K_ATP channel deficiency in mouse FDB causes an impairment of energy metabolism during fatigue

Abstract

The skeletal muscle ATP-sensitive K⁺ (K_ATP) channel is crucial in preventing fiber damage and contractile dysfunction, possibly by preventing damaging ATP depletion. The objective of this study was to investigate changes in energy metabolism during fatigue in wild-type and inwardly rectifying K⁺ channel (Kir6.2)-deficient (Kir6.2^−/−) flexor digitorum brevis (FDB), a muscle that lacks functional K_ATP channels. Fatigue was elicited with one tetanic contraction every second. Decreases in ATP and total adenylate levels were significantly greater in wild-type than Kir6.2^−/− FDB during the last 2 min of the fatigue period. Glycogen depletion was greater in Kir6.2^−/− FDB for the first 60 s, but not by the end of the fatigue period, while there was no difference in glucose uptake. The total amount of glucosyl units entering glycolysis was the same in wild-type and Kir6.2^−/− FDB. During the first 60 s, Kir6.2^−/− FDB generated less lactate and more CO₂; in the last 120 s, Kir6.2^−/− FDB stopped generating CO₂ and produced more lactate. The ATP generated during fatigue from phosphocreatine, glycolysis (lactate), and oxidative phosphorylation (CO₂) was 3.3-fold greater in Kir6.2^−/− than wild-type FDB. Because ATP and total adenylate were significantly less in Kir6.2^−/− FDB, it is suggested that Kir6.2^−/− FDB has a greater energy deficit, despite a greater ATP production, which is further supported by greater glucose uptake and lactate and CO₂ production in Kir6.2^−/− FDB during the recovery period. It is thus concluded that a lack of functional K_ATP channels results in an impairment of energy metabolism.

muscle fatigue is defined as a transient and recoverable decline in muscle force or power with repeated or continuous muscle contractions. It is a myoprotective mechanism that preserves cellular integrity via regulation of ATP levels and myoplasmic Ca²⁺ concentration ([Ca²⁺]_i) and, thus, survival (28). One myoprotective component is the ATP-sensitive K⁺ (K_ATP) channel, which is closed by ATP in the micromolar range under patch-clamp conditions (5), while increases in intracellular ADP activate the channel (18). More importantly, a decrease in intracellular pH to a level mimicking the change during fatigue activates the channel in intact, unfatigued muscle fibers when ATP is in the millimolar range (16). Extracellular adenosine, which is produced in a greater amount as the workload increases (23), activates K_ATP channels (5). Finally, there is strong evidence for K_ATP channel activation during metabolic stress in skeletal muscle (2, 13, 17, 39, 40, 42).

Under low-intensity workloads, e.g., 1-Hz twitch contractions for 5–7 min, the resting membrane hyperpolarizes, while action potential amplitude is reduced; these changes do not occur when K_ATP channel activity is abolished (51). During high-intensity workloads leading to fatigue, the channel becomes crucial in preventing contractile dysfunction and fiber damage (12, 13, 24, 47, 49). This myoprotection provided by K_ATP channels is clearly observed at 37°C, but not at lower temperatures (17). However, temperatures less than 37°C are not physiological, because muscle temperature exceeds 37°C, even during moderate exercise (8). There is one physiological condition for which skeletal muscle does not depend on the channel myoprotection: when a first fatigue bout acutely reduces the dependence on K_ATP channels to prevent contractile dysfunctions during a subsequent second fatigue bout, a phenomenon called fatigue preconditioning (7). However, because the duration of this phenomenon is only 3 h, for most of the time, muscles depend largely on the myoprotective effect of the K_ATP channel.

Two mechanisms have been elucidated for K_ATP channel myoprotection. The first mechanism involves a reduction in action potential amplitude, as an increased K⁺ efflux reduces the extent of the depolarization by the Na⁺ influx (19, 27). It is well known that a K⁺-induced membrane depolarization reduces tetanic force and sarcomere shortening (26, 38) because of lower Ca²⁺ release (26), which is caused by lower action potential amplitude (38). Reduction of Ca²⁺ release even occurs with small decreases in action potential amplitude as a consequence of K_ATP channel activation (51). Thus, as K_ATP channels reduce action potential amplitude, less Ca²⁺ is released, which in turn diminishes the ATP requirement by the Ca²⁺-ATPase pump and by myosin ATPase as less actomyosin links are formed, preserving ATP.

The second mechanism involves the resting membrane potential (E_m). Resting E_m normally depolarizes by <15 mV during fatigue and metabolic inhibition; in the absence of channel activity, the membrane depolarizes by up to 50 mV (12, 20). Concomitant to the depolarization, force increases in the absence of stimulation, termed unstimulated force. A low concentration of verapamil, a blocker of the voltage-gated Ca_V1.1 channel, prevents not only the increase in unstimulated force, but also an increase in unstimulated [Ca²⁺]_i; it also improves the capacity of muscle to recover tetanic [Ca²⁺]_i and force following fatigue, probably by preventing Ca²⁺-induced fiber damage (12, 42). These studies suggest that the large depolarization in the absence of K_ATP channel activity stimulates Ca_V1.1 channels, allowing for a Ca²⁺ influx. More importantly, better maintenance of resting E_m by K_ATP channels prevents any excessive increase in unstimulated [Ca²⁺]_i and force and, thus, large ATP demand by Ca²⁺-ATPase and myosin ATPase.

Glucose uptake (in vivo, but not in vitro) and insulin sensitivity (in vivo and in vitro) are greater in unfatigued skeletal muscles from inwardly rectifying K⁺ channel (Kir6.2)-deficient (Kir6.2^−/−) or sulfonylurea receptor type 2 (SUR2^-/-) mice, two K_ATP channel knockout models, than in wild-type mice (11, 32). During metabolic inhibition, ATP levels decrease and ADP levels increase faster in the absence than in the presence of K_ATP channel activity, but the final extent of the ATP depletion is not different (20). During 3 min of fatigue, ATP levels decrease to the same extent in mouse extensor digitorum longus in the absence or presence of pinacidil, a K_ATP channel agonist, or glibenclamide, a channel antagonist; in mouse soleus, pinacidil prevents the decrease in ATP, while ATP losses are greater in the presence of glibenclamide than in control conditions (27). For the soleus muscle, however, a pinacidil or glibenclamide effect on mitochondrial ATP production could not be eliminated. Treadmill running causes greater O₂ deficit and lactate production in Kir6.2^−/− than wild-type mice, but Alekseev et al. (1) suggest that a deficient circulatory system, due to a K_ATP channel deficiency in cardiac muscles, prevented proper O₂ delivery to skeletal muscles. A lack of O₂ may actually be further compounded by extensive fiber damage in Kir6.2^−/− diaphragm that occurs during treadmill running (49), which, in this case, reduces blood oxygenation. There is therefore a lack of knowledge about the impact of the K_ATP channel on energy metabolism during fatigue in skeletal muscle.

The objective of this study was to test the hypothesis that an impairment of energy metabolism occurs during fatigue in K_ATP channel-deficient muscles. To test this hypothesis, we measured, at various times during a fatigue bout, how phosphocreatine (PCr) and ATP contents change, how many glucosyl units from glycogen and extracellular glucose enter glycolysis, and how much is converted to lactate and CO₂ during fatigue in flexor digitorum brevis (FDB) muscles. From these changes, we calculated the extent of ATP production. The results suggest that a lack of K_ATP channel activity during fatigue impairs energy metabolism in terms of ATP production.

METHODS

Animals and muscle preparation.

K_ATP channel-deficient FDB muscles were obtained from mice null for the Kir6.2 gene (Kir6.2^−/−), which encodes for the protein making the pore of the K_ATP channel (32, 33). Kir6.2^−/− mice were generated using the 129ySv mouse line, as described by Miki et al. (33), and had been cross-bred with wild-type C57BL/6 mice for at least five generations. C57BL/6 and Kir6.2^−/− mice were bred in-house and fed ad libitum (no fasting prior to any experiments, because wild-type and Kir6.2^−/− mice have different fasting phenotypes) (31). The animals were housed according to the guidelines of the Canadian Council for Animal Care, and all experimental procedures were approved by the Animal Care Committee of the University of Ottawa.

Two- to 3-mo-old mice (20–25 g body wt, 58% male and 42% female for wild-type and 40% male and 60% female for Kir6.2^−/−) were anesthetized by an intraperitoneal injection of ketamine (0.3 mg/10 g body wt), xylazine (0.07 mg/g body wt), and acepromazine (0.03 mg/g body wt). FDB muscles were excised from the hindpaw, and the mice were euthanized with an overdose of anesthetics. The FDB muscle is separated by fascia into fiber bundles, with each bundle controlling the movement of a digit. All fibers in the bundle responsible for movement of the fourth digit were excised by cutting along the lateral fascia separating the fiber bundles for the third and fourth digits.

Force measurement.

FDB bundles were constantly immersed in physiological saline solution containing (in mM) 118.5 NaCl, 4.7 KCl, 2.4 CaCl₂, 3.1 MgCl₂, 25 NaHCO₃, 2 NaH₂PO₄, and 5.5 d-glucose. All solutions were continuously bubbled with 95% O₂-5% CO₂ to maintain pH 7.4. All experiments were carried out at 37°C, because muscle temperature exceeds 37°C, even with a moderate workload (8). FDB bundles were attached vertically to a force transducer (model FT03, Grass Technologies, Warwick, MA) and placed in an acrylic cuvette containing 1.7 ml of physiological saline solution. The physiological saline solution was constantly bubbled with 95% O₂-5% CO₂ and changed every 15 min.

Force was measured in all FDB bundles to ensure that the force changes during fatigue were the same for all measured metabolites. FDB bundles were allowed 30 min to equilibrate after the muscle length had been adjusted to give maximum tetanic force. During that time, tetanic contractions were elicited every 100 s with a 200-ms train of 0.3-ms, 10-V (supramaximal) pulses at 200 Hz. For the 3-min fatigue period, the time interval between contractions was reduced to 1 s. For the recovery period, the contraction interval was returned to 100 s. Pulses were generated with a Grass S88 stimulator and a Grass SIU5 isolation unit. The stimulating platinum wires (4 mm apart) were located on opposite sides of the bundles.

Tetanic force was monitored using a force transducer (model FT03, Grass Technologies) connected to a physiograph (model 79E, Grass Technologies). Tetanic force was defined as the force generated upon stimulation and calculated as the difference between force at the maximum height of contraction and force just prior to the stimulation. As FDB bundles were flash-frozen for metabolite measurements, a separate experiment was carried out to determine whether tetanic force differed between wild-type and Kir6.2^−/− mice. For that experiment, tetanic force was normalized as follows: F_normalized = F_measured/CSA∗CF and CSA = L_e/W_e∗M_density, where F_normalized is the normalized force (in N/cm²), F_measured is the measured force (in g), CF is the converting factor (0.00980665 N/g), CSA is cross-sectional area (in cm²), L_e is experimental muscle length (in cm), W_e is muscle weight (in g), and M_density is muscle density, taken as 1.06 g/cm³. During fatigue, the baseline increased (Fig. 1A), especially in Kir6.2^−/− FDB. This force, termed unstimulated force (13), was calculated as the difference between force measured just before a stimulation and force measured just before fatigue was elicited.

Fig. 1.

Decrease in tetanic force and increase in unstimulated force were significantly greater in inwardly rectifying K⁺ channel (Kir6.2)-deficient (Kir6.2^−/−) than wild-type flexor digitorum brevis (FDB). Fatigue was elicited with 1 tetanic contraction every second for 3 min. A and B: fatigue traces for wild-type (A) and Kir6.2^−/− (B) FDB. Horizontal line, 2 g; vertical line, 20 s. C and D: changes in mean tetanic force during fatigue and after 15 min of recovery for wild-type (WT) and Kir6.2^−/− FDB. Tetanic force was measured as increase in force upon stimulation and is expressed as percentage of prefatigue force. E and F: mean unstimulated force during fatigue and after recovery, measured as the difference between force just prior to a tetanic contraction and force prior to fatigue and expressed as percentage of prefatigue tetanic force. Values are means ± SE of 10 FDB bundles; error bars smaller than symbols are not shown. *P < 0.05 vs. WT (by ANOVA with least significant difference test).

Metabolite measurements.

At specific times before, during, or after fatigue, FDB bundles were freeze-clamped with clamps precooled in liquid nitrogen and kept at −80°C until they were used. Usually, metabolite extraction starts by pulverization of muscle tissues in a mortar filled with liquid nitrogen. However, wet weight of the FDB bundles was 1–2 mg, which included a significant amount of tendons, which can cause erroneous metabolite determinations. Therefore, bundles were lyophilized overnight (Freezemobile 6, VirTis) to allow us to separate muscle fibers from tendons at room temperature, and only muscle tissues were weighed using an ultrasensitive (μg range) balance (model ME30, Mettler). PCr, ATP, ADP, AMP, IMP, and lactate were extracted by addition of 500 μl of ice-cold 6% perchloric acid (PCA) to 100–500 μg of dry fibers. After sonication (sonic dismembrator, model 100, Fisher Scientific) at maximum power for 15 s, solutions were centrifuged for 30 min at 1,000 g and 4°C (Micromax 3590F2282, International Equipment). Supernatants were neutralized with ice-cold 3 M K₂CO₃, and the K⁺ salt was centrifuged for 15 min at 1,000 g and 4°C. PCr and lactate were determined enzymatically as described by Passoneau and Lowry (37). Briefly, PCr was determined using a creatine kinase-hexokinase-glucose-6-phosphate dehydrogenase system, and lactate was determined using a lactate dehydrogenase-pyruvate-glutamate transaminase (to push the reaction toward pyruvate formation) system. The increase in fluorescence from NADPH (produced by glucose-6-phosphate dehydrogenase) and NADH (produced by lactate dehydrogenase) was measured using a fluorometer (model LS50B, Perkin Elmer). Lactate was also determined in the physiological saline solution, which was first lyophilized (1.6 ml) and reconstituted into 500 μl of water to concentrate lactate. Solutions were centrifuged at 1,000 g to remove insoluble salt, and lactate was measured as described above.

ATP, ADP, AMP, and IMP were analyzed on a tandem mass spectrometer (MS/MS, Xevo XE, Micromass, Manchester, UK) coupled with the Acquity ultraperformance liquid chromatography system (Waters, Milford, MA) for solvent delivery and sample introduction. MassLynx software (version 4.1) running in the Microsoft Windows XP professional environment was used to control the instruments and to acquire data. Adenosine-[¹³C₁₀]5′-triphosphate was used as internal standard (Cambridge Isotope Laboratories, Tewksbury, MA). Quantification was achieved using calibration curves that were generated by plotting the analyte-to-internal standard peak ratio against known concentrations.

The electrospray ionization source was operated in the negative ion mode at a capillary voltage of 3.0 kV. The instrument settings were as follows: source temperature 120°C, desolvation temperature 350°C, cone voltage 20 V, and collision energy 26 eV (with argon used as collision gas). Scanning was in the multiple reaction monitoring mode using transitions of mass-to-charge ratio of 506 to 159 for ATP, 426 to 159 for ADP, 346 to 79 for AMP, 347 to 79 for IMP, and 516 to 159 for adenosine-[¹³C₁₀]5′-triphosphate with a dwell time of 0.05 s.

Chromatographic separation was performed on a Hypercarb column (2.1 × 100 mm, 5 μm; Thermo Scientific). Mobile phase A was 10 mM ammonium acetate, which was adjusted with ammonia to pH 10, and mobile phase B was 100% acetonitrile. The gradient program was as follows: 0–2 min at 95% phase A, 2–3.5 min from 95% to 88% phase A, 3.5–7.5 min at 88% phase A, 8.5–12.5 min at 50% phase A, and return to 95% phase A to recondition the column for 5.5 min. The flow rate was 0.3 ml/min, and the column temperature was maintained at 35°C. The injection-to-injection time was 18 min.

For glycogen determination, lyophilized FDB tissue was added to 100 μl of 1 N NaOH and incubated at 80°C for 15 min to degrade endogenous glucose. After neutralization with 66 μl each of 0.25 M HCl and 0.15 M sodium acetate buffer, glycogen was hydrolyzed to glucose by addition of 6 μl of 1.5 g/l amyloglucosidase and incubation at 37°C for 4 h. Glucose content was determined enzymatically, as described by Passoneau and Lowry (37) using a hexokinase, and glucose 6-phosphate and NADPH were measured as described above.

For glucose uptake and oxidation, FDB bundles were first exposed for 10 min to a physiological saline solution with zero glucose to wash out the interstitial glucose before exposure to a normal physiological solution containing a glucose marker. Glucose uptake was measured using [2-³H]deoxyglucose ([2-³H]DG), a nonmetabolized glucose marker, as described by Hansen et al. (21), with a few modifications. Briefly, FDB bundles were exposed to 2.0 μCi/ml [2-³H]DG (Perkin Elmer, Woodbridge, ON, Canada) and 0.9 μCi/ml [¹⁴C]sucrose (ARC Chemicals, St. Louis, MO). Cold glucose was not replaced by cold 2-DG to prevent a decrease in glycolysis due to inhibition by 2-DG. [¹⁴C]sucrose was used to determine the extracellular volume, so that the extracellular glucose could be subtracted from [2-³H]DG in the total bundle. FDB fibers were freeze-clamped, lyophilized, separated from tendons, weighed, added to 0.5 ml of 6% PCA, and sonicated as described above for PCr, ATP, and lactate. Solutions were centrifuged for 15 min at 1,000 g and 4°C; 480 μl of supernatant were added to 10 ml of Ecolume scintillation fluid (MP Biomedicals, Burlingame, CA). [2-³H]DG and [¹⁴C]sucrose were counted using a WinSpectral liquid scintillation counter (model 1414, Perkin-Elmer Life Sciences, Boston, MA). Quenching was corrected by addition of 100 μl of the [2-³H]DG-[¹⁴C]sucrose physiological saline solution to 380 μl of PCA solution to give a final PCA concentration of 6%.

Glucose oxidation to CO₂ was measured by exposure of FDB bundles to 2 μCi/ml d-[6-¹⁴C]glucose. The muscle chamber was immediately covered with a piece of Plexiglas, and solution bubbling was immediately stopped to prevent loss of ¹⁴CO₂ after addition of d-[6-¹⁴C]glucose. After they were freeze-clamped, the bundles were quickly placed into rubber-capped vials and exposed to 1 ml of ice-cold 1 M sulfuric acid that was injected through the vial cap using a syringe. Liberated ¹⁴CO₂ was trapped inside the vial using 0.5-ml tubes containing 300 μl of benzethonium hydroxide. Vials were kept at 4°C to prevent further metabolic reactions. Physiological solutions were also transferred to rubber-capped vials and injected with ice-cold PCA to give a final concentration of 6%. Vials were kept at room temperature, and liberated ¹⁴CO₂ was captured as described above. All 0.5-ml tubes containing trapped ¹⁴CO₂ were collected after 2 h, and 10 ml of CytoScint liquid scintillation fluid (MP Biomedicals) were added; a different liquid scintillation fluid was used, because addition of benzethonium hydroxide to Ecolume scintillation fluid resulted in a very large amount of chemiluminescence. ¹⁴CO₂ was counted with a WinSpectral liquid scintillation counter. Quenching was corrected by counting 100 μl of the d-[6-¹⁴C]glucose physiological solution dissolved in benzethonium hydroxide. The remaining bundles were lyophilized, and fibers were separated from tendons and weighed as described above.

Indirect calorimetry of mice.

Whole body O₂ consumption (V̇o₂) and CO₂ production (V̇co₂) were measured using a four-chamber Oxy-MAX system (version 5.11, Columbus Instruments, Columbus, OH). Mice were fed ad libitum and isolated in 2.5-liter chambers for 24 h. Mice were allowed 4 h of acclimatization before V̇o₂ and V̇co₂ were sampled for 1 min with a settle time of 2 min and an airflow rate of 0.5 l/min. All recording started in the afternoon. The temperature was maintained at 24°C, and lights were on from 0700 to 1900.

Fiber typing.

FDB fibers were embedded in a small amount of Tissue-Tek optimum cutting temperature compound (OCT, Sakura Finetek, Torrance, CA) and then frozen in isopentane precooled in liquid nitrogen. Muscles were stored at −80°C until analysis. A cryostat (Microm HM 500M, Leica) cooled to −18°C was used to cut serial 10-μm-thick cross sections from the midbelly of each muscle. Sections were mounted on Superfrost Plus slides. Myosin isoforms were stained using the Mouse on Mouse kit (Vector Laboratories) and the following primary antibodies (all from Developmental Studies Hybridoma Bank, Iowa City, IA): mouse IgM anti-type I myosin (catalog no. A4.840), mouse IgG anti-type IIA myosin (catalog no. SC71), mouse IgM anti-type IIX myosin (catalog no. 6A1), and mouse IgM anti-type IIB myosin (catalog no. BFF3).

Statistical analysis.

Values are means ± SE. Student's t-test was used to determine significant differences in fiber types. Split-plot ANOVA designs were used for force and V̇o₂ measurements, because measurements at different times came from the same muscle or mouse. Two-way ANOVA was used to determine significant differences for metabolite measurements: one factor was the mouse model, and the other was the time at which metabolites were measured. ANOVA was calculated using the general linear model (version 9.2) procedures of Statistical Analysis Software (SAS Institute). When a main effect or an interaction was significant, the least significant difference method was used to locate the significant differences (45). “Significant” refers only to a statistical difference (P < 0.05).

RESULTS

Fiber typing.

Energy use by cross bridges, or contractile ATP cost, varies three- to fourfold among the different fiber types in mammals (6, 46). It was therefore important to determine whether expression of the myosin isoform differed between wild-type and Kir6.2^−/− FDB. Here, we determined the proportion of fibers that express each of the four myosin isoforms. A large proportion of FDB fibers from wild-type C57BL/6 mice expressed myosin IIA (57%) and IIX (61%), while fewer fibers expressed myosin I (11%) and almost none (1%) expressed myosin IIB (Table 1). These differences in myosin isoform expression are similar to those previously reported for FDB muscles of CD-1 mice (3). Fewer Kir6.2^−/− FDB fibers expressed myosin I, 10% more expressed myosin IIA, and 25% more expressed myosin IIX than were expressed in C57BL/6 FDB.

Table 1.

Proportion of fibers expressing type I, IIA, IIB, and IIX myosin

	FDB
Fiber Type	Wild-type	Kir6.2−/−
I	10.5 ± 0.8	6.6 ± 0.7*
IIA	56.7 ± 4.0	65.2 ± 1.0*
IIB	1.0 ± 0.6	0.8 ± 0.5
IIX	60.5 ± 3.2	85.5 ± 3.3*

Values [means ± SE of 7 flexor digitorum brevis (FDB) muscles] represent the proportion of fibers that expressed each of the 4 myosin isoforms as a percentage of the total number of fibers, and not the fiber type per se, as many fibers expressed >1 myosin isoform. Proportion of fibers expressing type I, IIA, and IIB myosin was the same between wild-type and inwardly rectifying K⁺ (Kir6.2) channel-deficient (Kir6.2^−/−) FDB, while proportion of fibers expressing type IIX myosin was significantly greater in Kir6.2^−/− FDB.

^*P < 0.05 vs. wild-type (by t-test).

Tetanic and unstimulated force.

There was no significant difference in the initial tetanic force between wild-type and Kir6.2^−/− FDB: 29.3 ± 0.9 and 31.0 ± 0.9 N/cm², respectively (n = 5 muscles). Most of the decrease in tetanic force occurred during the first 60 s of fatigue for wild-type and Kir6.2^−/− FDB (Fig. 1). The decreases in mean tetanic force at 30 and 40 s were 10% greater for Kir6.2^−/− than wild-type FDB (Fig. 1C). By the end of the fatigue period, the final extent of the force decrease was not different between wild-type and Kir6.2^−/− FDB. After fatigue, FDB fibers were allowed 15 min of recovery, during which tetanic force returned toward prefatigue force; the recovery was significantly greater in wild-type than Kir6.2^−/− FDB: 90% vs. 79%. Unstimulated force occurred when FDB bundles failed to fully relax between contractions and was observed as an increase in baseline, especially for Kir6.2^−/− FDB (Fig. 1, A and B). In wild-type FDB, mean unstimulated force increased by ≤2% of the prefatigue tetanic force during the entire 180-s fatigue bout (Fig. 1D). In contrast, Kir6.2^−/− FDB displayed a sharp increase in mean unstimulated force to 30% by 30 s, decreasing to 22% at 60 s, where it plateaued until the end of the fatigue bout. After 15 min of recovery, unstimulated force of Kir6.2^−/− FDB was still significantly above prefatigue level by 4%. The transient faster decrease in tetanic force and greater increase in unstimulated force during fatigue, as well as the lower capacity to recover tetanic force following fatigue, in Kir6.2^−/− FDB were similar to previous findings (13) and are not further discussed.

Metabolite measurement.

Immediately after excision, FDB fibers released a large amount of lactate in the physiological solution (41). It was therefore important to remove this lactate prior to fatigue. Regardless of the metabolite being measured, the physiological solution was changed twice during the equilibrium preceding fatigue: after 15 min and again at 30 min. For ATP, PCr, glycogen, and lactate measurements, fatigue was elicited 10 min after a third change of the physiological saline solution. For glucose uptake and ¹⁴CO₂ production the fatigue was preceded by a 10-min wash, during which FDB fibers were exposed to physiological solution with no glucose, and a 10-min incubation in the presence of [2-³H]DG or [6-¹⁴C]glucose and 5 mM glucose.

PCr and adenylates.

Prior to fatigue, mean PCr levels were 11 μmol/g dry wt less in Kir6.2^−/− than wild-type FDB (Fig. 2A). When fatigue was elicited, the difference had disappeared by 30 s as PCr was depleted. Prior to fatigue, mean ATP levels were not significantly different between wild-type and Kir6.2^−/− FDB (Fig. 2B). In wild-type FDB, mean ATP contents fell significantly by 5.1 μmol/g dry wt during the first 30 s but returned to prefatigue levels by 120 s. In Kir6.2^−/− FDB, ATP levels also significantly dropped by 5.0 μmol/g dry wt during the first 30 s but, unlike wild-type FDB, continued to decrease throughout the fatigue period, becoming significantly less than in wild-type FDB by 180 s. Mean ADP contents did not change significantly in wild-type FDB (Fig. 2C). In Kir6.2^−/− FDB, mean ADP contents decreased from 60 to 180 s, becoming significantly less than at time 0 by 1.8 μmol/g dry wt at 180 s. Mean AMP content did not change significantly during fatigue and was not significantly different between wild-type and Kir6.2^−/− FDB (Fig. 2D). Mean IMP levels increased significantly during the first 30 s for wild-type and Kir6.2^−/− FDB, by 3.26 and 5.22 μmol/g dry wt, respectively (Fig. 2E). Thereafter, IMP remained elevated throughout the fatigue period. There was no significant difference between wild-type and Kir6.2^−/− FDB. Mean total adenylate contents, representing the sum of the ATP, ADP, AMP, and IMP contents, did not change significantly during fatigue in wild-type FDB (Fig. 2F). However, mean total adenylate content decreased from 60 to 180 s, becoming significantly less than at time 0 at 180 s.

Fig. 2.

ATP levels were significantly less in Kir6.2^−/− than wild-type FDB for most of the fatigue period. A–F: phosphocreatine (PCr), ATP, ADP, AMP, IMP, and total adenylate contents. Fatigue was elicited with 1 tetanic contraction every second for 3 min. Values are means ± SE of 5–7 FDB bundles. §P < 0.05 vs. 0 s; *P < 0.05 vs. WT (by ANOVA with least significant difference test).

Glucose uptake and glycogen breakdown.

Next, we measured how much glucose was transported from the extracellular space and how much glycogen was hydrolyzed to estimate glucosyl units entering glycolysis. Glucose uptake was measured using [2-³H]DG, a glucose marker, because, once phosphorylated, it becomes trapped in the sarcoplasm. However, 2-DG-P_i cannot be further metabolized and can inhibit hexokinase (21). To allow for normal glycolysis during fatigue, we could not replace 5 mM cold glucose with 5 mM cold 2-DG, as is normally done (32). GLUT1 and GLUT4 are the two glucose transporters expressed in skeletal muscle. To ensure that the affinity of both transporters is the same for glucose and 2-DG, we measured the [2-³H]DG uptake in the presence of 5 mM glucose or 5 mM 2-DG. One group of FDB muscles was from mice that had been fasted for 16 h to minimize the cell membrane GLUT4 content to best estimate glucose and 2-DG transport by GLUT1; another group of FDB muscles was exposed for 1 h to 1,000 μU/ml insulin to maximally increase the sarcolemmal GLUT4 content (32). As expected, insulin significantly increased [2-³H]DG uptake compared with control. However, [2-³H]DG uptake under control or insulin conditions was not significantly different between physiological solution containing 5 mM glucose and physiological solution containing 2-DG (Fig. 3). This suggests that GLUT1 and GLUT4 have similar affinity for glucose and 2-DG.

Fig. 3.

Uptake of [2-³H]deoxyglucose ([2-³H]DG) was the same whether physiological saline solution contained 5 mM glucose or 2-DG. FDB bundles from 16-h-fasted wild-type mice were incubated for 10 min with 2 μCi/ml [2-³H]DG and 0.9 μCi/ml [¹⁴C]sucrose in the presence of 5 mM glucose or 2-DG. Intracellular [2-³H]DG was calculated by subtraction of extracellular from total [2-³H]DG. When present, 1,000 μU/ml insulin was added 60 min prior to uptake measurements. Values are means ± SE of 5–6 FDB bundles. *P < 0.05 vs. control (by ANOVA with least significant difference test). There was no significant difference between glucose and 2-DG conditions for control and insulin conditions (P > 0.05).

At rest, glucose uptake was slightly, but not significantly, greater in wild-type than Kir6.2^−/− FDB: 12.1 vs. 11.2 μmol/g dry wt (Fig. 4). At the end of fatigue, glucose uptake was 20.6 μmol/g dry weight for Kir6.2^−/− FDB compared with 18.0 μmol/g dry weight for wild-type FDB, a difference that was, again, not significantly different. Glucose uptakes associated with the increased muscle activity, calculated from the difference in mean uptake before and after fatigue, were 5.9 and 9.3 μmol/g dry wt for wild-type and Kir6.2^−/− FDB, respectively, representing a difference of 3.4 μmol/g dry wt (Fig. 4B). However, if we consider that the mean glucose uptakes in unfatigued and fatigued FDB fibers were not different between wild-type and Kir6.2^−/− mice, it is unlikely that this difference in activity-dependent glucose uptake is significant. During the subsequent recovery, glucose uptake was significantly greater in Kir6.2^−/− than wild-type FDB: 33.3 vs. 27.0 μmol/g dry wt. The glucose uptake measured at the same time from unfatigued wild-type FDB was not much lower than that from fatigued FDB, whereas the difference was much larger for Kir6.2^−/− FDB.

Fig. 4.

Glucose uptake was not significantly different between wild-type and Kir6.2^−/− FDB during fatigue but was significantly greater in Kir6.2^−/− during recovery. A: glucose uptake was measured by incubation of FDB bundles in the presence of 2 μCi/ml [2-³H]DG as a glucose marker, 5 mM glucose, and 0.9 μCi/ml [¹⁴C]sucrose (to calculate extracellular volume and intracellular [2-³H]DG). In all experiments, [2-³H]DG and [¹⁴C]sucrose were added 10 min prior to elicitation of fatigue. One group of FDB fibers was freeze-clamped at the end of the fatigue (FAT) period (180 s); a second group was freeze-clamped at the same time, but without elicitation of fatigue, representing prefatigue uptake (0 s); a third group was freeze-clamped after 3 min of fatigue and 15 min of recovery (REC); and a fourth group was freeze-clamped at the same time as those allowed to recover for 15 min, but without a fatigue bout (NO FAT). Values are means ± SE of 4 FDB bundles. B: activity-dependent glucose uptake, calculated as the difference in mean uptake between 180 and 0 s for fatigue and between NO FAT and REC for the recovery period. SE was calculated as follows: √(S1²/n1) + (S2₂/n2), where S1 and S2 represent SE and n1 and n2 are sample sizes of each of the 2 means. §P < 0.05 vs. 0 s of fatigue; *P < 0.05 vs. WT (by ANOVA with least significant difference test).

Prefatigue glycogen content in terms of glucosyl units was 130 μmol/g dry wt for Kir6.2^−/− FDB, which was significantly greater than 112 μmol/g dry wt for wild-type FDB (Fig. 5A). After 60 s of fatiguing stimulation, wild-type FDB had lost 4 μmol/g dry wt compared with 10 μmol/g dry wt lost by the Kir6.2^−/− mice, a 2.5-fold difference (Fig. 5B). At the end of fatigue, glycogen content was still significantly greater in Kir6.2^−/− than wild-type FDB: 90 vs. 69 μmol/g dry wt (Fig. 5A). Interestingly, the final extent of the glycogen breakdown over 180 s became slightly less in Kir6.2^−/− FDB, which used 40 μmol/g dry wt, compared with wild-type FDB, which used 43 μmol/g dry wt (Fig. 5B). After 15 min of recovery, the glycogen level of wild-type FDB returned to 83 μmol/g dry wt, which was not significantly different from the 87 μmol/g dry wt measured at the same time but without a fatigue bout. Glycogen recovery, which reached 116 μmol/g dry wt, was significantly greater in Kir6.2^−/− than wild-type FDB. Kir6.2^−/− FDB had a greater glycogen store at rest and mobilized glycogen sooner than wild-type FDB but did not use more glycogen, despite lower ATP levels and greater unstimulated force.

Fig. 5.

Kir6.2^−/− FDB contained more glycogen than wild-type FDB, but the amount used during 180 s of fatigue was the same in wild-type and Kir6.2^−/− FDB. A: one group of FDB fibers was freeze-clamped after 60 s of fatiguing stimulation; a second group was freeze-clamped after 180 s; a third group was freeze-clamped at the same time as those at 180 s, but without elicitation of fatigue for prefatigue glycogen content (0 s); a fourth group was freeze-clamped after 3 min of fatigue and 15 min of recovery (REC); and a fifth group was freeze-clamped at the same time as those allowed to recover for 15 min, but without a fatigue bout (NO FAT). Values are means ± SE of 5–7 FDB bundles. B: activity-dependent glycogen breakdown during fatigue, calculated by subtraction of mean glycogen content at 0 s from mean content at 60 and 180 s. SE was calculated as shown in Fig. 4 legend. §P < 0.05 vs. 0 s; *P < 0.05 vs. WT (by ANOVA with least significant difference test).

Lactate and ¹⁴CO₂ generation.

To determine the fate of glucose entering glycolysis, we next measured lactate and ¹⁴CO₂ generation in FDB muscles and physiological solutions. After 60 s, mean lactate contents of wild-type and Kir6.2^−/− FDB were 56 and 49 μmol/g dry wt, respectively (Fig. 6A). With the prefatigue lactate content taken into consideration, the activity-dependent lactate production was 41 and 24 μmol/g dry wt for wild-type and Kir6.2^−/− FDB, respectively, representing a 1.7-fold difference (Fig. 6B). In wild-type FDB, lactate content increased much less during the last 120 s than during the first 60 s. Interestingly, the increases in lactate content during the last 120 s became greater in Kir6.2^−/− than wild-type FDB: 19 vs. 9 μmol/g dry wt. Lactate content continued to increase during the 15 min of recovery, reaching 181 μmol/g dry wt in Kir6.2^−/− FDB, which was significantly greater than 138 μmol/g dry wt in wild-type FDB.

Fig. 6.

Kir6.2^−/− FDB generated less lactate than wild-type FDB during the first 60 s, and the reverse was observed during the last 120 s. Lactate was measured in FDB muscles and physiological solution. A: total lactate content, calculated from the sum of lactate content in FDB muscles and physiological solutions. Muscle sampling at 0, 60, and 180 s of fatigue (FAT) and after 15 min of recovery with (REC) or without a fatigue bout (NO FAT) are described in Fig. 5 legend. Values are means ± SE of 5 FDB bundles. B: activity-dependent lactate production, calculated by subtraction of mean total lactate content between 0 and 60 s and 60 and 180 s and NO FAT and REC. SE was calculated as shown in Fig. 4 legend. §P < 0.05 vs. 0 s of fatigue; *P < 0.05 vs. WT (by ANOVA with least significant difference test).

For measurement of glucose oxidation, it was necessary to stop the bubbling of 95% O₂-5% CO₂ to prevent loss of ¹⁴CO₂. The amount of lactic acid generated by wild-type FDB was the same whether the muscles were fatigued with or without bubbling; i.e., there was no evidence for greater O₂ deficiency in the absence of bubbling (41). For wild-type FDB, ¹⁴CO₂ content did not change during fatigue (Fig. 7A). ¹⁴CO₂ content, however, significantly increased during the first 60 s in Kir6.2^−/− FDB by 1.9 μmol/g dry wt but failed to increase further during the last 120 s (Fig. 6B). During recovery, Kir6.2^−/− FDB significantly generated 3.0 μmol ¹⁴CO₂/g dry wt compared with 1.8 μmol ¹⁴CO₂/g dry wt generated by wild-type FDB, a 1.7-fold difference. Thus, Kir6.2^−/− FDB generated less lactate and more CO₂ than wild-type FDB during the first 60 s, while during the last 120 s, they generated more lactate and no CO₂.

Fig. 7.

More glucose was oxidized in Kir6.2^−/− than wild-type FDB during fatigue and recovery. Glucose oxidation was measured by incubation of FDB bundles in the presence of 2 μCi/ml d-[6-¹⁴C]glucose 10 min prior to elicitation of fatigue. ¹⁴CO₂ was measured in FDB fibers and physiological solution. A: total CO₂ content, calculated from the sum of ¹⁴CO₂ content in FDB muscles and physiological solutions. Muscle sampling at 0, 60, and 180 s of fatigue (FAT) and after 15 min of recovery with (REC) or without a fatigue bout (NO FAT) is described in Fig. 5 legend. Values are means ± SE of 5 FDB bundles. B: activity-dependent lactate production, calculated by subtraction of mean total CO₂ content between 0 and 60 s and 60 and 180 s and NO FAT and REC content. SE was calculated as shown in Fig. 4 legend. §P < 0.05 vs. 0 s of fatigue. *P < 0.05 vs. WT (by ANOVA with least significant difference test).

Whole body V̇o₂ and V̇co₂ in mice.

The activity-dependent ¹⁴CO₂ production in Kir6.2^−/− FDB suggests that Kir6.2^−/− FDB fibers may rapidly increase oxidative phosphorylation during fatigue. One study has reported greater V̇o₂ in Kir6.2^−/− than wild-type mice (1). Similarly, over a 24-h period, V̇o₂ was greater in Kir6.2^−/− than wild-type mice (Fig. 8A). We also determined V̇co₂ to calculate the respiratory exchange ratio (RER). Mean RER values were significantly lower in Kir6.2^−/− than wild-type mice when they were the least active, i.e., between 0800 and 1600 (Fig. 8B), which indicates a greater proportion of lipids than carbohydrates when mice are mainly inactive and not eating. The reverse situation was observed when mice became active, as RER was higher in Kir6.2^−/− than wild-type mice between 0000 and 0400, as well as between 1700 and 2200. This greater RER is indicative of a greater proportion of carbohydrate than fat oxidation at these times, when mice are physically active and eating.

Fig. 8.

Kir6.2^−/− mice have greater O₂ uptake (A) and lower respiratory exchange ratio (RER, B) over a 24-h period. White and black horizontal bars represent day (lights-on) and night (lights-off), respectively. Values are means ± SE of 5 mice. *P < 0.05 vs. WT (by ANOVA with least significant difference test).

DISCUSSION

The major findings of this study were as follows: compared with wild-type FDB and during fatigue, 1) ATP and total adenylate levels were significantly lower in Kir6.2^−/− FDB, 2) glucose uptake was similar in Kir6.2^−/− FDB, 3) glycogen depletion was greater at 60 s, but not at 180 s, 4) less lactate was generated and more glucose was oxidized during the first 60 s, while for the last 120 s, more lactate was generated but no further glucose was oxidized, 5) during recovery, glucose uptake, lactate and ¹⁴CO₂ generation, were significantly greater in Kir6.2^−/− FDB, and 6) Kir6.2^−/− mice had a wider range of RER.

Importance of myosin expression and myosin ATPase activity.

The myofibrillar ATPase activity during isometric contractions of myosin I is ∼25% of myosin IIX, while myosin IIA and IIB ATPase are 94% and 129%, respectively, of myosin IIX (6). The proportion of fibers expressing myosin IIB was very low (<1%) and not different between wild-type and Kir6.2^−/− FDB. On the other hand, myosin IIA and IIX were expressed in greater proportion in Kir6.2^−/− than wild-type FDB fibers, the differences being 10% and 25%, respectively. Such large increases, however, are unlikely due to a decrease in myosin I content, because 1) only 10% of FDB fibers expressed myosin I in wild-type mice and 2) the proportion was only 4% less in Kir6.2^−/− FDB fibers. It is therefore more likely that the increases in the proportion of fibers expressing myosin IIA and IIX resulted in an increase in the proportion of hybrid fibers expressing both isoforms. Considering the small differences in 1) myosin I isoform expression between wild-type and Kir6.2^−/− FDB and 2) the myofibrillar ATPase activity between myosin IIA and IIX, we suggest that any differences in energy metabolism (see below) are not related to any major difference in ATP costs during isometric contractions between wild-type and Kir6.2^−/− FDB.

Differences in metabolite content between resting wild-type and Kir6.2^−/− FDB.

Under resting conditions, glucose uptake was not significantly different between wild-type and Kir6.2^−/− FDB, as previously reported for in vitro basal glucose uptake between K_ATP channel-deficient (from Kir6.2^−/− and SUR2^−/− mice) and wild-type muscles (11, 32). Kir6.2^−/− FDB had significantly lower PCr content. These differences cannot be explained from the results of this study but may be related to Ca²⁺ metabolism. During moderate and high muscle activity, [Ca²⁺]_i tends to be higher in Kir6.2^−/− than wild-type muscles, possibly because of greater action potential amplitude in moderate exercise (51) or greater Ca²⁺ influx through Ca_V1.1 channels during fatigue (12, 42). The Kir6.2^−/− mouse is also associated with “excessive body heat production with low-intensity physical activity” (51). Taken together, these results suggest higher energy consumption in Kir6.2^−/− muscles, as supported by higher V̇o₂ (present study and Ref. 1), the final result being lower PCr levels.

Resting Kir6.2^−/− FDB fibers had more glycogen and better recovered their glycogen reserve after a fatigue bout than wild-type FDB. One possibility for this difference is that mice were not synchronized with respect to time of feeding. Such synchronization would require a fasting period followed by access to food prior to an experiment. However, because Kir6.2^−/− mice do not properly regulate their blood glucose level during fasting due to a defect in glucagon secretion (31), this approach was avoided to prevent an effect on the fatigue and metabolic kinetics. It is unlikely that a difference in feeding time is the major cause for the higher glycogen content in Kir6.2^−/− FDB, as both wild-type and Kir6.2^−/− mice are most active at night, as supported by the greater V̇o₂ at night. Furthermore, food consumption is similar in wild-type and Kir6.2^−/− mice (31).

A more likely explanation for the greater glycogen content in Kir6.2^−/− FDB is that in vivo glucose uptake is greater in Kir6.2^−/− than wild-type muscles (32), allowing for more glycogen storage. A second explanation may involve Ca²⁺. 5′-AMP-activated protein kinase (AMPK) can be activated by Ca²⁺-dependent calmodulin protein kinase-β (22). Activation of AMPK during a metabolic stress is expected to reduce glycogen synthesis and increase glycogen breakdown (25, 34, 50). However, the R225W mutation in the human AMPK γ₃-subunit results in a gain of function and is associated with increases in glycogen and decreases in triglyceride content in skeletal muscle (14, 15). If [Ca²⁺]i tends to be higher in Kir6.2^−/− than wild-type muscles, as discussed above, then one possible mechanism responsible for the greater glycogen in Kir6.2^−/− FDB is a higher level of Ca²⁺-activated AMPK, which, during nonmetabolic stress, favors glycogen synthesis, possibly because AMPK also increases glucose uptake (15, 34).

Glucose uptake and glycogen breakdown during fatigue.

Glucose uptake increased during the fatigue bout, and the calculated activity-dependent glucose uptake was greater in Kir6.2^−/− FDB. However, considering that glucose uptake was not significantly different between wild-type and Kir6.2^−/− FDB prior to and after 3 min of fatigue, one must be careful about a higher activity-dependent glucose uptake in Kir6.2^−/− FDB. Also, the increase in glucose uptake with muscle activity in this study was less than that previously reported, even during recovery. Hansen et al. (21) reported a nearly threefold increase in glucose uptake in epitrochlearis and soleus muscles following muscle activity. However, in their study, muscle activity consisted of 10-s contractions every minute for 15 min for a total of 150 s of contraction time compared with 200-ms contractions every second for 3 min, or 36 s of contraction time, in the present study. So, the lower glucose uptake in FDB is most likely due to less time or less muscle activity to translocate GLUT4 to the cell membrane.

During the first 60 s of fatigue, glycogen breakdown was 2.6-fold greater in Kir6.2^−/− than wild-type FDB. Large increases in unstimulated force occur within 20 s of fatigue (Fig. 1), which is most likely due to large increases in [Ca²⁺]_i (12, 13). Ca²⁺ is a known activator of glycogen phosphorylase (9); thus it is very likely that higher [Ca²⁺]_i resulted in greater activation of glycogen phosphorylase. The unexpected result was the lack of greater glycogen depletion in Kir6.2^−/− FDB during the last 120 s of fatigue, which was 29 μmol/g dry wt compared with 39 μmol/g dry wt in wild-type FDB. The lower glycogen mobilization in Kir6.2^−/− than wild-type FDB during the last 120 s, despite greater glycogen content prior to fatigue and at 60 s, is surprising for two reasons. 1) ATP demand is much greater because of the greater unstimulated [Ca²⁺]_i (13, 42) and force (present study) in Kir6.2^−/− than wild-type FDB. 2) It is also a time when ATP level continuously fell, instead of rising, as it did in wild-type muscle, which implies a greater energy deficit in Kir6.2^−/− FDB. In other words, a greater energy demand and lower ATP level should have resulted in greater glycogen mobilization in Kir6.2^−/− FDB. The lack of greater glycogen mobilization cannot be explained by glycogen content being too low, since at 60 and 180 s it was still greater in Kir6.2^−/− than wild-type FDB. We therefore suggest that, during the last 2 min of the fatigue period, glycogen mobilization in Kir6.2^−/− muscles became insufficient, possibly due to some impairment by a mechanism that remains to be elucidated.

With the activity-dependent glucose uptake and glycogen breakdown taken into account, it was possible to calculate the amount of glucosyl units entering glycolysis, which was 48.8 and 49.2 μmol/g dry wt for wild-type and Kir6.2^−/− FDB, respectively (Fig. 9A). Thus, despite the greater ATP demand associated with greater unstimulated [Ca²⁺]_i and force, Kir6.2^−/− FDB failed to mobilize more glucose.

Fig. 9.

Kir6.2^−/− and wild-type FDB used a similar amount of glucose, but Kir6.2^−/− mice generated more ATP. A: portions of glucosyl units entering glycolysis from extracellular glucose (GLU) and glycogen breakdown (GLY). B: fate of glucosyl units, in terms of amount converted to lactate (LAC) and amount oxidized to CO₂. Unexplained portion was calculated as the difference between the total amount of glucose entering glycolysis and the amount converted to lactate and oxidized. C: calculated amount of ATP produced from PCr and glucose when converted to lactate or CO₂ (see discussion for calculations). It is important to note that ATP production by oxidative phosphorylation (calculated from ¹⁴CO₂ production) occurred only during the first 60 s of fatigue, because as shown in Fig. 7, ¹⁴CO₂ did not increase further during the last 120 s.

Lactate and ¹⁴CO₂ generation: glucose fate.

Lactate content prior to and during fatigue was not significantly different between wild-type and Kir6.2^−/− FDB. However, most of the increase in lactate occurred during the first 60 s in wild-type FDB, while it was more gradual over the fatigue period in Kir6.2^−/− FDB. More importantly, ¹⁴CO₂ from [6-¹⁴C]glucose significantly increased during the first 60 s in Kir6.2^−/− FDB, while it remained unchanged in wild-type FDB. Furthermore, while daily locomotor activity and dwell time at rest are indistinguishable (1), daily V̇o₂ is greater in Kir6.2^−/− than wild-type mice (present study and Ref. 1). Finally, mean RER values were often higher in active Kir6.2^−/− than wild-type mice, reaching a value of 1 in some cases during activity at night, suggesting that Kir6.2^−/− mice favor glucose oxidation to a greater extent. We therefore suggest that Kir6.2^−/− FDB increased glucose oxidation within 60 s during our fatigue stimulation, which resulted in a greater proportion of pyruvate entering the tricarboxylic acid (Krebs) cycle and a 1.7-fold smaller lactate production.

Because the amount of ¹⁴CO₂ produced by Kir6.2^−/− FDB did not further increase from 60 to 180 s, the only possible fate for pyruvate became a conversion to lactate, which was twofold greater in Kir6.2^−/− than wild-type FDB. As discussed above for glycogen mobilization, this apparent sudden stop in ¹⁴CO₂ production during the last 120 s was also surprising, because it occurred when ATP demand was still elevated and ATP level was constantly dropping. One possible explanation for these results is the occurrence of an anoxic core after the first 60 s as O₂ is depleted, especially in the center of the FDB bundles. However, for 1- to 2-mg wet wt FDB bundles, this anoxic core is expected to be relatively small (4) and cannot fully explain a complete cessation of glucose oxidation. Another possibility is an impairment of oxidative phosphorylation. A small increase in [Ca²⁺]_i activates mitochondria (36), and this may have been a mechanism whereby glucose oxidation was increased during the first 60 s, as the increase in unstimulated force during that time is related to an increase in [Ca²⁺]_i, as discussed above. Excessive increases in [Ca²⁺]_i (>1 μM), on the other hand, suppress the activity of isolated mitochondria (36). Large mitochondrial membrane depolarizations and, possibly, impairment were reported following ischemic reperfusion, as unstimulated [Ca²⁺]_i increased from 0.2 to 0.7 μM in cardiac myocytes (48). Large increases in [Ca²⁺]_i do not occur in all fibers, as in one study, 55% of the tested Kir6.2^−/− FDB fibers partially or fully supercontracted as [Ca²⁺]_i reached levels similar to or higher than those during tetanic contractions (13). It is therefore possible that the lack of further increases in ¹⁴CO₂ during the last 2 min of fatigue was related to Ca²⁺ inhibition of mitochondrial activity in many fibers, constituting a second metabolic impairment.

The impairment was, however, transient, because ¹⁴CO₂ production became greater in Kir6.2^−/− than wild-type FDB during recovery. If impairment of oxidative phosphorylation during the last 120 s was due to high [Ca²⁺]_i, then the higher ¹⁴CO₂ production by Kir6.2^−/− FDB during recovery suggests that the Ca²⁺ effect was reversible and the mitochondrial impairment stopped as [Ca²⁺]_i decreased (as expected from the change in unstimulated force during recovery; see Fig. 1). It then allowed oxidative phosphorylation to become greater in Kir6.2^−/− FDB, possibly to recover from a greater energy deficit (see Energy balance during fatigue). Therefore, our results suggest that the K_ATP channel indirectly protects mitochondria from Ca²⁺ overload and function impairment in skeletal muscle as suggested by Storey et al. for cardiac myocytes (48).

Overall, during the 180 s of fatigue, wild-type FDB converted 52% of the glucose entering glycolysis to lactic acid and only 1% to CO₂, leaving 47% (or 23 μmol/g dry wt) of unaccounted glucose (Fig. 9B). Intermittent stimulation of human and rat muscles increases glycolytic intermediate content from 5–7 to 24–50 μmol/g dry wt (10, 43, 44). Therefore, it is suggested that the unaccounted glucose is mostly glycolytic intermediates. Kir6.2^−/− FDB converted less glucose to lactate (44%) and more to CO₂ (16%) than wild-type FDB, leaving a smaller proportion of unaccounted glucose (40%, or 20 μmol/g dry wt).

Energy balance during fatigue.

As previously reported, PCr levels decreased early during fatigue (29, 30). Here, we report that the decreases in Kir6.2^−/− FDB were not different from wild-type FDB. The ATP content of wild-type FDB decreased significantly to 65% of prefatigue levels during the first 30 s of fatigue, returned to 88% of prefatigue level by 120 s, and ended at 84% at 180 s. The subsequent increase in ATP while muscles continue to be stimulated (present study and Ref. 30) is interesting. The most likely explanation for the increase in ATP is a reduced ATP demand due to less Ca²⁺ released and force generation to the point that it becomes less than ATP production. Also, despite ATP recovery during fatigue, there is no sign of force recovery, suggesting that recovery of a normal ATP level is not sufficient for force recovery while muscles are continuously stimulated.

A similar 69% decrease in ATP content was also observed during the first 30 s for Kir6.2^−/− FDB. However, unlike wild-type FDB, the initial decrease in ATP was not followed by a return to prefatigue levels. Instead, ATP content continued to decrease after 60 s, reaching 38% of the prefatigue content by 180 s. Undoubtedly, the greater unstimulated [Ca²⁺]_i and force in Kir6.2^−/− FDB prevented the necessary decrease in ATP demand that allows for its recovery in wild-type FDB. We also suggest that the impairment of glycogen mobilization and oxidative phosphorylation, as discussed above, also contributed to the lack of ATP recovery, although Kir6.2^−/− FDB generated more ATP during the fatigue period than its wild-type counterpart (see below).

We also observed that the further loss of ATP during the last 2 min of fatigue in Kir6.2^−/− FDB was not associated with an increase in IMP content. Instead, there was a significant loss of adenylates (for reasons that cannot be explained from the data obtained in the present study). Thus the loss of adenylates constitutes another metabolic dysfunction, in addition to the dysfunctional glycogen mobilization and oxidative phosphorylation.

The amount of ATP generated from PCr was higher in wild-type than Kir6.2^−/− FDB: 57 vs. 44 μmol/g dry wt (Fig. 9C). This difference was due to significantly lower PCr content in Kir6.2^−/− FDB prior to fatigue, and not to a different final PCr level at the end of fatigue. The next source of ATP is conversion of glucose to lactate. Taking into consideration 1 ATP per lactate from glucose and 1.5 ATP per glucose moiety from glycogen and using the proportion of glycogen to extracellular glucose contributing to the amount of glucosyl units entering glycolysis, we estimated that wild-type FDB generated 36 μmol/g dry wt ATP compared with 30 μmol/g dry wt ATP in Kir6.2^−/− FDB.

The ATP generated from extracellular glucose oxidation was directly calculated from the ¹⁴CO₂ produced from d-[6-¹⁴C]glucose, but the ATP generated from glycogen cannot. The amount of glucosyl units entering glycolysis from glycogen was 7.2 and 4.3 times greater than the amount from extracellular glucose for wild-type and Kir6.2^−/− FDB, respectively. If it is assumed that the amounts of pyruvate from glycogen entering the tricarboxylic acid cycle were also 7.2 and 4.3 times greater than from extracellular glucose, then the calculated CO₂ from glycogen was 0.4 and 6.3 μmol/g dry wt for wild-type and Kir6.2^−/− FDB, respectively. With 36 ATP per CO₂ used for glucose and 39 ATP per CO₂ used for glycogen, the estimated ATP production from glucose oxidation was notably greater in Kir6.2^−/− than wild-type FDB: 300 vs. 19 μmol/g dry wt. The estimated total ATP production was 374 μmol/g dry wt for Kir6.2^−/− FDB compared with only 113 μmol/g dry wt for wild type FDB, a 3.3-fold difference. However, it must be kept in mind that generation of all 300 μmol ATP/g dry wt by oxidative phosphorylation (and, thus, most of the generated 374 μmol of ATP) occurred only during the first 60 s, as ¹⁴CO₂ production stopped thereafter. More importantly, production of 374 μmol of ATP still appears insufficient, because Kir6.2^−/− FDB fibers failed to recover their ATP during the 2nd min of fatigue. Our results, therefore, suggest that a greater energy deficit developed in Kir6.2^−/− than wild-type FDB. The notion of a greater energy deficit in Kir6.2^−/− FDB is further supported by the fact that glucose uptake, lactate and ¹⁴CO₂ generation, were greater in Kir6.2^−/− than wild-type fibers during recovery.

In conclusion, during the first 60 s of a fatiguing stimulation of one tetanic contraction per second, the K_ATP channel-deficient Kir6.2^−/− FDB was better at generating ATP, as it mobilized more glycogen and rapidly increased oxidative phosphorylation activity compared with wild-type FDB. However, during the last 120 s of fatigue, glycogen mobilization and oxidative phosphorylation in Kir6.2^−/− FDB appeared to become impaired, despite a lower level of ATP, significantly more glycogen, and greater energy demand due to greater unstimulated [Ca²⁺]_i and force in Kir6.2^−/− than wild-type FDB. Furthermore, there was a significant loss of adenylates. Our results, therefore, support the notion of Noma (35) that the K_ATP channel is important for the protection of energy metabolism. As a consequence of an impaired energy metabolism in the absence of K_ATP channel activity, Kir6.2^−/− skeletal muscles cannot sustain muscle activity as long as wild-type muscles, as they eventually and rapidly generate a greater energy deficit. This is another reason why Kir6.2^−/− mice are less efficient runners than wild-type mice (49).

GRANTS

This study was supported by grants from the National Science and Engineering Research Council of Canada to J.-M. Renaud and M.-E. Harper.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.S., N.D.C., L.Z., O.Y.A.-D., and J.-M.R. developed the concept and designed the research; K.S., M.B.J., N.D.C., M.P., L.Z., and O.Y.A.-D. performed the experiments; K.S., M.B.J., N.D.C., M.P., L.Z., O.Y.A.-D., and J.-M.R. analyzed the data; K.S., M.B.J., N.D.C., M.P., L.Z., M.-E.H., O.Y.A.-D., and J.-M.R. interpreted the results of the experiments; K.S., M.B.J., L.Z., and J.-M.R. prepared the figures; K.S., M.B.J., L.Z., and J.-M.R. drafted the manuscript; K.S., M.B.J., L.Z., M.-E.H., and J.-M.R. edited and revised the manuscript; K.S., M.B.J., N.D.C., M.P., L.Z., M.-E.H., and J.-M.R. approved the final version of the manuscript.