Regulatory, Integrative and Comparative Physiology

Resistance exercise decreases eIF2Bε phosphorylation and potentiates the feeding-induced stimulation of p70S6K1 and rpS6 in young men

Elisa I. Glover, Bryan R. Oates, Jason E. Tang, Daniel R. Moore, Mark A. Tarnopolsky, Stuart M. Phillips


We investigated the effect of resistance exercise and feeding on the activation of signaling proteins involved in translation initiation. Nine young men (23.7 ± 0.41 yr; BMI = 25.5 ± 1.0 kg/m2; means ± SE) were tested twice after they performed a strenuous bout of unilateral resistance exercise, such that their contralateral leg acted as a nonexercised comparator, in either the fasted and fed [1,000 kJ, each 90 min (3 doses): 10 g protein, 41 g carbohydrate, 4 g fat] states. Muscle biopsies were obtained 6 h postexercise from both legs, resulting in four experimental conditions: rest-fasted, rest-fed, exercise-fasted, and exercise-fed. Feeding increased PKB/Akt (Ser473) phosphorylation (P < 0.05), while exercise increased the phosphorylation of Akt and the downstream 70 kDa S6 protein kinase (p70S6K1, Thr389) and ribosomal protein S6 (rpS6, Ser235/236, Ser240/244; all P < 0.05). The combination of resistance exercise and feeding increased the phosphorylation of p70S6K1 (Thr389) and rpS6 (Ser240/244) above exercise alone (P < 0.05). Exercise also reduced phosphorylation of the catalytic epsilon subunit of eukaryotic initiation factor 2B (eIF2Bε, Ser540; P < 0.05). Mammalian target of rapamycin (mTOR, Ser2448), glycogen synthase kinase-3β (GSK-3β, Ser9), and focal adhesion kinase (FAK, Tyr576/577) phosphorylation were unaffected by either feeding or resistance exercise (all P > 0.14). In summary, feeding resulted in phosphorylation of Akt, while resistance exercise stimulated phosphorylation of Akt, p70S6K1, rpS6, and dephosphorylation eIF2Bε with a synergistic effect of feeding and exercise on p70S6K1 and its downstream target rpS6. We conclude that resistance exercise potentiates the effect of feeding on the phosphorylation and presumably activation of critical proteins involved in the regulation of muscle protein synthesis in young men.

  • hypertrophy
  • lean body mass
  • protein accretion
  • weightlifting

muscle protein synthesis (MPS) is synergistically increased by resistance exercise (3, 12, 17, 18, 48, 61) and feeding (4, 25, 51, 57). Although there are resistance exercise-induced changes in muscle protein breakdown (3, 48, 49), they are small by comparison to changes in MPS, which varies 3- to 4-fold between the fed and postabsorptive states and also with performance of resistive exercise (47, 52, 53). These observations point to the regulation of MPS as the primary locus of control in determining resistance exercise-induced changes in muscle protein mass. Moreover, while resistance exercise obviously stimulates transcription of genes relevant to adaptation and ultimately hypertrophy (for review, see Ref. 29), in the absence of a translational response (i.e., protein synthesis), transcriptional changes will not affect protein content (9, 34).

A number of studies in rodents (21, 22) and humans (5, 14, 17, 18, 20, 32, 36, 37) have characterized signaling proteins that are activated (i.e., dephosphorylated) in skeletal muscle with resistance exercise (18, 20, 37) and in combination with feeding branched-chain amino acids (5, 19, 32), or a mixture of essential amino acids/protein and carbohydrate (14, 17, 36). During recovery from resistance exercise, energy charge of the muscle is restored and phosphorylation of signaling proteins such as mTOR, the 70-KDa ribosomal protein kinase (p70S6K1), and ribosomal protein S6 (rpS6) are increased. Phosphorylation of these proteins can occur in response to contractile activity (18, 20, 37) but also occur with feeding (14, 17, 32, 36). In fact, feeding (i.e., amino acids and/or insulin) per se can increase phosphorylation of many signaling proteins involved in activation of MPS (25, 3941).

We examined how resistance exercise independently, and superimposed against the background of mixed meal feeding, impacted phosphorylation of proteins thought to be relevant in translation initiation (Akt/PKB, mTOR, p70S6K1, rpS6), ribosomal recycling (eIF2Bε, GSK-3β), and mechanotransduction (FAK). Several studies have demonstrated that the resistance exercise-induced increase in MPS persists for several hours (e.g., 8–48 h) (33, 43, 46, 48, 55). As feeding synergistically interacts with resistance exercise to further increase MPS (4, 46, 51, 57), it is likely that feeding at any time after acute resistance exercise would increase MPS to a greater degree than feeding alone (47). For this reason, a biopsy sampling time point of 6 h following resistance exercise was selected to examine how muscle contractile activity (or skeletal muscle loading) affects the anabolic response to feeding late into recovery. This muscle sampling time allows for a time that we know from previous work when MPS is still elevated (46, 48); however, the anabolic response of feeding alone may be reduced (7). Previous studies have demonstrated that muscle is able to respond to repeated boluses of amino acids, even when plasma levels remain elevated above resting levels (11), presumably because protein synthesis is modulated by changes in extracellular amino acid availability (6). Of note, however, is that MPS becomes refractory to provision of amino acids, when supplied by continuous infusion, after ∼2 h (7). Thus, in contrast to a single large bolus (17, 19) or intravenous feeding (4), we chose to provide aliquots of a mixed meal supplement every 90 min in an attempt to mimic a more realistic pattern of food consumption after exercise. Subjects were studied at rest and after resistance exercise in both the fed and fasted states. Because MPS is synergistically stimulated by feeding and resistance exercise (4, 46, 51, 57), we hypothesized that this response may be underpinned by a feeding and exercise-mediated synergistic stimulation of key signaling proteins involved in supporting enhanced MPS.



Subjects (n = 9 males; age = 23.7 ± 0.4 yr; mass = 80.8 ± 4.0 kg; height = 178 ± 2 cm; BMI = 25.5 ± 1.0 kg/m2; means ± SE) were recruited locally from the McMaster University campus via poster advertisements. All subjects completed a standard health questionnaire and reported being nonsmokers, were not taking any prescription medication, and were free from any medical conditions that would preclude their participation in a study of exercise and metabolic responses. Subjects were all habitually active and actively resistance trained (average 3 or 4 days/wk) at the time of the study. All subjects gave their written and informed consent prior to any participation in the research study. All procedures were approved by the Hamilton Health Sciences and local McMaster University Research Ethics Board and conformed to the Helsinki Declaration of 1983 on the use of human subjects in research.

Study design.

A week before any experimental trials, all subjects participated in a familiarization session where their unilateral 10 repetition maximum (RM) was determined for each leg for both leg press and knee extension exercises. Subjects reported, after having consumed no food after 2200 the evening prior, to the Exercise Metabolism Research Laboratory at 0700 on two different occasions separated by at least 2 wk. Each subject completed the fasted or fed condition trials in a randomized and counter-balanced manner. Subjects began each trial by performing a bout of unilateral leg press and knee extension exercise (4 sets of each at a workload equivalent to their previously determined 10 RM), which was again randomly assigned but counterbalanced based on voluntary strength. In this way, within two trials, we obtained legs that represented one of four conditions: rest-fasted (Rest-Fast), exercise-fasted (Ex-Fast), rest-fed (Rest-Fed), and exercise-fed (Ex-Fed).

Following the bout of exercise, subjects had a 20-gauge catheter inserted into a prominent dorsal hand vein; the catheter was kept patent by means of a 0.9% saline drip. Arterialized blood samples were obtained by warming the hand with a heating blanket (∼50°C) with all blood samples being drawn into heparinized tubes every 30 min postexercise (Fig. 1). During the fed trial, subjects consumed a mixed-meal drink (Boost, Novartis Nutrition, Mississauga, ON, Canada) containing 1,000 kJ and 10 g protein (as casein proteins), 41 g carbohydrate (as maltodextrin and sucrose), and 4 g of fat (as soy oil) at t = 90, 180, and 270 min postexercise. Muscle samples were obtained from the vastus lateralis of both legs within 10 min of each other at t = 360 min postexercise. Biopsies were taken under local anaesthetic (2% lidocaine) using a 5-mm Bergstrom needle modified for manual suction. Upon excision, samples were blotted free of blood and immediately snap frozen in liquid N2. Samples were stored at −80°C until analysis.

Fig. 1.

Schematic representation of the study protocol.

Blood analyses.

Whole blood was precipitated in 500 μl of 0.6 M perchloric acid (PCA) and neutralized by the addition of 1.25 M KHCO3 as previously described (45, 59). Plasma was obtained from the remaining whole blood by centrifugation (4500 g for 10 min at 4°C). All blood samples were stored at −20°C until analysis. Plasma was extracted in PCA, as described above and analyzed for amino acid content using HPLC methods that have been described previously (59). Whole blood glucose concentrations were determined using fluorometric methods (42). Plasma insulin was analyzed using a commercially available radioimmunoassay kit (Diagnostic Products, Los Angeles, CA). Inter-run and intra-run CVs for these assays were all less than 5%.

Western Blot analyses.

A small piece of wet muscle (∼20 mg) was homogenized by hand on ice in a 20 mM Tris (pH 7.2) buffer containing 1 mM Na3VO4, 50 mM NaF, 40 mM β-glycerolphosphate, 20 mM sodium pyrophosphate, 0.5% vol/vol Triton-X 100, and Complete Protease Inhibitor Mini-Tabs (Roche, Indianapolis, IN). Protein content of the homogenates was determined by the Bradford assay.

Samples (50 μg of protein) were loaded on 7.5 or 10% SDS-polyacrylamide gels and then transferred to a PVDF membrane. Membranes were blocked with 5% BSA (wt/vol) in Tris-buffered saline with 0.1% Tween (vol/vol) (TBST), and then incubated overnight in primary antibody overnight at 4°C: FAK Tyr576/577 (Santa Cruz Biotechnology, Santa Cruz, CA; no. 21831R, 1:1,000); total FAK (Santa Cruz Biotechnology, no. 558, 1:1,000); p70S6K1 Thr389 (Santa Cruz Biotechnology; no. 11759, 1:1,000); total p70S6K1 (Santa Cruz Biotechnology, no. 230, 1:1,000); GSK3β Ser9 (Cell Signaling Technology, Danvers, MA; 9336; 1:2,000); total GSK3β (Cell Signaling Technology, no. 9332; 1:2,000); rpS6 Ser235/236 (Cell Signaling Technology, no. 2211; 1:2,000); total rpS6 (Cell Signaling Technology; no. 2217; 1:1,000); mTOR Ser2448 (Cell Signaling Technology, no. 2971; 1:1,000); total mTOR (Cell Signaling Technology, no. 2972; 1:1,000); PKB/Akt Ser473 (Cell Signaling Technology, no. 9271; 1:1,000); total PKB/Akt (Cell Signaling Technology; no. 9272, 1:1,000), eIF2Bε Ser540 (Genetex, San Antonio, TX; no. GTX24775, 1:8,000); total eIF2Bε (Abcam, Cambridge, MA; no. ab32713, 1:1,500). After washing in TBST, membranes were incubated in HRP-linked anti-rabbit IgG secondary antibody (Amersham Biosciences, Piscataway, NJ; no. NA934V, 1:6,000), washed with TBST, developed using ECL (Amersham Biosciences; model no. RPN2106), and wrapped in Saran Wrap. Membranes were exposed to X-ray film (Kodak, Rochester, NY). Films were scanned with a CanoScan N1220U scanner, and bands were quantified with AlphaEase Fluorchem SP software (Alpha Innotech, San Leandro, CA). Membranes were probed with antibodies against the phosphorylated forms first, then incubated with stripping buffer (25 mM glycine-HCl pH 2.0, 1% SDS), and washed before incubation with primary antibody against total protein, with the exception of rpS6, for which the total form was probed first.

Statistical analyses.

Amino acid, glucose, and insulin data were analyzed using a two-factor (feeding vs. time) repeated-measures ANOVA. Data were analyzed using a 2-way ANOVA (feeding and exercise as factors). Significant effects were analyzed with post hoc preplanned comparisons using a Student's t-test. Significance was accepted at P < 0.05. All data are expressed as means ± SE.


Insulin and glucose.

Subjects ingested a total of 30 g of protein, 123 g of carbohydrate, and 12 g of fat (3,000 kJ or 37 kJ/kg) over the course of the protocol. Ingestion of the supplement drink every 90 min resulted in increments in blood glucose and plasma insulin above those seen in the fasted trial that coincided with meal consumption. Elevations in blood glucose and plasma insulin are shown in Fig. 2, A and B. The area under the insulin concentration by time curve in the fed state was 269% greater than in the fasted trial (which represents the response of insulin to exercise alone).

Fig. 2.

Plasma glucose (A) and plasma insulin (B) during the protocol. Values are expressed as means ± SE. *Significantly different (P < 0.05) from the same time point during the fasted-state trial. Time-dependent changes are omitted for clarity.

Amino acids.

Increases in whole blood amino acid concentration were seen following each bolus drink. Changes in essential amino acid (EAA) were greater in the fed than in the fasted trial (Fig. 3). Changes in leucine, branched chain, and total amino acids followed the same pattern (data not shown).

Fig. 3.

Plasma concentration of sum of essential amino acids during the protocol (leucine, isoleucine, valine, lysine, histidine, methionine, threonine, and phenylalanine). Values are presented as means ± SE. *Significantly different from the same time point during the fasted-state trial, P < 0.05. Time-dependent changes are omitted for clarity.

Signaling proteins.

Phosphorylation (Ser473) of PKB/Akt was increased in the rested condition due to feeding (Fig. 4A) and was also elevated by exercise in the fasted state. mTOR itself was not phosphorylated (Ser2448) to any significant degree due to either feeding or exercise (Fig. 5B). Downstream of mTOR, the phosphorylation of p70S6K1 (Thr389) was robustly increased by exercise, with a further increase seen with feeding (Fig. 4C). The p70S6K1 target protein rpS6 was also phosphorylated (Ser235/236) as a result of exercise, with no additional effect of nutritional provision on the Ser235/236 site; however, as with p70S6K1, we observed that nutritional provision did result in increased phosphorylation of rpS6 Ser240/244 beyond the activation induced by exercise alone (Fig. 4D). Phosphorylation of eIF2Bε at Ser540 strongly decreased in response to exercise regardless of nutritional state, with no effect of feeding at rest (Fig. 5A). Glycogen synthase kinase-3β phophorylation (Ser9) was also measured and found not to be affected by either nutrition or exercise or the combination of the two interventions (Fig. 5C). Similarly, we did not see any change in the phosphorylation status of focal adhesion kinase (FAK, Tyr576/577; Fig. 5D).

Fig. 4.

Phosphorylated/total of PKB/Akt (Ser473) (A), p70S6K1 (Thr389) (B), rps6 (Ser 235/236) (C), and rps6 (Ser 240/244) (D). Values are expressed as means ± SE. Significantly different from REST in the same nutritional condition, *P < 0.05, **P < 0.01; +significantly different from FAST in the same activity state, P < 0.05.

Fig. 5.

Phosphorylated/total of eIF2Bε (Ser540) (A), mTOR (Ser2448) (B), GSK-3β (Ser9) (C), and FAK (Tyr576/577) (D). Values are expressed as means ± SE. Significantly different from REST in the same nutritional condition, *P < 0.05, **P < 0.01.


We report here for the first time that resistance exercise is a potent stimulus for modifying the phosphorylation of eIF2Bε in humans. PKB/Akt, p70S6K1, and rpS6, known regulators of MPS, are activated in response to exercise, while a sustained (insulin mediated) feeding effect is seen for PKB/Akt. Amino acid provision following resistance exercise synergistically stimulates MPS (4, 17, 46), and this combined effect was observed for the phosphorylation of p70S6K1 and rps6, even 6 h after exercise and 1.5 h after feeding.

In the first stage of translation initiation, eIF2 recruits the initiatior methionyl-tRNA to the 40S ribosomal subunit, a process that requires GTP. eIF2B catalyzes the exchange of GDP for GTP on eIF2, renewing eIF2's tRNAi binding capacity (34). Recent data (35, 38) have suggested that “global” protein synthesis, which would appear to be stimulated by both feeding and resistance exercise (13, 14, 44), is primarily regulated by the activation of eIF2B. For instance, Farrell et al. (21, 22) has reported, in a rat model of plyometric resistance (approximating resistance exercise), that eIF2B activity was increased, whereas eIF4E-4G complex formation was not and thus eIF2B appeared to be more important than 4E binding 4G in regulating peptide chain initiation after resistance exercise. As such, we sought to examine the GSK-3β-mediated phosphorylation of eIF2B's catalytic ε subunit at Ser540 (rat Ser535) (58). Our results show for the first time in humans that resistance exercise results in a striking dephosphorylation of eIF2Bε at Ser, regardless of feeding status. Reduced phosphorylation at residue Ser540 releases eIF2Bε from an inhibitory state and would likely contribute to an increase in its guanine nucleotide exchange activity (58). Considering our measure of phosphorylation was made at a time when muscle protein synthesis would be markedly elevated (46, 48), these results suggest that, similar to rats (21, 22), eIF2B also plays an important role in the regulation of protein synthesis in response to resistance exercise in humans.

Interestingly, neither feeding nor resistance exercise impacted GSK3β's phosphorylation status. It may not be overly surprising that feeding did not affect the phosphorylation of this protein, since Liu et al. (40) did not see any effect of infused amino acids or low-dose insulin on GSK-3β phosphorylation. Blomstrand et al. (5) also reported little effect of exercise on GSK3β, either in the absence or presence of branched chain amino acids. It is possible that our failure to observe reciprocal changes in GSK3β and eIF2Bε phosphorylation is the result of the timing of our biopsy sample, whereby Akt-mediated changes in GSK3β may have returned to baseline with their effects still manifested in the phosphorylation state of eIF2Bε. Drummond et al. (19) recently reported a reduction in GSK3β phosphorylation 5 h after an EAA drink consumed 1 h postexercise, a finding that is difficult to interpret given that Akt phosphorylation transiently increased in that study, and a reduction in GSK3β's phosphorylation would predictably increase its activity. Inactivation of eIF2B by GSK3β first requires the phosphorylation of its epsilon subunit at a number of different residues (58). It is possible that the phosphorylation of these priming sites is reduced by resistance exercise, thereby interfering with the ability of GSK to repress the activity of eIF2Bε through phosphorylation at residue Ser540. Alternatively, Kimball et al. (34) has suggested that exercise may modulate eIF2B activity through dephosphorylation of the epsilon subunit. IGF-1-induced activation of eIF2Bepsilon in neuronal cells is mediated by dephosphorylation by protein phosphatase 1 (50). Future studies should investigate whether exercise-induced decreases in phosphorylation of eIF2Bε are mediated by a pathway independent of GSK3β.

Several studies in humans have recently examined the effect of resistive-type exercise on the phosphorylation of a number of proteins believed to regulate MPS, either in the fasted state (5, 17, 18, 20, 32, 37) or in conjunction with amino acids/protein provision with (16, 17, 36) and without (5, 19, 32) carbohydrate. Data from these studies indicate that mTOR signaling is elevated early (1–2 h) after a bout of resistance exercise (17, 18), with additional enhancement from amino acid and carbohydrate feeding (5, 17). In agreement with these studies, we found that the phosphorylation of p70S6K1 was robustly stimulated by exercise, with or without feeding. In light of the strong relationship between p70S6K1 phosphorylation and hypertrophy in rodents (2) and humans (56), our observation is perhaps not surprising. Rodent data suggest that leucine alone is stimulatory for the phosphorylation of p70S6K1 and its downstream target rpS6 (1) but that the phosphorylation of rpS6 is more persistent and present long after the phosphorylation of p70S6K1 has returned to baseline. This suggests that other pathways likely also affect the phosphorylation state of rpS6. Altered phosphatase activity, as may be the case for eIF2Bepsilon, could promote the prolonged hyperphosphorylated state of rps6 postexercise. We found that resistance exercise enhanced the phosphorylation of rpS6 at Ser240/244 in response to feeding. Moreover, our data also show that phosphorylation at Ser235/236 is increased in response to exercise alone. This effect could be mediated through p90rsk, which exclusively phosphorylates rps6 at Ser235/236 in an mTOR-independent manner (54), thereby providing a link to the upstream Ras/ERK signaling pathway, which is known to be activated by exercise (19, 32, 60).

We had hypothesized that the synergistic stimulation of MPS would be reflected in greater phosphorylation/dephosphorylation of the signaling proteins that are required to activate protein synthesis (8, 9, 17, 36, 38). In the present study, the effect of feeding was only evident at the level of PKB/Akt, suggesting that this sustained response was insulin mediated. Increased Akt phosphorylation after amino acid feeding has only been shown in studies when insulin increased or was elevated [compare (25, 28) with (27, 39)]. Our data for p70S6K1 and rps6 (Ser240/244) support the concept that resistance exercise plus nutritional provision results in an additive phosphorylation (i.e., greater than exercise or feeding alone). However, upstream members of the pathway (i.e., PKB/Akt and mTOR) did not display enhanced phosphorylation with the combined stimulus. That is not to say that the combination of resistance exercise and feeding do not synergistically activate PKB/Akt and mTOR, but rather, we acknowledge that our decision to obtain biopsies 6 h postexercise may have resulted in our failure to observe changes in phosphorylation status of these proteins (10). However, several studies have shown increased p70S6K1 Thr389 phosphorylation in the absence of elevated Ser2448 phosphorylation of mTOR (5, 20, 30, 32).

Focal adhesion kinase (FAK) has been shown to be responsive to loading and unloading in rodent models (23, 24, 26) and cyclic mechanical stretch in cell culture (62). As such, the acute regulation (i.e., phosphorylation) of FAK following intense resistance exercise could be a potential link in the mechanotransduction of a loading stimulus to induce increases in MPS. Such a possibility is predicted by the cellular “tensegrity” model, in which focal adhesion complexes anchor cytoskeletal proteins to the extracellular matrix of the cell to transmit forces and activate relevant signaling cascades, reviewed in Ref. 31. Our data showed that at 6 h postexercise, we could detect no changes in FAK phosphorylation (Tyr576/577). In the absence of any other human data examining this protein, we are unable to speculate why we observed this result since overload in animal models robustly increases not only FAK content but also FAK tyrosine phosphorylation, whereas unloading rapidly reduces FAK content and phosphorylation (15, 23, 26). It may be that increases in FAK activation occur in close temporal proximity to contraction itself or in response to chronic loading; thus, our stimulus may have been insufficient in magnitude or duration to elicit a response at 6 h postexercise.

Perspectives and Significance

Our data show, for the first time in humans, that resistance exercise reduces the phosphorylation of eIF2Bε, which is a potential mechanism that underlies the rise in muscle protein synthesis that occurs with this stimulus. Moreover, we have confirmed that resistance exercise is a potent activator of several key regulatory signaling proteins known to be critical in the regulation of MPS. Furthermore, resistance exercise potentiates the effect of feeding on the activation of p70S6k1 and rps6, up to 6 h postexercise; this may explain the resistance exercise-induced increase in sensitivity to amino acid feeding. Clearly, future studies would benefit from delineating the earlier time-dependent changes in phosphorylation of these proteins in addition to characterizing changes in MPS.


This research was supported by a grant from the Natural Sciences and Engineering Research Council of Canada (to S. M. Phillips). E. I. Glover, D. R. Moore, and J. E. Tang were all supported by doctoral graduate scholarships from the Canadian Institutes for Health Research (CIHR). B. R. Oates was supported by an Ontario Graduate Scholarship. S. M. Phillips was the recipient of a CIHR New Investigator career award and gratefully acknowledges that source of funding.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract