Calcium Induces Increases in Peroxisome Proliferator-activated Receptor γ Coactivator-1α and Mitochondrial Biogenesis by a Pathway Leading to p38 Mitogen-activated Protein Kinase Activation*

Previous studies have shown that raising cytosolic calcium in myotubes induces increases in peroxisome proliferator-activated receptor γ coactivator-1α expression and mitochondrial biogenesis. This finding suggests that the increases in cytosolic calcium in skeletal muscle during exercise may mediate the exercise-induced increase in mitochondria. The initial aim of this study was to determine whether raising calcium in skeletal muscle induces the same adaptations as in myotubes. We found that treatment of rat epitrochlearis muscles with a concentration of caffeine that raises cytosolic calcium to a concentration too low to cause contraction induces increases in peroxisome proliferator-activated receptor γ coactivator-1α expression and mitochondrial biogenesis. Our second aim was to elucidate the pathway by which calcium induces these adaptations. Raising cytosolic calcium has been shown to activate calcium/calmodulin-dependent protein kinase in muscle. In the present study raising cytosolic calcium resulted in increases in phosphorylation of p38 mitogen-activated protein kinase and activating transcription factor-2, which were blocked by the calcium/calmodulin-dependent protein kinase inhibitor KN93 and by the p38 mitogen-activated protein kinase inhibitor SB202190. The increases in peroxisome proliferator-activated receptor γ coactivator-1α expression and mitochondrial biogenesis were also prevented by inhibiting p38 activation. We interpret these findings as evidence that p38 mitogen-activated protein kinase is downstream of calcium/calmodulin-dependent protein kinase in a signaling pathway by which increases in cytosolic calcium lead to increases in peroxisome proliferator-activated receptor γ coactivator-1α expression and mitochondrial biogenesis in muscle.

Exercise induces an increase in the mitochondrial content of skeletal muscle (1,2). This adaptation results in increases in the capacities to oxidize carbohydrate and fat and generate ATP (3)(4)(5)(6). The transcription coactivator peroxisome proliferatoractivated receptor ␥ coactivator-1␣ (PGC-1␣) 4 regulates the coordinated expression of the mitochondrial genome and of the nuclear genes encoding mitochondrial enzymes (7)(8)(9). Overexpression of PGC-1␣ in muscle cells and in skeletal muscle results in an increase in mitochondria similar to that induced in muscle by exercise (10 -12). Exercise results in rapid increases in PGC-1␣ activation (13) and expression (14 -17) in skeletal muscle that are thought to mediate the exercise-induced increase in mitochondria. Studies on myotubes in culture have provided evidence that raising cytosolic Ca 2ϩ induces increases in PGC-1␣ expression (17,18) and mitochondrial biogenesis (17)(18)(19)(20)(21) and that this adaptation is prevented by inhibition of calcium/calmodulin-dependent protein kinase (CAMK) (18). These findings support the hypothesis that increases in cytosolic Ca 2ϩ resulting from release of sarcoplasmic (SR) Ca 2ϩ during excitation-contraction coupling is the signal or one of the signals generated in muscle during exercise that mediates the increase in mitochondrial biogenesis.
The initial purpose of this study was to determine whether raising cytosolic Ca 2ϩ in a more biologically relevant model, skeletal muscle, induces an increase in mitochondrial biogenesis similar to that found in myotubes. After finding that raising cytosolic Ca 2ϩ in our skeletal muscle preparation does result in increases in PGC-1␣ expression and mitochondrial biogenesis, we used this model to try to elucidate the pathway by which the Ca 2ϩ -induced increases in PGC-1␣ and mitochondrial biogenesis are mediated. Both the activity and expression of PGC-1␣ appear to be regulated by p38 MAPK (7,(22)(23)(24), which is activated in muscle by contractile activity (24 -26) and in myotubes by raising cytosolic Ca 2ϩ (27). p38 activates the transcription factor activating transcription factor 2 (ATF-2), which binds to the cAMP-response element binding site on the PGC-1␣ promoter and induces PGC-1␣ transcription (24,28). Our results provide evidence that Ca 2ϩ induces increases in PGC-1␣ and mitochondrial biogenesis in muscle by activating a pathway that leads from CAMK to p38 MAPK.

EXPERIMENTAL PROCEDURES
Materials-Reagents for SDS-PAGE were purchased from Bio-Rad. Reverse transcription kits were obtained from Promega (Madison, WI). Chemiluminescence reagents were a product of Amersham Biosciences. SB202190 and KN93 were obtained from Calbiochem. An antibody against cytochrome oxidase subunit I (COXI) was purchased from Molecular Probes (Eugene, OR). The anti-PGC-1 antibody was a product of Calbiochem. ␦-Aminolevulinate synthase (ALAS) and citrate synthase antibodies were generated by Alpha Diagnostics (San Antonio, TX) (16). Phosphospecific p38 MAPK, ATF2, and AMP-activated protein kinase antibodies were purchased from Cell Signaling (Beverly, MA). Antibody against threoninephosphorylated protein kinase B was purchased from Upstate Biotechnology (Lake Placid, NY). The horseradish peroxidase secondary antibodies were purchased from Jackson Immu-noResearch Laboratories (West Grove, PA). All other chemicals were obtained from Sigma.
Animals-This research was approved by the Animal Studies Committee of Washington University. Male Wistar rats weighing 51-75 g were obtained from Charles River and maintained on a diet of Purina rat chow and water ad libitum. Food was removed at 5:00 p.m. the day before an experiment. Rats were anesthetized with pentobarbital sodium (5 mg/100 g of body weight) given intraperitoneally, and the epitrochlearis muscles were removed. These muscles weighed ϳ6 -8 mg and are sufficiently thin so that oxygenation and substrate supply by diffusion are adequate, as evidenced by maintenance of normal ATP and phosphocreatine concentrations. The rat epitrochlearis contains ϳ76% type 2b fibers, 12% type 2a fibers, and 12% type 1 fibers (29).
Muscle Incubations-Rat epitrochlearis were incubated in oxygenated (95% O 2 , 5% CO 2 ) culture medium containing Dulbecco's modified eagle medium (␣-MEM), 10% fetal bovine serum, 50 microunits/ml purified pork insulin, 100 microunits/ml penicillin, 100 g/ml streptomycin, and 0.25 g/ml Fungizone. The flasks containing the muscles were maintained at 35°C in a shaking water bath and gassed continuously with 95% O 2 , 5% CO 2 . In one series of experiments, muscles were incubated in culture medium supplemented with 3.5 mM caffeine for 6 h to raise cytosolic Ca 2ϩ concentration by releasing Ca 2ϩ from the SR (20,30,31) and then allowed to recover for 18 h in oxygenated culture medium in the absence of caffeine. Some muscles were preincubated in culture medium supplemented with 10 M SB202190 or 10 M KN93 for 1 h before and during the 6-h caffeine exposure. Muscles recovered for 18 h in culture medium in the absence of KN93, SB202190, and caffeine. Because KN93 and SB202190 are lightsensitive, glass vials containing these compounds were wrapped in foil. The medium was changed every 6 h. At the end of the incubations, muscles were briefly washed in phosphate-buffered saline and clamp-frozen with tongs cooled to the temperature of liquid nitrogen. In another experiment, muscles were frozen after 2, 4, or 6 h of exposure to caffeine or after 6 h of exposure to caffeine and 2 h of recovery. Muscles were subsequently processed for Western blotting or mRNA extraction.
p38 MAPK Phosphorylation and Signaling-Epitrochlearis muscles were dissected out and allowed to recover in culture medium for 1 h in a shaking water bath maintained at 35°C with a gas phase of 95% O 2 , 5% CO 2 . Muscles were then incubated in the absence or presence of 10 M KN93 or 10 M SB202190 (32) for 1 h followed by a 15-min incubation with or without 3.5 mM caffeine in the continued absence or presence of inhibitors. Muscles were then clamp-frozen with tongs cooled to the temperature of liquid nitrogen and kept at Ϫ70°C until they were processed for Western analysis.
Western Blotting-Frozen epitrochlearis muscles were homogenized in modified radioimmune precipitation assay buffer (33), and protein concentration was measured by the method of Lowry et al. (34). Proteins were separated by SDS-PAGE using a 10% resolving gel, and Western blotting was performed as described previously (33,35) to determine changes in the phosphorylation status of p38 MAPK, ATF-2, 5Ј-AMP-activated protein kinase (Thr-172), and protein kinase B (Thr-308) and the levels of PGC-1␣, COXI, ALAS, and citrate synthase proteins. The blots were then incubated with the appropriate horseradish peroxidase-conjugated secondary antibody. Antibody-bound protein was detected by ECL.
Statistical Analysis-Data are presented as means Ϯ S.E. Differences between two groups were analyzed using unpaired t tests. Comparisons between the means of multiple groups were made using a one-way analysis of variance followed by a post hoc comparison using Fisher's least significant difference method.

Mitochondrial Marker Protein and ␤-Actin Levels in Fresh and Incubated
Muscles-A potential problem in studies on incubated epitrochlearis muscles is that decreases in the levels of mitochondrial and other proteins might occur as the result of accelerated proteolysis or reduced protein synthesis during the in vitro incubation. We, therefore, compared the levels of four mitochondrial marker proteins and ␤-actin in freshly dissected muscles and muscles incubated in vitro for 24 h. As shown in Fig. 1, the levels of these five proteins were similar in the fresh and the incubated muscles. Apparently, the anabolic effects of the growth factors in the fetal calf serum, the amino acids, and the insulin in the incubation medium are sufficient to maintain the muscles in good condition for up to 24 h.
Raising Cytosolic Ca 2ϩ Induces an Increase in Mitochondrial Biogenesis in Skeletal Muscle-Previous studies have shown that raising cytosolic Ca 2ϩ induces increases in mitochondrial proteins (17)(18)(19)(20) and respiratory capacity (20) in L6 myotubes. Exposure of rat epitrochlearis muscles to 3.5 mM caffeine for 6 h resulted in increases in the mRNAs of PGC-1␣, COXI, ALAS, and citrate synthase measured 18 h after caffeine exposure ( Fig.  2A). This concentration of caffeine raises cytosolic Ca 2ϩ to a concentration that is too low to cause muscle contraction or a decrease in high energy phosphate levels (37,38). PGC-1␣, COXI, and ALAS proteins, which have short half-lives, were also increased 18 h after caffeine exposure (Fig. 2B). In contrast, citrate synthase protein, which appears to have a half-life of ϳ6 -7 days (39), was not increased 18 h after caffeine treatment (Fig. 2B). These findings provide evidence that, as in myotubes (17)(18)(19)(20)(21), raising cytosolic Ca 2ϩ provides a signal that leads to an increase in mitochondrial biogenesis in skeletal muscle.
Inhibition of p38 MAPK Prevents the Ca 2ϩ Induced Increase in Mitochondrial Biogenesis-Both the activity (7,22,23) and expression (24) of PGC-1␣ are regulated by p38 mitogen-activated protein kinase, and p38 MAPK is activated in muscle by exercise and contractions (17,24,26) and by raising cytosolic Ca 2ϩ in myotubes (27). As shown in Fig. 3, treatment of muscles with caffeine to raise cytosolic Ca 2ϩ induced an increase in p38 phosphorylation that was blocked by the inhibitor of p38 activation SB202190 (40). Raising cytosolic Ca 2ϩ in muscle by FIGURE 2. Caffeine induces increases in the mRNA levels (A) and protein levels (B) of a number of mitochondrial enzymes. Rat epitrochlearis muscles were exposed to 3.5 mM caffeine for 6 h, and measurements were made on muscle frozen 18 h after caffeine treatment. Values are the means Ϯ S.E. for four pooled muscle samples with three muscles per sample for A and for five-seven muscles per group for B. *, p Ͻ 0.05 caffeine versus control. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.  JUNE 29, 2007 • VOLUME 282 • NUMBER 26 means of contractile activity or exposure to caffeine also results in activation of calcium/calmodulin-dependent kinase II (CAMKII) (38,41,42), which is the isoform of CAMK expressed in skeletal muscle (41)(42)(43). Inhibition of CAMKII activation using the CAMK inhibitor KN93 (32) prevents the increase in mitochondrial biogenesis induced by raising cytosolic Ca 2ϩ (18). p38 MAPK lies downstream of CAMK in a signaling pathway by which increases in cytosolic Ca 2ϩ lead to activation of p38 (44 -46). As shown in Fig. 4, the CAMK inhibitors KN62 and KN93 prevented the phosphorylation of p38 induced by raising cytosolic Ca 2ϩ , whereas the inactive KN93 analog, KN92, had no effect. Prevention of p38 activation with SB202190 completely blocked the Ca 2ϩ induced increases in expression of PGC-1␣, COXI, and ALAS proteins (Fig. 5). In contrast to the activation of p38, raising cytosolic Ca 2ϩ using caffeine had no effect on AMP kinase or protein kinase B/AKT phosphorylation (Fig. 6).

Calcium Increases in PGC-1␣ and Mitochondrial Biogenesis
Phosphorylation of ATF-2-The mechanism by which activation of p38 brings about an increase in PGC-1␣ expression appears to involve phosphorylation and activation of the transcription factor ATF-2, which binds to the cAMP-response element binding site on the PGC-1␣ promoter and induces PGC-1␣ transcription (24,28). As shown in Fig. 7, raising cytosolic Ca 2ϩ resulted in an increase in ATF-2 phosphorylation that is blocked by either the CAMK inhibitor KN93 or by SB202190, an inhibitor that prevents p38 phosphorylation (32).
Time Course of Ca 2ϩ -induced Increase in PGC-1␣ Protein-There was no appreciable increase in PGC-1␣ protein levels in muscles exposed to caffeine for 2, 4, or 6 h (Fig. 8). However, PGC-1␣ protein level was significantly increased 2 h after the caffeine treatment was completed (Fig. 8). An increase in COXI mRNA preceded the increase in PGC-1␣ protein, with significant increases detectable in muscles treated with caffeine for either 4 or 6 h (Fig. 9). This finding is compatible with the possibility that activation rather than increased expression of PGC-1␣ is involved in mediating the early phase of the adaptive response (13).

DISCUSSION
Our results provide evidence that p38 MAPK is downstream of CAMKII in a signaling pathway by which an increase in cytosolic Ca 2ϩ leads to increases in PGC-1␣ expression and mitochondrial biogenesis in skeletal muscle. It has been known for many years that endurance exercise induces an adaptive  increase in skeletal muscle mitochondria with increases in the capacity to oxidize carbohydrate and fat and generate ATP (1)(2)(3)(4)(5)(6). This adaptation plays a major role in the exercise-training-induced increase in endurance by decreasing the disturbance in homeostasis induced by muscle contractions/strenuous exercise, (47,48). It has been our long term goal to elucidate the mechanisms by which the exercise-induced increase in muscle mitochondria is mediated. However, for a long time there was no progress in this area because nothing was known regarding how expression of the mitochondrial genome and of the nuclear genes that encode mitochondrial proteins is regulated and coordinated. More recent research has provided much new information and insight regarding how this complex process is mediated. The initial breakthrough involved the discovery of the transcription factors NRF-1 and NRF-2, which regulate expression of various respiratory chain subunits, ALAS, and mitochondrial transcription factor A (Tfam) (49 -51). This finding was followed by discovery of other transcription factors and nuclear receptors that control expression of a range of mitochondrial constituents (52)(53)(54)(55). The key breakthrough that explained how the activity of these transcription factors is regulated and orchestrated was the discovery by Spiegelman's group (7, 10, 56 -58) of the transcription coactivator PGC-1␣, which binds to and coactivates the transcription factors involved in mitochondrial biogenesis.
Overexpression of PGC-1␣ in muscle cells or skeletal muscle increases mitochondrial biogenesis (10 -12), and there is considerable evidence that PGC-1␣ provides the link between adaptive stimuli and increased mitochondrial biogenesis (23,56,58). These stimuli include exercise, such as running and swimming, which results in increases in PGC-1␣ expression    and mitochondrial biogenesis (13,14,16,24,59). It is not feasible to use exercise as the stimulus for elucidating the signaling pathways by which exercise induces an adaptive increase in mitochondria because muscle contractions result in numerous disturbances in cellular homeostasis, many of which are potential candidates as inducers of mitochondrial biogenesis. One of these is the increase in cytosolic Ca 2ϩ due to release of Ca 2ϩ from the SR during excitation-contraction coupling. Following-up on earlier studies which suggested that raising cytosolic Ca 2ϩ in cultured myotubes can result in increases in mitochondrial proteins (19,21), we have conducted studies on the effect of raising cytosolic Ca 2ϩ in L6 myotubes (18,20). This model has the advantage that L6 myotubes are viable for many days, making it possible to determine whether raising cytosolic Ca 2ϩ results in an increase in functional mitochondria.
In our first study we raised cytosolic Ca 2ϩ by treating L6 myotubes with caffeine or W7, agents that release Ca 2ϩ from the SR, or with the Ca 2ϩ ionophore ionomycin for 5 h per day for 5 days (20). These treatments resulted in increased expression of a range of mitochondrial proteins and in the capacity to oxidize oleate, which we used as a marker of an increase in functional mitochondria (20). This adaptation was blocked by preventing the caffeine or W-7-induced release of Ca 2ϩ from the SR using dantrolene and the effect of ionomycin by use of Ca 2ϩ -free medium (20). In a subsequent study using the same approach, we found that raising cytosolic Ca 2ϩ in L6 myotubes induces increases in the expression of PGC-1␣ and Tfam as well as increased binding of NRF-1 to the ALAS promoter and of NRF-2 to the COXIV promoter (18). The increases in cytosolic Ca 2ϩ induced by caffeine result in increased phosphorylation of CAMKII, which is blocked by the CAMK inhibitor KN93 (38). Inhibition of CAMKII activation by KN93 also prevents the increase in mitochondrial biogenesis induced by raising cytosolic Ca 2ϩ (18). This finding provides evidence that the first step in the signaling pathway leading from an increase in cytosolic Ca 2ϩ to an increase in PGC-1␣ is mediated by CAMKII, which is the CAMK isoform present in skeletal muscle (41)(42)(43). Exercise also results in activation of CAMKII in muscle (41,42), suggesting that our findings in muscle in vitro have relevance to the normal adaptive response to exercise.
There has been considerable interest in the role of CAMKIV in the regulation of PGC-1␣ expression and mitochondrial biogenesis in muscle (60 -62). Overexpression of a constitutively active form of CAMKIV in skeletal muscle of mice induced an increase in mitochondria (60). It has also been reported that lowering muscle high energy phosphate levels, which results in activation of 5Ј-AMP-activated protein kinase and an increase in muscle mitochondria, results in increased CAMKIV expression and that this mediates the increase in mitochondria (61). In light of these findings, it has been proposed that activation of CAMKIV by increases in cytosolic Ca 2ϩ mediates the adaptive increase in muscle mitochondria induced by exercise (60 -62). However, CAMKIV is not expressed in skeletal muscle (41)(42)(43).
Activation of p38 MAPK results in both phosphorylation, with an increase in activity, of PGC-1␣ (7,22,23) and increased PGC-1␣ expression (24). The latter effect appears to be mediated through phosphorylation of ATF-2, which binds to and activates the cAMP-response element binding site on the PGC-1␣ promoter, resulting in increased PGC-1␣ expression (24,28). Exercise activates p38 MAPK (17, 24 -26) in skeletal muscle, and there is evidence that p38 MAPK lies downstream of CAMK in a pathway activated by Ca 2ϩ (44 -46). In this context, we evaluated the possibility that p38 MAPK lies downstream of CAMKII in the pathway by which raising cytosolic Ca 2ϩ leads to an increase in PGC-1␣ in skeletal muscle. In support of this possibility, we found that treatment of muscles with caffeine results in an increase p38 phosphorylation. We have shown that inhibition of CAMK activation using KN93 blocks the Ca 2ϩ -induced increase in mitochondrial biogenesis (18). In the present study, we found that KN93 also blocks the phosphorylation of p38 that occurs in response to raising cytosolic Ca 2ϩ . Furthermore, the inhibitor of p38 activation, SB202190 (32,40), blocks the increases in PGC-1␣, COXI, and ALAS expression induced by raising cytosolic Ca 2ϩ .
The final step in the Ca 2ϩ -activated pathway leading to increased PGC-1␣ expression is likely the phosphorylation and activation of ATF-2 by p38 (24). ATF-2 binds to the cAMPresponse element binding site on the PGC-1␣ promoter and appears to be involved in mediating the increase in PGC-1␣ expression (24,28). Inhibition of either CAMKII with KN93 or p38 with SB202190 prevented the increase in ATF-2 phosphorylation induced by raising cytosolic Ca 2ϩ in our muscle preparation, providing further evidence for involvement of CAMKII and p38 MAPK in mediating the increases in mitochondrial biogenesis. In this context, we interpret our finding as evidence that raising cytosolic Ca 2ϩ in muscle results in activation of CAMKII, that this is the first step in a pathway that leads to activation of p38 and ATF-2, and that activated ATF-2 then induces an increase in PGC-1␣ expression. p38 MAPK has also been shown to phosphorylate and activate PGC-1␣, resulting in its enhanced binding to, and activation of transcription factors (7,22,23). We have found that exercise results in increased DNA binding of NRF-1 and NRF-2 as well as increased expression of a number genes encoding mitochondrial enzymes before an increase in PGC-1␣ protein (13). These findings suggest that the initial phase of the increase in mitochondrial biogenesis induced by exercise or raising cytosolic Ca 2ϩ is mediated by activation of PGC-1␣ by p38 and that this adaptation is maintained and enhanced by a subsequent increase in PGC-1␣ expression. Support for this interpretation is provided by the finding in the present study that COXI mRNA is increased in muscles treated for 4 h with caffeine, whereas PGC-1␣ protein expression was not increased until 4 h later.
In conclusion, we interpret the present results together with previous findings (13,18,20,24) as evidence that raising cytosolic Ca 2ϩ in muscle induces an increase in mitochondrial biogenesis by both activation and increased expression of PGC-1␣ by a pathway that leads from CAMKII to p38 MAPK.