Digital Mitochondrial pyruvate carrier 2 hypomorphism in mice leads to Mitochondrial pyruvate carrier 2 hypomorphism in mice leads to defects in glucose-stimulated insulin secretion defects in glucose-stimulated insulin secretion

Carrier-facilitated pyruvate transport across the inner mitochondrial membrane plays an essential role in anabolic and catabolic intermediary metabolism. Mitochondrial pyruvate carrier 2 (Mpc2) is believed to be a component of the complex that facilitates mitochondrial pyruvate import. Complete MPC2 deficiency resulted in embryonic lethality in mice. However, a second mouse line expressing an N-terminal truncated MPC2 protein (Mpc2(Δ16)) was viable but exhibited a reduced capacity for mitochondrial pyruvate oxidation. Metabolic studies demonstrated exaggerated blood lactate concentrations after pyruvate, glucose, or insulin challenge in Mpc2(Δ16) mice. Additionally, compared with wild-type controls, Mpc2(Δ16) mice exhibited normal insulin sensitivity but elevated blood glucose after bolus pyruvate or glucose injection. This was attributable to reduced glucose-stimulated insulin secretion and was corrected by sulfonylurea KATP channel inhibitor administration. Collectively, these data are consistent with a role for MPC2 in mitochondrial pyruvate import and suggest that Mpc2 deficiency results in defective pancreatic β cell glucose sensing.


INTRODUCTION
Pyruvate is an important three-carbon intermediate in energy metabolism and is a central substrate in carbohydrate, fat, and amino acid catabolic and anabolic pathways. Pyruvate is generated in the cytoplasm through glycolysis and subsequently is transported into the mitochondrion for oxidation or carboxylation, which is required for several critical metabolic processes ( Figure 1A). For example, mitochondrial pyruvate carboxylation in the mitochondrial matrix results in the formation of oxaloacetate (an anaplerotic reaction), which is required for the biosynthesis of glucose via gluconeogenesis. Pyruvate entry into the mitochondrial matrix is also required for pyruvate oxidation, which is a prerequisite for the production of reducing equivalents in the tricarboxylic acid (TCA) cycle and production of citrate for de novo lipogenesis. Both pyruvate oxidation and carboxylation are critical for glucose-stimulated insulin secretion (GSIS) in pancreatic b cells (Jensen et al., 2008;Prentki et al., 2013;Sugden and Holness, 2011). Elevated blood glucose causes an increase in the intracellular ratio of ATP/ADP via increased mitochondrial pyruvate oxidation (Prentki et al., 2013;Sugden and Holness, 2011). ATP inhibits K ATP channels, leading to depolarization, Ca 2+ influx, and the release of insulin into the circulation (Huopio et al., 2002). Anaplerotic mitochondrial pyruvate metabolism activates pyruvate cycling pathways that alter NADPH, and regulates insulin secretion by inhibiting K ATP channel activity as well (Jensen et al., 2008).
The existence of carrier-assisted transport of pyruvate across the inner mitochondrial membrane (IMM) was demonstrated in the 1970s (Halestrap, 1975;Halestrap and Denton, 1975;Papa et al., 1971). However, the proteins that facilitate pyruvate import into the mitochondrial matrix have only recently been identified (Bricker et al., 2012;Herzig et al., 2012). New evidence has emerged that the mitochondrial pyruvate carrier (MPC) is composed of two proteins, MPC1 and MPC2, which form a hetero-oligomeric complex in the IMM, and that both proteins are required for complex activity and stability (Bricker et al., 2012;Herzig et al., 2012). The MPC protein complex has been determined to be essential for mitochondrial pyruvate transport in Drosophila and yeast (Bricker et al., 2012;Herzig et al., 2012). In humans, mutations in MPC1 have been identified and associated with defects in mitochondrial pyruvate metabolism, lactic acidosis, hyperpyruvatemia, severe illness, and failure to thrive (Bricker et al., 2012;Brivet et al., 2003). Since its discovery, interest in the MPC complex as a drug target for cancer, neurological disorders, and metabolic diseases has been extremely high. Recent work suggests that insulin-sensitizing thiazolidinedione compounds bind the MPC complex (Colca et al., 2013) and modulate mitochondrial pyruvate oxidation (Divakaruni et al., 2013;Colca et al., 2013). Thus, a better understanding of MPC function has the potential to advance our knowledge about intermediary metabolism and impact drug discovery for current public health problems.
Herein, we report the generation and characterization of two MPC2-deficient mouse strains. Although complete MPC2 deficiency resulted in embryonic lethality, mice expressing a truncated, partially functional MPC2 protein were viable. Diminished pyruvate metabolism led to elevated blood lactate concentrations, particularly when mice were challenged with stimuli that increased glucose metabolism. The mutant mice also exhibited elevated blood glucose during glucose tolerance tests (GTTs) and pyruvate tolerance tests (PTTs) that was likely due to reduced GSIS. Additionally, glucose intolerance could be corrected by administration of K ATP channel-inhibiting sulfonylurea drugs. These studies demonstrate the importance of mitochondrial pyruvate transport to glucose homeostasis in a mammalian system and support the continued exploration of mitochondrial function in complex metabolic diseases.

Tissue Expression of MPC Proteins
We first isolated various tissues from C57BL/6 mice to determine relative Mpc mRNA and MPC protein expression. We detected large variations in MPC protein (Figures 1B and 1C) and mRNA expression ( Figure 1D), with the highest levels observed in tissues with greater mitochondrial abundance. Isolation and equal loading of mitochondrial lysate from these tissues somewhat normalized MPC1 and MPC2 protein expression (Figures S1A and S1B). However, MPC protein was still most abundant in highly oxidative tissues even when corrected for mitochondrial protein content.

Generation of Mpc2-Deficient Mice
We targeted the Mpc2 allele for deletion by using zinc-finger nuclease (ZFN) technology. The first line of mice harbored a deletion of 2 nt just downstream of the start codon ( Figure S2A), resulting in a frameshift mutation and likely complete loss of protein product (Mpc2 À/À mice). The mutation results in embryonic lethality in homozygous embryos. However, the number of heterozygous mice present at weaning was as predicted by Mendelian ratios (p = 0.717; Table S1), and these mice were overtly indistinguishable from their wild-type (WT) littermates. Lethality in homozygous embryos likely occurs between embryonic day 11 (E11) and E13, a period when robust mitochondrial biogenesis takes place (Cá mara et al., 2011;Larsson et al.,  OMM, outer mitochondrial membrane; IMM, inner mitochondrial membrane; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; OAA, oxaloacetate; TCA, tricarboxylic acid. (B) Western blot depicting expression of the indicated proteins in whole-cell lysates from various C57Bl/6 mouse tissues. Cytochrome c (cyto c) is shown to demonstrate mitochondrial abundance in the whole-cell lysates.
A second attempted knockout strategy resulted in the deletion of 20 nt including the original start codon for MPC2 translation. Despite the deletion, Mpc2 mRNA was expressed at normal levels (data not shown) and MPC2 protein translation proceeded from a second downstream start codon ( Figure S2A). This resulted in a truncated protein that lacked the first 16 amino acids residues of the N terminus (MPC2 D16 ). Pups homozygous for the truncated allele were born at the expected Mendelian ratios (p = 0.995; Table S1) and were outwardly normal. The truncated protein, detected by western blotting with an antibody against the C terminus of MPC2, was less abundant than the WT protein in all tissues examined, except for brown adipose tissue (BAT; Figure 2A). The Mpc2 D16 mutation led to a marked reduction in the abundance of MPC1 protein, likely reflecting the maintenance of a stoichiometric ratio of MPC1 and MPC2 protein in the complex. Immunoprecipitation studies also indicated that the Mpc2 D16 mutant protein coimmunoprecipitated less MPC1 protein, suggesting that the interaction between the two proteins is significantly weakened by the loss of these 16 amino acids of MPC2 ( Figure 2B). The abundance of the pyruvate dehydrogenase (PDH) complex and pyruvate carboxylase (PC) proteins was not altered in various tissues of the Mpc2 D16 mice, suggesting a lack of significant downstream alterations in pyruvate metabolic enzymes ( Figure 2A). Western blotting of subfractionated kidney and heart demonstrated that the truncated Mpc2 D16 protein properly localized to mitochondria (Figure 2C) and specifically to the IMM in liver ( Figure 2D).

Isolated Mpc2 D16 Mitochondria Have Reduced Oxygen Consumption
The Mpc2 D16 mice were healthy and appeared outwardly normal, suggesting that the MPC complex was at least partially functional. Plasma triglyceride was decreased and plasma ketone concentrations were increased in Mpc2 D16 mice compared with WT control mice (Table 1), suggesting an increased reliance on fat oxidation. Possibly consistent with a reduced capacity for mitochondrial pyruvate import, blood lactate concentrations were also significantly elevated (Table 1). Therefore, we evaluated the effects of the MPC2 truncation on oxygen consumption in isolated mitochondria. We used heart and kidney tissue due to their high mitochondrial density and high MPC expression. Mpc2 D16 mitochondria from heart or kidney exhibited 25%-30% lower rates of pyruvate/ malate-stimulated respiration in 2,4-dinitrophenol (DNP)uncoupled conditions compared with WT controls (Figure 3A). Addition of methyl-pyruvate, a membrane-permeable pyruvate analog, rescued oxygen consumption to WT levels in Mpc2 D16 mitochondria. Additionally, maximal oxygen consumption rates in Mpc2 D16 mitochondria with glutamate/malate or (B) The ability of WT MPC2 and MPC2 D16 protein to coimmunoprecipitate MPC1 was tested by using an antibody against the C terminus of MPC2 and BAT mitochondrial lysates from WT and Mpc2 D16 mice. Negative control (irrelevant immunoglobulin G [IgG]) and input (nonprecipitated mitochondrial lysates) are also shown for comparison. (C) Western blots for MPC1 and MPC2 after subcellular fractionation to separate cytosolic and mitochondrial fractions from WT and Mpc2 D16 mice are shown. LDH is used to demarcate the cytosolic fraction, and VDAC1 demarcates the mitochondrial fraction. (D) Western blots after mitochondrial subfractionation to separate cytosolic, OMM, IMM, and matrix components by using liver from WT and Mpc2 D16 mice. LDH demarcates the cytosolic fraction, cytochrome c oxidase subunit 4 (Cox4) demarcates the IMM, and PC demarcates the matrix. succinate/rotenone were unaltered compared with WT controls. Together, these data suggest a specific effect of the Mpc2 D16 mutation on mitochondrial pyruvate oxidation.
Mitochondrial Membrane Potential Is Normal in Fibroblasts from Mpc2 D16 Mice Given the observed defect in pyruvate oxidation in Mpc2 D16 tissues, we investigated whether these mitochondria displayed any sign of overt dysfunction by assessing the mitochondrial membrane potential (DJ M ) in isolated adult fibroblasts of WT and Mpc2 D16 mice. Tetramethylrhodamine, ethyl ester, perchlorate (TMRE) staining revealed no difference in DJ M , a comprehensive measure of overall mitochondrial function, in Mpc2 D16 fibroblasts compared with WT controls ( Figure 3B). The lack of effect on mitochondrial membrane potential was likely due to the small decrease in pyruvate-stimulated respiration and the maintained ability to metabolize other substrates.
Exercise Endurance Is Normal in MPC2 D16 Mice During exercise, significant quantities of pyruvate, which is interconverted to lactate, are generated via anaerobic glycolysis in skeletal muscle. Lactate, transported by the blood to the liver, is converted back to glucose that can again be used anaerobically as an energy source in muscle. This process, known as the Cori cycle (Katz and Tayek, 1999;Pilkis et al., 1988), depends on pyruvate import in liver mitochondria. Therefore, we challenged Mpc2 D16 homozygote and Mpc2 +/À mice using an acute endurance exercise bout of gradually increasing intensity until exhaustion ( Figure 3C). The average blood glucose concentration of the Mpc2 D16 mice was not different from that of WT controls in the 60 min after exercise ( Figure 3C). However, the average blood lactate of the Mpc2 D16 was approximately 2-fold higher compared with WT controls immediately after exercise and remained significantly elevated for 30 min postexercise (Figure 3C). Postexercise blood lactate and glucose concentrations were not different between Mpc2 +/À and WT mice ( Figure 3C). Despite the robust differences in blood lactate concentrations postexercise, we observed no significant differences in acute endurance exercise capacity among the genotypes ( Figure 3C).

Elevated Blood Glucose and Lactate in Mpc2 D16 Mice following Pyruvate Injection
To test the effects of the Mpc2 mutations on pyruvate disposal in vivo, we subjected Mpc2 D16 homozygotes and  (C) Schematic of the exercise protocol. Shown are blood glucose and blood lactate curves for WT, Mpc2 +/À , and Mpc2 D16 mice following graduatedintensity exercise on a motor-driven treadmill (n = 6), and elapsed time until exhaustion for WT, Mpc2 +/À , and Mpc2 D16 mice during graduatedintensity exercise (n = 6). Data are presented as mean ± SEM. *p < 0.05, Mpc2 D16 versus WT.
Mpc2 +/À mice to an intraperitoneal (i.p.) PTT. Following pyruvate injection, blood lactate increased 5-fold in all groups. However, the decline in blood lactate concentration was significantly slower in Mpc2 D16 mice compared with WT controls ( Figure 4A). Blood glucose and lactate levels were not different between Mpc2 +/À and WT mice. Compared with WT mice, Mpc2 D16 mice also exhibited a significantly greater blood glucose excursion ( Figure 4A). These data are consistent with reduced pyruvate utilization resulting in elevated blood lactate concentrations.
To specifically evaluate the effect of the Mpc2 D16 mutation on pyruvate gluconeogenesis and other intracellular pathways of pyruvate metabolism, we repeated the pyruvate challenge in Mpc2 D16 mice with uniformly labeled [U-13 C]pyruvate. The mice were sacrificed, plasma and tissue were collected 60 min postinjection, and fractional 13 C enrichments were evaluated by mass spectrometry. There were no differences in hepatic oxaloacetate, malate, or fumarate concentrations or enrichments ( Figure S3), indicating that liver mitochondrial pyruvate anaplerosis is intact in Mpc2 D16 mice. Although the fractional enrichment of plasma glucose was unaffected by the Mpc2 D16 mutation ( Figure 4B), the plasma glucose concentration and therefore the total amount of 13 C-labeled plasma glucose were significantly elevated in Mpc2 D16 mice versus WT mice ( Figure 4B). These data suggest that the capacity for pyruvate-derived gluconeogenesis is intact in Mpc2 D16 mice. Incorporation of 13 C-label into plasma lactate was not different between Mpc2 D16 and WT control groups ( Figure 4C). However, due to the increased lactate concentration in the blood, 13 C-labeled plasma lactate was significantly elevated in Mpc2 D16 mice compared with WT controls ( Figure 4C), consistent with decreased mitochondrial pyruvate metabolism in extrahepatic tissues. Additionally, expression of liver gluconeogenic enzymes was unaltered in Mpc2 D16 mice, except that Pck2, which encodes the mitochondrial form of phosphoenolpyruvate carboxykinase, was significantly decreased ( Figure 4D), suggesting that compensatory activation of these genes does not explain the intact gluconeogenic flux. Altogether, these data suggest that hepatic gluconeogenesis is not impaired in Mpc2 D16 mice.
Mpc2 D16 Mice Are Sensitive to Insulin, but Have Reduced Plasma Insulin and Elevated Blood Glucose One possible explanation for the elevated blood glucose concentration in the PTT studies is that Mpc2 D16 mice are insulin resistant. We conducted insulin tolerance tests (ITTs) to examine insulin effectiveness. Although blood lactate during the ITT was significantly elevated in Mpc2 D16 mice compared with WT mice, blood glucose concentrations at all time points and the area under the curve (AUC) were not different, sug- gesting that insulin-stimulated glucose disposal is normal in Mpc2 D16 mice ( Figure 5A). Insulin-stimulated phosphorylation of Akt in liver was significantly enhanced ( Figure 5B) and was unaltered in gastrocnemius muscle (Figure 5C) of Mpc2 D16 mice compared with WT control mice. These data suggest that the Mpc2 D16 mice have normal sensitivity to the effects of insulin.
In contrast, Mpc2 D16 mice exhibited elevated blood lactate and trended toward a higher glucose AUC (p = 0.078) during an i.p. GTT ( Figure 5D). Plasma insulin concentration was significantly lower in Mpc2 D16 mice compared with WT controls at baseline and 30 min after glucose injection ( Figure 5E), suggesting that glucose intolerance in Mpc2 D16 mice was due to insufficient insulin secretion. Similarly, blood insulin concentration in Mpc2 D16 mice compared with WT controls was also significantly lower 15 min after a bolus pyruvate injection ( Figure 4E).

Impaired GSIS in Mpc2 D16 Mice Is Corrected by Glibenclamide
The data obtained in GTT studies suggested a defect in GSIS by pancreatic islet b cells of Mpc2 D16 mice. Both MPC1 and MPC2 are expressed in b cells, as demonstrated by colocalization with insulin in immunohistochemical staining of pancreatic sections ( Figure 6A). Western blotting analyses confirmed that MPC1 and MPC2 proteins are expressed in isolated pancreatic islets and demonstrated that the expression of the MPC2 D16 protein was markedly reduced compared with WT MPC2 ( Figure 6B). Islet insulin content in Mpc2 D16 mice was not different from that in WT controls ( Figure 6C). (E) Blood glucose and blood lactate curves during an i.p. GTT for WT and Mpc2 D16 mice treated with glibenclamide or saline vehicle (n = 11-13). (F) Plasma insulin from blood collected during the experiment described in (E) (n = 11-13). Data are presented as mean ± SEM. *p < 0.05, **p < 0.01 Mpc2 D16 versus WT.
We examined GSIS in isolated pancreatic islets from WT and Mpc2 D16 mice. Compared with WT islets, Mpc2 D16 islets exhibited reduced insulin secretion in response to 23 mM glucose ( Figure 6D). b cell glucose sensing requires mitochondrial pyruvate metabolism, resulting in closure of the K ATP channels (Huopio et al., 2002). We therefore examined the effects of glibenclamide, a sulfonylurea drug that inhibits K ATP channel activity, on GSIS in Mpc2 D16 islets. As we hypothesized, the GSIS defect was rescued by the addition of glibenclamide to high-glucose media ( Figure 6D). Additionally, Mpc2 D16 islets exhibited normal insulin secretion in response to KCl ( Figure 6D). Supplementing highglucose medium with glutamine, a metabolic substrate that does not require the MPC complex to enter the mitochondrion to be oxidized, also corrected the defect in GSIS ( Figure 6D). These findings collectively suggest that the Mpc2 D16 insulin secretion defect is specific to stimulation by mitochondrial pyruvate metabolism rather than to a general deficiency in insulin production or defects in the insulin secretory machinery.
To confirm the relevance of these findings in vivo, we administered glibenclamide to Mpc2 D16 mice with a bolus of glucose in GTT studies. Vehicle-treated Mpc2 D16 mice exhibited elevated blood glucose concentrations compared with WT mice at 30 and 60 min postinjection, but blood glucose concentrations in Mpc2 D16 mice treated with glibenclamide were not different from those in vehicle-treated WT mice ( Figure 6E). Administration of glibenclamide did not affect plasma lactate concentrations in mice of either genotype. Glibenclamide administration to Mpc2 D16 mice resulted in increased plasma insulin concentrations, which were reduced in Mpc2 D16 mice versus WT mice, at 15 and 30 min after injection ( Figure 6F). These data further support the model in which whole-body Mpc2 deficiency leads to a defect in GSIS that is improved by K ATP channel inhibition.

DISCUSSION
Carrier-mediated transport of pyruvate across the IMM is a critical step in intermediary metabolism, which is required for both anaplerotic and cataplerotic mitochondrial pyruvate metabolism. We found that complete loss of MPC2 led to embryonic lethality, but expression of a truncated MPC2 protein that exhibited hypomorphic activity was sufficient to support life. The most striking phenotypes of the Mpc2 D16 mice were elevated plasma lactate concentrations, particularly in the context of increased glucose concentrations, and a defect in b cell GSIS leading to relative glucose intolerance. These findings are consistent with a function for MPC2 in regulating mitochondrial pyruvate utilization and reveal important roles for MPC2 in regulating whole-body glucose homeostasis.
As mitochondrial pyruvate import is required in a number of critical metabolic pathways, it was not unexpected that complete ablation of Mpc2 resulted in embryonic lethality. Human mutations in MPC1 are associated with neonatal and juvenile mortality (Brivet et al., 2003). We were unable to recover viable Mpc2 À/À embryos after E11, which is when mitochondrial function begins to play a larger role in embryonic metabolism (Cá mara et al., 2011;Larsson et al., 1998;Metodiev et al., 2009;Park et al., 2007). Mpc2 +/À mice exhibited only a modest reduction in MPC2 and MPC1 protein abundance, and were phenotypically indistinguishable from WT littermates in every parameter we examined. Our inability to detect a phenotype from the Mpc2 +/À mice, even under conditions of metabolic challenge, likely highlights an excess transport capacity of the MPC complex.
We also serendipitously created a truncated mutant MPC2 protein as a consequence of the unpredictable results obtained by ZFN-mediated deletion. Although the Mpc2 D16 mutant has a nucleotide deletion larger than that of the Mpc2 À/À mouse (20 bp versus 2 bp, respectively), it utilizes a downstream methionine as an alternative start site for translation of a truncated MPC2 protein. Because the original start codon is ablated by the 20 bp deletion but is intact in the 2 bp deletion, this allows an alternative ribosomal entry sequence to be enforced. This start codon may be less efficient or the resulting protein less stable, given that MPC2 D16 protein is less abundant than fulllength MPC2. Most tissues in the Mpc2 D16 mice also had a significant reduction in abundance of the MPC1 protein. Previous work has shown that MPC1 and MPC2 must be coexpressed for proper assembly of the multimeric MPC complex, and that depletion of either protein leads to reduced abundance of the other (Bricker et al., 2012;Colca et al., 2013). It is likely that the reduced abundance of MPC1 in the Mpc2 D16 mice is due to the lower quantity of Mpc2 D16 compared with WT protein. A diminished or weakened protein-protein interaction between the two proteins is also a potential factor in the reduced complex stability, since less MPC1 was coimmunoprecipitated with MPC2 D16 compared with WT protein. However, the expression of MPC1 was only modestly affected by MPC2 truncation in BAT and islets, suggesting a tissue-specific regulation of MPC complex stability or the possible existence of homodimeric MPC protein complexes.
The truncated MPC2 protein is properly localized to the IMM and presumably, based on the phenotypes of the mutant mice, forms a transporter complex that is functionally hypomorphic. A number of mitochondrial carrier proteins of the IMM do not utilize canonical N-terminal mitochondrial localization sequences (Ferramosca and Zara, 2013;Pfanner and Geissler, 2001;Harbauer et al., 2014). Instead, these carriers utilize internal sequences that that have yet to be defined. Indeed, analysis by mitochondrial localization prediction programs, which predict mitochondrial targeting based on these canonical N-terminal localization sequences, did not produce a high-scoring prediction of mitochondrial localization for full-length or D16 MPC2 protein. However, the focus of this paper was not to define the targeting sequence, because that has been elusive for most mitochondrial proteins that lack the N-terminal canonical sequence.
Given that conversion of pyruvate into glucose by liver and kidney requires pyruvate to enter mitochondria for conversion to oxaloacetate, we expected that the Mpc2 D16 mice might be hypoglycemic in conditions requiring gluconeogenesis. Previous work using isolated liver preparations and a chemical inhibitor of MPC activity showed an acute inhibition of hepatic glucose output (Martin-Requero et al., 1986;Rognstad, 1983;Thomas and Halestrap, 1981). However, whether carrier-mediated pyruvate transport by MPC plays a significant role in controlling gluconeogenesis has been debated (Groen et al., 1983(Groen et al., , 1986Halestrap and Armston, 1984), and the hypomorphic function of the MPC transport complex was not a limiting factor in the gluconeogenic capacity of Mpc2 D16 mice. Plasma 13 C-labeled glucose levels were actually increased in [U-13 C]PTT studies and Mpc2 D16 mice were not hypoglycemic following exhaustive exercise, suggesting that the Cori cycle was intact. The observation that gluconeogenesis was not impaired is not entirely unprecedented, given that previous studies have shown that an up to 85% reduction in phosphoenolpyruvate carboxykinase expression does not impact gluconeogenic flux (Burgess et al., 2007). Further work with other, more complete, liverspecific loss-of-function models will be needed to define the role that MPC complex proteins play in regulating hepatic gluconeogenesis.
Another interesting phenotype of the Mpc2 D16 mice was the hyperglycemia detected during PTT and GTT studies. Given that the insulin-stimulated decrease in blood glucose concentration was normal in ITT experiments and pyruvate metabolism is important for b cell glucose sensing (Lu et al., 2002;Prentki et al., 2013;Sugden and Holness, 2011), we suspected that insulin secretion might be defective. Indeed, plasma insulin concentration was diminished in Mpc2 D16 mice compared with WT mice during the GTT and PTT studies, and isolated islets from Mpc2 D16 mice exhibited an attenuated insulin secretion response following the administration of 23 mM glucose. Glucose sensing and GSIS require mitochondrial pyruvate oxidation and carboxylation. b cells express very little lactate dehydrogenase (LDH), and most of the pyruvate generated from glycolysis enters the mitochondrion for further metabolism (Malmgren et al., 2009;Schuit et al., 1997). Pyruvate carboxylation and oxidation inhibit ATP-sensitive potassium channels (K ATP ), depolarizing the b cell membrane, opening voltagedependent calcium channels, and triggering the release of insulin into the circulation (Huopio et al., 2002). We found that the defects in insulin secretion by isolated islets and hypoinsulinemia/hyperglycemia in Mpc2 D16 mice were corrected by the sulfonylurea glibenclamide, which functions by binding and inhibiting the regulatory subunit of K ATP channels. It is likely that glibenclamide is able to circumvent the reduced capacity of Mpc2 D16 mutants for mitochondrial pyruvate metabolism. The observation that glutamine, which relies upon its own IMM carrier (Indiveri et al., 1998), is able to stimulate insulin secretion in Mpc2 D16 islets also supports the notion of a specific defect in mitochondrial pyruvate metabolism. Consistent with these data, while this paper was in review, Patterson et al. (2014) used chemical inhibitors and small interfering RNA (siRNA) methods to suggest that MPC inhibition suppressed pyruvate metabolism and GSIS in isolated islets and b cell lines. Taken together, these studies demonstrate that the hypoinsulinemia of Mpc2 D16 mice is a specific defect in GSIS and not a broader indication of b cell dysfunction.
The molecular identification of MPC proteins has stimulated interest in targeting this complex for the treatment of many chronic disease states. Pharmacologic approaches to modulate the activity of the MPC complex could yield treatments for broad categories of obesity-related metabolic diseases, neurodegenerative conditions, and neoplastic diseases associated with impaired pyruvate metabolism (Constantin-Teodosiu, 2013;Schell and Rutter, 2013;Stacpoole, 2012). For example, the MPC complex has recently been identified as a binding site for the thiazolidinedione class of insulin sensitizers (Colca et al., 2013), which acutely modulate mitochondrial pyruvate oxidation (Divakaruni et al., 2013;Colca et al., 2013). The extent to which regulation of mitochondrial pyruvate import by thiazolidinediones affects insulin sensitization remains unclear, but could be addressed by MPC loss-of-function mouse models. Screens to identify small molecules that enhance, inhibit, or selectively modulate MPC activity may identify novel therapeutics with applicability to myriad chronic diseases.

Animal Studies
Unless otherwise noted, all experiments were conducted with 10-to 16-weekold female mice generated and maintained in a pure C57BL/6 background. All animal experiments were approved by the Animal Studies Committee of Washington University School of Medicine.

Generation of Mpc2-Deficient Mice
Mpc2 transgenic mice were generated by SAGE Labs using ZFN targeting technology (Geurts et al., 2009). Specifically, the target site sequence was TCCCTAGGCCGCCGCGATggcagCTGCCGGCGCCCGAGGCC (NCBI Ref Seq: NC_000067.5. Exon 1,. ZFN pairs were microinjected into C57BL/6 zygotes and transferred to pseudopregnant C57BL/6 females at 0.5 days postcoitum. Western Blotting PAGE was performed on protein lysates utilizing Criterion precast gels (BioRad). Whole-cell lysates from tissue or cell culture were collected in HNET buffer containing protease and phosphatase inhibitors. For blots of mitochondrial lysates, mitochondria were obtained by differential centrifugation and solubilized in HNET buffer containing protease and phosphatase inhibitors. Blots on isolated pancreatic islets involved isolation of islets as described below, and 50 islets solubilized in 53 gel loading buffer per lane. For immunoprecipitation, 250 mg of BAT mitochondrial lysates was solubilized in lysis buffer containing 0.2% NP-40 and 10% glycerol. The lysates were precleared with 50 ml protein A beads and then incubated overnight with a rabbit MPC2 antibody (a gift from Michael Wolfgang). Protein A beads (50 ml) were then added 1 hr prior to immunoprecipitation. Liver mitochondrial fractionation was conducted as previously described (Benga et al., 1979) with slight modification. In brief, digitonin-treated mitochondria were centrifuged at 9,500 3 g. The digitonin supernatants were centrifuged at 100,000 3 g to isolate the outer membrane pellet (OMM) and the supernatant containing the inner membrane space. The digitonin pellet was loaded onto a discontinuous sucrose gradient to generate mitoplasts. The mitoplasts were treated with Lubrol WX and the IMM fractions were recovered from 100,000 3 g discontinuous sucrose gradients. The antibodies utilized were b-Actin (Sigma-Aldrich), Akt (Cell Signaling Technology), pS473-Akt (CST), CoxIV (CST), cytochrome c (Abcam), LDH (Abcam), MPC1 and MPC2 (gifts from Michael Wolfgang), PC (Santa Cruz), PDH cocktail (Abcam), a-tubulin (Sigma), and voltage-dependent anion channel (VDAC; Abcam).
Quantitative PCR and RT-PCR Total RNA was isolated using the RNAzol method (Tel-Test). Real-time PCR was performed using the ABI PRISM 7500 sequence detection system (Applied Biosystems) and the SYBR Green kit. Arbitrary units of target mRNA were corrected by measuring the levels of 36B4 RNA. The sequences of the oligonucleotides used in quantitative PCR (qPCR) analyses are listed in Table S4.

Isolation of Adult Murine Fibroblasts and Measurement of Mitochondrial Membrane Potential
Isolation of primary adult murine fibroblasts from tail clip was performed as previously described (Huss et al., 2004). Fibroblasts from WT and Mpc2 D16 mice were cultured in high-glucose Dulbecco's modified Eagle's medium supplemented with 20% fetal bovine serum. Cells were plated onto four-well glass chamber slides (Lab-Tek II) at a density of 30,000 cells/well and incubated overnight. The cells were then stained with 100 nM TMRE in Hank's balanced salt solution (HBSS) for 30 min, washed with HBSS, and imaged on an inverted fluorescence microscope (Nikon Instruments). The average TMRE fluorescence per cell was determined with ImageJ (NIH).

Acute Exercise Protocol
WT, Mpc2 D16 , and Mpc2 +/À mice performed a single bout of treadmill exercise to exhaustion. The mice were fasted for 4 hr prior to the exercise session. Following a 3 min familiarization period on the treadmill, the exercise protocol consisted of graduated intensity running at 5 m/min for 5 min, 10 m/min for 5 min, 15 m/min for 10 min, 25 m/min for 15 min, and 30 m/min until the mouse fatigued. The mice were encouraged to run with a mild electric stimulus (20 V) positioned at the end of the treadmill and were determined to be exhausted by their refusal to remain on the treadmill belt for 5 s. Postexercise blood glucose and lactate levels were determined at 0, 5, 10, 20, 30, 60, and 120 min after exhaustion.

Measurement of Baseline Plasma Parameters
Mice were fasted for 4 hr and tail blood glucose and lactate concentrations were determined using a One-Touch Ultra glucometer (LifeScan) and a Lactate Plus lactate meter (Nova Biomedical), respectively. Blood samples for plasma analyses were collected by mandibular bleeding. Plasma insulin was measured with an ultrasensitive ELISA (Crystal Chem). Plasma triglyceride and cholesterol concentrations were measured by Infinity assay kits (Thermo Fisher Scientific). Nonesterified fatty acids (NEFAs) and total ketone concentrations were measured using enzymatic assays (Wako Diagnostics).

GTT, PTT, and ITT
For the GTT and PTT, mice were fasted overnight for 16 hr and housed on aspen chip bedding. The mice were injected i.p. with 1 g/kg body weight glucose or with 2 g/kg body weight Na-pyruvate dissolved in sterile physiological saline. For the ITT, mice were fasted for 6 hr and injected i.p. with 0.75 U/kg body weight insulin. Tail blood glucose and lactate concentrations were determined at 0, 30, 60, and 120 min after challenge as described above. The total AUC was calculated using the trapezoidal rule. Blood samples for plasma insulin quantification were collected by mandibular bleeding as described above. A subset of mice were injected i.p. with 2 g/kg body weight Napyruvate and sacrificed at 15 min to determine plasma insulin concentrations during pyruvate challenge. To examine insulin-stimulated protein phosphorylation, a subset of mice were injected i.p. with 10 mU/g body weight human insulin 15 min before sacrifice to collect liver and muscle tissue.
The glibenclamide rescue studies were conducted in the context of a standard GTT, with sample time points adjusted to 0, 15, 30, and 60 min after challenge. Glibenclamide was administered at 0.1 mg/kg body weight in a glucose solution that contained 0.1% DMSO. Blood was collected at each time point by means of mandibular bleeding as described above.

C-Labeled Pyruvate Studies
Pyruvate labeling studies were conducted as previously described (Mé ndez-Lucas et al., 2013). Briefly, mice were fasted for 16 hr and injected i.p. with 2 g pyruvate per kg body weight with a 10% enrichment of uniformly labeled [U-13 C3]Na-pyruvate (Cambridge Isotope Laboratories). The mice were sacri-ficed at 60 min postinjection and plasma was analyzed by gas chromatography-mass spectrometry as previously described (Sunny and Bequette, 2010).

Pancreas Immunohistochemistry
Protein expression of MPCs in pancreatic islets was determined in paraffinembedded pancreatic sections from WT mice cut onto glass slides. The slides were rehydrated, permeabilized with trypsin, blocked in 1% BSA, and probed with guinea pig anti-insulin (Abcam) and either rabbit anti-MPC1 or rabbit anti-MPC2 (gifts from Michael Wolfgang). All antibodies were diluted 1:200 in buffer containing 1% BSA and incubations were conducted overnight. The sections were then washed three times for 5 min in PBS and probed with 1:1,000 donkey anti-rabbit green or 1:1,000 donkey anti-guinea pig red Alexa Fluor secondary antibodies (Invitrogen) for 2 hr. The slides were again washed three times for 5 min in PBS, mounted with Vectashield + DAPI, and imaged on an epifluorescence microscope (Nikon Instruments).

Islet Isolation and Insulin Release Experiments
Islet isolation and insulin release studies were performed as previously described (Remedi et al., 2004). Briefly, after the mice were euthanized, the pancreases were removed and injected with HBSS (Sigma) containing collagenase (0.5 mg/ml, pH 7.4). They were digested for 4 min at 37 C, and the islets were isolated by hand. Following overnight incubation in 3 mM glucose CMRL-1066 medium, islets (10 per well in 12-well plates) were incubated for 60 min at 37 C in CMRL-1066 + 1 mM glucose, 23 mM glucose, 23 mM glucose + 1 mM glibenclamide, 1 mM + 30 mM KCl, or 23 mM glucose + 10 mM glutamine, as indicated. After the incubation period, the medium was removed and assayed for insulin release. Experiments were repeated in triplicate. To estimate the islet insulin content, groups of ten islets were disrupted by ethanol-HCl extraction and sonicated on ice. Insulin secretion and content were measured using a rat insulin radioimmunoassay (Millipore) according to the manufacturer's instructions.

Statistical Analyses
Statistical comparisons were made using ANOVA, chi-square test, or t test. All data are presented as means ± SEM, with a statistically significant difference defined as p < 0.05.

SUPPLEMENTAL INFORMATION
Supplemental Information includes three figures and four tables and can be found with this article online at http://dx.doi.org/10.1016/j.celrep.2014.05.017.