Ulk1 phosphorylation at S555 is not required for endurance training-induced improvements in exercise and metabolic capacity in mice

Abstract

Endurance exercise training improves exercise capacity as well as skeletal muscle and whole body metabolism, which are hallmarks of high quality-of-life and healthy aging. However, its mechanisms are not yet fully understood. Exercise-induced mitophagy has emerged as an important step in mitochondrial remodeling. Unc-51-like autophagy-activating kinase 1, ULK1, specifically its activation by phosphorylation at serine 555, was discovered as an autophagy driver and to be important for energetic stress-induced mitophagy in skeletal muscle, making it a potential mediator of the beneficial effects of exercise on mitochondrial remodeling. Here, we used CRISPR/Cas9-mediated gene editing and generated knock-in mice with a serine-to-alanine mutation of Ulk1 on serine 555. We now report that these mice displayed normal endurance capacity and cardiac function at baseline with a mild impairment in energy metabolism as indicated by an accelerated increase of respiratory exchange ratio (RER) during acute exercise stress; however, this was completely corrected by 8 wk of voluntary running. Ulk1-S555A mice also retained the exercise-mediated improvements in exercise capacity and metabolic flux. We conclude that Ulk1 phosphorylation at S555 is not required for exercise-mediated improvements of exercise and metabolic capacity in healthy mice.

NEW & NOTEWORTHY We have used CRISPR/Cas9-mediated gene editing to generate Ulk1-S555A knock-in mice to show that loss of phosphorylation of Ulk1 at S555 blunted exercise-induced mitophagy and mildly impairs energy metabolism during exercise in healthy mice. However, the knock-in mice retained exercise training-mediated improvements of endurance capacity and energy metabolism during exercise. These findings suggest that exercise-induced mitophagy through Ulk1 activation is not required for the metabolic adaptation and improved exercise capacity in young, healthy mice.

INTRODUCTION

Regular exercise, also known as physical activity, has long been known as one of the most potent means of disease prevention and treatment throughout the history of mankind. In the past few decades, the molecular transducers of the benefits of exercise have been an important research focus (1, 2). Studies have unequivocally proven that physical activity can reduce all-cause mortality across all age or gender groups in broad populations, and that exercise prescription is a highly effective “medicine” in preventing and treating diseases (3, 4). However, the underlying molecular mechanisms of the benefits of exercise are not sufficiently understood.

In the past, the benefits of exercise were largely attributed to anabolic events in striated muscles (skeletal muscle and heart), for example, increases in muscle mass, mitochondrial content, capillary density, etc. However, there is a growing body of research findings in the past few years that suggest crucial roles of meticulously governed catabolic mechanisms in inducing the benefits of exercise, such as autophagy. Autophagy is defined as a physiological, intrinsic cellular process that removes the damaged/dysfunctional protein aggregates and organelles, such as mitochondria, and this process is highly conserved across species. It is now known that autophagy is highly regulated upon cellular or environmental cues, such as nutrient deprivation and energetic stress. Recent evidence indicated that post-translational control of autophagy may directly contribute to endurance exercise training-induced adaptations of exercise in skeletal muscle (5–7).

Mitophagy is a highly regulated, mitochondria-specific autophagic process that targets and removes damaged/dysfunctional mitochondria, which is crucial for mitochondrial quality control. To date, a few distinct mechanisms of autophagy and mitophagy regulation have been outlined. Unc 51-like autophagy activating kinase (ULK1 or ATG1) is one of the first molecules to be discovered as a crucial driver of autophagy (8, 9). An in vitro study by Egan and colleagues showed that Ulk1 can be directly phosphorylated at serine 555 by 5′ AMP-activated kinase (Ampk), a master kinase in sensing energetic stress (10–13), linking Ulk1-mediated autophagy and mitophagy to AMPK activation. This signaling pathway is central to energetic stress sensing during exercise (14). We and others have shown that an acute bout of endurance exercise activates ULK1 via phosphorylation of S555 (15–17). Our laboratory has previously reported that mitophagy can be induced by acute exercise in murine skeletal muscle through the Ampk-Ulk1 axis. We also found that acute exercise-induced mitophagy is blunted in muscle-specific Ulk1 knockout mice (17), and the benefit of exercise training on insulin sensitivity is impaired in the genetic model following 6 wk of voluntary wheel running (16). These findings suggest the importance of Ulk1 is exercise training-induced adaptations.

In the current study, we used CRISPR/Cas9-mediated gene editing and generated knock-in mice with nonphosphorylatable serine-to-alanine mutation of Ulk1 at S555 (S555A). We tested whether one phosphorylation of Ulk1 at S555 is required for normal exercise capacity and metabolic function and for the exercise training-induced benefits. Our data show that although the metabolic responses to acute exercise in untrained mice are mildly impaired in Ulk1-S555A knock-in (KI) mice compared with the wild-type (WT) littermate mice, phosphorylation of Ulk1 at S555 is apparently not required for the improved exercise capacity and metabolic adaptation after exercise training.

MATERIALS AND METHODS

Animals

All animal procedures were approved by the Institutional Animal Care and Use Committees at the University of Virginia and Virginia Tech. Ulk1-S555A knock-in mice (in C57BL/6J background) were generated at the Genetically Engineered Murine Model (GEMM) core facility at the University of Virginia. Homozygotic knock-in mice and littermate wild-type (WT) controls were generated from heterozygotic breeding parents, and genotypes were confirmed by somatic DNA sequencing. A Western blot analysis of using antibody against phosphorylated Ulk1 at S555 was performed at baseline to ensure the complete loss of phosphorylation (Fig. 1A). All mice were housed in 2–4 mice/cage except for voluntary running mice, which were single-housed, in temperature-controlled (21°C) quarters with 12:12-h light-dark cycle and ad libitum access to water and normal chow (Purina). Equal number of male and female mice were randomly assigned to either the exercise training group where they were housed individually in cages with free access to running wheels, or to the sedentary group where they were housed in cages not equipped with running wheels at 8 wk of age. All mice were provided food and water ad libitum. Daily running distance was recorded via computer monitoring. All subsequent metabolic and physiological measurements were conducted in a blind manner to the genotype and treatment (exercise or sedentary) of the mice to minimize potential bias. Other than the voluntary running activity, which showed a significant gender difference, all other parameters measured did not show significant gender differences. We therefore presented pooled data from both male and female mice for all the data presented in this article.

Figure 1.

Ulk1-S555A mice exhibit normal endurance capacity. A: schematic illustration of ULK1 protein structure. B: Western blot confirmation of loss of S555 phosphorylation and preservation of Ulk1 expression and moderately reduced phosphorylation at S757 in KI mice with quantification. Data presented as means ± SE as results of two-tailed t test. *P < 0.05; ****P < 0.0001. n = 3. C: representative images of FDB fibers co-transfected with pMitoTimer in wild-type (WT) or knock-in (KI) mice with (Running) or without (Con) 90-min treadmill running. Images are merged in red and green channels. Quantification of MitoTimer red:green fluorescence intensity and pure red puncta. n = 4–7 per group. Data presented as means ± SE as results of two-way ANOVA. D: schematic illustration of endurance exercise capacity test on treadmill. E: result of endurance capacity test in WT and KI mice at baseline. F: blood lactate taken before and after the acute exercise session. **P < 0.01 (n = 12 or 13). Data presented as means ± SE as results of two-way ANOVA.

Somatic Gene Transfer and Confocal Microscopy of MitoTimer Reporter Gene

Somatic gene transfer of pMitoTimer reporter gene in FDB muscles in WT and KI mice was performed as we described previously followed by a 90-min treadmill running and sample preparation for confocal microscopy (17, 18).

Glucose Tolerance Test and Insulin Tolerance Test

Glucose tolerance test (GTT) and insulin tolerance test (ITT) were performed in the last week of experiment, as described previously (17, 19, 20). All mice were placed in regular cages and fasted for 6 h starting at 9:00 AM. After taking a measure of fasting blood glucose by a glucometer (Ascensia, Bayer) from tail vein, a bolus of d-glucose (for GTT; 200 mg/kg) or insulin (for ITT; 2 U/kg) dissolved in normal saline was injected intraperitoneally. Tail vein blood glucose levels were measured at 30, 60, and 120 min after glucose or insulin injection. GTT and ITT were performed on separate days at least 48 h apart. Mice with less than 40 mg/dL of blood glucose at the end of ITT were rescued with a bolus injection of glucose at 200 mg/kg.

Dobutamine Stress-Echocardiography

At least 48 h after GTT and ITTs, the stress-echocardiography test was performed with a slightly modified protocol from a well-established method (21). Briefly, mice were maintained anesthetized via an inhalation mask using 2.5% isoflurane mixed with O₂ at a flow rate of 200 mL/min in a supine position on a warm plate maintained at 37°C. Two-dimensional (2-D)-guided M-Mode images were acquired in short-axis view with a 13 MHz linear transducer before and at 0 and 10 min after dobutamine (Sigma) injection (2.5 µg/g body wt, ip). All images were analyzed with ImageJ software (NIH).

Treadmill Metabolic Capacity Test

After echocardiography, mice were acclimatized to treadmill running on the Oxymax Metabolic Treadmill System (Columbus Instrument, OH) in the morning for 10 min at a speed of 10 m/min at 0% grade for 3 consecutive days. Running wheels in the exercise group were then locked before the endurance capacity test on the fourth day. For the metabolic capacity test, mice were placed in the treadmill system for 15 min at 5 m/min, 5% grade to acclimatize. Starting at minute 15, the treadmill speed increases for 3 m/min every 5 min at a 5% grade until exhaustion. An electric shock grid at the back of the treadmill set at 0.5 Amp, 1 Hz was used to stimulate running.

Treadmill Endurance Capacity Test

At least 48 h after the metabolic capacity test, mice were acclimatized to treadmill running for endurance capacity test. Each mouse ran on a treadmill in the morning for 10 min at a speed of 13 m/min at 0% grade for 3 consecutive days for acclimatization. Running wheels in the exercise group were then locked before the endurance capacity test on the fourth day. For the endurance capacity test, mice ran at 13 m/min for 30 min at a 5% grade followed by an increase in running speed of 3 m/min every 30 min until they showed perceived exhaustion as reported previously (17) (Fig. 2A). Manual prodding was used as stimulation in the endurance capacity test. Blood lactate was measured prior to and immediately after exercise cessation as a biochemical confirmation of full exhaustion.

Figure 2.

Ulk1-S555A mice retain normal adaptability of endurance capacity and whole body glucose homeostasis. A: schematic illustration of 8 wk of voluntary wheel running as endurance exercise training. B: daily running distances for the mice in the exercise groups (n = 17). C: endurance capacity test results of WT and KI mice after exercise training. Data presented as means ± SE as results of two-way ANOVA. ****P < 0.0001 (n = 16 or 17). D: blood lactate taken before and after the endurance running session. **P < 0.01; ***P < 0.001. Data presented as means ± SE, and two-way ANOVA was performed. Tukey’s post hoc test was performed when significant interaction between variables was observed as noted. E, F: glucose and insulin tolerance test. Data presented as means ± SE as results of two-way ANOVA. AUC (area under the curve) (n = 16 or 17) for GTT; AAC (area above the curve) for ITT (n = 13–17). *P < 0.05. GTT, glucose tolerance test; ITT, insulin tolerance test; KI, knock-in; WT, wild-type.

Western Blotting

Protein lysates from plantaris muscles collected at the end of all in vivo testing were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and wet transferred onto nitrocellulose membrane. Membranes were probed with the following primary antibodies at 1:1,000 dilution, targeting S555 p-Ulk1 (CST #5869), S757 p-Ulk1 (CST #6888), Ulk1 (Sigma-Aldrich #A7481 and CST #8054), p-AMPKα1/2(T172) (CST #2535), AMPKα1/2 (CST #2532), α/β Tubulin (CST #2148), p62 (Sigma # P0067), LC3 (CST #12741), OXPHOS cocktail (Abcam #ab19467), Vdac (CST #2148), Bnip3 (Sigma #N0399), and Gapdh (CST #2118). Secondary antibodies were goat anti-mouse IR680 and goat anti-rabbit IR800 (LI-COR). Membranes were scanned using the Odyssey infrared imaging system (LI-COR, USA). Proteins were analyzed in comparison to a common protein standard loaded on gel, prior to calculating phospho:total, signal:Ponceau or protein:Tubulin ratio.

Statistical Analyses

Data are presented as means ± SE. Experiment results in the WT and Ulk1-S555A mice with four groups are analyzed via two-way ANOVA with standard Tukey post hoc analyses. Where only one variable is present, data were analyzed using Student’s t test. Statistical significance was established a priori as P < 0.05.

RESULTS

Ulk1-S555A Knock-in Mice Exhibit Normal Endurance Capacity and Cardiac Function Compared with Wild-Type Littermates

Using CRISPR/Cas9, we generated mice with a non-phosphorylatable mutation of serine 555 to alanine (S555A) of Ulk1 in the C57BL/6J background (Fig. 1, A and B). Western blot analyses confirmed a complete loss of phosphorylation of Ulk1-S555 using the plantaris muscles in the KI mice, whereas the stoichiometry of the kinase was generally maintained, indicated by a normal level of expression of Ulk1. The site that is phosphorylation by mTORC1 at S757 was moderately decreased by the S555 mutation (Fig. 1B). We used somatic gene transfer of a mitochondrial reporter for mitochondrial oxidative stress and mitophagy (22, 23) in mouse flexor digitorum brevis (FDB) muscles and showed evidence that a 90-min treadmill running resulted in increased mitochondrial oxidative stress in wild-type (WT) mice but not in KI mice. Treadmill running led to a trend of increased mitophagy in WT mice, whereas sedentary KI mice showed accumulation of mitochondria containing autolysosome but not further induced by exercise (Fig. 1C). This later finding is consistent with a similar recent finding in the heart in KI mice (24). We then subjected WT and KI mice to a treadmill running test for endurance capacity (running scheme shown in Fig. 1D) and found that despite having phosphorylation of Ulk1 at S555 ablated (Fig. 1B) and acute exercise-induced mitophagy blunted (Fig. 1C), sedentary KI mice had normal exercise capacity compared with WT mice (Fig. 1E). Blood lactate levels were measured before and after the exhaustive running session, and the increase of blood lactate was elevated in both WT and KI mice, indicating exhaustion biochemically. However, there was a trend of greater elevation of blood lactate in KI mice, suggesting a moderate impairment of metabolism during exercise in KI mice (Fig. 1F). To further confirm the preservation of cardiorespiratory fitness in the KI mice, stress-echocardiography with bolus dobutamine injection (2.5 µg/g ip) was performed. This measurement indicated no major cardiac dysfunction in both WT or KI mice (Table 1). Baseline and post-dobutamine (10 min) fractional shortening (FS) ejection fraction (EF), left-ventricle end-systolic diameter (LVESD), left-ventricle end-diastolic diameter (LVEDD), and posterior wall thickening (PWT) did not show statistically significant differences among these groups despite the fact there were clear responses to dobutamine stimulation (Table 1). These findings suggest that Ulk1 S555 phosphorylation is not required for the maintenance of baseline cardiorespiratory fitness and endurance capacity.

Table 1.

Stress-echocardiography of WT and Ulk1-S555A (KI) mice before and after exercise training

		FS, %		EF, %		LVESD, mm		LVEDD, mm		PWT, %
	Mice (n)	Baseline	+Dobutamine	Baseline	+Dobutamine	Baseline	+Dobutamine	Baseline	+Dobutamine	Baseline	+Dobutamine
WT-Sed	10	39.43 ± 2.55	44.74 ± 2.10	71.35 ± 2.97	82.47 ± 1.99	2.42 ± 0.12	1.96 ± 0.10	3.72 ± 0.09	3.54 ± 0.09	36.20 ± 2.88	39.72 ± 4.00
WT-Ex	12	34.72 ± 0.97	43.34 ± 1.58	71.98 ± 1.25	81.35 ± 1.50	2.48 ± 0.08	2.01 ± 0.08	3.79 ± 0.09	3.54 ± 0.08	37.15 ± 3.10	35.89 ± 2.35
KI-Sed	8	33.44 ± 2.14	40.67 ± 2.10	69.86 ± 3.15	78.57 ± 2.12	2.49 ± 0.12	2.18 ± 0.10	3.73 ± 0.09	3.66 ± 0.07	35.65 ± 3.88	43.68 ± 2.79
KI-Ex	9	33.42 ± 1.91	43.69 ± 3.28	69.91 ± 2.45	80.7 ± 3.11	2.49 ± 0.13	2.06 ± 0.16	3.73 ± 0.11	3.64 ± 0.10	32.43 ± 3.50	37.83 ± 3.78

Data presented as means ± SE as results of two‐way ANOVA. Dobutamine was injected at 2.5 µg/g body wt, ip, and echocardiography was performed before and 10 min after the initial injection. EF, ejection fraction; FS, fractional shortening; KI, knock‐in; LVESD, left‐ventricle end‐systolic diameter; LVEDD, left‐ventricle end‐diastolic diameter; PWT, posterior wall thickening; WT, wild‐type. No statistical significance was reported among groups at respective time points.

Ulk1-S555A Mice Exhibit Normal Response in Blood Glucose Tolerance and Insulin Sensitivity after Exercise Training

We then sought to investigate the necessity of phosphorylated Ulk1 at S555 in improving exercise capacity and metabolic adaptation to endurance exercise training. Ulk1-S555A mice and wild-type littermates were subjected to voluntary wheel running for 8 wk, and various physiological and metabolic tests were performed at end point (Fig. 2A). WT and KI mice achieved a similar but significant daily running distance (Fig. 2B) although female knock-in mice were more active during the training period compared with both male knock-in mice and wild-type female mice (Supplemental Fig. S1A). Exercise training resulted in moderately reduced body weight with no difference between WT and KI mice (Supplemental Fig. S1B). Tibia length measurement showed a very subtle genetic difference (Supplemental Fig. S1C). Importantly, exercise training improved exercise capacity as shown by the profound increase of treadmill running distance in both WT and KI mice (Fig. 2C). Exercise training also led to significantly reduced blood lactate elevation (Δlactate) at exhaustion in both WT and KI mice (Fig. 2D). These findings showed clear evidence that KI mice have normal improvement in exercise capacity following endurance exercise training, suggesting that Ulk1 S555 phosphorylation is not required for endurance exercise training-induced improvement of exercise capacity. Finally, glucose tolerance showed no significant difference between WT and KI mice before and at after training but the insulin tolerance test showed significant positive impact of exercise independent of genotype (Fig. 2, E and F). There was no clear difference in these parameters between male and female mice (Supplemental Fig. S1D).

Exercise-Induced Adaptations in Metabolic and Exercise Capacity Are Independent of Phosphorylation of Ulk1-S555

An important adaptation to exercise training is the capacity to sustain a low respiratory exchange ratio (RER, defined as CO₂ production/O₂ consumption) during exercise (25). We investigated the metabolic capacity in both WT and KI mice via assessment of gas exchanges during an acute bout of exercise on a metabolic treadmill system (Olympus, OH). Metabolic treadmill test again confirmed the improvement of exercise capacity in both WT and KI mice by showing significantly increased running distance, speed, and duration along with a significant increase in V̇o_2max in both WT-Ex and KI-Ex groups (Fig. 3, B and C). Consistent with these findings are findings for RER and heat production at exhaustion as well as estimated anaerobic threshold and fatty acid to carbohydrate metabolism crossover time, the time it takes for RER to reach 0.85, showing a significant impact of exercise training independent of genotype (Supplemental Fig. S2). KI-Sed mice showed an early increase of RER, lower fraction of fatty acid oxidation and higher fraction of carbohydrate metabolism during the acute bout of exercise compared with WT-Sed (Fig. 3, E and F), suggesting an impaired baseline metabolism in the KI mice. This data is consistent with the finding of a slightly more profound increase of blood lactate at exhaustion during endurance capacity test in KI mice (Fig. 1F). These differences were completely abolished after 8 wk of training, suggesting that Ulk1 (S555A) mice have normal adaptation in response to exercise training (Fig. 3, E and F). When we compared muscle and heart mass normalized by tibia length, we found that endurance exercise training promoted increased muscle mass for oxidative soleus muscle but decreased muscle mass for predominantly glycolytic gastrocnemius muscle independent of the genotype (Supplemental Fig. S1E). The loss of muscle mass for glycolytic muscles appeared profound in male mice while increased oxidative muscle and heart mass were more evident in female mice (Supplemental Fig. S1E). None of these changes showed any genotype dependency.

Figure 3.

Exercise training improves metabolic capacity during running in both wild-type and Ulk1-S555A mice. A: schematic illustration of metabolic treadmill running test. B: running distance, speed, and duration of metabolic running test. Data presented as means ± SE as results of two-way ANOVA. ****P < 0.0001 (n = 7–9). C: V̇o_2max results. **P < 0.01 Sed vs. Ex. D: V̇o₂ and V̇co₂ during the running test. E: respiratory exchange ratio (RER). F: estimated fraction of fatty acid oxidation (%FAO) and carbohydrate utilization (%CHO). ***P < 0.001 between genotypes. n = 5–9. Data presented as means ± SE, and two-way ANOVA was performed. Ex, exercised-trained; KI, knock-in; Sed, sedentary; WT, wild-type

Exercise Training Promotes Oxidative Adaptation Independent of Phosphorylation of Ulk1-S555

To gain better understanding of the underlying mechanism(s), we perform Western blot analysis for muscle contractile and mitochondrial proteins as well as autophagy/mitophagy markers. The 8-wk voluntary wheel running resulted in a significantly greater protein expression of myosin heavy chain IIa (Myh2) in plantaris muscle in both WT and KI mice (Fig. 4A). The impact of voluntary wheel running on mitochondrial biogenesis was minimal as quantified by immunoblotting of mitochondrial electron transport chain protein cytochrome oxidase 4 (Cox4), cytochrome c (Cycs) and voltage-dependent anion channel (Vdac) (Fig. 4, A and B) as well as proteins in OXPHOS complexes (Supplemental Fig. S3). Although KI mice showed an almost complete loss of Ulk1(S555) phosphorylation in plantaris muscles, there was no impact on baseline AMPK phosphorylation except for a very moderate reduction of total AMPK (Fig. 4, C and D). A significant decreased p62 expression was observed in both WT and KI mice following exercise training, indicating increased autophagy by exercise training. Total Lc3 and Lc3-II/I ratio did not show significant changes whereas Bnip3 was significantly higher in WT and KI mice following training (Fig. 4, C and D). These findings suggest that WT and KI mice had equally normal skeletal muscle adaptation along with signs of enhanced autophagy flux and potentially enhanced mitophagy-mediated by Bnip3.

Figure 4.

Exercise training induces adaptations in recruited skeletal muscle in wild-type and Ulk1-S555A mice. A: representative Western blot images for Myh2, Cox4, Cycs, and Vdac in plantaris muscle in sedentary (Sed) and exercise-trained (Ex) wild-type (WT) and knock-in (KI) mice. M stands for molecular marker. A line between lanes is for gel images from the same gel. B: quantification of the protein levels normalized by Gapdh. Data presented as means ± SE as results of two-way ANOVA. * and **P < 0.05 and P < 0.01, respectively. C: representative Western blot images for p-Ulk1(S555), p62, Lc3, and Bnip3 in plantaris muscle in sedentary and exercise-trained WT and KI mice. D: quantification of the protein levels normalized by Gapdh. Data presented as means ± SE as results of two-way ANOVA. *, **, and ****P < 0.05, P < 0.01, and P < 0.0001, respectively.

DISCUSSION

Mounting evidence supports that regular exercise training is the best intervention for quality-of-life improvements in all gender, race, or age groups; however, the molecular transducers of the benefits of exercise are yet to be unveiled. Posttranslational regulation of autophagy has been extensively studied in the past few decades, opening new avenues of research in understanding the molecular mechanism of diseases and interventions, including the effects of exercise. Ulk1 was first discovered as the mammalian homolog of Atg1 in yeast (26, 27), which has then been extensively studied as a major autophagy initiator (28). One of the first studies on exercise-induced autophagy by He and Colleagues showed that light-chain 3-II (LC3-II) puncta in murine muscle and heart was significantly increased during recovery following an acute episode of treadmill running, and this process was dependent on phosphorylation of B cell lymphoma 2 (BCL-2), which can recruit and form complex with Ulk1 (5, 29). An important mechanistic breakthrough was the discovery of Ampk-dependent Ulk1 phosphorylation at S555, linking Ulk1-mitophagy to energetic stress sensing (high AMP/ADP-to-ATP ratio) (14). Our group then showed that phosphorylated Ulk1 at S555 is induced by acute exercise in skeletal muscle, and knockout of Ulk1 and inhibition of its upstream AMPK blocks exercise-induced mitophagy in skeletal muscle (17). Muscle-specific deletion of the Ulk1 gene impairs exercise training-induced improvement of glucose homeostasis (16). Taken together, there is a fast-growing interest in understanding the importance of Ulk1 and Ulk1-mediated autophagy/mitophagy in exercise training-mediated benefits. Here, we showed that whole body ablation of Ulk1 phosphorylation of serine 555 does not affect baseline glucose tolerance and insulin sensitivity, but only mildly impair metabolic flux while exercise capacity is preserved. Long-term (8-wk) voluntary wheel running improved exercise capacity and energy metabolism during exercise in both WT and Ulk1-S555A mice to the same levels, suggesting Ulk1 phosphorylation at S555 is not required for endurance exercise training-induced exercise capacity and metabolic adaptations.

We and others have extensively used skeletal muscle-specific Ulk1 knockout mice to show that Ulk1 is required for exercise-induced mitophagy and improvement of insulin sensitivity (16, 30, 31). In the current study, we did not observe apparent impairments in the Ulk1-S555A mice at baseline other than very mild impairment of metabolism during acute exercise. This can be partly due to the young age of animals (8–16 wk for the duration of exercise training), but it also implicates that the baseline exercise capacity and metabolic function are independent of phosphorylation at S555. It is understood that apart from autophagy regulation, Ulk1 also acts as a regulatory hub for important metabolic functions, like glycolysis (32–34). It is so far unclear about the mechanisms of Ulk1 activation regarding the Ulk1-glycolysis pathway, though some studies have shown potential roles of AMPK in inducing glycolytic flux (33, 35). This could potentially reconcile why mutating the activation site important of autophagy did not abolish the benefits of exercise, whereas loss of whole protein is unfavorable. Therefore, further dissection in different regulatory mechanisms of Ulk1 for different pathways is an important future direction.

To our knowledge, the present study is the first to investigate the requirement of Ulk1 phosphorylation at S555 for the benefits of exercise using mice with a genetically mutated phosphorylation site of Ulk1 with the kinase stoichiometry remained intact. Despite the fact that the KI mice showed signs of impaired mitophagy induced by acute exercise (Fig. 1C), they showed nearly completely normal functional adaptation to endurance exercise training as indicated by the profound improvements of exercise capacity (Figs. 2C and 3B). All the parameters measured during the V̇o_2max treadmill running test are consistent with this conclusion (Fig. 3, B and C and Supplemental Fig. S2). These findings were further corroborated by findings of RER, %FAO and %CHO during metabolic treadmill running test (Fig. 3D), the enhanced Myh2 protein expression and reduced p62 protein expression in plantaris muscle (Fig. 4, A–D). These findings provide strong evidence that S555 phosphorylation is not required for endurance exercise training-induced improvements of exercise capacity, metabolic flux and at least some aspects of skeletal muscle adaptation.

It is worth noticing that we have observed accumulation of mitochondria containing autolysosome in sedentary KI mice, which was not further induced by an acute bout of treadmill running (Fig. 1C). This finding suggests that KI mice have impaired exercise-induced mitophagy. However, following eight weeks of voluntary running, KI mice showed significantly increased expression of Bnip3 (Fig. 4, C and D). These findings on one hand suggest that exercise-induced Ulk1 activation by S555 phosphorylation and the consequent mitophagy are not required for the functional adaptations in response to exercise training, but on the other hand may suggest the importance of Bnip3-mediated mitophagy.

Although the achieved point-mutation is a novel approach in dissecting molecular mechanism, it brings up a limitation in studying organ/tissue-specific impacts. One recent study by Kuramoto and colleagues has shown that exercise induces hepatic autophagy through circulatory factors released from skeletal muscle, emphasizing the idea that exercise-induced autophagy in non-contractile organs can be an important transducer of metabolic adaptations of exercise (36). It is an increasing topic in exercise physiology to study multiorgan response and impacts of exercise, therefore future dissection in this regard is warranted to understand the regulatory components of exercise-induced autophagy.

In conclusion, using CRISPR/Cas9-mediated Ulk1-S555A knock-in model we observed a mild impairment of metabolic homeostasis during exercise in sedentary Ulk1-S555A mice, but this defect was completely abolished by 8 wk of voluntary wheel running to a comparable level compared with trained WT mice. Therefore, we have shown for the first time that Ulk1 phosphorylation at S555 is not required for exercise training-mediated improvements of endurance capacity and energy metabolism during exercise.

DATA AVAILABILITY

The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S3: https://doi.org/10.6084/m9.figshare.25762623.

GRANTS

The present study was supported by NIH-R01AR050429 and NIH-R01AR077440 (to Z.Y.).

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.G., H.S., and Z.Y. conceived and designed research; Y.G., H.S., Q.Y., M.Z., O.W., and J.C.D. performed experiments; Y.G., H.S., Q.Y., M.Z., O.W., J.C.D., and Z.Y. analyzed data; Y.G., H.S., O.W., J.C.D., and Z.Y. interpreted results of experiments; Y.G., O.W., J.C.D., and Z.Y. prepared figures; Y.G. drafted manuscript; Y.G., O.W., J.C.D., and Z.Y. edited and revised manuscript; Y.G., H.S., Q.Y., M.Z., O.W., J.C.D., and Z.Y. approved final version of manuscript.