Specific amino acids are now known to acutely and chronically regulate insulin secretion from pancreatic β-cells in vivo and in vitro. Understanding the molecular mechanisms by which amino acids regulate insulin secretion may identify novel targets for future diabetes therapies. Mitochondrial metabolism is crucial for the coupling of amino acid and glucose recognition to the exocytosis of the insulin granules. This is illustrated by in vitro and in vivo observations discussed in the present review. Mitochondria generate ATP, which is the main coupling factor in insulin secretion; however, the subsequent Ca2+ signal in the cytosol is necessary, but not sufficient, for full development of sustained insulin secretion. Hence mitochondria generate ATP and other coupling factors serving as fuel sensors for the control of the exocytotic process. Numerous studies have sought to identify the factors that mediate the amplifying pathway over the Ca2+ signal in nutrient-stimulated insulin secretion. Predominantly, these factors are nucleotides (GTP, ATP, cAMP and NADPH), although metabolites have also been proposed, such as long-chain acyl-CoA derivatives and the key amino acid glutamate. This scenario highlights further the importance of the key enzymes or transporters, glutamate dehydrogenase, the aspartate and alanine aminotransferases and the malate/aspartate shuttle, in the control of insulin secretion. Therefore amino acids may play a direct or indirect (via generation of putative messengers of mitochondrial origin) role in insulin secretion.
- amino acids
- gene expression
- signal transduction
Amino acid metabolism, under appropriate conditions, is known to enhance insulin secretion from primary islet cells and β-cell lines [1–5]. In vivo, L-glutamine and L-alanine are quantitatively the most abundant amino acids in the blood and extracellular fluids followed closely by the branched-chain amino acids . The positive effect of two amino acids, leucine plus phenylalanine, on insulin secretion was reported in a recent clinical assessment of the effect of leucine and phenylalanine administered in the presence of a protein hydrolysate to Type II diabetic patients and suitable controls, which resulted in a 3-fold increase in insulin secretion compared with carbohydrate alone . As in vivo insulin secretion is normally determined by administration of an oral or intravenous glucose load, it is probable that in vivo insulin secretion measurements are an underestimate of that possible from a mixed nutritional load. From in vitro studies, L-alanine metabolism is known to enhance insulin secretion alone or synergistically enhance glucose-stimulated insulin secretion from the clonal β-cell line BRIN-BD11 [8,9]; however, L-glutamine, although readily consumed by β-cells, does not stimulate insulin secretion in the absence of allosteric activation of GDH (glutamate dehydrogenase). L-Leucine may enhance L-glutamine-stimulated insulin secretion by acting as a GDH allosteric activator . Alternatively, in islets overexpressing GDH, glutamine enhances both mitochondrial metabolism and insulin secretion . Although many investigators have used high concentrations of individual amino acids in vitro, a study by Bolea et al.  has demonstrated that specific amino acid mixtures at physiological concentrations potently stimulate insulin secretion. Four amino acids were found to be particularly important for stimulating β-cell electrical activity, essential for insulin secretion: leucine, isoleucine, alanine and arginine.
Thus a relatively small number of amino acids promote or synergistically enhance insulin release from pancreatic β-cells [13,14]. The mechanisms by which amino acids enhance insulin secretion are varied. The cationically charged amino acid L-arginine does so by direct depolarization of the plasma membrane at neutral pH, but only in the presence of glucose. Other amino acids, which are co-transported with Na+, can also depolarize the cell membrane as a consequence of Na+ transport and thus induce insulin secretion by activating voltage-dependent Ca2+ channels. Metabolism of amino acids, resulting in partial oxidation, e.g. L-alanine , may initially increase the cellular content of ATP, leading to closure of the ATP-sensitive K+ channel, depolarization of the plasma membrane, activation of voltage-activated Ca2+ channels, Ca2+ influx and insulin exocytosis. However, it is possible to additionally stimulate insulin secretion via allosteric effects on regulatory proteins such as GDH (as described for leucine) . In general, metabolism of key amino acids appears to be necessary for appropriate generation of regulatory ‘signals’ which then impact upon insulin secretion. The identity of these regulatory signals has remained elusive, but evidence suggests that mitochondrial signals are important, including generation of ATP, citrate and glutamate . The potential roles of some key amino acids in the pancreatic β-cell are summarized in Figure 1.
STIMULUS–SECRETION COUPLING CIRCUITRY IN THE β-CELL
In vivo, the β-cell is constantly monitoring nutrient availability and metabolic status and can generate appropriate secondary stimulus-coupling signals in response to the most minor changes in the concentration of specific metabolites. This is coupled with regulatory input from other signalling pathways, including neuropeptides. The β-cell is metabolically distinct from almost all other mammalian cell types in several respects: (i) it can utilize glucose in the physiologically relevant range (2–20 mM) as it expresses a combination of GLUT-2 (high Km glucose transporter) and glucokinase, (ii) it possesses low lactate dehydrogenase and plasma membrane mono-carboxylate pyruvate/lactate transporter activity and correspondingly high activity in the mitochondrial malate/aspartate shuttle, so ensuring mitochondrial oxidation of NADH, (iii) it has a high activity of both pyruvate dehydrogenase and pyruvate carboxylase, ensuring that anaplerotic and oxidative metabolism of glucose/pyruvate can co-exist. All these specific metabolic adaptations are geared to enhancing mitochondrial TCA (tricarboxylic acid) cycle activity, oxidative phosphorylation and efficient ATP production. An enhancement of the ATP/ADP ratio results in closure of the ATP-sensitive K+ channel, depolarization of the plasma membrane, opening of voltage-activated Ca2+ channels, influx of Ca2+ and finally fusion of insulin-containing granules with the plasma membrane. The opening of Ca2+ channels is intermittent, oscillating with the membrane potential, and therefore results in oscillations of [Ca2+]i (intracellular Ca2+ concentration) that, in turn, trigger oscillations of insulin secretion .
In addition to this metabolic complexity, the β-cell can metabolize a number of key amino acids which, via mitochondrial metabolism, can generate further stimulus–secretion ‘coupling’ factors. Lipid metabolism, via long-chain acyl-CoA formation, may also impact upon insulin secretion . Indeed, it was reported that citrate exported from the mitochondria to the cytosol forms malonyl-CoA with CoA, promoting the accumulation of long-chain acyl-CoAs , thereby enhancing Ca2+-evoked insulin exocytosis . Amino acids also play a role as modulators of lipid metabolism. Acetyl-CoA carboxylase, responsible for malonyl-CoA synthesis, is activated by glutamate-sensitive PP2A (protein phosphatase 2A) , an effect demonstrated in islet β-cells . Recently, we have demonstrated that addition of 10 mM L-alanine to the BRIN-BD11 β-cell line increased expression of ATP-citrate lyase by 2-fold (G. A. Cunningham, N. H. McClenaghan, P. R. Flatt and P. Newsholme, unpublished work). ATP citrate lyase will convert citrate into acetyl-CoA in the cytosol, thus providing the key step in fatty acid synthesis, acetyl-CoA carboxylase, with substrate.
In the mitochondrial matrix, Ca2+ increases the activity of several dehydrogenases. In this manner, increased cytosolic Ca2+ occurring during cell activation is relayed to the mitochondria via a Ca2+ uniporter . Such Ca2+ entry is favoured by activation of the respiratory chain, for instance by glucose in the β-cell. Therefore hyperpolarization of the mitochondrial membrane permits the rise in mitochondrial Ca2+ further activating NADH-generating dehydrogenases . The primary actions of glucose are via potentiation of ATP concentration by enhanced TCA cycle substrate (oxidative and anaplerotic) supply. Generation of other additive factors derived from glucose metabolism might also be promoted by mitochondrial Ca2+ elevation .
SIGNALLING ROLE OF AMINO ACIDS
Certain amino acids are now known to play important nutrient-sensing roles involving the mTOR (mammalian target of rapamycin)-mediated signalling pathway . mTOR is a component of a signalling pathway which couples insulin receptor stimulation and nutrient availability with protein synthesis via activation and phosphorylation of the ribosomal protein S6. This is accomplished via activation of the protein kinase p70S6K, which is sensitive to both mTOR and insulin signalling pathways. Leucine is the most effective amino acid in this regard. The activation of the mTOR pathway is probably important in the β-cell where mTOR and insulin signalling are likely to synergize to stimulate both protein synthesis and insulin secretion [26,27]. It is not known how amino acids activate mTOR, but it is probable that stimulation of a kinase or inhibition of a phosphatase that act upon mTOR as a substrate is involved [25,28].
EFFECT OF AMINO ACIDS ON GENE EXPRESSION IN THE β-CELL
Although glucose and fatty acids are now known to regulate insulin secretion and β-cell integrity via changes in gene expression, the role of amino acids has, up to now, been neglected. However, we have recently demonstrated that key genes in the areas of signal transduction, metabolism and apoptosis are regulated by alanine or glutamine (Figure 2). Analysis performed using the Affymetrix rat genome RGU34A microarray revealed that a total of 66 genes were increased 1.8-fold or greater after 24 h culture with L-alanine, and a total of 148 genes were increased 1.8-fold or greater after 24 h culture with 10 mM L-glutamine. These altered genes are grouped according to molecular function in Figure 2. Thus, although the latter two amino acids may acutely regulate insulin secretion as described below, they also play a role in regulating β-cell gene expression, which will impact upon the ability of the β-cell to chronically respond to nutrient availability, metabolism, insulin secretion and functional integrity.
MECHANISMS OF AMINO-ACID-DEPENDENT STIMULATION OF INSULIN SECRETION
In a recent study, we  reported that both rat islets and BRIN-BD11 β-cells consumed L-glutamine at high rates. Islets may have a high rate of protein turnover even under basal conditions, which would require L-glutamine for purine and pyrimidine synthesis, subsequent mRNA production and, in addition, protein synthesis. Despite the fact that L-glutamine is rapidly taken up and metabolized by islets, it alone does not stimulate insulin secretion or potentiate glucose-induced insulin secretion . However, it enhances L-leucine-induced insulin secretion . This action has been attributed to activation of GDH by L-leucine, which leads to an increased entry of L-glutamine into the TCA cycle and subsequent oxidation. Matschinsky and co-workers  have shown that glucose inhibits glutaminolysis in β-cells and, as a result, blocks leucine-stimulated insulin secretion (see Figure 3 for details).
It has been reported that L-glutamine is converted into GABA (γ-aminobutyric acid) in islets . A recent study  has suggested the conversion of glutamine into GABA as a means to explain the paradox that L-glutamine alone does not stimulate insulin release. Fernandez-Pascual et al.  have shown that L-glutamine is metabolized preferentially to GABA and L-aspartate in islets and that, in the presence of L-leucine, there is an increased metabolism of L-glutamate and GABA. The production of 14CO2 from L-[U-14C]glutamine was attributed mainly to the formation of GABA, and the accumulation of L-aspartate was suggested to be as a result of oxaloacetate transamination with glutamate where the oxaloacetate was produced by the GABA shunt pathway. Under this scheme there is no oxidation of L-glutamine via the TCA cycle and the authors suggest  that this may explain the lack of its ability to induce insulin secretion. GABA released from pancreatic β-cells may regulate pancreatic α-cell glucagon release via GABAA receptor occupation, glucagon and glutamate release and subsequent glutamate receptor (AMPA) activation on β-cells, as suggested by Moriyama and Hayashi in a recent review .
Using 13C-labelled glutamine and NMR spectroscopy, we  have shown that the major products of L-[1,213C]glutamine metabolism are L-[1,213C]glutamate and L-aspartate labelled at positions C1 and C4 in BRIN-BD11 cells. L-Aspartate is formed after entry of L-glutamate into the TCA cycle. Additionally, L-glutamate produced from glutamine entered the γ-glutamyl cycle and resulted in an increased production of glutathione. There was no NMR-based evidence for the production of GABA in our experimental conditions. Addition of glucose caused an increase in glutamate concentration but no increase in glutamine consumption, agreeing with previous reports that glucose can inhibit glutaminolysis. Using 13C-isotopomer analysis of anaplerotic flux in INS-1 cells, Cline and co-workers  have demonstrated that glutamine addition increased flux through GDH, but this was not correlated with insulin secretion.
Recently a signalling role for glutamine in insulin secretion has been proposed . Using normal islets Li et al.  used a potential glutamine synthetase inhibitor MSO (methionine sulphoximine) to investigate the role of intracellularly generated glutamine in insulin secretion. They reported that, in the presence of this inhibitor, the insulin released in response to a glucose ramp was abolished and that this inhibition could be reversed by addition of L-glutamine or its non-metabolizable analogue 6-diazo-5-oxo-L-norleucine. However, caution should be exercised when interpreting the results from this study as MSO is not a specific glutamine synthetase inhibitor and, indeed, inhibits a number of glutamate-utilizing enzymes, including γ-glutamylcysteine synthetase, thus blocking potential metabolism of glucose-derived glutamate via the γ-glutamyl cycle. Although this issue was partially addressed by use of BSO (buthionine sulphoximine), an inhibitor of γ-glutamylcysteine synthetase, the mechanisms involved in glutamine-based ‘signalling’ have yet to be identified.
To date, the role of L-glutamate in the stimulation of insulin secretion is still controversial. Intracellular generation of L-glutamate has been proposed to participate in nutrient-induced stimulus–secretion coupling, as an additive factor in the amplifying pathway of glucose-stimulated insulin secretion . During glucose stimulation, total cellular glutamate levels have been reported to increase in human, mouse and rat islets as well as in clonal β-cells [8,37–40], whereas, in other studies, no change was detected [41,42]. The finding that mitochondrial activation in permeabilized β-cells directly stimulates insulin exocytosis  pioneered the identification of glutamate as a putative intracellular messenger [37,43]. Possibly, glutamate could be transported into secretory granules, thereby promoting Ca2+-dependent exocytosis [37,43]. Such a model has been substantiated by the demonstration that clonal β-cells express vesicular glutamate transporters and that glutamate transport characteristics are similar to neuronal transporters . The mechanisms by which glutamate promotes insulin exocytosis might involve pH changes in the secretory vesicles . In order to challenge the glutamate hypothesis, we  overexpressed GAD (glutamate decarboxylase) in INS-1E cells and rat islets to reduce cytosolic glutamate levels. GAD will catalyse the conversion of glutamate into GABA, which plays an important paracrine role in the islet, as described above. In control cells, glucose (15 mM) caused a 2.3-fold rise in glutamate concentration, whereas GAD overexpression resulted in a significant reduction in the glutamate response. GAD overexpression also reduced secretory responses to high glucose in INS-1E cells as well as rat islets.
Others have robustly challenged the role of glutamate in insulin secretion [39,42]. MacDonald and Fahien  did not observe an increase in intracellular glutamate concentration on addition of glucose (16.7 mM) in rat islets. Incubation with L-glutamine (10 mM) increased the glutamate concentration 10-fold but did not stimulate insulin release, leading the authors to cast doubt on the proposed role of L-glutamate. In a separate study, Bertrand et al.  demonstrated that, on incubation with glucose, a significant increase in glutamate concentration occurred in depolarized mouse and rat islets. However, they argued against the glutamate hypothesis on the basis of experiments with L-glutamine: L-glutamine caused an increase in glutamate content with no effect on insulin secretion. Additionally in this study, activation of GDH by BCH (2-amino 2-norbornane carboxylate) lowered glutamate levels but increased insulin secretion. However, L-glutamine as a precursor for glutamate may lead to saturating concentrations of glutamate without activation of the K+ATP-dependent pathway and thus not result in an increase in insulin secretion .
L-Alanine is consumed at high rates in both BRIN-BD11 cells and rat islets . Addition of 16.7 mM glucose significantly enhanced L-alanine consumption in both BRIN-BD11 β-cells and primary islet cells, suggesting a critical role for L-alanine in β-cell function. Remarkably, L-alanine consumption by both BRIN-BD11 cells and primary rat islets was similar to that of D-glucose (Table 1).
Early studies have shown that L-alanine is taken up and oxidized by ob/ob mouse islets . Recently, L-alanine has been shown to have insulinotropic effects both in cell lines and in rat islets [9,14]. Dunne et al.  suggested that in RINm5F cells the insulinotropic action of L-alanine was due to co-transport with Na+, which resulted in membrane depolarization and led to the generation of Ca2+ spike potentials and an increase in intracellular Ca2+. McClenaghan et al. , using the pancreatic β-cell line BRIN-BD11, demonstrated that L-alanine influenced glucose-induced insulin secretion by electrogenic Na+ transport. More recently, L-alanine was shown to be metabolized by BRIN-BD11 cells using 13C NMR  to give end-products that included glutamate, aspartate and lactate. Additionally, by use of the respiratory poison oligomycin, the metabolism of alanine was shown to be important for its ability to stimulate insulin secretion . It is indeed likely that a combination of co-transport with Na+ and ATP generation via metabolism results in the insulinotropic action of this amino acid.
In contrast with our own work, Sener and Malaisse  reported that addition of L-alanine did not stimulate insulin secretion from rat islet cells. However, in the presence of L-leucine or 2-ketoisocaproate, alanine promoted insulin secretion. Additionally, Sener and Malaisse  showed that L-alanine induced an increase in Ca2+ uptake and was oxidized by the β-cell. They concluded that L-alanine could stimulate insulin secretion under specific conditions of nutrient availability and that the mode of induction of insulin secretion may be a combination of increased ATP production and Na+ co-transport.
L-Leucine stimulates insulin release in pancreatic β-cells by a process that involves: (i) increased mitochondrial metabolism by activation of GDH, and (ii) an increase in ATP production by transamination of leucine to α-ketoisocaproate and subsequent entry into the TCA cycle via acetyl-CoA [10,50,51]. However, leucine-induced insulin secretion is inhibited in the presence of high glucose , as high glucose inhibits flux through glutaminase and GDH. There has been renewed interest in L-leucine metabolism as a result of the observation of hyperinsulinism in patients with activating mutations in the regulatory site of GDH [53,54]. Affected patients have increased β-cell responsiveness to leucine and develop hypoglycaemia following a protein meal. Key mutations in the inhibitory allosteric site in GDH (GTP binding) result in the loss of negative allosteric regulation.
Although one of the proposed mechanisms by which L-leucine induces insulin secretion is the conversion of L-leucine into α-ketoisocaproate, a recent report  has shown that leucine and α-ketoisocaproate stimulated insulin release via distinct mechanisms. α-Ketoisocaproate was proposed to stimulate insulin release by a combination of mechanisms, including its own catabolism and transamination to leucine with production of α-ketoglutarate.
Previous studies have shown that leucine can activate the translational regulators PHAS-I (phosphorylated heat- and acid-stable protein regulated by insulin) and p70 S6 kinase via mTOR [28,56,57]. Mitochondrial signals generated by metabolism of leucine have been suggested to be important for the activation of the mTOR mitogenic signalling pathway in the β-cell .
The stimulation of insulin release by L-arginine has been proposed to involve the transport of the cationic amino acid into the cell which leads to membrane depolarization [4,59,60]. A detailed study by Malaisse and co-workers  agreed with this argument. L-Arginine was shown to cause an elevation in [Ca2+]i as a result of its electrogenic transport into the β-cell via the amino acid tranporter mCAT2A. Depolarization of the plasma membrane will then result in activation of voltage-dependent Ca2+ channels, an intracellular increase in Ca2+ and a subsequent stimulation of insulin secretion. Clinical assessment of administered L-arginine has revealed only limited beneficial effects, possibly due to rapid removal of the amino acid in the epithelial cells of the intestine, where it can be rapidly converted into ornithine and citrulline, then exported to the kidney [62,63] or the liver, and converted into proline for export .
Alternatively, L-arginine metabolism in the β-cell can give rise to glutamate production and thus can influence insulin secretion as described above . However, no studies have yet explored L-arginine metabolism in detail in the β-cell, thus the potential for L-glutamate generation remains to be determined.
ROLE OF AMINO ACIDS IN NADH MITOCHONDRIAL SHUTTLES
Amino acids are also implicated in NADH mitochondrial shuttle activity, in particular the malate/aspartate shuttle and its member aralar1 . Aralar1 is an aspartate/glutamate exchanger isoform, activated by cytosolic Ca2+ and expressed in excitatory tissues, including pancreatic islets [65–68]. In the course of glucose-stimulated insulin release, electrons of the glycolysis-derived reduced form of NADH are transferred to mitochondria through the NADH shuttle system. Accordingly, the NADH shuttle couples glycolysis to activation of mitochondrial energy metabolism to trigger insulin secretion. In β-cells, the NADH shuttle system is composed essentially of the glycerophosphate and the malate/aspartate shuttles . It was suggested that the malate/aspartate shuttle prevails over the glycerophosphate shuttle  as opposed to other tissues .
Inhibition of the malate/aspartate shuttle by amino-oxyacetate attenuates the secretory response to nutrients . Amino-oxyacetate inhibits transamination reactions and hence will impair the activity of the malate/aspartate shuttle. Therefore the malate/aspartate shuttle is essential for both mitochondrial metabolism and cytosolic redox state. As evidenced by the lactate/pyruvate ratio , the malate/aspartate shuttle has an important role in nutrient sensing  and glucose-stimulated insulin secretion. Additionally, the pyruvate/citrate shuttle regenerates cytosolic NAD+ and might play a role in the formation of the candidate coupling factor malonyl-CoA in glucose-induced insulin secretion .
Inhibition of cytosolic NADH re-oxidation by mitochondria, using amino-oxyacetate, lowers mitochondrial membrane potential to an extent similar to glucose deprivation  and inhibits insulin secretion . In a β-cell model with mitochondrial DNA damage, NADH accumulation decreases mitochondrial membrane potential and abrogates insulin secretion . Low activity of NADH shuttles in β-cells has been found in aging [78,79] and in models of Type II diabetes .
In the mitochondria, the transfer of NADH electrons to the respiratory chain creates a proton electrochemical gradient that drives ATP synthesis. The formation of a proton gradient limits the production of mitochondrial coupling factors . In addition to ATP generation, the mitochondrial membrane potential drives the transport of metabolites between mitochondrial and cytosolic compartments and the production of mitochondrial factors that couple secretagogue metabolism to insulin secretion. Therefore the malate/aspartate NADH shuttle plays a central role in the coupling of glucose recognition to insulin secretion. In a recent report, we  observed that adenoviral-mediated overexpression of Aralar1 in insulin-secreting cells increased glucose-induced mitochondrial activation and the secretory response. This was accompanied by enhanced glucose oxidation and reduced lactate production. Therefore aspartate–glutamate carrier capacity appears to set a limit for NADH shuttle function and mitochondrial metabolism.
Amino acid metabolism is essential to normal pancreatic β-cell function, as highlighted in this review. Acutely, key amino acids such as alanine and glutamine can regulate β-cell function and insulin secretion. The mechanisms by which these amino acids confer their regulatory effects are complex and involve mitochondrial metabolism. L-Glutamine metabolism is important for L-glutamate and glutathione production. D-Glucose conversion into L-glutamate can occur in specific conditions of glutamine limitation, via 2-oxoglutarate in a transamination reaction. L-Glutamine metabolism appears to be especially important in the β-cell, due to optimization of mitochondrial function. Thus L-glutamine is utilized by the β-cell not only for oxidation, but also for glutathione production. We speculate that, in conditions where activation or overexpression of GDH activity is promoted, the metabolism of L-glutamine will contribute to a rise in the ATP/ADP ratio, thus inactivating the K+ATP channel and depolarizing the plasma membrane, resulting in intracellular Ca2+ elevation and insulin secretion. Glutamate may play a key role in stimulation of insulin secretion directly, via metabolism or via its role in enhancing malate/aspartate shuttle activity. Other amino acids such as leucine or arginine may play a role in enhancing insulin secretion by allosteric activation of metabolism or membrane depolarization or a combination of these two possibilities. Chronic effects of changes in amino acid concentration in vivo and in vitro on β-cell function and integrity have not yet been investigated in detail. Marliss et al.  originally reported large changes in glutamine and branched-chain amino acid concentrations (2–10-fold) during progression to diabetes in the BB Wistar rat (an animal model of diabetes). We have determined significant changes in plasma glutamine concentration in newly diagnosed Type I diabetic patients (A. Kerr, G. Dixon and P. Newsholme, unpublished work). The effects of the changes in amino acid concentration on β-cell gene expression in vivo are unknown, but we have recently determined specific functional, signalling and metabolic gene expression changes in response to L-alanine or L-glutamine in vitro. Importantly, L-leucine, which is a key regulator of the mTOR signalling pathway (which regulates protein synthesis via activation of the protein kinase p70S6K), may play an important role in the maintenance of β-cell mass in vivo as described by others . Defects in this pathway result in diminished β-cell size, hypoinsulinaemia and glucose intolerance as demonstrated in S6K1-deficient mice . Understanding the mechanisms by which amino acids regulate insulin secretion in vivo may reveal novel sites for targeting drugs for the therapy of Type II diabetes in the future.
The Health Research Board of Ireland and the Swiss National Science Foundation supported our research.
Abbreviations: [Ca2+]i, intracellular Ca2+ concentration; EST, expressed sequence tag; GABA, γ-aminobutyric acid; GAD, glutamate decarboxylase; GDH, glutamate dehydrogenase; MSO, methionine sulphoximine; mTOR, mammalian target of rapamycin; TCA, tricarboxylic acid
- The Biochemical Society