Differential Effect of Endurance Training on Mitochondrial Protein Damage, Degradation, and Acetylation in the Context of Aging

Abstract

Acute aerobic exercise increases reactive oxygen species and could potentially damage proteins, but exercise training (ET) enhances mitochondrial respiration irrespective of age. Here, we report a differential impact of ET on protein quality in young and older participants. Using mass spectrometry we measured oxidative damage to skeletal muscle proteins before and after 8 weeks of ET and find that young but not older participants reduced oxidative damage to both total skeletal muscle and mitochondrial proteins. Young participants showed higher total and mitochondrial derived semitryptic peptides and 26S proteasome activity indicating increased protein degradation. ET however, increased the activity of the endogenous antioxidants in older participants. ET also increased skeletal muscle content of the mitochondrial deacetylase SIRT3 in both groups. A reduction in the acetylation of isocitrate dehydrogenase 2 was observed following ET that may counteract the effect of acute oxidative stress. In conclusion aging is associated with an inability to improve skeletal muscle and mitochondrial protein quality in response to ET by increasing degradation of damaged proteins. ET does however increase muscle and mitochondrial antioxidant capacity in older individuals, which provides increased buffering from the acute oxidative effects of exercise.

The aging phenotype in skeletal muscle is characterized by accumulation of oxidative damage to DNA, proteins, and lipids, together with decreased resistance to stress (1–5). Reports indicate that aging interferes with both gene transcription and the translation of the transcripts (synthesis) into skeletal muscle (6,7). Although the age-related decline in skeletal muscle protein synthesis can be reversed by aerobic exercise training (ET) (8), there is no increase in muscle mass following ET. Here, we hypothesized that ET increases not only protein synthesis but also degradation thereby improving the quality of proteins damaged by reactive oxygen species.

Stable-isotope based studies show that acute aerobic exercise in young and older men can increase whole body protein turnover (9,10), which if occurs in skeletal muscle could improve skeletal muscle quality. It remains to be determined whether ET increases muscle protein degradation and results in reductions to oxidative damage to proteins. Alternatively increased endogenous antioxidant buffering could protect skeletal muscle from oxidative insult and maintain protein quality. For example, 9 weeks of ET older rats increased DNA repair and decreased its susceptibility to an oxidative insult in skeletal muscle (11), whereas a recent cross-sectional study of young and older sedentary and trained participants exposed to a single bought of exercise found training protected skeletal muscle DNA from oxidation through elevations in 8-oxoguanine-DNA glycosylase activity independent of age (12). In addition, we have previously shown in the context of aging that those who chronically perform ET have higher levels of the mitochondrial deacetylase, SIRT3 (12), which was recently shown to deacetylate key proteins involved in the cellular antioxidant stress response (13).

Here, we investigate the impact of 8-weeks of ET previously sedentary young and older participants on skeletal muscle and mitochondrial protein oxidation and degradation to determine whether age has modulated the muscle protein quality control system in the context of ET. We further investigate the effect of ET on antioxidant defense, and acetylation status of specific proteins in skeletal muscle in comparison to age-matched sedentary controls (CONs).

Experimental Procedures

Characteristics of the Participants and Endurance Training

Participants were randomized to either ET or CON for the 8-week intervention. Eleven young (5 women and 6 men) and 10 older (5 women and 5 men) participants completed the 8-week ET program, while 12 young (6 men and 6 women) and 10 older (6 men and 4 women) participants completed the CON after consenting to the research study approved by the Mayo Clinic Institutional Review Board conducted in accordance with the principles of the Declaration of Helsinki. All participants were weight-stable and reported less than 30 minutes of exercise 2 days a week prior to the study. The supervised ET program consisted of 8 weeks of cycling at 65% of peak oxygen consumption for 60 minutes. Participants started with 3 days of training per week and progressed to 5 days per week by the end of the first month outlined in detail previously (14). Participants randomized to the CON group were instructed to maintain their normal activities of daily living over the 8-week period. In addition, because skeletal muscle nicotinamide phosphoribosyltransferase (NAMPT) content showed an attenuated response to ET in older participants (see Results section), we evaluated NAMPT and its upstream regulator, NAD+ concentration, in a cohort of young and older chronically trained or sedentary participants (>4 years) to evaluate a longer-term training effect. Detailed participant characteristics regarding the long-term chronically trained and sedentary participants have been published in detail previously (15).

Experimental Protocol

Healthy, nonsmoking, nondiabetic young (18–30 years) and elderly (≥65 years) people with a body mass index <32kg/m² and who did not exercise on a regular basis (<two times per week), or were taking medications known to affect metabolic measurements or contraindicated for muscle biopsy (ie, corticosteroids, tricyclics, benzodiazepines, opiates, barbiturates, and anticoagulant medications) were chosen for screening. After informed consent was obtained, potential participants were chosen based on physical examination, history, hematological and biochemical tests, where any participants with history of substance abuse, abnormalities of kidney (elevated serum creatinine), liver (serum transaminase elevation ≥3 times the upper limit of normal range), cardiovascular disease, or muscle function, including orthopedic problems that precluded regular ET were excluded. Participants were then provided an Actigraph Accelerometer (Actilife) to wear for the next 7 days. Body composition was measured using dual energy x-ray absorptiometry (DPX-IQ; Lunar, Madison). During VO_2peak measurements participants were instructed to continue until volitional fatigue, with all but one completing within 10% of age predicted maximum heart rate. Participants were admitted to the Mayo CTSA Clinical Research Unit 7 or more days after baseline and follow-up VO_2peak measurement to minimize effects of maximal exercise on study parameters. Participants received a standardized weight-maintaining diet (20% protein, 50% carbohydrate, and 30% fat) during the 3 consecutive days before the muscle biopsy. Participants were admitted to the clinical research unit at 17:00 hours the night before the study day and given a standard meal (16 kcal/kg of fat free mass with 20% protein, 50% carbohydrate, and 30% fat) at 22:00 hours. At 08:00 hours after an overnight fast a muscle biopsy (300–400 mg) was taken from the vastus lateralis muscle under local anesthesia (lidocaine, 2%) with a percutaneous needle as previously described (15) and immediately frozen in liquid nitrogen for future analyses. Participants then undertook an 8-week endurance training program as previously detailed (14) after which the measurements were repeated 48 hours after the final exercise bout.

The experimental protocol for the long-term endurance training cross-sectional analysis has been described in detail previously (15).

Proteasome Activity

Proteasome activity was determined from muscle homogenates using a fluorogenic substrate for peptidylglutamyl-like activity of the 26S Proteasome (16) with detailed methods outlined in Supplementary Methods.

Superoxide Dismutase-2 and Catalase Activity

Muscle total superoxide dismutase-2 (SOD2) activity was measured in muscle lysate as the consumption of xanthine oxidase-generated superoxide radical by SOD in a competitive reaction with a tetrazolium salt with the addition potassium cyanide to inhibit SOD1 (Cayman Chemical Company, Ann Arbor, MI). SOD1 activity was estimated by subtracting total SOD activity from SOD2 activity. Catalase (CAT) activity was determined in muscle lysate spectrophotometrically by measuring hydrogen peroxide removal. This is a direct kinetic assay that follows the action of catalase on hydrogen peroxide and is based on measurement of the ultraviolet absorption of peroxide at 240nm every 30 seconds for 5 minutes (17).

Gene Expression

Total RNA was isolated from ~20mg of skeletal muscle tissue using the Qiagen RNeasy Mini Kit according to the manufacturer’s instructions. Complementary DNA was prepared using the Taqman reverse transcription kit (Life Technologies) per the manufacturer’s instructions. Real-time PCR was performed on a Viia7 Real Time PCR System (Life Technologies) using the Taqman gene expression assays. The following human gene expression assays (Life Technologies) were used for quantitative real-time PCR: SOD1 (Hs00533490_m1), SOD2 (Hs00167309_m1), and CAT (Hs00156308_m1). Reactions were performed in a 384-well assay format with a relative standard curve. All samples were plated in duplicate, and each plate contained one experimental gene and an endogenous CON, human β2-microglobulin.

Mass Spectrometry-Based Proteomics

Mitochondria from the muscle biopsy were isolated by differential centrifugation following a previously described protocol (18). A two-step workflow to detect and quantify posttranslational modifications including semitryptic peptides was followed as previously published (19–21) and detailed in Supplementary Methods. Mass spectrometry (MS)1 quantification of acetylated proteins was performed on filtered mitochondrial proteins using Skyline software (version 1.4.0.4421), and intensities were normalized to the corresponding nonacetylated form of the same mitochondrial peptide present in each data file (22) and are detailed in Supplementary Methods.

Statistics

Statistics were performed in Prism 6.0c (GraphPad Software Inc.). Power analyses revealed that with 10 young and 10 older participants allocated equally to the treatment groups, we can expect to be able to detect a minimum within group difference of ~20%–25% effect size and a minimum between group difference of ~30%–35% effect. We used the standard sample size formula for the test of equality of two normally distributed means to determine the minimum detectable effect size. A two-sided experimental type I error rate of .05 and a type II error rate of .05 (power = 0.95) was specified in each calculation. Quantitative data are presented as mean ± SEM. Unpaired t tests using a sequential Bonferonni procedure were used to compare means between groups for age at baseline and then the delta differences between training and age-matched CONs. Bioinformatics analyses are detailed in the Supplementary Methods. Significance was set at p < .05. Baseline comparisons were performed and are reported with data from both the ET and CON before intervention.

Results

Participant Characteristics

Participant characteristics for the intervention trial are detailed in Supplementary Tables 1 and 2. A total of 23 young (11 women and 12 men) and 20 older (9 women and 11 men) previously sedentary men and women completed the study. A limited number of young participants were reported in a previous article investigating muscle PGC-1α expression (14). At baseline no significant differences were observed for height, weight, body mass index, or daily physical activity count (data not shown) between young and old participants. Older participants had increased body fat percentage (p = .01) and lower maximal oxygen consumption (VO_2peak) based on a graded cycling exercise test (p < .01) at baseline. Eight weeks of ET did not alter bodyweight in either group, but decreased body fat percentage in the young compared with CON (p < .01). The ET program increased VO_2peak (p < .01) compared with CON a 10±2.7% change in the young and 15±6.2% in the older participants compared to their respective CON groups. The increase in VO_2peak after ET was a similar absolute change and therefore attenuating the age difference at baseline.

Endurance Training Reduces Damage to Skeletal Muscle and Mitochondrial Proteins in Young Only

We first determined whether 8 weeks of ET previously sedentary young and older participants influenced oxidation of methionine, phenylalanine, and tryptophan residues and deamidation of asparagine and glutamine residues in skeletal muscle and mitochondrial proteins. Baseline comparison revealed no effect of age on total muscle oxidation or deamidation, whereas 8-weeks of ET reduced protein oxidation (p < .001) in young but not in older participants (Figure 1A and Supplementary Table 3), with no change in deamidation to total skeletal muscle proteins in either group (Supplementary Table 3). We then cross-referenced the proteomic results with the MitoCarta (23) to evaluate the effects of ET on oxidation to mitochondrial proteins. We found that ET decreased oxidation to mitochondrial proteins in the young only (p < .001), with no change in older participants (Figure 1B and Supplementary Table 3).

Figure 1.

Protein damage and degradation after exercise training (ET). Total (A) and mitochondrial (B) protein oxidative damage in skeletal muscle decreased after ET in young but not older participants. Total (C) and mitochondrial (D) semitryptic peptide spectral counts increased in young, but decreased in older participants after ET. Proteasome activity, measured in proteasome enriched whole muscle homogenates, did not differ at baseline between groups (E), whereas ET significantly increased proteasome activity in the young but not older participants compared with their matched controls (CON). No difference in lon protease content at baseline and there was no effect after ET compared with matched CON (F). Blind detection of posttranslational modifications (PTMs) was performed using mass spectrometry in muscle tissues (n = 6 per group). To compare PTM abundance across groups, spectral counts for oxidation and semitryptic peptides were normalized, scaled, and compared in a pairwise fashion using chi-square test p value cutoff of p < .05 and fold-change ≥0.20. See Supplementary Table 3 for differential expression p values and fold changes. Error bars indicate ±SEM, significance indicated for age (baseline comparison) and training effects (delta difference comparison of ET with CON) by unpaired t tests with sequential Bonferonni correction (p < .05).

ET Increases Protein Degradation in Young But Not Older Participants

We previously showed that ET increased muscle protein synthesis without any increase in muscle mass suggesting that protein degradation is increased by exercise (8). Therefore, we investigated whether the differential results on protein damage could be explained by protein degradation. First, using MS, we searched for semitryptic peptides from total and mitochondrial proteins: a qualitative measure of in vivo proteolytic activity (20). At baseline there was no difference in semitryptic peptides concentration, whereas 8 weeks of ET increased total and mitochondrial protein semitryptic peptides in young (p < .001) while older participants exhibited a decline (Figure 1C and D and Supplementary Tables 3 and 4, p < .001). We then measured 26S proteasome activity, which selectively degrades oxidized proteins (24,25), and found that its maximal activity increased after ET in the young (p < .05) but not in older participants (Figure 1E). Although the proteasome selectively degrades oxidized proteins in skeletal muscle, it does not degrade mitochondrial proteins. Therefore, we measured the protein content of lon protease (Figure 1F). We found no difference in lon protease content at baseline or after ET in either group compared to their matched CON.

Endurance Training Enhances Mitochondrial Antioxidant Buffering Capacity in Old

Each bout of endurance exercise transiently increases reactive oxygen species production (13,26), and therefore we determined ET also enhances the endogenous antioxidant system. At baseline we did not observe a difference in mRNA levels of SOD1 or CAT, whereas SOD2 showed an age-related decline although ET decreased mRNA transcripts for SOD1 and 2 in the young only with no change in the older participants in comparison to CON (Figure 2A–C). At the protein level age-related reductions at baseline were observed for both SOD2 and CAT, whereas ET significantly increased SOD2 protein in the young only (Figure 2D–F). There were no baseline differences in SOD1, SOD2, or CAT activity between the groups (Figure 2A–C). SOD1 activity was not different with training (Figure 2A), but SOD2 and CAT activity increased (p = .04, p = .03) following ET in the older participants (Figure 2B and C). At baseline there were no differences between ET and CON groups within their age groups for mRNA, protein content, or activity measures.

Figure 2.

Endogenous antioxidant response to exercise training (ET). At baseline there was no significant difference between young and older participants at the mRNA level for superoxide dismutase-1 (SOD1) and catalase, whereas SOD2 showed an age-related decline. ET significantly decreased levels of SOD1 and SOD2 with no change in catalase transcript levels (A–C). At the protein level, Western blot analysis showed a significant effect of age on mitochondrial based SOD2 and catalase, whereas the ET program increased levels of SOD2 in the young only (D–F). Activity level measurements showed similar levels of SOD1, 2 and catalase activity at baseline (G–I). Although ET did not affect SOD1 activity, SOD2 and catalase activity increased in older participants. Error bars indicate ±SEM, significance indicated for age (baseline comparison) and training effects (delta difference comparison of ET with CON) by unpaired t tests with sequential Bonferonni correction (p < .05).

ET Increases SIRT3 Content

At baseline both skeletal muscle NAMPT and SIRT3 contents were reduced in older participants (p = .004 and p = .013, respectively), whereas 8 weeks of ET increased NAMPT content in young (p = .043) but not older participants therefore exaggerating the baseline age differences. Although SIRT3 content in both young and older participants increased (p = .012, p = .03, Figure 3), a lower absolute increase was apparent in older participants. There were no differences at baseline between ET and CON groups within their age groups for NAMPT or SIRT3 contents.

Figure 3.

NAMPT and SIRT3 protein levels show age-related decline and training increase: sedentary aging resulted in a significant decline of SIRT3 content at baseline, whereas exercise training (ET) resulted in a significant increase in SIRT3 content in both young and older participants. Error bars indicate ±SEM, significance indicated for age (baseline comparison) and training effects (delta difference comparison of ET with CON) by unpaired t tests with sequential Bonferonni correction (p < .05).

Isocitrate Dehydrogenase 2 Acetylation Decreases After ET

Because SIRT3, a mitochondrial localized deacetylase increased after ET in both young and older participants, we analyzed the acetylation status of our isolated mitochondrial samples by liquid chromatography tandem MS. At baseline, initial mitochondrial protein identification of acetylated lysines revealed a total of 168 acetylated proteins and a total of 277 acetylated lysine residues (Supplementary Acetylation Report). Ingenuity Pathway Analysis revealed acetylation occurred on proteins from multiple mitochondrial pathways including the electron transport chain, citric acid cycle and fatty acid β-oxidation with the majority of proteins exhibiting only one or two acetylation sites (Supplementary Figure 1A and B, Supplementary Table 5).

We next created a database of acetylated proteins and acetylation sites by tandem MS in our samples. We then searched for and quantified the initial MS signal (termed MS1) for acetylated peptides based on having a tryptic cleavage at arginine or a nonacetylated lysine (Skyline v1.4.0.4421). The acetylated MS1 signal was normalized within each sample to its nonacetylated form. Because of the low abundance of acetylated mitochondrial proteins in comparison to structural and membrane proteins, we were able to accurately quantify acetylations across samples (young, old, baseline, and post ET) on only a subset of proteins and subset of samples. Therefore, we were unable to make a baseline comparison of acetylated peptides but after ET, lysine 106 on isocitrate dehydrogenase 2 (IDH2) was found to be significantly lower (Supplementary Figure 1C).

Discussion

This study shows that young, but not older, previously sedentary participants exhibit reduced oxidative damage to total and mitochondrial proteins after 8 weeks of ET. Moreover, in response to ET younger participants increased proteasome activity and semitryptic peptides indicating increased protein degradation. Although protein expressions of the endogenous antioxidants, CAT, and SOD2 were significantly lower in older people, ET increased activities of these proteins in older participants in comparison to a sedentary matched CON. We also provide evidence that independent of age, ET increased mitochondrial deacetylase SIRT3 content and decreased acetylation of IDH2, a mitochondrial protein involved in regulating the cellular antioxidant stress response (27).

A key finding of this study is the demonstration that following ET oxidative damage to total and mitochondrial skeletal muscle proteins was reduced in young but not in older participants. Previous studies have shown that skeletal muscle aging is associated with increases in oxidative damage to DNA, lipids, and proteins (1,3,12). Although strong evidence suggests that independent of age regular ET protects skeletal muscle DNA by increased 8-oxoguanine-DNA glycosylase activity (11,12), even long-term vigorous aerobic ET does not fully normalize the age effect on mitochondrial protein content (20). This intervention study demonstrated that although younger people reduced oxidative damage to both total and mitochondrial proteins following ET, older people failed to do so. These results are in agreement with the cross-sectional report of Safdar and coworkers (28) who reported that compared with active young participants, both active and sedentary older participants showed elevations in protein carbonyl content, a marker of oxidative damage, in whole muscle homogenates. We measured both indices of endogenous antioxidant defenses and protein degradation to understand why older people did not improve protein quality after ET. The results demonstrated that older participants failed to increase protein degradation after ET, an adaptation observed in the younger participants.

Importantly our results offer a potential mechanism for the continued elevation of oxidative damaged proteins in older people. This finding may explain the attenuated increase in absolute VO_2peak in this cohort of older participants, a finding consistent with studies of similar intensity and duration (29–31), although longer training of higher intensity does produce significant improvement in the VO_2peak of older participants (32–35). In the absence of any single validated and sensitive measure of muscle protein degradation, we assessed muscle protein degradation by two different indices. Proteasome activity was increased following ET in young people but not in the older. Moreover, increased semitryptic peptides (products of endogenous protein degradation) further support increased protein degradation following ET. The older participants did not increase their proteasome activity, and semitryptic peptides are supportive of a hypothesis that aging impairs the proteolytic adaptive response as originally proposed in 1956 (36). It is highly likely that this increased protein degradation explains the decreased oxidative damage to total skeletal muscle proteins observed after ET in the young, as the 26S proteasome selectively degrades oxidized proteins (37).

Our MS based oxidative damage data for mitochondrial proteins showed a decrease in mitochondrial oxidative damage in young participants. However, mitochondrial proteins are not degraded by the 26S proteasome, and therefore we measured lon protease content, a mitochondrial specific protease (38), but found no significant change observed. Although this finding is consistent with those in young rats after both 2 and 12 weeks of ET (39,40), these results do not provide an explanation for reduced damage to mitochondrial proteins. It is possible that lon protease activity increased transiently following exercise or independent of measurable changes in content by immunohistochemistry as was reported in the livers of young and older rats (41), an interpretation supported by the increase in semitryptic peptide concentration derived from mitochondria.

Older participants increased the activity of mitochondrial localized SOD2, a finding that is consistently seen in most ET studies in rodents (42–44), but not all (45). However, because only the young decreased oxidative damage to mitochondrial proteins, these results suggest increased SOD2 activity is not sufficient to affect levels of mitochondrial protein oxidation; a finding in agreement with lifelong overexpression of SOD2 in mice (46). Our results show an interaction between age and exercise for CAT activity. No change in CAT activity seen in the younger participants is consistent with findings of ET in both humans (47) and rodents (42,45). The increase in CAT activity after ET in older participants is not consistent with rodent models of aging and ET. However, the result is consistent with the only previous human exercise study in older adults we are aware of where skeletal muscle CAT activity was measured (48).

A decrease in acetylation of IDH2, an enzyme that converts NADP⁺ to NADPH in mitochondria, after ET provides evidence that ET is associated with changes in acetylation status of mitochondrial proteins identified as SIRT3 substrates independent of age. The loss of SIRT3 in mice increases the acetylation of mitochondrial proteins including IDH2 (49) and increases oxidative damage to multiple tissues (27). Overexpression of SIRT3 in cultured cells protects against oxidative stress-induced cell death in an IDH2 dependent manner (27). Our finding that SIRT3 content increased in both young and older participants after ET extends our previous cross-sectional data (15) and is in agreement with intervention data in rats (40). The NAMPT, the upstream regulator of SIRT3 increased in young but not older participants prompted us to evaluate whether it might take a longer training period to increase NAMPT in older participants. Therefore, we analyzed NAMPT content and NAD+ concentration in whole muscle homogenates from a previously published chronic ET study (15) and found that chronic ET increased NAMPT content and NAD+, which had decreased with age (Supplementary Figure 2).

In conclusion and summarized in Figure 4, ET exhibits a differential response in young and older adults to protein quality and stress resistance. Although younger participants responded by decreasing oxidative damage and increasing protein degradation, older participants increased skeletal muscle antioxidants CAT and SOD2. The aforementioned finding indicates an age-related deficiency in protein quality in response to ET. Finally, we provide evidence for changes in acetylation status of a key mitochondrial protein in association with changes in SIRT3 content suggesting ET may exhibit similar effects to caloric restriction on skeletal muscle.

Figure 4.

Summary figure: endurance training exhibits differential response in young and older adults to protein quality and stress resistance. Younger, but not older, participants demonstrated a reduction in oxidative damage to skeletal muscle and mitochondrial proteins, which appears largely controlled by processes specific to protein degradation. Although older, but not young, adults exhibited increased catalase and superoxide dismutase-2 (SOD2) antioxidant activity after exercise training (ET), which may afford greater reactive species buffering capacity and protection from oxidative stress, both young and older participants showed an increase in the SIRT3 pathway with ET compared to an age-matched control group.

Supplementary Material

Supplementary material can be found at: http://biomedgerontology.oxfordjournals.org/

Funding

We are grateful for support from the National Institutes of Health grants AG09531 (to K.S.N.), T32 DK007198 (M.L.J.), KL2 TR000136-07 (M.L.J.), T32 DK007352 (M.M.R. and A.R.K.), KL2-RR024151 (B.A.I.), and UL1 TR000135. Additional support was provided by the Mayo Foundation and the Murdock-Dole Professorship (to K.S.N.) and the Mayo Clinic Center for Individualized Medicine (S.D.).

Conflict of Interest

The authors have declared that no conflict of interest exists.