Ablating the protein TBC1D1 impairs contraction-induced sarcolemmal glucose transporter 4 redistribution but not insulin-mediated responses in rats

TBC1 domain family member 1 (TBC1D1), a Rab GTPase-activating protein and paralogue of Akt substrate of 160 kDa (AS160), has been implicated in both insulin- and 5-aminoimidazole-4-carboxamide ribonucleotide formyltransferase/IMP cyclohydrolase-mediated glucose transporter type 4 (GLUT4) translocation. However, the role of TBC1D1 in contracting muscle remains ambiguous. We therefore explored the metabolic consequence of ablating TBC1D1 in both resting and contracting skeletal muscles, utilizing a rat TBC1D1 KO model. Although insulin administration rapidly increased (p < 0.05) plasma membrane GLUT4 content in both red and white gastrocnemius muscles, the TBC1D1 ablation did not alter this response nor did it affect whole-body insulin tolerance, suggesting that TBC1D1 is not required for insulin-induced GLUT4 trafficking events. Consistent with findings in other models of altered TBC1D1 protein levels, whole-animal and ex vivo skeletal muscle fat oxidation was increased in the TBC1D1 KO rats. Although there was no change in mitochondrial content in the KO rats, maximal ADP-stimulated respiration was higher in permeabilized muscle fibers, which may contribute to the increased reliance on fatty acids in resting KO animals. Despite this increase in mitochondrial oxidative capacity, run time to exhaustion at various intensities was impaired in the KO rats. Moreover, contraction-induced increases in sarcolemmal GLUT4 content and glucose uptake were lower in the white gastrocnemius of the KO animals. Altogether, our results highlight a critical role for TBC1D1 in exercise tolerance and contraction-mediated translocation of GLUT4 to the plasma membrane in skeletal muscle.

Skeletal muscle glucose uptake in response to either insulin or contraction is a highly regulated process that results in facilitated diffusion through glucose transporter 4 (GLUT4). 3 The effects of these metabolic stimuli are additive (1,2), ultimately resulting in the translocation of GLUT4 from intracellular depots to the sarcolemmal membrane. Akt2 (also known as protein kinase B (PKB)) is a proximal member of the insulin signaling cascade, and its phosphorylation is an essential step in insulin-stimulated glucose transport (3), whereas contractionmediated translocation of GLUT4 is believed to occur via a distinct molecular mechanism (2, 4 -7). Although the terminal downstream factors of both of these cascades are currently not fully understood, the binding of Rab-GTPases to GLUT4-containing vesicles is thought to be critical for translocation in both processes (8).
Rab GTPase-activating proteins (GAPs) are members of the Ras superfamily of monomeric G-proteins that regulate numerous membrane trafficking events, including vesicle formation and movement along actin and tubulin networks, as well as membrane fusion. Two protein members of the TBC1 (Tre-2, BUB2, CDC16) domain family member 1 (TBC1D1) and TBC1D4 (also known as AS160) are functional GAPs that are believed to prevent GLUT4 translocation by maintaining Rabs in their inactive GDP-bound state (8). Although AS160 and TBC1D1 are only 61% similar in sequence over their entire length, their Rab GAP domains are 91% similar and have the same Rab substrate specificity (9). Upstream intracellular signaling events (e.g. insulin stimulation or contraction) result in the phosphorylation of AS160 and TBC1D1 and binding of these proteins to the molecular adaptor protein 14-3-3. Ultimately, these events enable GAPs to return to an active GTPbound state and initiate vesicle exocytosis and trafficking of GLUT4 transporters (8). Work in cell lines has shown that deletion of TBC1D1 increases GLUT4 translocation to the plasma membrane (10), whereas overexpression inhibits it (9,11), further supporting the notion that TBC1D1 plays a role in the control of GLUT4 translocation. Paradoxically, results in animals lacking TBC1D1 show no difference in whole-body glucose or insulin tolerance, whereas in vivo maximal insulin-stimulated skeletal muscle glucose uptake is impaired (12,13), which is contradictory to the proposed mechanism of action for this protein in glucose handling and GLUT4 trafficking. However, both congenic mice lacking TBC1D1 (12) and conventional TBC1D1 knock-out (KO) mice (13,14) display an ϳ50% reduction in total GLUT4 protein, which by itself has been shown to impair in vitro glucose uptake without altering whole-body glucose tolerance (15). Therefore, one possible explanation for the decrease in glucose transport in TBC1D1-deficient mice is the decrease in GLUT4 protein content, making interpretations on the functional role of TBC1D1 with respect to intracellular trafficking difficult. Nevertheless, the decreased GLUT4 content suggests that TBC1D1 influences glucose homeostasis and may represent a compensatory attempt by the cell to maintain glucose homeostasis and limit glucose flux into the muscle.
More recently, work in both rodents (16 -18) and humans (17,19) has shown that TBC1D1 may play a critical role in the signaling cascade in response to skeletal muscle contraction with TBC1D1 a downstream target of 5Ј-AMP-activated protein kinase (AMPK). Contraction, but not insulin stimulation, has been shown to increase binding of 14-3-3 to TBC1D1 in an AMPK-dependent manner in both rodents (16) and humans (17), a process that has been shown to be critical in the regulation of GLUT4 trafficking by AS160 (20,21). Furthermore, phosphorylation of TBC1D1 via AMPK leads to a greater decrease in electrophoretic mobility compared with Akt, suggesting that AMPK is a more robust regulator of this protein than insulin (18).
Although the precise role of TBC1D1 in glucose transport appears uncertain, one consistent finding among cell and animal studies is its effect on lipid oxidation. Several models have shown alterations in lipid oxidation as a result of TBC1D1 deficiency. Specifically, TBC1D1 ablation in vitro and in vivo increases fatty acid oxidation (13,22), and conversely, overexpression of TBC1D1 reduces skeletal muscle palmitate oxidation (22,23) and ␤-hydroxyacyl-CoA dehydrogenase (␤-HAD) activity (23). However, the mechanism of action resulting in up-regulated fatty acid oxidation remains unknown. The fatty acid transporter FAT/CD36 has been shown to translocate in response to both insulin and contraction (24) and is critical in the regulation of skeletal muscle fuel selection (25), but whether this occurs via a vesicle-dependent mechanism involving TBC1D1 remains to be determined.
Altogether, state-of-the art studies have provided compelling evidence that TBC1D1 ablation impairs glucose uptake and increases fatty acid oxidation. However, the current KO mouse models display a marked reduction in GLUT4 content, which by itself may also convey a similar phenotype (i.e. impaired glucose uptake, greater reliance on fatty acids, and impairment in exercise capacity), making interpretations difficult. Therefore, we aimed to characterize the role of TBC1D1 in a previously unexamined KO rat model that lacks a compensatory reduction in GLUT4 content. Using this model, we provide evidence that TBC1D1 regulates contraction but not insulin-mediated GLUT4 translocation in glycolytic muscles and subsequently decreases exercise capacity. These results appear specific to GLUT4 as FAT/CD36 trafficking was not altered. Taken together, these results provide support for a role of TBC1D1 in intracellular GLUT4 trafficking in response to contraction but not insulin stimulation.

Tissue screen and animal characterization
We first characterized TBC1D1 KO animals to determine whether they displayed phenotypic differences compared with wild-type (WT) animals (Table 1). There were no differences in animal or tissue weights or in heart size normalized to tibial length. TBC1D1 deficiency was confirmed by screening for TBC1D1 protein in skeletal muscle homogenate. KO animals did not express TBC1D1 protein in red or white muscle; however, they had similar levels of AS160, Akt2, and FAT/CD36 (Fig. 1A). We also assessed GLUT4 protein content, which was not different compared with WT rats in red or white muscle (Fig. 1, B and C).

Skeletal muscle response to insulin
Given the unaltered GLUT4 content in this KO model and the proposed role of TBC1D1 in glucose homeostasis, we first examined the effect of ablating this protein on glucose handling at the whole-body level. In support of previous work demonstrating that TBC1D1 influences beta cell mass/stability (26,27), TBC1D1 KO animals displayed impaired glucose clearance during an intraperitoneal glucose tolerance test (GTT) ( Fig. 2A) as demonstrated by an increase in the area under the curve (p Ͻ 0.05; Fig. 2B). However, there was no difference between TBC1D1 WT and KO animals during an intraperitoneal insulin tolerance test (ITT), suggesting that peripheral insulin sensitivity is intact in this model (Fig. 2, C and D). Consistent with similarwhole-bodyinsulinsensitivity,insulinincreasedthephosphorylation of Akt2 by ϳ3-fold in both red and white muscles in both WT and KO animals (Fig. 3, A and B). Total muscle

TBC1D1 is required for contraction-induced GLUT4 trafficking
GLUT4 expression and insulin-mediated GLUT4 trafficking in giant sarcolemmal vesicles was also unaltered between genotypes ( Fig. 3, C and D). Collectively, these data suggest that TBC1D1 is not required for insulin-stimulated glucose uptake in skeletal muscle.

Substrate utilization
As alterations in TBC1D1 protein have been associated with changes in fat metabolism, we next examined fuel utilization in vivo. In the present study and consistent with previous reports in murine models (13,22), TBC1D1 KO rats displayed a

TBC1D1 is required for contraction-induced GLUT4 trafficking
decreased respiratory exchange ratio (RER), and therefore calculated fat oxidation was increased whereas carbohydrate oxidation was concomitantly decreased in both the light and dark cycle (Fig. 4, A-C; p Ͻ 0.05). These responses represent a shift in fuel preference as there was no change in total energy expenditure (Fig. 4D). The increase in fatty acid oxidation was also observed in ex vivo experiments in isolated EDL muscle strips. Specifically, palmitate oxidation was increased by 36% ( Fig. 4E; p Ͻ 0.05). In contrast, there was no change in glucose oxidation ( Fig. 4F) in KO animals.

Mitochondrial respiration
Given this alteration in skeletal muscle fatty acid oxidation, we next examined whether the KO animals displayed a change in mitochondrial content or respiratory function. TBC1D1 KO animals displayed increased ADP-stimulated oxygen consumption in permeabilized muscle fibers ( Fig. 5A; p Ͻ 0.05); however, there was no change in lipid-supported respiration (palmitoyl-CoA (P-CoA); Fig. 5A, right y axis). To determine whether this functional increase was due to a change in mitochondrial protein content, we digested the fibers used for oxygen consumption measurements for assessment via Western blotting. Similar to previous work in TBC1D1 KO mice (14), mitochondrial content in both digested fibers (Fig. 5B) and muscle homogenate (Fig. 5, C and D) was similar between WT and KO animals. Furthermore, there was no difference in basal ␤-HAD enzyme activity in white (WT, 5.10 Ϯ 1.04 mol⅐min Ϫ1 ⅐g Ϫ1 wet weight; KO, 4.33 Ϯ 0.63 mol⅐min Ϫ1 ⅐g Ϫ1 wet weight; n ϭ 6, p ϭ 0.30) or red muscle (WT, 14.56 Ϯ 1.77 mol⅐min Ϫ1 ⅐g Ϫ1 wet weight; KO, 13.95 Ϯ 1.37 mol⅐ min Ϫ1 ⅐g Ϫ1 wet weight; n ϭ 6, p ϭ 0.77) between genotypes. To further examine the effects of ablating TBC1D1 on oxidative phosphorylation, we determined the kinetic properties of ADP to stimulate respiration in the presence of saturating substrates (Fig. 6, A and C). Maximal respiration was higher in KO animals when ADP was titrated both with and without creatine (ϳ37%, p Ͻ 0.05 versus WT; Fig. 6, A and C, respectively); however, this did not alter estimated ADP sensitivity as assessed by the apparent K m (Fig. 6, B and D, respectively). Similarly, maximal respiration in KO animals was also increased (ϳ30%, p Ͻ 0.05) when pyruvate was titrated in the presence of saturating ADP concentrations with no effect on the apparent K m (Fig. 6, E and F).

Skeletal muscle contraction
We next sought to determine whether the up-regulation of fatty acid oxidation seen at both the whole-body and skeletal muscle levels in KO animals could be explained by altered substrate delivery and/or metabolic inflexibility in response to contraction. Following electrical stimulation, AMPK and ACC were phosphorylated regardless of genotype or muscle (Fig. 7, A  and B), confirming successful contraction of the muscles and suggesting that these signaling cascades remained intact in the absence of TBC1D1. We therefore isolated giant sarcolemmal vesicles for the determination of plasma membrane transport-

TBC1D1 is required for contraction-induced GLUT4 trafficking
ers. Although FAT/CD36 increased in red muscle following contraction (Fig. 7C), there was no difference between genotypes and no effect of contraction in either group in white mus-cle (Fig. 7D). This suggests that the increase in whole-body fatty acid oxidation in KO animals is not due to increased transport into skeletal muscle. We next sought to determine whether

TBC1D1 is required for contraction-induced GLUT4 trafficking
there was an impairment in contraction-mediated glucose uptake and GLUT4 trafficking. Electrically induced contraction resulted in a significant increase in 2-deoxyglucose (2-DG) uptake in both genotypes in red and white muscles (p Ͻ 0.05; Fig. 8, A and B); however, compared with WT, this effect was significantly reduced in KO animals in white muscle (p Ͻ 0.05; Fig. 8B). The increase in glucose uptake from basal levels (⌬ change) following stimulation was also significantly reduced in white muscle (p Ͻ 0.05; Fig. 8A, inset) and tended to be lower in red muscle (p ϭ 0.07; Fig. 8B, inset) of KO animals. These effects were mirrored in sarcolemmal GLUT4 protein content as there was a main effect for contraction to increase GLUT4 in red muscle with a strong trend for a decreased response to contraction in KO animals (p ϭ 0.10; Fig. 8C, inset). In addition, both absolute plasma membrane GLUT4 content and the ability of contraction to stimulate plasma membrane GLUT4 content were lower in white muscle of KO animals following con-traction (Fig. 8, D and inset), further suggesting that ablation of TBC1D1 impairs intracellular trafficking and glucose uptake. Given this finding, we analyzed glycogen utilization during contraction to determine whether there was a compensatory increase in endogenous substrate utilization. However, there was no change in glycogen content or in skeletal muscle metabolite production ( Table 2) between genotypes following contraction.

Time to exhaustion
Given our finding of reduced GLUT4 trafficking and glucose uptake in KO animals in response to contraction and the prominent role of exogenous substrates during contraction in rats, we next examined whether this would result in impaired running performance. Consistent with reductions in glucose uptake in situ, running time decreased in TBC1D1 KO animals by ϳ18% at low intensity and ϳ25% at the higher intensity (Fig. 9).

Discussion
In this study, we utilized a novel TBC1D1 KO rat model to elucidate the role of this protein in intracellular GLUT4 trafficking. We have shown that, in contrast to the TBC1D1-deficient murine model, there is no decrease in total GLUT4 protein content or impairment in GLUT4 accumulation on the plasma membrane of muscle in response to insulin. These data therefore suggest that TBC1D1 is not required for insulin-mediated GLUT4 trafficking. In contrast, we provide evidence that contraction-mediated glucose uptake and GLUT4 translocation in KO animals are impaired in TBC1D1 KO animals, and this is associated with compromised exercise capacity. These findings demonstrate that TBC1D1 plays an important role in contraction-mediated but not insulin-stimulated GLUT4 translocation in skeletal muscle.

TBC1D1 KO animals display intact insulin signaling
In the current study, we have shown that TBC1D1 ablation results in impaired glucose tolerance, which is consistent with previous reports implicating this protein in maintaining beta cell mass (26,27). However, the current work also demonstrates that TBC1D1 is not required for intracellular GLUT4 trafficking in response to insulin as there were no differences in ITT or in total or phosphorylated (Ser-473) Akt2 or plasma membrane GLUT4 content. Our findings are supported by previous work showing that mutation of the canonical Akt phosphorylation motif on TBC1D1 (Thr-596) to Ala had no effect on insulinstimulated glucose transport (28). We have also shown that total GLUT4 protein is unaltered between WT and KO rats, which is critical to the interpretations of the functional role of TBC1D1. This is in contrast to previous work utilizing TBC1D1-deficient mice that showed impaired insulin-stimulated glucose transport (12,13). However, these studies were confounded by an ϳ50% reduction in GLUT4 content, thus making it difficult to ascertain the exact role of TBC1D1 in glucose transport. It is therefore possible that the reduced glucose uptake in TBC1D1-deficient mouse models is a consequence of lowered GLUT4 content as opposed to a direct role of TBC1D1 in mediating GLUT4 trafficking.

TBC1D1 is required for contraction-induced GLUT4 trafficking Increased fatty acid oxidation in TBC1D1 KO animals
We have shown that ablation of TBC1D1 resulted in an increase in whole-body fat oxidation, which occurred in the absence of any change in energy expenditure. This is consistent with previous studies utilizing TBC1D1-deficient mice (13,22) as well as cell line models with reduced TBC1D1 (22). We have also shown an increase in palmitate oxidation in KO animals in isolated EDL strips; however, there was no change in glucose oxidation between genotypes. This suggests that, in the absence of other substrates, the ability to utilize fatty acids is increased in TBC1D1 KO animals, whereas the basal capacity to oxidize carbohydrate is not impaired.
To elucidate the mechanisms underlying the change in substrate utilization, we evaluated a proximal step in lipid oxidation, FAT/CD36 trafficking to the plasma membrane. FAT/ CD36 has previously been shown to respond to similar metabolic stimuli as GLUT4, including insulin and skeletal muscle contraction (24), and is a key component for the regulation of skeletal muscle fuel selection (25). It was therefore possible that TBC1D1 plays an important role in intracellular trafficking of this transporter and that ablation of the TBC1D1 gene would result in a compensatory increase in localization of FAT/CD36 to the plasma membrane. However, there was no difference between KO and WT animals in FAT/CD36 protein content measured in sarcolemmal vesicles at rest (p ϭ 0.45) and following contraction in red or white muscle. This therefore indicates that increases in fatty acid transport cannot explain the shift in fuel selection in these animals.
Previous work utilizing TBC1D1 overexpression in mice showed reduced activity of ␤-HAD concomitant with a decrease in fatty acid oxidation (23). However, neither ␤-HAD activity nor rates of P-CoA supported mitochondrial respiration were altered in KO rats and therefore cannot explain the observed increase in fatty acid oxidation seen at the wholebody or muscle level. In contrast, TBC1D1 KO animals display increased mitochondrial respiration in the absence of any change in mitochondrial protein content, suggesting improved respiratory function. This cannot be explained by increased sensitivity to substrates as there was no change in the apparent K m for pyruvate or for ADP either in the presence or absence of creatine. It is therefore possible that the increase in mitochondrial function is a compensatory adaptation in response to TBC1D1 ablation; however, further work is required to elucidate the exact mechanism behind this. An increase in maximal mitochondrial respiration supports our finding of increased reliance on fatty acid oxidation and may indicate tighter metabolic control as increased respiration, and therefore ATP production, should result in decreased accumulation of mitochondrial-free ADP, although this remains speculative and requires further examination.

Impaired contraction-mediated GLUT4 translocation
Consistent with previous reports, we have shown that run time to exhaustion is impaired in TBC1D1 KO animals (14). However, in the previous study, total GLUT4 protein was decreased, again making it difficult to determine whether the impairment in exercise capacity was due to a decrease in transporter protein content or the ablation of TBC1D1. In the current study, we have shown that total GLUT4 protein content was not reduced in the KO animals, and therefore the exercise impairment is likely a result of decreased contraction-mediated GLUT4 trafficking and therefore decreased glucose uptake. This clearly highlights TBC1D1 as having a critical role in glucose homeostasis during exercise, especially given the increased reliance on exogenous glucose as a fuel source in rodents. Interestingly, the decrease in contraction-mediated glucose uptake and plasma membrane GLUT4 content did not result in increased intramuscular glycogen breakdown or an increase in cellular energy perturbation as measured by metabolite accumulation. This suggests that despite the apparent improvement in mitochondrial respiratory function TBC1D1 KO animals are not able to adequately compensate for the decrease in glucose uptake to rescue exercise capacity.

Conclusion
Altogether, we have shown that TBC1D1 KO rats displayed increased mitochondrial respiratory function, which may explain the increase in fatty acid oxidation seen both ex vivo and at the whole-body level. Despite these alterations, KO animals displayed decreased exercise capacity likely due to the decrease in GLUT4 translocation and glucose uptake seen following contraction. Combined with the findings that TBC1D1 expression is highly specific to skeletal muscle (11,18) and appears to be more robustly regulated by AMPK than Akt-mediated phosphorylation (18), these results support the theory that TBC1D1 is downstream of AMPK and represents a likely terminus for the contraction-mediated signaling cascade responsible for GLUT4 translocation.

Chemicals
All chemicals were purchased from Sigma-Aldrich unless stated otherwise.

TBC1D1 is required for contraction-induced GLUT4 trafficking
Rodents TBC1D1 KO rats on a Sprague-Dawley background were provided by Dr. Hamra (29). Animals were housed in a temperature-and humidity-regulated room on a 12/12-h light-dark cycle. Rats were maintained on a chow diet (Harlan Laboratories Inc.) with ad libitum access to food and water. All protocols were approved by and performed in accordance with the Committee on Animal Care guidelines at the University of Guelph.

Glucose and insulin tolerance tests
Four-hour-fasted rats underwent intraperitoneal GTT (2 g/kg of body weight) and ITT (1 unit/kg of body weight; Humulin, Lilly) tolerance tests. Blood glucose was measured using a glucose meter (Freestyle Lite, Abbott Diabetes Care) for both GTT and ITT before and up to 120 and 45 min post-injection, respectively.

Indirect calorimetry
Oxygen consumption (VO 2 ), carbon dioxide (VCO 2 ), and RER were measured over a 24-h period in a CLAMS (Comprehensive Laboratory Animal Monitoring System) open circuit system (Columbus Instruments). Fat and carbohydrate oxidation and total energy expenditure were calculated as described previously (30).

Substrate oxidation
To determine the rates of substrate oxidation, glucose and palmitate oxidation assays were performed in EDL muscle strips as described previously (31,32). Briefly, palmitate oxidation was measured in pregassed (95% O 2 , 5% CO 2 ) modified Krebs-Henseleit buffer supplemented with 4% BSA and palmitate (0.5 mM; 0.5 Ci/l [1-14 C]palmitate) (PerkinElmer Life Sciences) at 30°C for 40 min by measuring the 14 CO 2 released in addition to determining the water-soluble 14 C-labeled intermediates. Glucose oxidation was determined in a separate set of experiments as described previously (33). Briefly, muscle strips were incubated in modified Krebs-Henseleit buffer supplemented with 5.5 mM glucose and 0.2 Ci D-[U-14 C]glucose (American Radiolabeled Chemicals) at 30°C for 40 min. Gaseous 14 CO 2 was captured as above and counted using standard liquid scintillation techniques. Substrate oxidation rates were calculated using the specific activity of labeled palmitate or glucose within the Krebs-Henseleit buffer and normalized to muscle weight.

Skeletal muscle response to insulin
To determine skeletal muscle-specific insulin signaling, rats were anesthetized (sodium pentobarbital; 60 mg/kg of body weight), and then muscle samples were excised before (basal) and 30 min after an intravenous insulin injection (1 unit/kg of body weight). Harvested tissues were collected for plasma membrane preparations or rapidly frozen in liquid nitrogen for Western blotting as described below.

Electrical stimulation-induced muscle contraction
For muscle contraction experiments, the sciatic nerve of the experimental limb was exposed, and stimulating electrodes were placed around the nerve. Electrical stimulation was applied for 15 min (train delivery, 100 Hz/3 s at 6 -8 V; train duration, 200 ms; pulse duration, 10 ms). Following stimulation, both experimental and control hind limb muscles (red and white gastrocnemius and red and white tibialis anterior) were excised and either immediately frozen in liquid nitrogen for further analysis or used for the preparation of giant sarcolemmal vesicles.

In vivo glucose clearance
Muscle-specific glucose uptake was performed by tail vain injection. Specifically, 50 Ci of tritiated 2-DG was injected 5 min before commencing 15 min of sciatic nerve stimulation (as described above). Thereafter, muscle was immediately taken and rates of muscle-specific glucose uptake were determined by analyzing radioactivity from red and white gastrocnemius as described previously (31,32).

Muscle enzymatic activity
Red and white tibialis anterior muscles were excised from both experimental and control limbs following stimulation experiments and immediately frozen in liquid nitrogen. A small piece of frozen muscle (6 -10 mg) was taken for the spectrophotometric determination of ␤-HAD at 37°C as described previously (33).

Glycogen and metabolite analysis
Red and white gastrocnemius muscles from both the control and experimental limb in stimulation experiments was freezedried, dissected free of visible blood and connective tissue, and powdered for metabolite and glycogen analyses. An aliquot of freeze-dried muscle (10 -12 mg) was extracted with 0.5 M perchloric acid (HClO 4 ) containing 1 mM EDTA and neutralized with 2.2 M KHCO 3 . The supernatant was used to determine creatine, PCr, ATP, and lactate with enzymatic spectrophotometric assays (33,34). Muscle glycogen content was determined from an aliquot of freeze-dried muscle (2-3 mg) as described elsewhere (33).

Isolation of plasma membranes
Giant sarcolemmal vesicles were prepared as described previously (24). Briefly, red and white gastrocnemius muscles excised following contraction and insulin stimulation experiments were cut into thin strips and incubated (1 h, 34°C, 100 rpm) in 140 mM KCl, 10 mM MOPS (pH 7.4) containing collagenase VII (150 units/ml) and aprotinin (10 mg/ml). The muscle was washed with KCl/MOPS containing 10 mM Na 2 EDTA (pH 7.4), and the supernatants were combined. Percoll (3.5%), KCl (28 mM), and aprotinin (10 g/ml) were added to the supernatant. This solution was layered under 3 ml of 4% Nycodenz and 1 ml of KCl/MOPS and centrifuged (60 ϫ g, 45 min, 25°C). The vesicles were harvested from the interface of Nycodenz and KCl/MOPS and pelleted by centrifugation (9000 ϫ g, 10 min, 25°C), and the resulting pellet was resuspended in KCl/MOPS and stored at Ϫ80°C for Western blotting.

Preparation of permeabilized fibers
A small portion of muscle was placed in ice-cold BIOPS (50 mM MES, 7.23 mM K 2 EGTA, 2.77 mM CaK 2 EGTA, 20 mM im-TBC1D1 is required for contraction-induced GLUT4 trafficking idazole, 0.5 mM DTT, 20 mM taurine, 5.77 mM ATP, 15 mM PCr, and 6.56 mM MgCl 2 ⅐H 2 O (pH 7.1)) and separated under a microscope into bundles using fine-tipped forceps as described previously (35). Fiber bundles were then treated with 30 g/ml saponin for 30 min at 4°C and then washed for 15 min in MiR05 respiration buffer (0.5 mM EGTA, 10 mM HH 2 PO 4 , 110 mM sucrose, 1 mg/ml fatty acid-free BSA (pH 7.1)) for respiration analysis as described below.

Mitochondrial respiration
Measurements of O 2 consumption were performed in MiR05 respiration buffer on prepared permeabilized fibers using an Oxygraph-2K (Oroboros Instruments, Innsbruck, Austria) at 37°C in the presence of 25 M blebbistatin as described previously (36). ADP-stimulated respiration was determined in the presence of 5 mM pyruvate and 2 mM malate as well as 60 M P-CoA. Maximal state III respiration was determined by the addition of 5 mM ADP with 10 mM glutamate and 10 mM succinate added to determine maximal complex I and maximal complex I ϩ II respiration, respectively. Uncoupled respiration was measured in the presence of 50 M 2,4-dinitrophenol. Mitochondrial kinetics in permeabilized fibers were analyzed using three separate titration protocols. ADP kinetics were analyzed both in the presence and absence of 20 mM creatine. Both ADP titrations were initiated with 2 mM malate and 10 mM pyruvate, and ADP was titrated at various concentrations. Pyruvate kinetics were determined in the presence of 5 mM ADP, 5 mM malate, and 20 mM creatine, and pyruvate was titrated in at various concentrations. Then 10 mM glutamate and 10 mM succinate were added following the pyruvate and ADP titrations to determine maximum mitochondrial respiration. Finally, 10 M cytochrome c was added with Ͻ10% increase in respiration in all measurements indicating outer mitochondrial membrane integrity. The apparent K m values for pyruvate and ADP were determined as described previously (35). All fibers were recovered from the respirometer, freezedried, and weighed for normalization to muscle bundle weight. Freeze-dried fibers were then digested in 5 g⅐l lysis buffer as described previously (37) for determination of mitochondrial proteins via Western blotting.

Time to exhaustion
Animals were familiarized to the treadmill at slow speeds (10 m⅐min Ϫ1 ) for 3 consecutive days followed by a 48-h rest before the start of the exercise bout. Animals completed a time to exhaustion trial at two different intensities (low; 20 m⅐min Ϫ1 , 5% grade, and high, 20 m⅐min Ϫ1 , 15% grade) separated by Ն48 h.

Statistics
Results are expressed as means Ϯ S.E. and were analyzed by Student's t test or one-or two-way analysis of variance with a Fisher's least significant difference post hoc test where appropriate. Significance was set at p Ͻ 0.05 where NS indicates not significant. All graphs and statistical analyses were performed using GraphPad Prism 5.0 (GraphPad Software).