The transport of cytosolic adenosine diphosphate (ADP) into the mitochondria is a major control point in metabolic homeostasis, as ADP concentrations directly affect glycolytic flux and oxidative phosphorylation rates within mitochondria. A large contributor to the efficiency of this process is thought to involve phosphocreatine (PCr)/Creatine (Cr) shuttling through mitochondrial creatine kinase (Mi-CK), whereas the biological importance of alterations in Cr-independent ADP transport during exercise remains unknown. Therefore, we utilized an Mi-CK knockout (KO) model to determine whether in vivo Cr-independent mechanisms are biologically important for sustaining energy homeostasis during exercise. Ablating Mi-CK did not alter exercise tolerance, as the time to volitional fatigue was similar between wild-type (WT) and KO mice at various exercise intensities. In addition, skeletal muscle metabolic profiles after exercise, including glycogen, PCr/Cr ratios, free ADP/adenosine monophosphate (AMP), and lactate, were similar between genotypes. While these data suggest that the absence of PCr/Cr shuttling is not detrimental to maintaining energy homeostasis during exercise, KO mice displayed a dramatic increase in Cr-independent mitochondrial ADP sensitivity after exercise. Specifically, whereas mitochondrial ADP sensitivity decreased with exercise in WT mice, in stark contrast, exercise increased mitochondrial Cr-independent ADP sensitivity in KO mice. As a result, the apparent ADP Km was 50% lower in KO mice after exercise, suggesting that in vivo activation of voltage-dependent anion channel (VDAC)/adenine nucleotide translocase (ANT) can support mitochondrial ADP transport. Altogether, we provide insight that Cr-independent ADP transport mechanisms are biologically important for regulating ADP sensitivity during exercise, while highlighting complex regulation and the plasticity of the VDAC/ANT axis to support adenosine triphosphate demand.
The transport of adenosine diphosphate (ADP) from the cytosol into the mitochondria represents a major control point in metabolic homeostasis, as cytosolic ADP is a potent allosteric activator of glycolytic flux , whereas the movement of ADP into mitochondria influences oxidative phosphorylation rates . As a result, the improvement in exercise performance, muscle glycogen sparing, attenuated production in lactate, and the increased reliance on aerobic metabolism following training has been attributed to an improvement in mitochondrial ADP sensitivity [3–8]. Historically, this response has entirely been accredited to the induction of mitochondrial biogenesis [5,6]; however, external regulation on the proteins involved in mitochondrial ADP transport likely exists.
ADP transport between the cytosolic and mitochondrial compartments involves three major protein complexes: voltage-dependent anion channel (VDAC) on the outer mitochondrial membrane (OMM), mitochondrial creatine kinase (Mi-CK) in the intermembrane space (IMS), and adenine nucleotide translocase (ANT) on the inner mitochondrial membrane [9,10]. While ANT is required for ADP/adenosine triphosphate (ATP) exchange, Mi-CK is thought to concentrate ADP within the intermembrane space to optimize diffusion of ADP into the mitochondria . Moreover, phosphocreatine (PCr) and Creatine (Cr) are estimated to diffuse ∼2000 times faster across the OMM/through the cytosol than ADP/ATP , and therefore, phosphate transfer via creatine kinase reactions is believed to contribute substantially to metabolic homeostasis. In particular, during muscle contraction where ATP requirements increase ∼100-fold , the Cr-dependent pathway (via Mi-CK) is considered a major contributor to match energy demands , as it regenerates PCr, while enhancing ADP availability for mitochondrial respiration [9,12]. This is supported in vitro, as the presence of Cr in permeabilized muscle fibers improves mitochondrial ADP sensitivity, whereas PCr has the opposite effect . However, Mi-CK ablation does not alter resting skeletal muscle cytosolic ADP, PCr, or Cr concentrations . This could suggest that in vivo phosphate shuttling is not required, or alternatively, that Cr-independent transport compensates to sustain energy homeostasis.
ANT has several potential regulatory mechanisms, including acetylation of lysine 23 , tyrosine 194 phosphorylation , and glutathionylation/carbonylation [18,19]. While the functional roles for these regulatory points remain unknown, previous work has demonstrated that Cr-independent ADP sensitivity, which highlights the VDAC/ANT axis, can be externally regulated. For instance, ADP sensitivity in permeabilized muscle fibers is increased following high-intensity exercise , yet reduced following acute moderate-intensity exercise  and chronic exercise training [22–26]. Moreover, exercise training attenuates the inhibitory effect of palmitoyl-CoA (P-CoA) on Cr-independent ADP sensitivity , whereas omega-3  and resveratrol  supplementation improve Cr-independent ADP sensitivity. However, the biological importance of Cr-independent ADP transport remains unknown, as counterintuitively type I muscle fibers display decreased Cr-independent ADP sensitivity post-exercise despite a higher reliance on oxidative metabolism . Since PCr/Cr shuttling is believed to contribute substantially to metabolic homeostasis [9,12], the previously observed changes in Cr-independent ADP sensitivity remain biologically questionable.
Therefore, we utilized mice devoid of Mi-CK to determine whether Cr-independent mechanisms could sustain metabolic control, particularly during acute exercise. While cytosolic creatine kinase (MM-CK) or double-knockout (KO) (MM-CK and Mi-CK) mice have known compensatory up-regulation of mitochondrial proteins [29,30], fiber type shifts , and severe impairments in ‘burst’ activity [29–32], force production , and muscular relaxation , single Mi-CK KO mice do not have alterations in these parameters . Moreover, while previous work has indicated that ablating Mi-CK does not alter ADP sensitivity at rest , whether this pathway can be regulated to sustain energetic requirements during exercise is unknown. Therefore, altogether, utilization of Mi-CK KO mice provides a means to directly study the in vivo importance of Cr-independent regulation, and to determine whether this alone can support ADP sensitivity during exercise. It was hypothesized that in the absence of phosphate shuttling, Cr-independent ADP sensitivity would be improved to sustain energy homeostasis during exercise.
All experimental procedures were approved by the Animal Care Committee at the University of Guelph, and conformed to the guide for the care and use of laboratory animals (US National Institutes of Health). Cryopreserved embryos from wild-type (WT) and Mi-CK (Ckmt1tm2Bew) null mice were generously provided by Dr Be Wieringa from Dr Craig Lygate's repository. Breeding pairs were generated on a C57BL/6 background at the Toronto Centre for Phenogenomics and bred on site at the University of Guelph. Age-matched (12 weeks of age) male and female Mi-CK WT and KO mice were group-housed with a 12:12 h light–dark cycle, and were provided standard chow and water ad libitum. Animals were randomly selected to characterize the resting phenotype or to examine exercise responses (n = 6 sedentary and n = 6 acute exercise per genotype). Equal numbers of male and female animals were used in each group, and all data were pooled, provided that no sex differences were evident. All animals were anesthetized with an intraperitoneal injection of sodium pentobarbital (60 mg/kg; MTC Pharmaceuticals, Cambridge, ON) before skeletal muscle tissue collection.
Resting whole-body metabolic measurements
At 12 weeks of age, Mi-CK WT and KO mice were placed in metabolic caging (Columbus Instruments, Columbus, OH). Resting oxygen consumption (VO2) and carbon dioxide production (VCO2) were recorded. VO2 and VCO2 values were used to calculate total carbohydrate (CHO) and fat oxidation, and energy expenditure as follows :
Values were divided by 60 and multiplied by 16.19 (CHO) and 40.80 (fat) to convert into kilojoules per hour.
Acute exercise protocol
Mice were acclimated to motorized treadmills for 10 min, at 15 m/min and a 5% grade. Four days later, mice were subjected to a run test to exhaustion, beginning at 15 m/min at a 5% grade for 1 h, then increased by 2 m/min every 5 min until volitional cessation. Run times to exhaustion were also determined at higher intensities, specifically 20 m/min (20% grade) and 22 m/min (20% grade). We utilized a moderate intensity for our metabolic analyses to optimize aerobic fuel utilization and ensured that all mice could complete the running protocol to limit variability. Specifically, 1 week after an exercise tolerance test, mice ran for 1 h at a moderate intensity (15 m/min at a 5% grade) or remained sedentary as a comparison. Mice were anesthetized using sodium pentobarbital, cervical dislocation was performed, and the red quadriceps and red gastrocnemius were removed immediately and frozen in liquid nitrogen for subsequent analyses. The red quadriceps and gastrocnemius are predominately oxidative muscles and contain similar fiber types . These muscles yielded similar responses for all measurements; therefore, values were treated as duplicates and averaged for each animal.
Determination of skeletal muscle metabolites
Muscles were snap frozen from sedentary and exercised mice, dried, and powdered, and metabolites were extracted using perchloric acid (0.5 M) and neutralized with potassium bicarbonate (2.2 M). The concentrations of ATP, PCr, Cr, and lactate were determined using enzymatic spectrophotometric assays by assuming equilibrium of the creatine kinase and adenylate kinase reactions . Free ADP and adenosine monophosphate (AMP) concentrations were calculated using ATP, PCr, Cr, and lactate concentrations, as previously described . For glycogen analysis, glucose monophosphates were degraded using sodium hydroxide (0.1 M) at 80°C (10 min). Samples were neutralized in a buffer containing HCl (0.1 M), citric acid (0.2 M), and Na2HPO4·7H2O (0.2 M). Total glycogen was analyzed using spectrophotometric methods .
Preparation of permeabilized muscle fibers
Muscles were separated in BIOPS buffer containing CaK2-EGTA (2.77 mM), K2-EGTA (7.23 mM), Na2-ATP (5.77 mM), MgCl2·6H2O (6.56 mM), Na2-PCr (15 mM), imidazole (20 mM), dithiothreitol (0.5 mM), and MES (50 mM). Fiber bundles were treated with 40 µg/ml saponin for 30 min, a cholesterol-specific detergent that permeabilizes the sarcolemma. Fibers were then washed in mitochondrial respiration medium (MIRO5) containing EGTA (0.5 mM), MgCl2·6H2O (3 mM), potassium lactobionate (60 mM), KH2PO4 (10 mM), Hepes (20 mM), sucrose (110 mM), and fatty acid-free bovine serum albumin (BSA; 1 g/l). Fibers remained in MIRO5 until respiration analysis.
Permeabilized muscle fiber respiration
Mitochondrial respiration experiments using permeabilized muscle fibers (PMFs) were performed using high-resolution respirometry (Oroborus Oxygraph-2K, Innsbruck, Austria) at 37°C. Previous work has demonstrated physiologically unrealistic ADP sensitivity in the absence of the myosin II inhibitor blebbistatin, as these experiments suggest that >50% of maximal mitochondrial respiration occurs in resting muscle [20,21]. Therefore, the majority of our experiments occurred in the presence of blebbistatin  to better reflect the in vivo environment, while a few experiments in WT animals were also conducted in the absence of blebbistatin to model the ability of Cr/PCr and P-CoA to modulate ADP sensitivity in a contracted situation. Exercise-based experiments were conducted in the presence or absence of Cr (20 mM) or PCr (20 mM), and ADP (0, 10, 25, 100, 175, 250, 500, 1000, 2000, 4000, and 6000 µM) stimulated respiration was determined in the presence of 10 mM pyruvate and 5 mM malate. Glutamate (10 mM) and succinate (10 mM) were added following the ADP titration to determine maximum mitochondrial respiration. Experiments were also conducted in contracted PMFs (no blebbistatin) in the presence and absence of P-CoA (60 µM) with and without Cr (20 mM) or PCr (20 mM). Cytochrome c (10 µM) did not significantly elevate the respiration rate of experiments. Respiratory control ratios (RCRs) were determined for all experiments (state 3/state 4 respiration). All PMFs were recovered, freeze-dried, and data normalized to bundle weight.
Preparation of isolated muscle mitochondria and respiration
Skeletal muscle mitochondria were isolated using differential centrifugation . Muscle was removed, minced in isolation buffer (100 mM sucrose, 100 mM KCl, 50 mM Tris–HCl, 1 mM KH2PO4, 0.1 mM EGTA, 0.2% BSA, and 1 mM ATP; pH 7.4), weighed, and homogenized using a motorized Teflon pestle (750 rpm). Mitochondria underwent centrifugation at 800 g for 10 min, were resuspended in 4 ml of isolation buffer, and supplemented with 0.025 µg/mg tissue protease for 5 min. Thereafter, 10 ml of isolation buffer was added and the sample immediately spun at 5000 g for 5 min . The pellet was repeatedly resuspended in isolation buffer and pelleted at 10 000 g for 10 min. Mitochondria were prepared immediately for western blot analyses or mitochondrial respiration using an ADP clamp (i.e. 5 units/ml hexokinase; 5 mmol/l 2-deoxyglucose), as previously described . Experiments were conducted using 10 mM pyruvate, 2 mM malate, and an ADP titration (2.5, 5, 10, 20, 40, and 100 µM) in the presence and absence of Cr (20 mM) at 25°C. To calculate P/O ratios, the change in oxygen following the addition of 50 µM ADP was used in the presence of 10 mM pyruvate and 2 mM malate, followed by the addition of 2 mM ADP, 10 mM glutamate, and 10 mM succinate to determine maximal respiration rates.
Whole muscle red gastrocnemius was homogenized, diluted to 1 µg/µl, and protein was loaded equally for α-tubulin (5 µg protein; 1:5000, Abcam, Cambridge, MA, USA; ab7291), OXPHOS (5 µg protein; 1:500, MitoSciences, Eugene, OR, USA; ab110413), complex IV (subunit 4) (5 µg protein; 1:3000, Invitrogen; A21347), VDAC (5 µg protein; 1:5000, Abcam; ab14734), ANT1 (20 µg protein; 1:2000, Abcam; ab110322), ANT2 (25 µg protein; 1:2000, Abcam; ab118076), Mi-CK (5 µg protein; 1:5000, Abcam; ab131188), cytosolic creatine kinase (5 µg protein; 1:5000, MM-CK; Abcam; ab193292), PDH (5 µg protein; Molecular Probes), CPT1 (20 µg protein; Alpha Diagnostic Intl., Inc., TX, USA; CPT1M11-A), and SERCA (25 µg protein; Affinity Bioreagents, CO, USA). Isolated mitochondria were diluted to 0.5 µg/µl and protein was loaded equally (7.5 µg protein) for complex IV (subunit 4), VDAC, PDH, ANT1, Mi-CK, MM-CK, and SERCA. Proteins were separated using SDS–PAGE, blocked, and incubated with appropriate primary and secondary antibodies. All membranes were detected using enhanced chemiluminescence (ChemiGenius2 Bioimaging System, SynGene, Cambridge, UK).
The apparent Km for ADP was determined by Michaelis–Menten kinetics using Prism software (GraphPad Software, Inc., La Jolla, CA, USA), as previously described . Maximal respiration (Vmax) was determined as the highest respiratory value obtained . Statistical analyses were performed using SigmaPlot (version 12.0, Systat Software, Inc.). Exercise tolerance measurements were determined using independent samples t-tests (two-tailed). Whole-body measurements and all sedentary/exercise parameters were analyzed using two-way ANOVAs, followed by Student–Newman–Keuls (SNK) post hoc analyses. Contracted PMF experiments were analyzed using one-way ANOVAs and SNK post hoc analyses. All data are expressed as mean ± SEM. Significance was determined as P ≤ 0.05.
Verification of non-compensatory changes in the Cr-independent pathway and metabolism in Mi-CK KO mice
Although previous work has indicated that the MM-CK does not compensate in terms of elevated gene expression in Mi-CK KO mice , we determined whether Mi-CK KO mice had compensatory changes in proteins involved in Cr-independent ADP/ATP transport, MM-CK, and mitochondrial content. First, we confirmed the absence of Mi-CK in the KO mice (Figure 1A). Despite the loss of Mi-CK, MM-CK and proteins involved in mitochondrial ADP transport (ANT1, ANT2, and VDAC) and metabolism (complexes I–V, CPT1, and PDH) remained unaltered (Figure 1B,C). This was accompanied by similar mitochondrial respiration in the presence and absence of ADP, maximal complex I respiration, maximal complex I and II respiration, and the RCR of PMFs between WT and KO mice (Figure 1D). To verify that the unaltered protein contents did not alter resting metabolism, through impaired ADP flux into the mitochondria, we also confirmed that CHO, fat oxidation, and energy expenditure at the whole-body level were similar between groups (Figure 1E–G). Altogether, these data confirm the absence of compensatory changes in Mi-CK KO mice, and therefore allowed us to study whether the Cr-independent pathway (i.e. VDAC and ANT) can regulate energy homeostasis to sustain exercise capacity, fuel selection, and ADP sensitivity during exercise.
Mi-CK KO mice do not have compensation of protein markers of mitochondrial content, ADP/ATP transport, MM-CK, maximal respiration rates, or resting fuel selection.
Mi-CK KO mice do not have impaired exercise tolerance or altered muscle metabolic profiles
Given the loss of PCr/Cr energy transfer associated with ablating Mi-CK, we next examined whether the reliance on the Cr-independent pathway could sustain exercise tolerance in KO mice. The exercise capacity of Mi-CK WT and KO mice was not different, as the time to voluntary exhaustion was similar between genotypes at both moderate and high intensities of exercise (Figure 2A–C). Regardless of genotype, some mice had difficulty maintaining higher intensities of exercise (i.e. ran <10 min); therefore, we chose a moderate intensity that all mice could sustain for the remainder of our analyses to decrease variability. In this regard, we re-ran mice at 15 m/min (5% grade) for 1 h to examine steady-state metabolic profiles. While exercise resulted in lower muscle glycogen (∼50% decrease) and PCr (∼20% decrease), and increased muscle lactate and Cr (∼20% increase), these responses were not altered in the absence of the Mi-CK (Figure 3A–D). Moreover, while exercise did not alter muscle total Cr or ATP concentrations, free ADP and AMP concentrations were increased similarly with exercise in WT and KO mice (Figure 3E–H). Altogether, these data suggest that either Mi-CK PCr/Cr energy transfer is not essential to maintain ADP delivery to mitochondria during exercise, or alternatively, that Cr-independent processes are activated in vivo to maintain ADP transport and metabolic homeostasis.
Mi-CK KO mice do not have impaired exercise tolerance at moderate- or high-intensities of exercise.
Mi-CK KO mice do not have altered fuel selection during exercise.
Improved ADP sensitivity in Mi-CK KO mice immediately post-exercise
We next aimed to determine whether Cr-independent ADP transport was activated in Mi-CK KO mice. At rest, the apparent ADP Km was unaltered in sedentary mice in the absence of the Mi-CK enzyme (i.e. no Cr or PCr in the media; Figure 4A,B), supporting the consistent VDAC and ANT proteins reported (Figure 1B,C). However, acute exercise increased the apparent Km to ADP ∼25% in WT mice, while in stark contrast, the apparent Km was decreased ∼33% in KO mice (Figure 4A,B). As a result, relative to WT mice, the apparent Km was ∼45% lower in KO mice after acute exercise, suggesting a dramatic increase in Cr-independent ADP transport (Figure 4A,B). Combined, these data indicate that in the absence of in vivo Cr-dependent energy transfer (i.e. via Mi-CK), mitochondrial ADP sensitivity is improved to maintain oxidative metabolism during exercise. This highlights a complex regulatory system involving a Cr-independent pathway that has largely remained understudied.
We next examined the influence of acute exercise on ADP sensitivity in the presence of phosphate shuttling. In contrast with the responses observed with Cr-independent ADP sensitivity, acute exercise did not alter the apparent ADP Km in either genotype in the presence of Cr or PCr (Figure 4C–F). While the apparent ADP Km was higher in KO mice in the presence of Cr (Figure 4D), Cr still decreased the apparent Km ∼33% in Mi-CK KO mice (Figure 4C,D relative to sedentary Km values in Figure 4A,B). Moreover, PCr similarly impaired mitochondrial ADP respiratory sensitivity in WT and KO mice, as the apparent Km for ADP in the presence of PCr increased almost 2-fold in both genotypes (Figure 4E,F relative to sedentary Km values in Figure 4A,B). Combined, these data further validate the Mi-CK genotype while also suggesting that in situ assessments of Cr/PCr sensitivity in PMFs are partially confounded by a MM-CK contribution. Alternatively, MM-CK relocation to the mitochondrial intermembrane space in KO mice could occur to maintain Cr and PCr sensitivity, explaining the responses observed in KO mice. However, mitochondria isolated from KO mice did not contain either CK enzyme, whereas mitochondrial proteins were comparably concentrated compared with WT mice (Figure 5A). This is similar to a previous finding reported using MM-CK gene expression in diaphragm muscle . Moreover, isolated mitochondrial respiration experiments in the presence and absence of Cr demonstrated similar mitochondrial integrity (Figure 5B) and maximal respiration rates (Figure 5C), without significant improvements in ADP sensitivity in WT or KO mice (Figure 5D). However, when assessed within each genotype (i.e. t-test instead of a two-way ANOVA), Cr decreased the apparent ADP Km ∼10% (P = 0.01) in WT mice, whereas Cr did not stimulate ADP sensitivity in KO mice. Combined, these data suggest an absence of MM-CK redistribution to mitochondria, supporting the importance of the observed Cr-independent improvement in ADP sensitivity following acute exercise. In addition, these data indicate that the retained response to Cr/PCr in PMFs from Mi-CK KO mice reflects an MM-CK contribution; therefore, the in vivo contribution of Cr-dependent energy transfer may be overestimated using this model.
Mi-CK KO mice exhibit improved ADP sensitivity in the absence of Cr or PCr immediately post-exercise.
MM-CK does not relocate to the intermembrane space of Mi-CK KO mice.
Muscle contraction reveals a lack of Cr-stimulated respiration despite maintained P-CoA inhibition of ANT
To further study the importance of Cr-dependent and -independent ADP delivery to the mitochondria during exercise, we next modeled muscle contraction in an in situ situation. In this regard, previous research has demonstrated that PMFs remain in a relaxed state when experiments are conducted either at 25°C or in the presence of blebbistatin (myosin ATPase inhibitor) at 37°C , as performed in the current study. In contrast, we allowed PMFs from WT mice to contract by withholding blebbistatin. As previously reported , contraction of PMFs increased Cr-independent ADP sensitivity ∼4-fold (relative to sedentary Km values in Figure 4B). However, while Cr decreased the apparent ADP Km ∼50% in relaxed PMFs (Figure 4D), once PMFs were contracted, Cr did not improve ADP sensitivity (Figure 6A). One potential explanation in this model could be a diffusion limitation preventing Cr from accumulating in the mitochondrial intermembrane space. We therefore also used P-CoA in an attempt to modulate ADP sensitivity in contracted PMFs, as P-CoA is known to inhibit ANT and reduce ADP sensitivity [23,43–45]. Importantly, in the presence and absence of Cr, P-CoA dramatically attenuated ADP sensitivity (Figure 6A), verifying the absence of a diffusion limitation as an explanation for the lack of Cr-mediated drive on respiration in contracted PMFs. In addition, the provision of PCr still attenuated ADP sensitivity, a response that was additive with P-CoA (Figure 6B) — indicating that the creatine kinase reaction was still functional in contracted fibers, but Cr simply did not affect ADP sensitivity. One possible explanation could be that muscle contraction alters the compartmentalization of the IMS (i.e. decreased IMS volume); therefore, Cr is not required to concentrate ADP, as transport under these circumstances may already be optimal. Given the inability of Cr to alter respiration in contracted fibers from WT mice, and therefore the absence of an Mi-CK-mediated stimulation on respiration in this situation, we did not repeat these measurements in KO mice. Together, these data suggest that Cr-dependent phosphate shuttling is not required for contracting muscle, while further emphasizing the potential importance of Cr-independent ADP provision during exercise.
ADP sensitivity in contracted PMFs in the presence and absence of Cr, PCr, and P-CoA.
The transport of ADP into mitochondria represents a key control point for skeletal muscle metabolism; however, the regulation of this process is poorly defined and the in vivo importance of Cr-independent processes remains largely understudied. We hypothesized that Cr-independent ADP transport would be activated during exercise to sustain metabolic profiles and exercise tolerance. Our data supports this hypothesis, as we provide evidence that Cr-independent ADP sensitivity is improved during exercise in KO mice, suggesting that activation of VDAC or ANT maintains exercise tolerance and metabolic outcomes in KO mice. Altogether, these data highlight that activation of Cr-independent ADP sensitivity is important to sustain relative energetic requirements during exercise, emphasizing the biological importance of this process.
Regulation of ADP sensitivity by the Cr-independent pathway during acute exercise
Our current data suggest that external regulation of Cr-independent mechanisms are not required to sustain energy homeostasis at rest, given the unaltered protein contents and ADP sensitivity in the absence of Cr/PCr. Since elevated ADP transport is required during exercise to match energy demands, we challenged these mice with an acute bout of exercise in an attempt to establish whether the Cr-independent pathway can be regulated in the absence of Cr/PCr shuttling in vivo. During muscle contraction, ATP requirements increase ∼100-fold , particularly to support myosin ATPase , SERCA [47,48], and Na+/K+ -ATPase . Therefore, ADP provision to the mitochondria must increase to enable oxidative phosphorylation to support these increased ATP demands. However, ablation of the Cr-dependent pathway, via Mi-CK loss, did not alter run time to exhaustion at several intensities of exercise. In addition, muscle glycogen breakdown and metabolic profiles in skeletal muscle, including free ADP and AMP concentrations, were not different in KO mice after running at a moderate intensity of exercise, which should maximize the contribution of aerobic metabolism to ATP provision . These data indicate that ADP transport for oxidative phosphorylation was not impaired in the absence of IMS phosphate shuttling and ADP transport, and energy utilization was not compromised based on similar reductions in PCr and glycogen content. However, we provide evidence that ADP transport was maintained during exercise through activation of Cr-independent processes in Mi-CK KO mice, as the apparent ADP Km was ∼50% lower than WT mice post-exercise. While the exact mechanism to account for the improvement in ADP sensitivity is currently unknown, VDAC and ANT are required for ADP/ATP movement across mitochondrial membranes [10,21], and therefore should be examined for possible regulation. Since sedentary ADP sensitivity, maximal respiration rates, and VDAC/ANT protein contents were similar between WT and KO mice, this strongly suggests external regulation of these proteins independent of Cr/PCr energy transfer and diffusion of ADP into the mitochondria. Previous work supports our current findings in WT mice, as reduced ADP sensitivity has been observed in humans following acute moderate-intensity exercise  and endurance training [22–24,26,51]. In contrast, acute and chronic high-intensity exercise increased ADP sensitivity . Together with the current findings in Mi-CK KO mice, this demonstrates that when sufficiently metabolically challenged (i.e. high-intensity exercise or ablation of phosphate shuttling), Cr-independent ADP transport sensitivity can drastically be improved. Although we cannot fully differentiate in the current study whether the regulation in WT and KO mice occurs strictly on VDAC, ANT, or both, since ANT provides substrate to drive the Mi-CK reaction (i.e. ATP), oxidative phosphorylation (i.e. ADP), and is a convergent point in both Cr-dependent and Cr-independent mechanisms, it is plausible that external regulation occurs at this step. Moreover, ANT contains possible sites for post-translational modification, including lysine 23 acetylation , phosphorylation of tyrosine 194 , and glutathionylation/carbonylation [18,19]. Although the influence of exercise on these sites is not well established, it is conceivable that ATP demand at various relative exercise intensities can activate or suppress ADP transport through ANT, allowing for tighter control of energy homeostasis. As such, high-intensity exercise has been shown to decrease ANT acetylation of lysine 23, which is predicted to increase ADP affinity . Interestingly, this aligns with previous evidence of increased ADP sensitivity after high-intensity exercise in situ in humans . This may also explain the similar response in Mi-CK KO mice from the current study, provided exercise induced a greater relative ‘stress’ on Cr-independent ADP transport mediated through the VDAC/ANT complexes. Although it is unclear whether moderate-intensity exercise increases ANT acetylation, given that high-intensity exercise has been shown to reduce ANT acetylation , as well as improve ADP sensitivity , an increase in ANT acetylation is a plausible explanation for the reduction in ADP sensitivity following moderate-intensity exercise (WT mice from the current study and others ). Additionally, previous work has highlighted the ability of the cytoskeleton to regulate OMM proteins, including VDAC . Therefore, another possible mechanism to improve ADP sensitivity during exercise could involve post-translational modifications of upstream proteins, such as tubulin or desmin, resulting in altered VDAC/ANT regulation of adenine nucleotide transport across mitochondrial membranes [10,52]. Regardless of the absence of a mechanism to explain the observed alterations in ADP sensitivity, the present data provide support that Cr-independent mechanisms can regulate ADP transport during exercise to sustain energetic requirements, emphasizing the plasticity of Cr-independent ADP transport in vivo.
In situ Cr-dependent experiments contain an MM-CK contribution
In the presence of Cr, in situ ADP sensitivity is increased through mechanisms attributed solely to Mi-CK [21,22,34]. However, the current study and others  have shown that although the Cr-stimulated response is attenuated in Mi-CK KO mice, these mice still exhibit Cr-mediated improvements in ADP sensitivity despite the loss of Mi-CK. Moreover, we demonstrate that PCr inhibition is similar between WT and KO mice. The absence of MM-CK relocation to the IMS suggests that the observed Cr-dependent improvements in ADP sensitivity appear to be confounded by cytosolic MM-CK in situ; therefore, research examining Cr/PCr influences on mitochondrial ADP sensitivity may not be mitochondrial specific [21,23,28]. While the extent of interference by MM-CK is unclear, these data demonstrate that Cr/PCr-dependent ADP sensitivity is influenced by MM-CK in the absence of Mi-CK, a response unaltered following exercise. Although our data in isolated mitochondria indicate the absence of MM-CK in the IMS, previous work has suggested that MM-CK could relocate closer to the OMM following Mi-CK ablation , as MM-CK is capable of binding near sites of energy utilization including the myofilbrils  and SERCA . Thus, although we cannot identify the exact location of MM-CK in the current study, the partially retained response to Cr and PCr may reflect an MM-CK contribution that has become proximal to the OMM. In addition, the present data suggest that external regulation of either MM-CK or Mi-CK may not be required to maintain metabolic control in vivo, as the apparent Km values within each genotype remain unchanged after exercise despite elevated energy demands. In contrast, compared with WT mice, Mi-CK KO mice had similar resting and improved post-exercise Cr-independent ADP sensitivity, respectively. Altogether, these data suggest that Cr-independent ADP transport is biologically important and can be regulated in vivo.
Conclusions and perspectives
The current data highlight the plasticity of Cr-independent mechanisms to sustain exercise tolerance and fuel selection during a metabolic challenge, suggesting that both Cr-dependent and -independent mechanisms are important for optimizing ADP transport to meet energy demand in vivo. While Cr-dependent energy transfer is estimated to contribute to ∼80% of transport , ANT and VDAC are thought to be required for ADP/ATP movement across mitochondrial membranes. In the present study, the conserved exercise tolerance following Mi-CK ablation suggests that this enzyme is dispensable for energy production, as it appears that activation of VDAC/ANT is sufficient to overcome the absence of Mi-CK. Furthermore, we have previously shown that Cr-independent ADP sensitivity is altered with acute exercise , chronic exercise training , omega-3 supplementation , diabetes , resveratrol , and in the presence of reactive lipids . However, the biological importance of these observations remained questionable, as phosphate shuttling via Cr-dependent mechanisms was only modestly changed. However, the present data strongly suggest that alterations in Cr-independent ADP sensitivity are biologically relevant and can influence oxidative phosphorylation . These data may also help explain fiber type differences in resting muscle, as type I fibers, which are more oxidative than type II fibers, display a 3-fold higher apparent ADP Km in the absence of Cr . Combined with the present findings, these data may suggest that exercise is required to stimulate Cr-independent ADP transport in type I fibers. In contrast, given the greater potential to hydrolyze ATP in type II fibers, and the lower apparent ADP Km in the absence of Cr , it is possible that in vivo Cr-independent ADP transport is activated in resting type II fibers to maintain metabolic homeostasis at the onset of exercise. While these arguments remain plausible, clearly they are currently speculative and await further investigation. Moreover, elucidating the regulatory ‘control points’ underlying this process may have implications for muscle performance, especially as the apparent ADP Km increases ∼3-fold with training [25,26], suggesting a massive reserve potential for ADP sensitivity that could be utilized in non-steady conditions to better control metabolic homeostasis. Ultimately, our results demonstrate that Cr-independent ADP sensitivity is drastically improved with exercise in the absence of phosphate shuttling, as the apparent ADP Km was ∼50% lower in KO mice following treadmill running. These data highlight that Cr-independent ADP sensitivity can be activated in vivo, and therefore, may be a future target to improve mitochondrial bioenergetics in diverse metabolic situations. Moreover, our data in KO mice demonstrating partially retained Cr-stimulated improvements in ADP sensitivity highlight the importance of considering cellular location of creatine kinases when utilizing PMF experiments, as these analyses can be confounded by the MM-CK. Therefore, caution should be taken to avoid overinterpretation of Mi-CK's contribution to these experiments, as the possibility that MM-CK influences ADP sensitivity in permeabilized fibers remains possible.
Overall, while the current study establishes that Cr-independent ADP transport can sustain energy homeostasis during exercise, the exact mechanisms involved remain unknown. Future work should aim to address this knowledge gap, including direct regulation on VDAC/ANT and indirect regulation through interactions with the cytoskeletal network. This knowledge could have implications beyond understanding exercise responses, as diabetic rodents display impairments in Cr-independent ADP sensitivity ; therefore, activation of the VDAC/ANT axis in vivo could modulate disease progression as well.
ADP, adenosine diphosphate; AMP, adenosine monophosphate; ANT, adenine nucleotide translocase; ATP, adenosine triphosphate; BSA, bovine serum albumin; CHO, carbohydrate; CPT1, carnitine palmitoyltransferase 1; Cr, creatine; IMS, intermembrane space; KO, knockout; Mi-CK, mitochondrial creatine kinase; MM-CK, cytosolic creatine kinase; OMM, outer mitochondrial membrane; P-CoA, palmitoyl-CoA; PCr, phosphocreatine; PDH, pyruvate dehydrogenase; PMFs, permeabilized muscle fibers; RCR, Respiratory control ratio; SERCA, sarcoendoplasmic reticulum calcium ATPase; SNK, Student–Newman–Keuls; VDAC, voltage-dependent anion channel; WT, wild type.
G.P.H. conceptualized the study rationale. P.M.M. and G.P.H. designed experiments, whereas P.M.M. conducted all experiments and analyzed data. Both authors interpreted data and wrote the manuscript.
This work was funded by the Natural Sciences and Engineering Research Council of Canada (G.P.H.), and infrastructure was purchased with assistance from the Canadian Foundation for Innovation/Ontario Research Fund. P.M.M. is supported by an NSERC graduate scholarship.
The Authors declare that there are no competing interests associated with the manuscript.