AMP-activated protein kinase is required for exercise-induced peroxisome proliferator-activated receptor co-activator 1 translocation to subsarcolemmal mitochondria in skeletal muscle.

In skeletal muscle, mitochondria exist as two subcellular populations known as subsarcolemmal (SS) and intermyofibrillar (IMF) mitochondria. SS mitochondria preferentially respond to exercise training, suggesting divergent transcriptional control of the mitochondrial genomes. The transcriptional co-activator peroxisome proliferator-activated receptor γ co-activator 1α (PGC-1α) and mitochondrial transcription factor A (Tfam) have been implicated in the direct regulation of the mitochondrial genome in mice, although SS and IMF differences may exist, and the potential signalling events regulating the mitochondrial content of these proteins have not been elucidated. Therefore, we examined the potential for PGC-1α and Tfam to translocate to SS and IMF mitochondria in human subjects, and performed experiments in rodents to identify signalling mechanisms regulating these translocation events. Acute exercise in humans and rats increased PGC-1α content in SS but not IMF mitochondria. Acute exposure to 5-aminoimidazole-4-carboxamide-1-ß-ribofuranoside in rats recapitulated the exercise effect of increased PGC-1α protein within SS mitochondria only, suggesting that AMP-activated protein kinase (AMPK) signalling is involved. In addition, rendering AMPK inactive (AMPK kinase dead mice) prevented exercise-induced PGC-1α translocation to SS mitochondria, further suggesting that AMPK plays an integral role in these translocation events. In contrast to the conserved PGC-1α translocation to SS mitochondria across species (humans, rats and mice), acute exercise only increased mitochondrial Tfam in rats. Nevertheless, in rat resting muscle PGC-1α and Tfam co-immunoprecipate with α-tubulin, suggesting a common cytosolic localization. These data suggest that exercise causes translocation of PGC-1α preferentially to SS mitochondria in an AMPK-dependent manner.

Differential susceptibility of subsarcolemmal and intermyofibrillar mitochondria to apoptotic stimuli.

Apoptosis can be evoked by reactive oxygen species (ROS)-induced mitochondrial release of the proapoptotic factors cytochrome c and apoptosis-inducing factor (AIF). Because skeletal muscle is composed of two mitochondrial subfractions that reside in distinct subcellular regions, we investigated the apoptotic susceptibility of subsarcolemmal (SS) and intermyofibrillar (IMF) mitochondria. SS and IMF mitochondria exhibited a dose-dependent release of protein in response to H2O2 (0, 25, 50, and 100 microM). However, IMF mitochondria were more sensitive to H2O2 and released a 2.5-fold and 10-fold greater amount of cytochrome c and AIF, respectively, compared with SS mitochondria. This finding coincided with a 44% (P < 0.05) greater rate of opening (maximum rate of absorbance decrease, V(max)) of the protein release channel, the mitochondrial permeability transition pore (mtPTP), in IMF mitochondria. IMF mitochondria also exhibited a 47% (P < 0.05) and 60% (0.05 < P < 0.1) greater expression of the key mtPTP component voltage-dependent anion channel and cyclophilin D, respectively, along with a threefold greater cytochrome c content, but similar levels of AIF compared with SS mitochondria. Despite a lower susceptibility to H2O2-induced release, SS mitochondria possessed a 10-fold greater Bax-to-Bcl-2 ratio (P < 0.05), a 2.7-fold greater rate of ROS production, and an approximately twofold greater membrane potential compared with IMF mitochondria. The expression of the antioxidant enzyme Mn2+-superoxide dismutase was similar between subfractions. Thus the divergent protein composition and function of the mtPTP between SS and IMF mitochondria contributes to a differential release of cytochrome c and AIF in response to ROS. Given the relatively high proportion of IMF mitochondria within a muscle fiber, this subfraction is likely most important in inducing apoptosis when presented with apoptotic stimuli, ultimately leading to myonuclear decay and muscle fiber atrophy.

Regulation of CPT I activity in intermyofibrillar and subsarcolemmal mitochondria from human and rat skeletal muscle.

Carnitine palmitoyltransferase I (CPT I) is considered the rate-limiting enzyme in the transfer of long-chain fatty acids (LCFA) into the mitochondria and is reversibly inhibited by malonyl-CoA (M-CoA) in vitro. In rat skeletal muscle, M-CoA levels decrease during exercise, releasing the inhibition of CPT I and increasing LCFA oxidation. However, in human skeletal muscle, M-CoA levels do not change during moderate-intensity exercise despite large increases in fat oxidation, suggesting that M-CoA is not the sole regulator of increased CPT I activity during exercise. In the present study, we measured CPT I activity in intermyofibrillar (IMF) and subsarcolemmal (SS) mitochondria isolated from human vastus lateralis (VL), rat soleus (Sol), and red gastrocnemius (RG) muscles. We tested whether exercise-related levels ( approximately 65% maximal O2 uptake) of calcium and adenylate charge metabolites (free AMP, ADP, and Pi) could override the M-CoA-induced inhibition of CPT I activity and explain the increased CPT I flux during exercise. Protein content was approximately 25-40% higher in IMF than in SS mitochondria in all muscles. Maximal CPT I activity was similar in IMF and SS mitochondria in all muscles (VL: 282 +/- 46 vs. 280 +/- 51; Sol: 390 +/- 81 vs. 368 +/- 82; RG: 252 +/- 71 vs. 278 +/- 44 nmol.min-1.mg protein-1). Sensitivity to M-CoA did not differ between IMF and SS mitochondria in all muscles (25-31% inhibition in VL, 52-70% in Sol and RG). Calcium and adenylate charge metabolites did not override the M-CoA-induced inhibition of CPT I activity in mitochondria isolated from VL, Sol, and RG muscles. Decreasing pH from 7.1 to 6.8 reduced CPT I activity by approximately 34-40% in both VL mitochondrial fractions. In summary, this study reports no differences in CPT I activity or sensitivity to M-CoA between IMF and SS mitochondria isolated from human and rat skeletal muscles. Exercise-induced increases in calcium and adenylate charge metabolites do not appear responsible for upregulating CPT I activity in human or rat skeletal muscle during moderate aerobic exercise.