05) different from control. Thus blockade of the MAPK pathway does not occlude leptin activation of K ATP channels in CRI-G1 insulinoma cells.
Cellular stresses (e.g. UV irradiation and hydrogen peroxide) and certain cytokines have been demonstrated to signal via two other MAPK subfamilies, stress-activated protein kinase and p38 MAPK, pathways partly controlled by PI 3-kinase (45,46). Consequently, the effects of SB 203580, a p38 MAPK (47), and at higher concentrations (3-10 M) an in vitro stress-activated protein kinase (48) inhibitor were examined on the leptin response. CRI G1 cells dialyzed with an electrode solution containing 5 mM ATP and 10 M SB 203580 had a mean resting membrane potential and slope conductance of Ϫ41.0 Ϯ 3.2 mV and 0.72 Ϯ 0.12 nS (n ϭ 3), respectively. Application of leptin (10 nM) hyperpolarized the cells to Ϫ74 Ϯ 4.8 mV and increased the slope conductance to 6.85 Ϯ 2.30 nS (n ϭ 3; data not shown) with a reversal of Ϫ79.0 Ϯ 1.4 mV (n ϭ 3) indicating an increase in K ϩ conductance. Tolbutamide (100 M) reversed the actions of leptin resulting in membrane potential and conductance values of Ϫ42.0 Ϯ 4.3 mV and 0.70 Ϯ 0.07 nS, respectively (n ϭ 3). Consequently, these studies indicate that the common signaling pathways, associated with insulin and growth factor stimulation of cells and that utilize PI 3-kinase as a signaling intermediate, appear not to be functionally connected to the leptin-induced increase in K ATP channel activity in this cell line.

DISCUSSION
Inositol phospholipids, applied directly to the cytoplasmic aspect of membrane patches, have been shown to activate native and cloned K ATP channels (29 -32). Specifically, PtdIns(4,5)P 2 has been proposed as a physiological activator of K ATP channels that functions by disconnecting the channel and PKB␥ (gray bars) were immunoprecipitated and assayed as described previously (35). B, cells were stimulated with 10 nM leptin for the times indicated and lysed as described under "Materials and Methods." The PKB isoforms were immunoprecipitated and assayed as described above. C, cells were preincubated with 10 nM leptin for 10 min then stimulated with 100 nM insulin for the times indicated and lysed as described under "Materials and Methods." The PKB isoforms were immunoprecipitated and assayed as described above. The data in A-C are presented as the mean fold activation (relative to the activity in the absence of stimuli) Ϯ S.E. for three separate experiments each performed in triplicate. from the constant inhibition produced by the high cytoplasmic concentration of ATP (31,32). A suggested mechanism to explain this phenomenon is that PtdIns(4,5)P 2 binds directly to the C terminus of Kir6.2 and, by charge neutralization, stabilizes the open state of the channel by reducing the inhibitory effect of ATP (31). The present study also demonstrates that PtdIns(4,5)P 2 can activate K ATP channels of the CRI-G1 cell line when presented directly to the cytoplasmic aspect in the presence of ATP. Surprisingly, however, this effect, like that of leptin, was blocked by inhibitors of PI 3-kinase and was mimicked by the direct application of PtdIns(3,4,5)P 3 , a key product of type I PI 3-kinases. By contrast, an inhibitor of phosphoinositide-specific PLCs did not prevent the effect of PtdIns(4,5)P 2 on K ATP channel activation. Taken together these results suggest that channel activation results not from a direct action of PtdIns(4,5)P 2 but via a product of its metabolism by a PI 3-kinase, most likely PtdIns(3,4,5)P 3 .
Recently, it has been shown that the ob gene product leptin activates K ATP channels in pancreatic beta cells (8) and CRI-G1 insulin-secreting cells (9), an effect consistent with suppression of insulin secretion. This action of leptin was shown to be blocked by inhibitors of PI 3-kinase (23). These observations, together with the present results, suggested that leptin may exert its effects on K ATP channels by stimulating a type I PI 3-kinase to synthesize PtdIns(3,4,5)P 3 . If correct it would be expected that leptin should produce a detectable increase in the level of PtdIns(3,4,5)P 3 present in CRI-G1 cells. Whereas insulin stimulated the production of PtdIns(3,4,5)P 3 , leptin was without apparent effect using our assay system. One possible explanation for this paradox is that leptin does activate PI 3-kinase to a limited degree, but the resulting increase in PtdIns(3,4,5)P 3 may be too small to detect reliably using a measure of the total cell PtdIns(3,4,5)P 3 content. Recent studies have shown the existence of a considerable receptor reserve for the PI 3-kinase-dependent activation of PKB by insulin such that substantial activation of PKB occurs at concentrations of insulin that cause barely detectable changes in the total cell PtdIns(3,4,5)P 3 content. 2 Thus the activity state of PKB, and perhaps other downstream targets of PI 3-kinase signaling pathways, may provide a more sensitive indicator of small changes in PtdIns(3,4,5)P 3 content. CRI-G1 cells were shown to contain the ␣ and ␥ isoforms of PKB both of which could be rapidly and potently activated by insulin. Leptin, however, had no detectable effect on PKB activity in these cells. Furthermore, the lack of sensitivity of K ATP channel activation to rapamycin establishes that p70 S6k , another protein/serine kinase that is frequently activated downstream of PI 3-kinase, 2 I. H. Batty and C. P. Downes, unpublished data. , and tolbutamide (q). PD 98059 did not prevent leptin-induced increase in K ϩ conductance. C, current clamp record of a separate cell dialyzed with 5 mM ATP and 50 nM rapamycin. Dialysis of cells with rapamycin for at least 10 min had no effect on the resting membrane potential and did not prevent the leptin-induced hyperpolarization. D, plot of current-voltage relationships for the voltage clamped currents obtained in C: rapamycin (‚), rapamycin and leptin (E), and tolbutamide (q). Rapamycin failed to prevent the leptin-induced increase in K ϩ conductance.
is not required for this leptin-mediated response.
There is clear evidence of cross-talk between insulin and leptin receptor-signaling pathways. For example, leptin attenuates some insulin-induced signals, including gluconeogenesis in human hepatocytes, even though both insulin and leptin stimulate PI 3-kinase in these cells (49). In addition, exposure of CRI-G1 cells to insulin prevents leptin activation of K ATP channels (23). It was therefore possible that leptin also activates an inhibitory pathway that attenuates some aspects of insulin signaling.
The sensitivity of K ATP channel activation by leptin to low concentrations of wortmannin and LY294002, however, strongly suggests that a PI 3-kinase is involved in this response. PI 3-kinases have long been known to exhibit both lipid and protein kinase activities that are functions of the same, wortmannin-sensitive active site. A recent study, which made use of a p110 mutant that retained protein kinase activity in the absence of detectable lipid kinase activity, suggested that the protein kinase activity of PI 3-kinase ␥ mediates the activation of MAPK (41). Furthermore, leptin activates MAPK in Chinese hamster ovary cells stably expressing leptin receptor isoforms (26), a mouse embryonic cell line (24), and the pancreatic beta cell line MIN6 (25). Well characterized inhibitors of the extracellular signal-regulated kinase pathway and pathways involving the related MAPK isoforms, p38 and stressactivated protein kinase, however, all failed to prevent the activation of K ATP channels by leptin. This suggests that the protein kinase activity of PI 3-kinase does not account for the activation of K ATP channels by leptin via any of these pathways.
Another possible explanation of our results is that leptin receptor activation may regulate a dedicated pool of PI 3-kinase leading to a localized increase in PtdIns(3,4,5)P 3 that has little effect on the total PtdIns(3,4,5)P 3 content of the cells. In this scenario, PtdIns(3,4,5)P 3 could bind directly to K ATP channel subunits as proposed originally for PtdIns(4,5)P 2 (31,32), subsequently lowering ATP sensitivity and causing channel activation. Alternatively, a localized pool of PtdIns(3,4,5)P 3 could act indirectly via its binding to protein targets other than those addressed in this report. Whatever the identity of the putative effector in the pathway for K ATP channel activation it must display a high degree of selectivity for PtdIns(3,4,5)P 3 over other phosphoinositides, because PtdIns(4,5)P 2 was ineffective in this system when PI 3-kinase was blocked. The majority of known targets for PtdIns(3,4,5)P 3 have a common structural theme, because they all possess one or more pleckstrin homology domains with a high degree of selectivity for PtdIns(3,4,5)P 3 over PtdIns(4,5)P 2 (14,50).
An important feature of lipid signals is that they are likely to remain for some time in the membrane in which they are synthesized. The possibility that PtdIns(3,4,5)P 3 might be generated in different subcellular compartments or membrane microdomains in response to different stimuli may help to explain why signaling by both insulin and leptin can be blocked by PI 3-kinase inhibitors yet these hormones accomplish quite different signaling outcomes. Insulin receptors are not thought to recruit PI 3-kinase directly but instead phosphorylate cytosolic insulin receptor substrate-proteins which in turn bind and activate PI 3-kinases via phosphotyrosine/Src homology 2 domain interactions. Although the leptin receptor, like many other cytokine receptors, is generally considered to signal via janus-tyrosine kinase 2 to various signal transducers and activators of transcription (2), there is evidence that some cytokine receptors (e.g. p55 tumor necrosis factor and epidermal growth factor receptors) can associate directly with phosphoinositide kinases (52). Perhaps leptin receptors can also recruit PI 3-ki-nases directly leading to a spatially limited production of PtdIns(3,4,5)P 3 and the selective activation of K ATP channels.
One final possibility to be considered is that the putative PI 3-kinase product elicited by leptin action that is responsible for K ATP channel activation is not PtdIns(3,4,5)P 3 itself but some other 3-phosphoinositide. Although this will ultimately require direct measurement of these lipids in radiolabeled cells, it is unlikely that phosphatidylinositol 3,4-bisphosphate accounts for these effects, because it is an effective activator of PKB (51,52), which was not significantly stimulated by leptin in CRI-G1 cells. Other possibilities include PtdIns3P, a product of type II and type III PI 3-kinases, and phosphatidylinositol 3,5,bisphosphate, which is produced by the consecutive phosphorylation of PtdIns by a 3-kinase and the resulting phosphatidylinositol 3-phosphate by a 5-kinase (53). These latter lipids are implicated in vesicle trafficking events and their production is sensitive to nM concentrations of wortmannin.
In conclusion, our results show that activation of K ATP channels by leptin likely requires PI 3-kinase activity. In support of such a mechanism, PtdIns(3,4,5)P 3 mimicked the effect of leptin by increasing K ATP channel activity in whole-cell and inside-out current recordings, whereas the effects of PtdIns(4,5)P 2 in these systems were blocked by inhibitors of PI 3-kinase. Surprisingly, however, leptin did not increase the total PtdIns(3,4,5)P 3 content nor did it activate any of several components of established PI 3-kinase signaling pathways. It is possible that either a localized pool of PtdIns(3,4,5)P 3 or some other 3-phosphoinositide might account for these observations. Cell labeling studies, which might allow the detection of alternative lipid mediators, and the use of intracellular probes to monitor the subcellular distribution of PtdIns(3,4,5)P 3 should help to resolve these issues in our future work.