Leucine culture reveals that ATP synthase functions as a fuel sensor in pancreatic beta-cells.

Our goal was to investigate whether leucine culture affects beta-cell glucose sensing. One-day culture of rat islets with 10 mM leucine had no effect on glucose-induced insulin secretion. One-week leucine culture decreased the threshold for glucose-induced insulin secretion and increased maximal insulin secretion at 30 mM glucose. Glucose-induced cytosolic free Ca(2+) was increased at 1 week but not at 1 day of leucine culture. Without glucose, ATP content was not different with or without leucine culture for 1 week. With 20 mM glucose, ATP content was higher by 1.5-fold in islets cultured for 1 week with leucine than those without leucine. Microarray experiments indicated that culture of RINm5F cells with leucine increased expression of ATP synthase beta subunit 3.2-fold, which was confirmed by real time reverse transcription-PCR analysis (3.0- +/- 0.4-fold) in rat islets at 1 week but not after 1 day with leucine culture. Down-regulation of ATP synthase beta subunit by siRNA decreased INS1 cell ATP content and insulin secretion with 20 mM glucose. Overexpression of ATP synthase beta subunit in INS1 cell increased insulin secretion in the presence of 5 and 20 mM glucose. In conclusion, one-week leucine culture of rat islets up-regulated ATP synthase and increased ATP content, which resulted in elevated [Ca(2+)] levels and more insulin exocytosis by glucose. Depletion of ATP synthase beta subunit with siRNA produced opposite effects. These data reveal the fuel-sensing role of mitochondrial ATP synthase in the control of ATP production from glucose and the control of glucose-induced insulin secretion.

Glucose is the main secretagogue of insulin secretion from pancreatic ␤-cells. The mechanisms of glucose-induced insulin secretion have been studied extensively. Through glycolysis and oxidation, glucose increases pancreatic ␤-cell ATP/ADP ratio, which closes ATP-sensitive potassium (K ATP ) channels and depolarizes the cell membrane. This results in an influx of extracellular Ca 2ϩ and increase of free cytosolic [Ca 2ϩ ] that stimulates exocytosis of insulin granules (1)(2)(3)(4)(5). Another mechanism is K ATP channel-independent and involves increased effectiveness of [Ca 2ϩ ] (6,7). In pancreatic ␤-cells, most of the intracellular ATP comes from the oxidation of glucose-derived pyruvate and oxidation of NADH in the mitochondria via the electron transport chain. Damage or inhibition of ATP synthesis results in ␤-cell dysfunction and impairs glucose-stimulated insulin secretion (8,9). Some recent publications indicate that superoxide produced by hyperglycemia activates uncouplingprotein-2 (UCP2) and destroys the proton gradient between inner and outer mitochondrial membranes. This negatively affects the activity of ATP synthase, decreases ATP production, and impairs glucose-stimulated insulin secretion of pancreatic ␤-cells resulting in diabetes (8,9). More recently, it is shown that mitochondrial metabolism sets the maximal limit of fuel stimulated-insulin secretion in ␤-cells (10). These findings imply that mitochondrial ATP synthesis may play a vital role in fuel-stimulated insulin secretion of pancreatic ␤-cells.
Some amino acids, particularly leucine and its non-metabolizable analogue 2-amino-2-norbornanecarboxylic acid, have been known to stimulate insulin secretion from pancreatic ␤-cells by activation of glutamate dehydrogenase (11)(12)(13)(14)(15). More recently, the branched-chain amino acids, including leucine, isoleucine, and valine, have been reported to activate the mammalian target of rapamycin (mTOR) 1 signaling pathway (16 -19) in ␤-cells. Leucine stimulates protein synthesis and pancreatic ␤-cell proliferation via the mTOR signaling pathway at physiological concentrations (16). These studies indicate a new role of branched-chain amino acids in pancreatic ␤-cell biology in addition to serving as fuels or residues for protein synthesis.
As the rate-limiting enzyme of glucose metabolism, it is believed that glucokinase sets a strict control on glucose metabolism in pancreatic ␤-cells. However, overexpression of glucokinase or hexokinase I fails to increase the maximal insulin output induced by glucose, although it decreases the threshold for glucose-induced insulin secretion in pancreatic ␤-cells (10, 20 -22). Thus, glucokinase may not be the only rate-limiting step in glucose-induced insulin secretion of pancreatic ␤-cells.
Other factor(s) may also be rate-limiting and contribute to the tight control of glucose-induced insulin secretion. As the key enzyme catalyzing the conversion of ADP and P i to ATP in the electron transport chain, ATP synthase (complex V) may play an important role in ATP synthesis and hence glucose-induced insulin secretion. Thus, the aim of this study was to investigate the role of ATP synthase in glucose-induced insulin secretion by attempting to change its expression level with leucine culture and siRNA.

EXPERIMENTAL PROCEDURES
Isolation and Culture of Rat Islets-Islets were isolated from 4 -6 male Sprague-Dawley rats (220 -300 g) by collagenase P digestion and * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  Ficoll gradient centrifugation as described in detail previously (14). Before culturing, the islets were washed at least three times with leucine-free RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 g/ml streptomycin, and 100 IU/ml penicillin in one sterile conical tube. Batches of 50 -100 islets were cultured in 10 ml of RPMI 1640 with or without 10 mM leucine at 37°C in 5% CO 2 , 95% air in a 6-cm dish for 1 day or 1 week. Medium was changed every three days.
Free Cytosolic [Ca 2ϩ ] Measurement-Islets were loaded in 3 ml of KRB-G0 with 2.0 M fura-2 AM and 0.2 mg/ml pluronic F-127 at 37°C in 5% CO 2 , 95% air for 30 min. Islets were placed in the center of a perifusion chamber and fixed onto the tip of a glass micropipet by slight suction. They were perifused with KRB at a flow rate of 1 ml/min at 37°C. The method for measuring [Ca 2ϩ ] of rat islets was described in detail previously (14). The perifusion protocol was as follows: 5 min with G0, G2, G5, G10, G30, and 30 mM KCl, respectively. Two to three islets of each condition were analyzed for each islet preparation. Data were collected from at least four islet preparations, and each preparation used four different rats. When calculating the increase of free cytosolic [Ca 2ϩ ] in each experiment, the average concentration of cytosolic [Ca 2ϩ ] between 0 -5 min under KRB-G0 was set as baseline. For a given condition, the absolute value of cytosolic [Ca 2ϩ ] was calculated as the mean concentration during the entire 5-min period of stimulation, whereas the net increase of [Ca 2ϩ ] was obtained by subtracting the baseline value from the absolute value.
Insulin Secretion and ATP Content Measurement in Incubated Islets-Islets were washed once with 10 ml of prewarmed KRB-G0 and then incubated in 10 ml of KRB-G0 at 37°C in 5% CO 2 , 95% air for 30 min. Islets were then divided into batches of five islets and transferred into tubes containing 1 ml of KRB with 0, 5, or 20 mM glucose. Islets were incubated at 37°C in 5% CO 2 , 95% air for 1 h in a shaking water bath. After incubation, 950 l of supernatant was removed for insulin radioimmunoassay. Lysis buffer (450 l) for ATP extraction (Roche Applied Science) was added into each tube containing islets. Islets were sonicated on ice for 40 pulses at 60% of maximal output. Sonicated samples were then frozen and thawed twice with dry ice to ensure complete lysis of cells. Islet lysate insulin content was assayed by radioimmunoassay. Islet lysate ATP content was measured using the ATP Bioluminescence Assay Kit HS II (Roche Applied Science). Fifty-l samples were measured in triplicate. Data were collected from at least 15 batches of islets in four independent experiments.
Quantitative Real Time RT-PCR assay-After treatments, islets were transferred to a 1.5-ml sterile tube and washed once with 1 ml of sterile diethyl pyrocarbonate-treated water. Total RNA was extracted from the islets using the RNeasy mini kit (Qiagen, Inc.), and agarose gel electrophoresis was used to analyze the quality of total RNA. Real-time RT-PCR was performed with One-step RT-PCR Master Mix reagents kit (TaqMan®). The cycling conditions for RT-PCR were as follows: 48°C for 30 min and 95°C for 10 min followed by 45 cycles of 95°C for 15 s and 60°C for 1 min. Quantitative values were obtained as threshold PCR cycle number (Ct) when the increase in fluorescent signal of PCR product showed exponential amplification. Target gene mRNA level was normalized to that of ␤-actin in the same sample. In brief, the relative expression level of the target gene compared with that of ␤-actin was calculated as 2 Ϫ⌬Ct , where ⌬Ct ϭ Ct target gene Ϫ Ct ␤-actin . The ratio of relative expression of the target gene in leucine-treated islets to that of untreated islets was then calculated as 2 Ϫ⌬⌬Ct , where ⌬⌬Ct ϭ ⌬Ct treated islet Ϫ ⌬Ct non-treated islet (23)(24)(25). Each sample was measured in duplicate for each experiment.
Down-regulation of ATP Synthase by siRNA Cassette-Four siRNA cassettes targeting rat ATP synthase ␤ subunit were designed and synthesized as PCR product (Genescript, Inc.). The target sequence of cassette A, which is under the control of human U6 promoter, is 5Ј-AGATGAGTGTTGAACAGGAAA-3Ј, and that of cassette B, which is under the control of human H1 promoter, is 5Ј-ATGGCACTGAAGGCT-TGGTTA-3Ј. The target sequence of cassette C is 5Ј-TGGTGGTGAT-TCTTTCCTGCA-3Ј under the control of U6 promoter and that of cassette D is 5Ј-CACCTGCTAAGATCTGCTGG-3Ј under control of H1 promoter. The day before transfection, the INS1 cells were seeded in 24-well plates with antibiotic-free RPMI 1640 (10% fetal bovine serum) so that the cells could grow more quickly in culture and reach 80% confluence the next day. For transfection of each well, 0.15 or 0.30 g of siRNA cassette DNA was diluted into 25 l of Opti-MEM I (serum-and antibiotic-free) and mixed gently. Then, 4 l of Plus TM Reagent (Invitrogen) was added and mixed gently. The mixture was incubated at room temperature for 15 min. At the same time, 1 l of Lipofectamine TM 2000 (Invitrogen) was added to 25 l of Opti-MEM I and mixed gently before the mixture was incubated at room temperature for 15 min. After incubation, the two mixtures were combined to generate DNA-Lipofectamine 2000 complex and incubated at room temperature for 15 min. During the incubation, the culture medium in 24-well plates was replaced with 200 l of Opti-MEM I. The DNA-Lipofectamine 2000 complex was added into the well, and the plate was rocked gently back and forth. Cells were then incubated at 37°C in 5% CO 2 -95% air for 3 h before 1 ml of complete RPMI 1640 was added into each well. The cells were cultured for 2 days.
The cells were washed twice with 1 ml of KRB-G0 before they were preincubated in 1 ml of KRB-G0 for 1 h. They were then incubated in 1 ml of KRB with 0 mM or 20 mM glucose for 1 h. All KRB was transferred into a 1.5-ml tube and centrifuged at 15,000 rpm at 4°C for 10 min. The supernatant was collected for insulin measurement by radioimmunoassay. Lysis buffer (200 l/well) (Roche Applied Science) was added, and the cells were scraped on ice. The cell lysate was transferred into a 1.5-ml tube and centrifuged at 4°C at 15,000 rpm for 15 min. The ATP content in the supernatant was assayed as above, and the data was normalized to its protein content. For real time RT-PCR, 48 h after transfection the cells were washed twice with 1 ml of ice-cold phosphate-buffered saline followed by extraction of total RNA. RT-PCR was performed as described above.
Cloning of ATP Synthase ␤ Subunit from INS1 Cells-Two specific primers, 5Ј-GAC GGG CCC AA ATG TTG AGT CTT GTG GGG CG T GTG-3Ј containing an ApaI site and 5Ј-GAC GCG GCC GCA TCA CGA CCC ATG CTC CTC TGC CAG-3Ј containing a NotI site, were synthesized. The full-length ATP synthase ␤ subunit cDNA (GenBank TM accession number P19044) of 1.6 kb was amplified using rat INS1 cell mRNA as a template and cloned into pShuttle vector with the restriction enzymes ApaI and NotI.
Overexpression of ATP Synthase ␤ Subunit in INS1 Cells-Fortyeight h after transfection, cells were washed once with cold phosphatebuffered saline. Total RNA was extracted from the cells, and real time RT-PCR was performed as above. For insulin secretion, INS1 cells were seeded in a 24-well plate and transfected with pShuttle-ATP-synthase-␤ subunit. 48 h after transfection, the cells were washed twice with 1 ml of prewarmed KRB and preincubated in 1 ml of KRB-G0 for 1 h. Cells were then incubated in 1 ml of KRB containing 0, 5, or 20 mM glucose for 1 h, respectively. The cells transfected with empty pShuttle vector were used as control.

Culturing Islets with Leucine Increases Glucose-induced
Insulin Secretion-To investigate whether leucine can affect glucose-induced insulin secretion, rat islets were cultured in complete RPMI 1640 (11 mM glucose, 10% fetal bovine serum) with or without 10 mM leucine for 1 day or 1 week. The time course of perifusion is shown in Fig. 1, and the averaged insulin secretion during each condition is presented in Fig. 2. After a 1-day culture, both untreated (Leu0) and leucine-treated (Leu10) islets had similar basal insulin secretion in the absence of glucose ( Fig. 2A). Thus, leucine treatment for 1 day failed to enhance glucose-induced insulin secretion in rat islets under all glucose concentrations tested and 30 mM KCl (Fig. 1A).
Culturing islets had a smaller Ca 2ϩ response and a smaller insulin secretion peak response to KCl than did fresh islets (Figs. 1A and 3A). Therefore, there was no significant difference in the glucose-induced change of free cytosolic [Ca 2ϩ ] between 1-daytreated islets and untreated islets.
Basal [Ca 2ϩ ] without glucose was the same in islets cultured for 1 week with (72 Ϯ 4 nM) or without (68 Ϯ 6 nM) 10 mM leucine (p Ͼ 0.05). With 2 mM glucose, free cytosolic [Ca 2ϩ ] of treated islets began to increase by 3 Ϯ 1 nM (p Ͻ 0.05) compared with basal level, whereas that of untreated islets was the same as baseline (68 Ϯ 6 nM) (Fig. 4B). With 5 mM glucose, free cytosolic [Ca 2ϩ ] of treated islets increased by 16 Ϯ 4 nM (p Ͻ 0.05) when compared with baseline, whereas that of untreated islets was still the same as baseline (Fig. 4B). Glucose at 10 mM increased free cytosolic [Ca 2ϩ ] by 33 Ϯ 6 nM in treated islets but only 3 Ϯ 1 nM in untreated islets. Glucose at 30 mM increased free cytosolic [Ca 2ϩ ] by 47 Ϯ 6 nM (p Ͻ 0.05) in treated islets but only by 9 Ϯ 1 nM in untreated islets. Treated islets also had a significantly larger increase of [Ca 2ϩ ] (80 Ϯ 7 nM) than did untreated islets (33 Ϯ 2 nM) when stimulated with 30 mM KCl compared with baseline (Figs. 3B and 4B). Therefore, there was a significant elevation of glucose-induced free cytosolic [Ca 2ϩ ] in 1-week-treated islets compared with untreated islets (Figs. 3, B and C, and 4B). The dose-response curve of glucose-induced cytosolic Ca 2ϩ rise was shifted to the left by leucine treatment.
Leucine Culture Increases ATP Synthase ␤ Subunit Expression-Most of the ATP in pancreatic ␤-cells comes from oxidation of NADH in the mitochondria by the electron transport chain. ATP is synthesized directly by complex V (ATP synthase). Our preliminary microarray experiments show that treatment with leucine up-regulated ATP synthase ␤ subunit by 3.2-fold in the RINm5F cell line. In the present study, we focused on analyzing ATP synthase (␤ chain) expression by real time RT-PCR. Our results showed that in 1-day-treated islets, the ATP synthase ␤ subunit mRNA level was 1.4-Ϯ 0.2-fold higher than that of untreated islets, which was not statistically significant (p Ͼ 0.05, Fig. 8A). In 1-week-treated islets, the ATP synthase ␤ subunit mRNA level was 3.0-Ϯ 0.2-fold (p Ͻ 0.01) higher than that of untreated islets. Our results also showed that culture with 10 mM D-leucine for 1 week failed to affect the ATP synthase ␤ subunit mRNA level in rat islets (Fig. 8B). To test if leucine affects gene expression via the mTOR signaling pathway, rat islets were cultured with 10 nM rapamycin or 10 nM rapamycin plus 10 mM leucine for 1 week. The results showed that 10 nM rapamycin alone had no significant effect on ATP synthase ␤ subunit mRNA level. Rapamycin failed to block the leucine effects on gene expression when tested at 10 nM (Fig. 8B) or 100 nM (data not shown).

FIG. 5. Effects of 1-day leucine culture on glucose-induced ATP content and insulin secretion of rat islets.
Batches of five cultured islets were incubated in 10 ml KRB-G0 for 30 min and then divided into different tubes containing 1 ml of KRB with 0 mM, 5 mM, and 20 mM glucose, respectively. The tubes were incubated at 37°C in 5% CO 2 , 95% air for 1 h in a shaking bath. After incubation, the supernatant was taken for insulin measurement (B). ATP was extracted from the islets and measured as described under "Experimental Procedures" (A) in duplicate or triplicate. The data are mean Ϯ S.E. of more than 12 batches from four independent experiments using four different rats.

Down-regulation of ATP Synthase ␤ Subunit Using siRNA
Cassettes-Real time RT-PCR analysis showed that cassette A reduced significantly (p Ͻ 0.05) the ratio of Ct value of ATP synthase ␤ subunit mRNA level by about 20% when compared with control. The effect of cassette B was not significant (p Ͼ 0.05). The transfection efficiency of GFP-containing plasmid in INS1 cells was about 30 -40%. Assuming that the transfection of siRNA cassette DNA had similar efficiency and that the ATP synthase ␤ subunit was decreased by 50% in the transfected cells, the difference of Ct values between these cells and the control cells would not be more than 1. It is difficult to detect such small differences by real time RT-PCR. At 48 h post transfection, cassette A reduced glucose-induced ATP content significantly (p Ͻ 0.05) by 42 and 48% at 0.15 and 0.30 g/well, respectively (Fig. 9B). Cassette B reduced glucose-induced ATP content significantly (p Ͻ 0.05) by 48 and 43% at 0.15 and 0.30 g/well, respectively (Fig. 9B). The siRNA cassette effects on ATP content also seemed to reach a maximum at 0.15 g/well. In the same cells, siRNA cassette A reduced 20 mM glucoseinduced insulin secretion significantly (p Ͻ 0.05) by 27 and 24% at 0.15 and 0.30 g/well, respectively (Fig. 9A). Cassette B reduced glucose-induced insulin secretion significantly by 27 and 31% at doses of 0.15 and 0.30 g/well, respectively. The effect seemed to reach a maximum at 0.15 g/well. The effects on ATP content correlated well with those on insulin secretion. Cassettes C and D did not cause significant changes in ATP content and insulin secretion.

Role of ␤-Cell Mitochondria in Fuel Sensing-Pancreatic
␤-cells secrete insulin in response to glucose through metabolism and production of ATP. Though glycolysis produces small amounts of ATP, the vast majority of ATP is produced through the mitochondrial oxidation of pyruvate, which is produced as the glycolytic product of glucose. The increase in ATP/ADP ratio closes ATP-sensitive potassium channels and causes membrane depolarization. This opens voltage-dependent calcium channels to allow Ca 2ϩ influx. The increase of cytosolic Ca 2ϩ triggers exocytosis of insulin granules. The amount of secreted insulin is controlled tightly by the glucose concentration. The mechanisms of this control (or glucose sensing) have been the subject of extensive studies (1-7). It is now believed that the high K m hexokinase IV (glucokinase) is the ratelimiting enzyme for glucose phosphorylation and glucose metabolism in the ␤-cell. However, the sensors for other fuels, such as amino acids, are still unclear. It has recently been shown that a large increase of glycolytic intermediates that bypass the restriction of glucokinase does not affect insulin secretion, whereas mitochondrial membrane hyperpolarization strongly correlates with limitations of insulin secretion (10). It has been suspected that factors downstream of glucokinase play important roles in the restriction of metabolism and insulin secretion in the presence of large amounts of glycolytic intermediates. It is suggested that limitations imposed by the mitochondrial membrane potential are likely related to the thermodynamics of generation of the proton gradient across the mitochondrial inner membrane, which are created by the mitochondrial electron transport chain complex. The limitations of proton gradient formation thus limit ATP production.

FIG. 6. Effects of 1-week leucine culture on glucose-induced ATP content and insulin secretion in rat islets.
Batches of five islets were washed once with prewarmed KRB-G0, and the washed islets were incubated in 10 ml of KRB-G0 for 30 min and then divided into different tubes containing 1 ml of KRB with 0 mM, 5 mM, and 20 mM glucose, respectively. The islets were incubated at 37°C in 5% CO 2 , 95% air for 1 h in a shaking bath. The supernatant was taken for insulin measurement (B). ATP was extracted from the islets and measured as described in detail under "Experimental Procedures" (A) in duplicate or triplicate. The insulin content in cell lysate was measured (C). Fractional insulin secretion (D) was calculated by dividing secreted insulin (B) by insulin content (C). The data are mean Ϯ S.E. of more than 18 batches of islets from five independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01 when compared with untreated control.
It has been shown that inhibitors of the electron transport chain and ATP production all impair insulin secretion (10, 26 -30). An increase in mitochondrial metabolism by a Na ϩ -Ca 2ϩ exchanger inhibitor also potentiates glucose-induced insulin secretion in rat islets (31). It is thus proposed that the mitochondrial proton gradient sets the limit of insulin secretion.
For the first time, our study demonstrates that culture of pancreatic ␤-cells with leucine increases their glucose sensitivity as demonstrated by the leftward shift in the threshold of insulin secretion and of cytosolic Ca 2ϩ rise and by the increase in the maximal response of insulin secretion and of cytosolic Ca 2ϩ rise. Our findings contradict those reported by Anello et al. (32) that 24-h culture with 20 mM leucine impairs glucoseinduced insulin secretion and increases ADP level in rat islets. Their findings also show a decrease in ATP/ADP ratio. The lack of changes in the ATP level and glucose utilization and oxidation in their study is difficult to explain. However, we did not detect any changes in glucose-induced insulin secretion, cytosolic Ca 2ϩ , or ATP level after 24 h culture with leucine in rat islets. In contrast, changes in glucose effects on insulin secretion, cytosolic Ca 2ϩ , and ATP content were detected after 1 week of culture with leucine. The insulin secretion induced by high potassium immediately after glucose stimulation was also increased dramatically. This may indicate the accumulation of large amounts of ATP and glucose metabolites inside the cells, which amplifies the effects of depolarization.
To pinpoint the mechanisms for the change in glucose sensitivity, a series of experiments was performed in our study to test various steps of the glucose signal transduction pathways. In agreement with the changes in glucose-induced insulin se-cretion in leucine treated islets, cytosolic [Ca 2ϩ ] response to glucose was also increased after treatment for 1 week but not for 1 day. The threshold was decreased from 10 mM to 2 mM glucose (Fig. 3C). The maximal [Ca 2ϩ ] response in the presence of 30 mM glucose was increased dramatically. Thus, the alteration of the signal transduction is located more upstream of Ca 2ϩ influx or Ca 2ϩ channels. Islet ATP was thus measured in the presence and absence of glucose with or without leucine treatment. Leucine culture for 1 day did not affect ATP levels; however, after 1 week, ATP levels were increased (Fig. 6) suggesting that leucine culture can regulate glucose metabolism and/or the mitochondrial electron transport chain. Numerous enzymes and regulatory proteins are involved in the regulation of glucose metabolism and mitochondrial functions, making investigation of all of them in one study impractical. Instead, a genome-wide screening of changes in gene expression has been used. Our microarray experiments demonstrate that the mitochondrial ATP synthase ␤ subunit was up-regulated by 3.2-fold in leucine-treated RINm5F cells. Changes of other metabolic enzymes have not been detected in these experiments. This enzyme is also called mitochondrial electron transport chain complex V and plays an important role in the production of ATP (33,34). Energy from the proton gradient is used to generate ATP as the collapse of the gradient through ATP synthase drives the ATP synthetic machinery (8,34). Though the screening experiments were first performed in RINm5F cells, the increase in the ATP synthase ␤ subunit mRNA level was subsequently confirmed in rat islets. Rat islet ATP synthase ␤ subunit expression was increased by leucine treatment for 1 week but not for 1 day, which is in agreement with insulin secretion, cytosolic Ca 2ϩ , and ATP level measure-  (Leu30min, striped bars). All four groups of islets were washed with KRB and pre-incubated in 10 ml of KRB-G0 for 30 min as described above. ATP content (A) and insulin secretion (B) were measured as described above. The data are mean Ϯ S.E. of more than 12 batches of four independent experiments using four different rats. *, p Ͻ 0.05 when compared with the untreated group; #, p Ͻ 0.05 when compared with D-leucine treated group and 30-min leucine-treated group before experiment.

FIG. 8. Real time RT-PCR assays of ATP synthase gene.
Batches of 80 -100 islets were cultured in RPMI 1640 with indicated conditions for 1 day or 1 week. Total RNA was extracted and RT-PCR was performed as described in the methods procedure. A, effect of one-day and one-week leucine culture on ATP synthase mRNA level in rat islets. *, p Ͻ 0.01 when compared with the control value; #, p Ͻ 0.05 when compared with 1 day's value. B, islets were cultured without leucine, with 10 nM rapamycin, with 10 mM D-Leucine, with 10 mM leucine, and with 10 mM leucine plus 10 nM rapamycin for 1 week, respectively. Total RNA was extracted, and RT-PCR was performed. *, p Ͻ 0.01 when compared with the control; &, p Ͻ 0.05 when compared with the Dleucine value. Data are mean Ϯ S.E. of at least six independent experiments, and each used six rats. ments. Our results thus establish the important role of mitochondrial regulation of ATP production in glucose-induced insulin secretion.
The data of leucine culture are consistent with the hypothesis that ATP synthase functions as a rate-limiting step in glucose sensing. However, leucine culture may also change the expression of other proteins and their functions, which may also contribute to the changes of glucose sensitivity. Thus we used siRNA technique to specifically down-regulate ATP synthase. Two of four siRNA cassettes caused ϳ30% decrease of 20 mM glucose-induced insulin secretion in INS1 cells without affecting the basal insulin secretion without glucose. ATP content in the presence of 20 mM glucose was decreased by 40 -50%, but ATP content in the absence of glucose was not affected. The ATP synthase ␤ subunit mRNA level was decreased significantly by siRNA cassette A, but non-significantly by cassette B. This is mainly because of the low sensitivity of real time RT-PCR technique. Overexpression of ATP synthase ␤ subunit alone using plasmid also increased glucose induced insulin secretion in INS1 cells. These data thus confirm the rate-limiting feature of ATP synthase in ␤ cell glucose sensing.
Type-2 Diabetes and Mitochondrial Function-Type-2 diabetes is characterized by hyperglycemia, insulin resistance, and impaired glucose-stimulated insulin secretion from pancreatic ␤-cells. The exact mechanisms of the pathogenesis of type-2 diabetes are still unclear and likely involve many different genes. As the rate-limiting enzyme of glucose metabolism, glucokinase is considered the glucose sensor in pancreatic ␤-cells. Loss of function mutations of glucokinase causes diabetes (35,36). However, recent studies demonstrated that the metabolic impairment in type-2 diabetics may also involve other enzymes. It has been shown that NADH-O 2 oxidoreductase activity and citrate synthase activity of skeletal muscle are lower in type-2 diabetics than lean subjects and non-diabetic obese subjects (37). The same study demonstrates that the skeletal muscle mitochondria are smaller in type-2 diabetic and obese subjects than those in muscle from lean volunteers (37). It has been shown that type-2 diabetic subjects have decreased oxidative enzyme activity and increased lipid content in skeletal muscles (38). A recent study (39) using a proteomics approach has identified eight potential protein markers for type-2 diabetes in human skeletal muscle biopsies in the fasting state. The observed changes in protein expression indicate increased cellular stress (up-regulation of two heat shock proteins) and perturbations in ATP synthesis and mitochondrial metabolisms (down-regulation of ATP synthase ␤ subunit and creatine kinase B). It is demonstrated that the catalytic ␤ subunit of ATP synthase is phosphorylated in vivo and that the levels of a down-regulated ATP synthase ␤ subunit phosphoisoform in diabetic muscle correlated inversely with fasting plasma glucose levels. These data suggest that ATP synthase ␤ subunit plays a role in the regulation of ATP synthesis and that alterations in the regulation of ATP synthesis and cellular stress proteins may contribute to the pathogenesis of type-2 diabetes (39). The human gene is identical to the rat gene (P19044) under investigation in this study except for three residues. If it is true that the decrease of ATP synthase protein level and the decrease of phosphorylation level of this enzyme are related to the development of type-2 diabetes (39), the same enzyme in the pancreatic ␤-cells of type-2 diabetic human may have similar defects, which may result in impaired glucose-induced insulin secretion. A recent study by Deng et al. shows that the threshold of glucose-induced insulin secretion is higher in the islets of type-2 diabetic human islets compared with non-diabetic islets. The amount of insulin secretion during a ramp of glucose stimulation is much smaller in diabetic than non-diabetic islets (40).
Target(s) of Leucine-Recently, the role of amino acids in regulating protein translation by metabolism-linked signaling pathways rather than their well known role in serving as precursors for protein synthesis has become the subject of extensive studies (16,18,19,41). Leucine has been reported to mediate the phosphorylation of eukaryotic initiation factor 4-Ebinding protein-1 (eIE-E-BP1), which is also designated as phosphorylated heat-and acid-stable protein regulated by insulin (PHAS-1). Leucine also causes the phosphorylation of p70S6 kinase (p70 s6k ) through mTOR signaling pathway in a FIG. 9. Down-regulation of ATP synthase gene by siRNA cassette in INS1 cells. For insulin secretion and ATP content measurement, INS1 cells were transfected with a siRNA cassette at 0.15 or 0.3 g/well in 24-well plates. Two days after transfection, the cells were washed twice with 1 ml of KRB-G0 and preincubated in 1 ml of KRB-G0 for 1 h before being incubated in 1 ml of KRB with 0 or 20 mM glucose for 1 h. The supernatant was collected for insulin secretion (A), and the cells were lysed. The ATP content in the supernatant was assayed and the data was normalized to its protein content (B). *, p Ͻ 0.05 when compared with control. For real time RT-PCR, 48 h after transfection the cells were washed twice with ice-cold phosphate-buffered saline, and total RNA was extracted. RT-PCR was performed as described under "Experimental Procedures." *, p Ͻ 0.05; **, p Ͻ 0.01 when compared with control, which was treated without siRNA cassette. Data are mean Ϯ S.E. of at least three independent experiments with more than nine measurements. variety of cell lines and primary cells (18,19). Leucine enhances protein synthesis via the mTOR signaling pathway in pancreatic ␤-cells. Some important transcriptional regulators might be among those affected proteins. We thus propose that leucine might regulate ␤-cell glucose sensitivity by regulating some key metabolic genes involved in glucose metabolism and ATP synthesis. A genome-wide screening of 40,000 genes in RINm5F cells using microarray analysis showed that leucine culture up-regulated ATP synthase at the mRNA level. This is a key enzyme controlling ATP synthesis within the mitochondria. ATP synthase might not be the only gene affected by leucine. The direct target of leucine is still unknown. However, its effects are structure-dependent, because D-leucine has no effect. Presumably, leucine may interact with one or multiple transcription factors or regulatory protein(s) to affect the transcription of multiple genes.
To preclude the possibility that the increase of ATP content and insulin secretion in leucine-treated islets resulted from the fuel effects of leucine, the islets were cultured for 1 week without leucine and then treated with 10 mM leucine for 30 min before experiments. Then, ATP content and insulin secretion were assayed. The experimental results showed that a short term leucine treatment did not affect glucose sensing and thus exclude the fuel effects of leucine (Fig. 7).
In conclusion, leucine culture up-regulates ATP synthase ␤ subunit and ATP production in a time-dependent manner. Mitochondrial ATP synthesis is under the control of ATP synthase and plays a vital role in glucose-induced insulin secretion in pancreatic ␤-cells.