Exercise-induced alterations in extracellular signal-regulated kinase 1/2 and mammalian target of rapamycin (mTOR) signalling to regulatory mechanisms of mRNA translation in mouse muscle

Abstract

The present study examined the effects of an acute bout of treadmill exercise on signalling through the extracellular signal-regulated kinase (ERK)1/2 and mammalian target of rapamycin (mTOR) pathways to regulatory mechanisms involved in mRNA translation in mouse gastrocnemius muscle. Briefly, C57BL/6 male mice were run at 26 m min⁻¹ on a treadmill for periods of 10, 20 or 30 min, then the gastrocnemius was rapidly removed and analysed for phosphorylation and/or association of protein components of signalling pathways and mRNA translation regulatory mechanisms. Repression of global mRNA translation was suggested by disaggregation of polysomes into free ribosomes, which occurred by 10 min and was sustained throughout the time course. Exercise repressed the mTOR signalling pathway, as shown by dephosphorylation of the eukaryotic initiation factor (eIF)4E-binding protein-1 (4E-BP1), enhanced association of the regulatory-associated protein of mTOR with mTOR, and increased assembly of the tuberin–hamartin complex. In contrast, exercise caused no change in phosphorylation of either Akt/PKB or tuberin. Upstream of mTOR, exercise was associated with an increase in cAMP, protein kinase A activity, and AMP-activated protein kinase phosphorylation. Simultaneously, exercise caused a rapid and sustained activation of the MEK1/2–ERK1/2–p90RSK pathway, resulting in increased phosphorylation of downstream targets including eIF4E and the ribosomal protein (rp)S6 on S235/S236. Overall, the data are consistent with exercise-induced repression of mTOR signalling and global rates of mRNA translation, accompanied perhaps by up-regulated translation of selected mRNAs through regulatory mechanisms such as eIF4E and rpS6 phosphorylation, mediated by activation of the ERK1/2 pathway.

Previous studies have shown that during and immediately after acute endurance exercise, e.g. running on a treadmill, global rates of protein synthesis are reduced in skeletal muscle (Gautsch et al. 1998). In part, the decline in protein synthesis is a result of repressed signalling through a protein kinase referred to as the mammalian target of rapamycin (mTOR) (Gautsch et al. 1998). mTOR regulates a variety of processes involved in cell growth through phosphorylation of downstream targets. Of particular relevance to mRNA translation, mTOR phosphorylates eukaryotic initiation factor (eIF)4E binding protein-1 (4E-BP1) and ribosomal protein S6 kinase-1 (S6K1) (Hay & Sonenberg, 2004). 4E-BP1 binds to eIF4E and prevents it from binding to eIF4G to form the eIF4F complex. During translation initiation, eIF4F binds to mRNA through the association of eIF4E with the m⁷GTP cap structure of the mRNA (Kapp & Lorsch, 2004). The eIF4F–mRNA complex then binds to the 40S ribosomal subunit through the association of eIF4G with the eIF3–40S ribosomal subunit complex. Phosphorylation of 4E-BP1 by mTOR is a precursor event that allows subsequent phosphorylation by other, as yet unidentified, protein kinases that ultimately result in release of 4E-BP1 from the inactive 4E-BP1–eIF4E complex, permitting eIF4E to bind to eIF4G and form the active eIF4F complex. Phosphorylation of S6K1 on T389 by mTOR generates a docking site for another protein kinase, phosphoinositide-dependent kinase 1 (PDK1), which subsequently phosphorylates S6K1 on T229, resulting in its activation (Alessi et al. 1998; Frodin et al. 2002). S6K1 phosphorylates at least three proteins that have been implicated in regulating mRNA translation, including eIF4B (Raught et al. 2004; Holz et al. 2005), ribosomal protein S6 (rpS6) (Krieg et al. 1988; Ferrari et al. 1991), and eukaryotic elongation factor (eEF)2 kinase (Redpath et al. 1996). As a consequence, phosphorylation of S6K1 by mTOR indirectly modulates both the initiation and elongation phases of mRNA translation.

Signalling through mTOR is enhanced in response to signalling through the protein kinase B (PKB; also known as Akt) (Fingar & Blenis, 2004) and mitogen-activated protein (MAP) kinase signalling pathways (Ma et al. 2005; Rolfe et al. 2005), and repressed in response to activation of the AMP-activated protein kinase (AMPK) (Bolster et al. 2002; Inoki et al. 2003b; Shaw et al. 2004). All three signalling pathways converge at an upstream regulator of mTOR function, the tuberous sclerosis complex (Tee et al. 2002), which consists of the proteins tuberin, the product of the TSC2 gene, and hamartin, the product of the TSC1 gene. Tuberin is a GTPase activator protein (GAP) for a ras homologue referred to as ras homologue enriched in brain (Rheb), that binds to and regulates mTOR function (Castro et al. 2003). Binding of Rheb–GTP enhances signalling through the kinase, whereas binding of Rheb–GDP inhibits mTOR function (Inoki et al. 2003a). Akt/PKB phosphorylates tuberin on multiple residues including T1462, resulting in decreased tuberin GAP activity, and enhanced mTOR signalling (Potter et al. 2002). Similarly, tuberin is phosphorylated by both the extracellular-regulated protein kinases (ERK)1 and ERK2 (Ma et al. 2005) and by p90RSK, a downstream target of ERK1/2 signalling (Lazar et al. 1995; Dalby et al. 1998; Roux et al. 2004). Phosphorylation by either ERK1/2 or p90RSK is associated with decreased tuberin GAP activity and therefore enhanced signalling through mTOR (Roux et al. 2004; Ma et al. 2005). Previous studies showing that skeletal muscle contraction produced during physical activity or exercise results in activation of ERK1/2, and kinases upstream of ERK1/2 such as MEK1/2, provide a likely mechanism through which p90RSK is activated during exercise (Goodyear et al. 1996; Widegren et al. 2000, 2001).

AMPK regulates mTOR function through two mechanisms. In the first case, AMPK directly phosphorylates mTOR on T2446 (Cheng et al. 2004) which attenuates phosphorylation of S2448, a target of PKB signalling (Chiang & Abraham, 2005). Phosphorylation on S2448 usually correlates with increased signalling downstream of mTOR (Bolster et al. 2002; Chiang & Abraham, 2005), suggesting that its phosphorylation might activate mTOR protein kinase activity. Thus, phosphorylation on T2446 by AMPK may indirectly repress mTOR function by inhibiting PKB-mediated phosphorylation on S2448. The second mechanism through which AMPK regulates mTOR function is through phosphorylation of tuberin on T1227 and S1345 (Inoki et al. 2003b). In contrast to phosphorylation by PKB, ERK1/2 and p90RSK, phosphorylation of tuberin by AMPK promotes tuberin activation (Inoki et al. 2003b). Thus, by stimulating the intrinsic GTPase activity of Rheb, tuberin promotes the conversion of Rheb–GTP to Rheb–GDP and thereby represses signalling through mTOR.

In the present study, the effect of an acute bout of treadmill exercise on signalling through ERK1/2 and mTOR to downstream targets involved in the regulation of mRNA translation was examined. Exercise promoted phosphorylation of AMPK, MEK1/2 and ERK1/2, as well as phosphorylation of rpS6 on S235/S236, a target of both S6K1 and p90RSK. These exercise-induced changes were associated with a disaggregation of polysomes into free ribosomes suggesting an inhibition of global rates of mRNA translation. It also repressed signalling through mTOR, as assessed by decreased phosphorylation of 4E-BP1, enhanced association of mTOR with the regulatory-associated protein of mTOR (raptor), and dissociation of the tuberin–hamartin complex. Overall, the results suggest that endurance exercise activates signalling pathways that, individually, would have either positive (ERK1/2, p90RSK) or negative (AMPK) effects on mTOR function, but that in vivo the negative influence of AMPK activation predominates with a net reduction in signalling through mTOR and a repression of mRNA translation and protein synthesis.

Methods

Materials

Polyvinylidene difluoride (PVDF) membranes were purchased from Pall Life Sciences Corporation. Anti-S6K1 and anti-raptor antibodies were purchased from Bethyl Laboratories. Anti-tuberin was purchased from Santa Cruz Biotechnologies. All other antibodies were purchased from Cell Signalling Technology. Enhanced chemiluminescence (ECL) detection kits were purchased from Amersham Biosciences, and donkey anti-rabbit and sheep anti-mouse horseradish peroxidase-conjugated IgG were purchased from Bethyl Laboratories.

Treadmill exercise

All experimental procedures were approved by the Institutional Animal Care and Use Committee of the Pennsylvania State University College of Medicine. C57BL/6 male mice (∼19–21 g, 3 month old; Charles River Laboratories, Wilmington, MA, USA) were maintained on a 12: 12 h light–dark cycle, with free access to water and rodent chow (Harlan-Teklad Rodent Chow, Madison, WI, USA). Initially, mice were acclimatized to the motorized treadmill (Eco3/6, Columbus Instruments, Columbus, OH, USA) by running 10 min per day for 3 days. The grade was held constant at 10%, while the speed was gradually increased to 26 m min⁻¹ over the 3 day period. On the day of the experiment, food was taken away from the mice 3–4 h prior to exercise. Then the mice were run on a 10% grade at 26 m min⁻¹ for up to 30 min. Following the treadmill run, animals were anaesthetized by breathing a 95% O₂, 5% CO₂ gas mixture via a nose cone connected to an isoflurane vaporizer (3.0–4.0% isoflurane). Animals were deemed deeply anaesthetized only if they did not respond to numerous tactile stimuli (e.g. tail or foot pinch response and eye reflex response). These assessments were made frequently during the surgical protocol, and animals remained in a deeply anaesthetized state during all described procedures. Once the appropriate anaesthetic state was achieved, the gastrocnemius muscle was excised, immediately frozen in liquid nitrogen, and then stored at −80°C for subsequent analyses (see below). Animals were then killed by severing the diaphragm and heart while still deeply anaesthetized with isoflurane. Control mice were treated in a similar fashion except they did not run on the treadmill.

Analysis of Western blots

Protein immunoblots were visualized via enhanced chemiluminescence (ECL), as previously described (Kimball et al. 2002), then quantified by measuring the luminescent signal using a GeneGnome Bio-Imaging System (SynGene).

Quantification of AMPK, PKB, 4E-BP1, rpS6, MEK1/2, ERK1/2, p90RSK and eIF4E phosphorylation state

Frozen muscle (∼0.1 g) was homogenized in buffer containing 20 mm Hepes (pH 7.4), 2 mm EGTA, 50 mm NaF, 100 mm KCl, 0.2 mm EDTA, 50 mmβ-glycerolphosphate, 1% Triton X-100, 1% deoxycholate, 0.1 mm phenylmethylsulphonyl fluoride (PMSF), 1 mm benzamidine, 1 mm DTT, and 0.5 mm sodium vanadate, using a Polytron homogenizer PT10-35. The homogenate protein concentration was measured using a kit from Bio-Rad. Aliquots of the homogenate were combined with an equal volume of 2 × SDS sample buffer(SDS buffer composition: 125 nm Tris, pH 6.8, 25% glycerol, 2.5% SDS, 2.5%β-melcaptoethanol and 0.2% bromophenol blue), and proteins were resolved by electrophoresis on a 7.5% (S6K1), 10% (AMPK, PKB, p90RSK), 12.5% (MEK1/2, ERK1/2), or 15% (4E-BP1, rpS6, eIF4E) polyacrylamide gel. The proteins were transferred to polyvinylidene difluoride (PVDF) membranes, which were then incubated with anti-phosphopeptide antibodies directed against AMPK(T172), PKB(S473), 4E-BP1(T37/T46), rpS6(S235/236), rpS6(S240/244), MEK1/2(T217/221), ERK1/2(T202/Y204), p90RSK(S380), or eIF4E(S209). The immunoblots were developed and analysed as described above. The blots were then stripped of antibody and reprobed with antibodies that recognize AMPK, PKB, rpS6, MEK1/2, ERK1/2, p90RSK or eIF4E independent of phosphorylation state.

Measurement of cAMP content

The tissue content of cAMP was measured using a cAMP assay kit (RPA538) from Amersham Biosciences/GE Healthcare according to the manufacturer's protocol. Briefly, frozen muscle (∼0.05 g) was weighed and then homogenized in ice-cold 6% TCA. The homogenate was centrifuged and the resultant pellet was washed five times in water-saturated diethyl ether, then dried by centrifugation under vacuum. Samples were suspended in the assay buffer provided with the kit and then aliquots were added to tubes containing [¹²⁵I]cAMP, rabbit anti-cAMP antibody, and scintillation proximity assay (SPA)-fluomicrospheres containing bound anti-rabbit antibody. A separate set of tubes, containing known concentrations of cAMP, were similarly prepared. The tubes were incubated overnight at room temperature with constant mixing. The amount of [¹²⁵I]cAMP bound to the SPA-fluomicrospheres was measured by liquid scintillation spectrometry and the data were expressed as the per cent of [¹²⁵I]cAMP bound for each standard and sample. Tissue cAMP concentrations were calculated based upon the standard curve.

Measurement of PKA activity

The phosphotransferase activity of PKA was measured using a PKA assay kit (Upstate Cell Signalling Solutions). Frozen muscle (∼0.05 g) was homogenized in 7 volumes of 1 × assay dilution buffer consisting of 20 mm Mops (pH 7.2), 25 mmβ-glycerophosphate, 5 mm EGTA, 1 mm sodium vanadate, 0.1 mm dithiothreitol, 1 μm microcystin and 10 μl ml⁻¹ Sigma protease inhibitor mixture. The homogenate was centrifuged at 1000 g for 3 min at 4°C, and the protein concentration in the supernatant was measured using a kit from Bio-Rad. Supernatant containing 50 μg of protein was diluted to 5 μl with 1 × assay dilution buffer and then 25 μl of the assay mix (16 mm Mops (pH 7.2), 20 mmβ-glycerophosphate, 4 mm EGTA, 0.8 mm sodium vanadate, 80 μm dithiothreitol (DTT), 0.8 μm microcystin, 8 μl ml⁻¹ Sigma protease inhibitor mixture, 100 μm Kemptide, 7.5 mm magnesium chloride, 100 μm[³²P]ATP (1 μCi μmol⁻¹), 0.4 μm PKC inhibitor peptide and 4 μm compound R24571 (a calmodulin-dependent protein kinase inhibitor)) were added. A duplicate sample of diluted supernatant was assayed using the same assay mix containing 1 μm PKA inhibitor peptide. Both assay mixtures were incubated at 30°C for 5 min, and then 15 μl of the assay mixture were spotted onto a P81 phosphocellulose filter. The phosphocellulose filters were washed sequentially with 0.75% phosphoric acid and acetone, as described in the manufacturer's instructions, and the amount of radioactivity bound to the filters was measured by liquid scintillation spectrometry. PKA activity was calculated as the difference between the sample assayed in the absence of PKA inhibitor peptide and the sample assayed with the inhibitor.

Analysis of hamartin association with tuberin, tuberin phosphorylation state, and mTOR and raptor with mTOR

Frozen muscle (∼0.1 g) was homogenized in CHAPS buffer (40 mm Hepes (pH 7.5), 120 mm NaCl, 1 mm EDTA, 10 mm pyrophosphate, 10 mmβ-glycerolphosphate, 40 mm NaF, 1.5 mm sodium vanadate, 0.3% CHAPS, 0.1 mm PMSF, 1 mm benzamidine, and 1 mm DTT) and the homogenate was mixed on a platform rocker for 20 min at 4°C, and then clarified by centrifugation at 1000 g for 3 min (4°C). An aliquot of supernatant containing either 500 μg or 700 μg of protein was combined with 5.8 μl of anti-tuberin antibody or 1.4 μl of anti-mTOR antibody, respectively, and mixed on a platform rocker overnight at 4°C. Immune complexes were isolated with a goat anti-rabbit BioMag IgG (PerSeptive Diagnostics) bead slurry. Prior to use, the beads were blocked with 0.1% non-fat dry milk in CHAPS buffer and then washed in CHAPS buffer. After a 1 h incubation at 4°C, the beads were collected using a magnetic stand, washed twice with CHAPS buffer, and once in CHAPS buffer containing 200 mm instead of 120 mm NaCl and 60 mm instead of 40 mm Hepes. The immunoprecipitates were solubilized in 1 × SDS sample buffer, and then boiled for 5 min. The beads were removed by centrifugation and the supernatant was collected and subjected to SDS-PAGE. The proteins in the gel were transferred to PVDF membranes, which were then incubated in anti-tuberin antibody, anti-hamartin antibody, anti-raptor antibody, or anti-mTOR antibody overnight at 4°C. The blots were visualized by ECL, and then the ratios of hamartin to tuberin or raptor to total mTOR were calculated.

Analysis of S6K1 hyperphosphorylation state

Aliquots of muscle homogenate in 2 × SDS sample buffer were resolved by electrophoresis on a 7.5% (S6K1) polyacrylamide gel. The proteins were transferred to PVDF membrane, which was then incubated with anti-S6K1 polyclonal antibody; the blots were then visualized by ECL. Typically, S6K1 resolves into multiple bands following electrophoresis on SDS-polyacrylamide gels whereby the electrophoretic mobility of S6K1 is inversely proportional to the degree of phosphorylation. The fastest migrating, hypophosphorylated forms of the proteins are designated the α-bands. The relative phosphorylation of S6K1 was estimated as the proportion present in the hyperphosphorylated β-, γ- and δ-forms relative to the total amount of the protein.

Analysis of polysome aggregation

Sucrose density gradient centrifugation was employed to analyse muscle polysome aggregation state following treadmill exercise. Gastrocnemius samples were powdered (Bio-Pulverizer, BioSpec Products Inc., Bartlesville, OK, USA) under liquid nitrogen, and connective tissue was carefully removed. Powdered muscle tissue (∼0.1 g) was homogenized for 20 s in 10 volumes of resuspension buffer containing 50 mm Hepes (pH 7.4), 75 mm KCl, 5 mm MgCl₂, 250 mm sucrose, 1% Triton X-100, 1.3% deoxycholate, 100 μg ml⁻¹ cycloheximide, 25 μl ribonuclease inhibitor SUPERasin (per 5 ml) using a Polytron homogenizer PT10-35. Homogenates were incubated on ice for 5 min, and then 150 μl ml⁻¹ Tween–deoxycholate mix (1.34 ml Tween 20, 0.66 g deoxycholate, 18 ml sterile water) was added, and the samples were thoroughly mixed. Samples were incubated on ice for 15 min and then centrifuged at 1000 g for 15 min at 4°C. The resulting supernatant (600 μl) was layered on a 20–47% linear sucrose density gradient (50 mm Hepes (pH 7.4), 75 mm KCl, 5 mm MgCl₂) and centrifuged in a SW41 rotor at 260000 g for 180 min at 4°C. Following centrifugation, the gradient was displaced upward (2 ml min⁻¹) using Fluorinert (Isco, Lincoln, NE, USA) through a spectrophotometer, and the optical density at 254 nm was continuously recorded (chart speed, 150 cm h⁻¹).

Statistics

Data are expressed as means ±s.e.m. All comparisons were made to control conditions using a one-way ANOVA analysis (time course data) or two-tailed t test analysis (30 min data) using Prism v. 3.0 software (GraphPad Software). The significance level was set at P < 0.05.

Results

Previous studies (Gautsch et al. 1998) have shown that endurance exercise represses protein synthesis in skeletal muscle. However, the mechanism(s) through which treadmill exercise acts to repress muscle protein synthesis is poorly defined. In the present study, sucrose density gradient centrifugation was utilized to identify changes in polysome aggregation that would indicate whether exercise preferentially affected the initiation or elongation phase of mRNA translation. As shown in Fig. 1, a shift in the distribution of ribosomes from polysomes to monosomes was observed after 10 min (Fig. 1A) of treadmill running, with a further shift occurring following runs of 20 min (Fig. 1C) and 30 min (Fig. 1E). A shift in ribosomes from polysomes to monosomes indicated that translation initiation was inhibited relative to elongation.

Figure 1. Effect of acute treadmill exercise on polysome aggregation in mouse gastrocnemius muscle.

C57BL/6 male mice were exercised on a motorized treadmill at 26 m min⁻¹, 10% grade for 10 min (A), 20 min (B) or 30 min (C). Immediately following the exercise bout, the gastrocnemius muscle was removed, frozen between aluminium blocks pre-cooled in liquid nitrogen, and stored at −80°C. The frozen muscle was powdered under liquid nitrogen, then homogenized in lysis buffer containing cycloheximide, SUPERasin and detergent. The homogenate was clarified by centrifugation at 3000 g for 15 min, layered onto a 20–47% sucrose gradient, then centrifuged at 40 000 g for 5 h 20 min. Following centrifugation, the gradient was displaced upward through a spectrophotometer and the absorbance at 254 nm was continuously recorded. A representative profile from each condition is presented. The peaks corresponding to 40S and 60S ribosomal subunits, 80S monomers, and polysomes are denoted on the figure (n = 6–9 per group).

The initiation of mRNA translation is regulated in part through modulation of signalling through the mTOR pathway (Holz et al. 2005). To assess changes in mTOR signalling in response to treadmill running, the phosphorylation state of two downstream targets, 4E-BP1 and S6K1, was measured. As shown in Fig. 2A, phosphorylation of 4E-BP1 on T37/T46, a site directly phosphorylated by mTOR (Gingras et al. 1999), exhibited a continual decline over the 30 min time course, suggesting that mTOR signalling was repressed in response to exercise. Unfortunately, it was not possible to assess exercise-induced changes in phosphorylation of the mTOR site on S6K1 (T389) because phosphorylation of that site was too low to measure in muscle from control animals, most probably due to the 4 h fast before the exercise (Fig. 2B). In fact, essentially all of S6K1 was present in the hypophosphorylated α-form.

Figure 2. Effect of acute treadmill exercise on phosphorylation of 4E-BP1 and S6K1 in mouse gastrocnemius muscle.

C57BL/6 male mice were exercised as described in the legend to Fig. 1. Immediately following the exercise bout, the gastrocnemius muscle was removed, homogenized and clarified by centrifugation at 1000 g for 3 min. The resulting muscle homogenate was combined with an equal amount of 2 × SDS sample buffer. Equal amounts of protein from each sample were assessed for the phosphorylation of 4E-BP1 on T37/T46 (A), and the hyperphosphorylation status and phosphorylation on T389 of S6K1 (B) by Western blot analysis. Representative Western blots are shown. *P < 0.05 versus control conditions (n = 6–9 per group)

Further evidence supporting an exercise-induced repression of mTOR signalling was obtained by examination of the effect of treadmill running on upstream components of the mTOR signalling pathway. Recent studies suggest that the raptor–mTOR complex can adapt at least two conformations, one being a closed, less active conformation and the other an open, more active one (Kim et al. 2002). When the complex is in the open, active conformation, it is easily dissociated by the detergent CHAPS, whereas the closed, less active conformation is resistant to dissociation. To determine whether or not endurance exercise altered the association of raptor with mTOR, mTOR immunoprecipitates were washed with buffer containing CHAPS and then analysed by Western blot for raptor and mTOR content. As shown in Fig. 3A, the amount of raptor present in mTOR immunoprecipitates from muscle of exercised animals was significantly increased compared with sedentary controls, suggesting that exercise promotes a change in the association of raptor with mTOR such that the complex is present in the less active conformation. Signalling through mTOR is regulated in part by changes in assembly of the tuberin–hamartin complex where dissociation of the complex is associated with enhanced mTOR signalling (Inoki et al. 2002). In the present study, the amount of hamartin present in tuberin immunoprecipitates was increased in response to treadmill running (Fig. 3B), suggesting that exercise represses mTOR through activation of a signalling pathway(s) upstream of tuberin–hamartin.

Figure 3. Effect of acute treadmill exercise on the association of raptor with mTOR and hamarin with tuberin.

C57BL/6 male mice were exercised as described in the legend to Fig. 1. Immediately following the exercise bout, the gastrocnemius muscle was removed and homogenized. A, mTOR was immunoprecipitated as described in Methods and the immunoprecipitates (IP) were assessed for raptor and mTOR content by immunoblot (WB) analysis, and the ratio of raptor to mTOR was calculated. The results are expressed as a percentage of control. B, tuberin was immunoprecipitated as described in Methods and the immunoprecipitates were assessed for hamartin and tuberin content by immunoblot analysis, and the ratio of hamartin to tuberin was calculated. The results are expressed as a percentage of control. *P < 0.05 versus control conditions (n = 6 per group).

Signalling through tuberin is negatively regulated by the Akt/PKB (Fingar & Blenis, 2004) and MEK/ERK (Ma et al. 2005; Rolfe et al. 2005) signalling pathways whereas signalling through PKA/AMPK promotes tuberin function (Inoki et al. 2002). To determine whether treadmill running might repress mTOR signalling through inhibition of Akt/PKB, phosphorylation of Akt/PKB on S473 was measured by protein immunoblot analysis. As shown in Fig. 4A, no change in Akt/PKB phosphorylation was observed in muscle from mice run for 30 min compared with sedentary controls. Moreover, phosphorylation of tuberin on T1462, a site reported to be directly phosphorylated by Akt/PKB (Manning et al. 2002), was unchanged (Fig. 4B), suggesting that Akt/PKB activity was unaffected by treadmill running. Overall, the results suggest that treadmill running did not repress mTOR signalling through changes in the Akt/PKB signalling pathway.

Figure 4. Effect of acute treadmill exercise on PKB phosphorylation on S473 and tuberin phosphorylation on T1462.

C57BL/6 male mice were exercised as described in the legend to Fig. 1. Immediately following the exercise bout, the gastrocnemius muscle was removed, frozen between aluminium blocks pre-cooled in liquid nitrogen, and stored at −80°C until analysed. A, PKB phosphorylation on S473 was measured by protein immunoblot analysis of muscles homogenized in 2 × SDS sample buffer as described in Methods. Results were normalized for PKB content and are expressed as a percentage of control condition. B, phosphorylation of tuberin on T1462 was measured in tuberin immunoprecipitates as described in Methods. Values were normalized for tuberin content and are expressed as a percentage of the control value. *P < 0.05 versus control (n = 6 per group).

An alternative pathway through which exercise might repress signalling through mTOR involves activation of AMPK. An important step in activation of AMPK is phosphorylation of its α-subunit on T172 (Hawley et al. 1996). As shown in Fig. 5A, phosphorylation of AMPK on T172 was rapidly increased in response to treadmill running. One pathway through which AMPK activity may be regulated is through activation of PKA (Collins et al. 2000). Therefore, the cAMP content and PKA activity were measured in muscle from control and exercised mice. As shown in Fig. 5B, cAMP content was significantly increased after a 30 min run. In addition, PKA activity was increased over two-fold in response to 30 min of exercise (Fig. 5C).

Figure 5. Effect of acute treadmill exercise on AMPK phosphorylation, cAMP concentration and PKA activity in mouse gastrocnemius muscle.

C57BL/6 male mice were exercised as described in the legend to Fig. 1. A, immediately following the exercise bout, the gastrocnemius muscle was removed, homogenized and clarified by centrifugation at 1000 g for 3 min. The resulting muscle homogenate was combined with an equal volume of 2 × SDS sample buffer and equal amounts of protein from each sample were assessed for the phosphorylation of AMPK on T172 by Western blot analysis. The results are expressed as a percentage of the control conditions. Representative Western blots are shown (CON; control). For analysis of cAMP concentration (B) and PKA activity (C), gastrocnemius muscle was quickly removed and frozen between aluminium blocks pre-cooled in liquid nitrogen and then stored at −80°C until analysed. *P < 0.05 versus control (n = 6–9 per group).

In addition to repressing mTOR signalling, activation of AMPK leads to enhanced signalling through the MEK/ERK signalling pathway (Kim et al. 2001; Chen et al. 2002). As shown in Fig. 6A and B, phosphorylation of both MEK1/2 on S217/221 and ERK1/2 on T202/Y204 was rapidly increased in response to treadmill running. To assess the functional significance of ERK1/2 activation, phosphorylation of two downstream targets of the MEK/ERK signalling pathway, p90RSK and eIF4E, was assessed. Phosphorylation of p90RSK on S380 was unchanged after a 10 min treadmill run, but was significantly increased in response to exercise of longer duration (Fig. 6C). Similarly, phosphorylation of eIF4E was unchanged after a 10 or 20 min run, but was increased two-fold after 30 min of exercise.

Figure 6. Effect of acute treadmill exercise on phosphorylation of MEK1/2 on S217/S221, ERK1/2 on T202/Y204, p90RSK on S380, and eIF4E on S209 in mouse gastrocnemius muscle.

C57BL/6 male mice were exercised as described in the legend to Fig. 1. Immediately following the exercise bout, the gastrocnemius muscle was removed, homogenized, and clarified by centrifugation at 1000 g for 3 min. The resulting muscle homogenate was combined with an equal volume of 2 × SDS sample buffer and equal amounts of protein from each sample were assessed for the phosphorylation of MEK1/2 on S217/S221 (A), ERK1/2 on T202/Y204 (B), p90RSK on S380 (C), and eIF4E on S209 (D) by Western blot analysis. Results are expressed as a percentage of the control. Representative Western blots are shown. *P < 0.05 versus control (n = 6–9 per group).

Phosphorylation of rpS6 is mediated by both S6K1 and p90RSK, whereby S6K1 phosphorylates two sets of sites, S235/236 and S240/244, and p90RSK phosphorylates only S235/236 (Pende et al. 2004). As shown in Fig. 7, phosphorylation of S240/244 was undetectable in muscle from either control or exercised mice, a result consistent with the finding that S6K1 phosphorylation was low. In contrast, phosphorylation of rpS6 on S235/236 was rapidly enhanced in response to treadmill running, suggesting that phosphorylation of these sites is mediated by p90RSK in response to exercise.

Figure 7. Effect of acute treadmill exercise on phosphorylation of ribosomal protein (rp) S6 in mouse gastrocnemius muscle.

C57BL/6 male mice were exercised as described in the legend to Fig. 1. Immediately following the exercise bout, the gastrocnemius muscle was removed, homogenized, and clarified by centrifugation at 1000 g for 3 min. The resulting muscle homogenate was combined with an equal volume of 2 × SDS sample buffer and equal amounts of protein from each sample were assessed for the phosphorylation of rpS6 on S235/S236 and S240/S244 by Western blot analysis. Results are expressed as a percentage of the control. Representative Western blots are shown. *P < 0.05 versus control (n = 6–9 per group).

Discussion

The results of the present study suggest that during acute endurance exercise, signalling through mTOR is repressed, as assessed by decreased phosphorylation of the mTOR target 4E-BP1 (Fig. 8). Thus, decreased mTOR signalling probably accounts, in part, for the previously reported decline in protein synthesis associated with treadmill running (Gautsch et al. 1998). In contrast, mTOR signalling is enhanced following a single bout of resistance exercise in association with increased rates of global protein synthesis (Bolster et al. 2003; Kubica et al. 2005). Moreover, in isolated muscle preparations, high frequency, but not low frequency, electrical stimulation enhances mTOR signalling and increases protein synthesis (Atherton et al. 2005). Thus, fundamental differences exist between high-intensity, resistance-type exercise and low-intensity, endurance-type exercise in regard to activation of signal transduction pathways and regulation of mRNA translation.

Figure 8. Signal transduction pathways through which endurance exercise regulates mRNA translation in skeletal muscle.

Endurance exercise promotes the activation of multiple signalling pathways in skeletal muscle including the extracellular signal-regulated kinase (ERK)1/2 and protein kinase A (PKA) pathways. Both ERK1/2 and the downstream effector, p90 ribosomal protein S6 kinase (p90RSK), phosphorylate and inhibit the mTOR repressor complex consisting of tuberin and hamartin. In contrast, PKA phosphorylates the AMP-activated protein kinase (AMPK), which subsequently phosphorylates and activates the tuberin–hamartin complex. Thus, activation of ERK1/2 and p90RSK activate, whereas protein kinase A/AMPK inhibit mTOR signalling to eIF4E binding protein (4E-BP1) and promotes the ribosomal protein S6 kinase (S6K1). Simultaneously, ERK1/2 phosphorylation of eIF4E. A continuous line represents direct signalling whereas a dashed line represents indirect signalling involving intermediate effectors that are not shown for simplicity. A line with an arrowhead represents activation whereas a line ending in a perpendicular line represents inhibition.

The finding in the present study that the time course of changes in 4E-BP1 phosphorylation mirrors that for alterations in AMPK phosphorylation, together with results from previous studies showing that AMPK phosphorylates both tuberin (Inoki et al. 2003b) and mTOR (Cheng et al. 2004), suggest that a likely pathway through which endurance exercise causes decreased signalling through mTOR involves activation of AMPK. The other two known signalling pathways through which mTOR function is regulated, i.e. the Akt/PKB and MEK/ERK signalling pathways, do not seem to be involved in the exercise-induced repression of mTOR signalling. Thus, in the present study, Akt/PKB phosphorylation and phosphorylation of Akt/PKB-directed sites on tuberin were unchanged immediately after cessation of treadmill running. Moreover, rather than a reduction in MEK/ERK signalling, which would be expected to promote repression of mTOR-dependent signalling, phosphorylation of both MEK1/2 and ERK1/2 was increased by treadmill running.

Signalling from AMPK, PKB and ERK1/2-p90RSK converges upstream of mTOR on tuberin. Tuberin functions as a binary complex with hamartin to stimulate the GTPase activity of Rheb and thereby convert it to the GDP-bound form (Castro et al. 2003; Inoki et al. 2003a). Binding of guanine nucleotides to Rheb is essential for mTOR function, because a Rheb variant that cannot bind to GDP or GTP still binds to mTOR, but the resulting mTOR–Rheb complex has no detectable protein kinase activity (Tee et al. 2003; Long et al. 2005). In contrast, binding of either Rheb–GDP or Rheb–GTP to mTOR yields a functional complex, but the mTOR–Rheb–GTP complex exhibits significantly greater activity compared with the GDP complex. Thus, activation of AMPK during exercise would be expected to promote conversion of the active mTOR–Rheb–GTP complex to the inactive GDP complex. Unfortunately, it was not technically feasible to perform such measurements in the present study. Measurement of GDP and GTP bound to Rheb is typically performed in cells in culture by radiolabelling nucleotide pools using ³²P_i, followed by immunoprecipitation of Rheb, and resolution of Rheb-bound GDP and GTP by thin layer chromatography. Because of the impracticality in using ³²P_i in animals in vivo, such measurements were not attempted in the studies described herein.

In addition to Rheb, mTOR function is regulated through its association with at least three other proteins. For example, raptor binds to mTOR and formation of the mTOR–raptor complex greatly enhances phosphorylation of downstream targets, such as 4E-BP1 and S6K1, that bear a mTOR signalling (TOS) motif (Schalm et al. 2003). Association of mTOR with raptor is regulated by nutrients and hormones through an incompletely defined process, although another mTOR interacting protein, G protein β-subunit-like protein (GβL; also known as LST8) is probably involved in the effect (Kim et al. 2002, 2003). Similarly, binding of mTOR to eIF3 is also enhanced by nutrients and hormones (Holz et al. 2005). eIF3 is thought to act as a scaffold that brings together the mTOR–raptor complex and the mTOR substrates, 4E-BP1 and S6K1, permitting their phosphorylation. In the present study, the amount of raptor associated with mTOR in mTOR immunoprecipitates washed with buffer containing CHAPS was increased in response to treadmill running. This finding suggests that exercise promotes a conformational change in the raptor–mTOR complex, resulting in a shift into the closed, less active conformation. Whether or not endurance exercise might modulate the association of mTOR with GβL or eIF3 is unknown, but is an obvious question for future studies.

In cells in culture, activation of either the Akt/PKB or the MEK/ERK signalling pathway in the absence of AMPK activation enhances signalling through mTOR (Shaw et al. 2004). However, results from a recent study (Hahn-Windgassen et al. 2005) reveal that activation of AMPK overwhelms insulin-induced activation of Akt/PKB, and results in repression of mTOR signalling. The results of the present study suggest that a similar scenario occurs in skeletal muscle during exercise. Thus, treadmill running enhanced phosphorylation of both AMPK and MEK/ERK. However, as assessed by phosphorylation of downstream targets, signalling through mTOR was reduced, suggesting that the negative influence of AMPK prevailed over the positive input from ERK and p90RSK. Yet, even though mTOR function was repressed, phosphorylation of rpS6 on S235/S236 was increased in exercised compared with control muscle. Typically, phosphorylation of rpS6 is used as an index of S6K1 function. However, in addition to phosphorylation by S6K1, the S235/S236 sites are also phosphorylated by p90RSK (Pende et al. 2004). The finding that phosphorylation of p90RSK is increased immediately after a bout of treadmill running may explain, in part, the enhanced phosphorylation of rpS6 on S235/S236.

For a number of years, phosphorylation of rpS6 by S6K1 was thought to enhance selectively the translation of mRNAs with an uninterrupted string of pyrimidine residues, referred to as a TOP motif, immediately downstream of the mRNA m⁷GTP cap structure (Terada et al. 1994; Loreni et al. 2000). However, the individual roles of the five phosphorylation sites on rpS6 in mRNA translation is as yet undefined. In particular, a possible role for phosphorylation of S235/S236 in the absence of phosphorylation of S240/S244 in the selective translation of mRNAs encoding specific proteins is unexplored. Moreover, results of recent studies (Tang et al. 2001; Stolovich et al. 2002; Pende et al. 2004; Ruvinsky et al. 2005) suggest that phosphorylation of rpS6 is not required for enhanced translation of TOP mRNAs. Whether or not the translation of other mRNAs might be selectively regulated by rpS6 phosphorylation is unknown.

Another process through which ERK1/2 might regulate mRNA translation is through phosphorylation of eIF4E. eIF4E is phosphorylated on S209 by the MAP kinase-interacting kinases (MNK)1 and MNK2 (Fukunaga & Hunter, 1997; Waskiewicz et al. 1999). MNK1/2 are phosphorylated by both ERK1/2 and p38 MAP kinases (Waskiewicz et al. 1997). However, in the present study, phosphorylation of p38 MAP kinase was unchanged in response to treadmill running (data not shown), suggesting that the observed increase in eIF4E phosphorylation was due to activation of ERK1/2 rather than p38 MAP kinase. A number of studies have shown a correlation between enhanced phosphorylation of eIF4E and increased rates of protein synthesis (Scheper & Proud, 2002). However, more recent studies show that phosphorylation of eIF4E on S209 reduces the affinity of the protein for the m⁷GTP cap structure (Scheper et al. 2002), a finding that implies that eIF4E phosphorylation should inhibit cap-dependent mRNA translation. In contrast, another study (Knauf et al. 2001) shows that increased eIF4E phosphorylation mediated by exogenous expression of either wild-type or a constitutively active variant of MNK2, but not a kinase-dead variant, has no effect on cap-dependent mRNA translation, but instead, enhances cap-independent (i.e. internal ribosome entry site (IRES)-dependent) mRNA translation. Thus, rather than affecting global rates of protein synthesis, phosphorylation of eIF4E may lead to selective translation of mRNAs encoding specific proteins. The idea that eIF4E phosphorylation has no effect on global rates of protein synthesis is supported by the finding that in mice lacking both MNK1 and MNK2, eIF4E phosphorylation is undetectable in all tissues, yet the animals are viable and develop normally (Ueda et al. 2004). In addition, in fibroblasts lacking both kinases, eIF4E phosphorylation is completely absent, but global rates of protein synthesis and cap-dependent mRNA translation are the same as in wild-type cells, even under conditions in which eIF4E phosphorylation should be enhanced.

The activation of AMPK in response to endurance exercise is probably dependent upon LKB1 because, in contrast to wild-type mice, in mice lacking LKB1, muscle contraction induced by electrical stimulation of the sciatic nerve has no effect on AMPKα2 activity or phosphorylation (Sakamoto et al. 2004). In part, LKB1-mediated phosphorylation of AMPKα is probably a result of an increase in the ratio of AMP: ATP. However, phosphorylation of LKB1 by PKA may also play a role in the effect. At the onset of exercise, the plasma concentration of catecholamines (e.g. adrenaline (epinephrine)) and other hormones that act through PKA (e.g. glucagon) increase while growth-promoting hormones such as insulin, decline (Galbo et al. 1975; Goldfarb et al. 1989; Sheldon et al. 1993). The increase in plasma adrenaline concentrations results in activation of β-adrenergic receptors and subsequently GTP-dependent adenylate cyclase in muscle (Chasiotis et al. 1983). Activation of adenylate cyclase results in production of cAMP, thereby activating PKA. Studies showing that protein kinase A (PKA) phosphorylates LKB1 (Collins et al. 2000; Sapkota et al. 2001) and that in liver glucagon enhances phosphorylation of both LKB1 and AMPK (Kimball et al. 2004) suggest that the activation of PKA observed in the present study might be involved in up-regulation of AMPK phosphorylation during treadmill running.

Overall, the results of the present study are consistent with a model wherein endurance exercise promotes activation of AMPK in skeletal muscle leading to enhanced association of tuberin with hamartin and repressed mTOR signalling to biomarkers of translation initiation. In this model, inhibition of mTOR signalling occurs despite enhanced signalling through the MEK/ERK/p90RSK signalling pathway, suggesting that AMPK-mediated repression of mTOR signalling overcomes the stimulatory effect of ERK- and p90RSK-mediated signalling. Moreover, even though mTOR signalling to downstream targets relevant to mRNA translation is repressed, signalling through the MAP kinase signalling pathway enhances phosphorylation of eIF4E on S209 and rpS6 on S235/S236. Although the functional consequence of such phosphorylation events has not been addressed herein, it is tempting to speculate that they might lead to preferential translation of mRNAs encoding specific proteins that might perhaps be involved in recovery during the post-exercise period. Future studies will be required to examine that possibility.