FABP3-mediated membrane lipid saturation alters fluidity and induces ER stress in skeletal muscle with aging

Sarcopenia is characterized by decreased skeletal muscle mass and function with age. Aged muscles have altered lipid compositions; however, the role and regulation of lipids are unknown. Here we report that FABP3 is upregulated in aged skeletal muscles, disrupting homeostasis via lipid remodeling. Lipidomic analyses reveal that FABP3 overexpression in young muscles alters the membrane lipid composition to that of aged muscle by decreasing polyunsaturated phospholipid acyl chains, while increasing sphingomyelin and lysophosphatidylcholine. FABP3-dependent membrane lipid remodeling causes ER stress via the PERK-eIF2α pathway and inhibits protein synthesis, limiting muscle recovery after immobilization. FABP3 knockdown induces a young-like lipid composition in aged muscles, reduces ER stress, and improves protein synthesis and muscle recovery. Further, FABP3 reduces membrane fluidity and knockdown increases fluidity in vitro, potentially causing ER stress. Therefore, FABP3 drives membrane lipid composition-mediated ER stress to regulate muscle homeostasis during aging and is a valuable target for sarcopenia.

A ged skeletal muscle exhibits loss of muscle mass and function, defined as sarcopenia 1 , contributing to a number of unfavorable consequences, including falls, fractures, frailty, and mortality in the elderly 2,3 . Owing to the importance of age-related muscle weakness for overall health of the elderly, the World Health Organization has established ICD-10-CM disease code for sarcopenia 4 . Sarcopenia can be triggered by a variety of conditions, including hormonal imbalances, loss of motor neurons, inflammation, defective mitochondrial homeostasis, and altered body composition 5 . In addition, intramuscular lipid and metabolite accumulation increases with age, possibly leading to skeletal muscle insulin resistance and systemic inflammation [6][7][8] . However, little is known about the underlying mechanism, key players, or downstream effectors of lipid remodeling dysfunction in skeletal muscles during aging.
Fatty acid binding proteins (FABPs) are 14-15 kDa proteins that bind to hydrophobic ligands such as long chain fatty acids 9 . FABPs include nine members with tissue-specific expression patterns 10 . Among FABPs, the role of FABP4 and FABP5 are relatively well-studied. FABP4/5 are highly expressed in adipose tissue and macrophages and contribute to metabolic diseases with adipose inflammation and insulin resistance [11][12][13] . In mice, Fabp4 deficiency in macrophages induces monounsaturated fatty acid production via stearoyl-CoA desaturase 1 (SCD-1) upregulation, which protects against atherosclerosis 14 . Moreover, Fabp4/5 double-deficient mice exhibit higher tolerance to high fat dietinduced obesity, insulin resistance, and diabetes 13 . FABP3 is highly expressed in the heart and skeletal muscles 15 , and its expression in skeletal muscles is elevated on consumption of a high fat diet. Fabp3 partially deficient mice reportedly exhibit reduced fatty acid utilization 16 . FABP3 is considered to serve as a lipid "chaperone" regulating solubility, mobility, and utilization of fatty acids 10 . In our previous study, proteomic analysis identified that FABP3 is highly expressed in aged muscle 17 . However, how FABP3 contributes to muscle aging has not been addressed.
Here, we focus on elucidating age-associated FABP3 function as a lipid chaperone in skeletal muscle. Using comparative lipidomic analyses in young versus aged muscle, and in FABP3overexpressing or knockdown muscle, we identify a unique ageassociated lipid composition signature, which is altered by FABP3 abundance. Using fluorescence recovery after photobleaching (FRAP) analysis for myotubes, we investigate whether lipid composition affects membrane fluidity. We further evaluate FABP3-mediated downstream signaling and assesse muscle mass and contractility allowing us to propose a mechanism for skeletal muscle aging.

Results
Lipidomic signature of acyl chains saturation in aged skeletal muscle. FABP3 expression was dramatically increased in skeletal muscle but not in cardiac muscle with age (Fig. 1a). Since FABP3 has been reported as a lipid "chaperone" which regulates fatty acid metabolism 10 , we first investigated a lipidomic signature in "young" (3-month-old) and "aged" (24-month-old) mouse tibialis anterior (TA) muscle using UPLC/QTOF mass spectrometry. Our analysis identified 203 lipid species. Approximately 53% of total identified lipid species were significantly changed, with 45 lipid species increased and 62 decreased (Fig. 1b). We next analyzed the distribution and alteration of the lipid classes. The major lipid species identified in skeletal muscle were the membrane lipids phosphatidylcholine (PC, 75% of total identified lipid), phosphatidylethanolamine (PE, 11%), lysophosphatidylcholine (LPC), sphingomyelin (SM), and phosphatidylinositol (PI) (Supplementary Table 1). Aged muscle showed marked changes in the content of lipid classes (Supplementary Table 1 and Supplementary Fig. 1a). Notably, aged muscle had elevated SM and LPC, showing 2.4-fold and 2.8-fold increases over young muscle, respectively ( Fig. 1c and Supplementary Fig. 1a). We next analyzed acyl chain composition of detected PC, PE, SM, and LPC in aged muscle compared to young muscle. In aged muscle, 16:0/20:4 PC significantly decreased by 24% and 16:0/16:0 PC increased 2.1-fold over young muscle. Other PC species changed slightly or insignificantly ( Fig. 1d and Supplementary Fig. 1b). When individual contents of acyl chains were summed, polyunsaturated PC acyl chains decreased in aged muscle, but saturated species increased ( Fig. 1h and Supplementary Fig. 1c). The PE species containing 40:6, 38:4, and 40:8 acyl chains decreased in aged muscle by 45,42, and 48%, respectively. However, 40:4, 34:1, and 36:1 PE increased more than 2-fold ( Fig. 1e and Supplementary Fig. 1d). Identification of individual acyl chains was not possible for all lipid species, due high isomeric or isobaric lipid levels. When PE species containing the same number of double bonds were combined, PEs with more double bonds decreased in aged muscle, but PEs containing fewer double bonds increased ( Supplementary Fig. 1e). These data indicate that PC and PE acyl chains were shifted from polyunsaturated to saturated in aged muscle. Interestingly, all identified SM and LPC species (which mainly contain saturated or monounsaturated acyl chains 18 ) increased simultaneously in aged muscle (Fig. 1f, g and Supplementary Fig. 1f, g). In parallel, we evaluated PC acyl chain length. Changes in C18 and >C18 fatty acid levels reflect very long chain fatty acid (VLCFA) elongase activity. Chain length is associated with the degree of unsaturation in VLCFAs 19 . Physical interactions between the VLCFA-elongase complex and desaturase have been reported in yeast 20 . Functional crosstalk between the elongase complex and desaturase has been reported in myoblasts 21 . As expected, the acyl chain length and the degree of unsaturation were correlated. Aged muscle increased in C18 acyl chain levels, but >C18 acyl chains decreased (Fig. 1i). Taken together, aged skeletal muscle exhibits a unique lipidomic signature of acyl chain saturation. These results prompted us to investigate whether FABP3 upregulation in aged skeletal muscle might be involved in such lipid remodeling.
FABP3-overexpressing young muscle exhibits a lipid composition similar to aged muscle. To study the impact of FABP3 on intramuscular lipid composition, we transfected plasmid constructs encoding HA-cherry FABP3 or HA-cherry control into young TA muscles using an electroporation gene delivery system ( Supplementary Fig. 2a). Using lipidomic analysis, we found that FABP3 overexpression increased 53 lipid species and decreased 32 lipid species (Fig. 2a). Lipid species with significant changes comprised nearly half of the total identified lipid species, indicating that FABP3 plays a major role in muscle lipidome remodeling. We next analyzed lipid class distribution and alteration (Supplementary Table 1). Compared to young muscle controls, FABP3 overexpression increased total SM and LPC contents by 29 and 42%, respectively (Fig. 2b, Supplementary  Fig. 2b and Supplementary Table 1). Interestingly, this pattern was similar in aged muscle (Fig. 1c).
FABP3 executes PERK-eIF2α-mediated inhibition of protein translation. Given that increased membrane lipid saturation induces ER stress and the unfolded protein response 22 , we investigated whether ER stress was induced in aged mouse muscle with elevated membrane SFA content. Phosphorylation of an ER stress sensor, PERK (pancreatic ER kinase), increased in aged TA muscle (Fig. 3a). Consistent with evidence that PERK phosphorylates eIF2α, thereby inhibiting protein translation 23 , we detected increased eIF2α phosphorylation and~50% decreased de novo protein synthesis in aged muscle (Fig. 3a). Therefore, we next investigated whether lipid remodeling via FABP3 overexpression could induce ER stress in vivo. Mouse TA muscle overexpressing FABP3 had increased PERK and eIF2α phosphorylation, and~40% reduced de novo protein synthesis (Fig. 3b). These results suggest that FABP3-induced lipid remodeling might induce an ER stress response and defective protein synthesis in aged skeletal muscle. To further confirm the role of FABP3 on ER stress signaling and protein synthesis, we established a stable C2C12 cell line that overexpressed FABP3 under the control of Cre recombinase. Acutely overexpressing FABP3 via Cre recombinase-expressing adenovirus increased PERK phosphorylation in differentiated C2C12 myotubes. Further, FABP3 overexpression increased eIF2α phosphorylation and inhibited de novo protein synthesis ( Fig. 3c and Supplementary Fig. 3a), which was more severe than in myotubes treated with the well-known ER stress inducer, palmitate. PERK-eIF2α regulates cell death via ATF4 and CHOP activation upon ER stress 24 , and inhibits protein synthesis via the 43S preinitiation complex 25 . We excluded a possibility that FABP3 induced cell death via PERK-eIF2α, because FABP3 expression did not induce Atf4 and Chop mRNA expression ( Supplementary  Fig. 3b). Since we found that PERK inhibition almost totally rescued FABP3-induced inhibition of protein synthesis ( Supplementary   Fig. 3c), we postulated that FABP3-driven eIF2α phosphorylation inhibits protein synthesis by directly inhibiting the 43S preinitiation complex.
FABP3 overexpression deteriorates muscle mass and force. We next investigated the physiological significance of FABP3 levels in skeletal muscle. Muscle recovery after immobilization in aged mice is incomplete or delayed compared to young mice 30 . We, therefore, assessed muscle atrophy after 5 days of immobilization and muscle recovery after 5 days of remobilization in FABP3-overexpressing TA muscle (Fig. 4a). Although the atrophic morphology between FABP3-overexpressing TA muscle and control TA muscle was similar, control young muscle regained muscle mass during the remobilization, while FABP3-overexpressing muscle did not (Fig. 4b). Moreover, immunohistochemical analysis revealed that FABP3-overexpressing muscle displayed significantly smaller muscle fibers than control muscle after remobilization, similar to aged muscle (Fig. 4c, d); meanwhile, western blot analysis revealed that FABP3 maintained high levels of PERK and eIF2α phosphorylation and inhibited protein synthesis during remobilization (Fig. 4e). However, transient FABP3 overexpression per se did not induce a significant difference in muscle mass before immobilization (Supplementary Fig. 4). FABP3 overexpression also did not induce expression of atrophy-related ubiquitin ligases, such as Atrogin-1 and MuRF1, nor impair autophagy by Atg5, 7, 12, 16L, and SQSTM1 expression ( Supplementary Fig. 3i-k). Next, to evaluate the effect of FABP3 on skeletal muscle function in intact muscles independent of other parameters, we compared ex vivo contractile properties, force, and fatigability, of FABP3-overexpressing and control young TA muscles. The maximum twitch forces were not different between FABP3-overexpressing and control TA muscles (Fig. 4f). However, at increasing stimulation frequencies in the tetanic force test (10-200 Hz, 100 V), the tetanic force was lower in the FABP3-overexpressing muscle than in the control muscle (Fig. 4g, h), similar to the aged muscle 31 . When subjected to fatigueinducing repetitive stimulations at 1 Hz, 100 V for 10 min, FABP3overexpressing muscles were more fatigue-sensitive than control muscles (Fig. 4i). These results together suggest that FABP3 upregulation during aging reduces both muscle mass and force.
FABP3 inhibition ameliorated membrane fluidity and alleviated ER stress. To investigate whether inhibition of FABP3 could protect against ER stress, we examined FABP3 knockdown in aged mouse muscle with ER stress. FABP3 knockdown decreased PERK and eIF2α phosphorylation and improved protein synthesis (Fig. 6a). Consistently in C2C12 myotubes, FABP3 knockdown reduced palmitate-induced PERK-eIF2α phosphorylation and improved protein synthesis (Fig. 6b). To investigate whether FABP3 inhibition could re-establish membrane fluidity that was lost after palmitate treatment, we performed FRAP analysis in FABP3-knockdown C2C12 myotubes. The FRAP signal was significantly increased in FABP3 knockdown myotubes in the presence of palmitate compared to control cells (light blue vs. deep blue symbol in Fig. 6c). Taken together with FABP3 overexpression data showing decreased FRAP signal (Fig. 3), we propose that increased FABP3 in aged muscle may reduce membrane fluidity and the ER stress response.
To test whether lipid remodeling is an underlying mechanism for FABP3-induced membrane rigidity and ER stress, we evaluated the effect of polyunsaturated fatty acids (PUFAs) in FABP3-overexpressing C2C12 myotubes. Addition of docosahexaenoic acid (22:6, DHA) into the culture media markedly enhanced the FRAP signal that was lost both in FABP3overexpressing (light red vs. deep red symbol) and palmitate-treated myotubes (light blue vs. deep blue symbol) (Fig. 7a). These results suggest that PUFAs restore membrane fluidity in such myotubes. DHA supplementation also significantly suppressed FABP3-induced PERK and eIF2α phosphorylation, while improving protein synthesis (Fig. 7b), resulting in improved myotube recovery that was defective in FABP3overexpressing myotubes (Fig. 7c). This result is consistent with previous reports showing DHA inhibits palmitate-induced ER stress in C2C12 myotubes 26 . We suggest that FABP3-induced lipid remodeling may be an underlying mechanism in agerelated membrane rigidity and ER stress, which is partially rescued by increased PUFA content.
FABP3 knockdown ameliorated age-associated impairment of muscle recovery. We investigated whether FABP3 knockdown ameliorates impaired TA muscle recovery in aged mice (Fig. 8a). FABP3-knockdown tended to slow the muscle loss during immobilization, and significantly increased muscle mass during remobilization. (Fig. 8b). Immunohistochemical analysis revealed significantly larger muscle fibers in FABP3-knockdown aged muscle than in control aged muscle after remobilization (Fig. 8c,  d). FABP3-knockdown maintained low levels of PERK and eIF2α phosphorylation and improved protein synthesis during remobilization (Fig. 8e). Next, to evaluate the effect of FABP3 inhibition on aged skeletal muscle function, we compared the ex vivo muscle contractility of FABP3-knockdown and aged muscle. While the maximum twitch force remained unchanged (Fig. 8f), the tetanic force at high frequencies (150-200 Hz) was slightly higher in FABP3-knockdown muscle than in control aged muscle (Fig. 8g, h). While aged muscles displayed a gradual reduction in tetanic forces over 120 Hz, FABP3-knockdown muscle retained tetanic force up to 200 Hz (Fig. 8g). When subjected to fatigueinducing repetitive stimulations, FABP3-knockdown muscle was more fatigue-resistant than aged muscles (Fig. 8i). Together, these results indicate that FABP knockdown ameliorates age-related decline in muscle mass and strength, thus serving as a potentially valuable therapeutic target.

Discussion
In this study, we demonstrated that FABP3 upregulation in aged muscle could be an intrinsic cue contributing to the age-associated loss of muscle mass and strength via regulation of membrane lipid composition. FABP3-overexpressing young muscle exhibited lipid composition similar to aged muscle and had delayed recovery after muscle immobilization, with defective protein synthesis induced by ER stress. Interestingly, FABP3 inhibition improved intrinsic ERstress-induced muscle dysfunction in aged muscle, with a noticeable restoration of young muscle-like lipid composition. We revealed that proper FABP3 regulation is critical for maintenance of membrane fluidity. Collectively, our results suggest that FABP3 is a key target that modulates age-dependent lipid remodeling and subsequent muscle homeostasis. Lipidomic analyses of mouse hindlimb muscles revealed that PUFA/SFA levels were decreased among the membrane phospholipids in both aged muscles and FABP3-overexpressing young muscles in comparison with the control young muscle, yet higher in the FABP3-knockdown aged muscle (Figs. 1, 2, and 5). We excluded the possibility that FABP3 alters membrane lipid composition   by regulating mRNA expression of fatty acid synthesizing enzymes ( Supplementary Fig 6a, b). Therefore, we believe that the composition is altered owing to the higher affinity of FABP3 for SFAs compared to PUFAs 15 , leading to subsequent delivery of SFAs to the membrane 32 . In humans, increasing evidence indicates the beneficial effect of fish oil-derived n − 3 PUFA supplements on muscle mass and function 33 , probably by exerting antiinflammatory effects, through mTOR activation, and through an increase in insulin sensitivity. Anti-inflammatory effects would potentially result from a reduction in pro-inflammatory cytokine levels, such as CRP, IL-6, IL-1β, and TNFα. Activated insulin and mTOR signaling pathways are probably involved in the increase in muscle protein synthesis; however, the underlying mechanism is yet unclear 34 . We observed that DHA supplementation in FABP3overexpressing myotubes alleviated PERK-mediated ER stress and restored protein synthesis (Fig. 7). Taken together, we suggest that age-related FABP3 upregulation could disrupt muscle homeostasis by decreasing PUFA/SFA content of muscle membrane. We investigated how phospholipid saturation induces ER stress and subsequent dysfunction in aged muscle. The PUFA/SFA ratio is a key factor determining membrane fluidity 27 . Membrane lipid saturation induces IRE1α and PERK activation and subsequent ER stress responses 22 . Combining these reports, we hypothesized that FABP3-mediated membrane lipid saturation in aged muscle might induce membrane rigidity, which in turn leads to ER stress. In this study, we demonstrate that FABP3 overexpression decreases membrane fluidity, while FABP3 knockdown increases membrane fluidity (Figs. 3e and 6c), which suggests a membrane fluidity-related mechanism in FABP3-induced ER stress in aged muscle.
Another noticeable lipid change was that SM and LPC content increased in aged and FABP3-overexpressing young muscle but decreased in FABP3-knockdown aged muscle. Sphingolipids promote muscle wasting and contractile dysfunction through induction of apoptosis and the inhibition of progenitor cell proliferation, insulin receptor signaling, glucose and amino acid uptake, and protein synthesis 35,36 . LPC inhibits myoblast fusion and decreases muscle fiber growth by reducing membrane fluidity 21,37 . Considering that both SM and LPC contain mostly saturated or monounsaturated acyl chains (see ref. 18 and Figs. 1f, g, 2e, f, and 5e, f), it is likely that increased SM and LPC content in aged or FABP3overexpressing muscle (Figs. 1c, 2b) contributes to membrane saturation and subsequent reduction in membrane fluidity.
Here, we showed the mechanical link between membrane fluidity and ER stress by temperature shift assay and showed that activation of PERK, an ER transmembrane protein, is necessary for this link. However, we still do not know the exact mechanism of PERK activation. The lipid composition reportedly influences membrane proteins (e.g., channels or receptors) by modifying their conformation, distribution, or oligomerization [38][39][40] . We hypothesize that a physical membrane state induced by lipid remodeling causes clustering of PERK and subsequent activation by autophosphorylation, which requires further investigations. Other membranous organelle(s) can be an alternative target of lipid remodeling. To better understand lipid-mediated muscle aging, it is necessary to identify an unknown target and its molecular and physiological functions.
In this study, while we found that increased FABP3 expression in aged muscle is critical for muscle recovery, the molecular mechanism of its increase with aging still remains unclear. We found that SIRT1 negatively regulates FABP3 expression. SIRT1 inhibition by gene silencing, or treatment with chemical inhibitors, such as EX-527 and sirtinol, increased both protein and mRNA levels of FABP3 (unpublished data). It was reported that SIRT1 activity depends on NAD+ levels, which decline with age in skeletal muscle 41 . Therefore, we speculate that decreasing activity of SIRT1 with aging gradually increases FABP3 expression in the skeletal muscle. However, further studies are required to better understand the regulatory mechanism.
Based on our present study showing FABP3-induced alterations in membrane lipid composition and membrane fluidity, we propose a role of FABP3 as a lipid chaperone for the regulation of skeletal muscle aging. Modified membrane lipid composition might change membrane properties, promote ER stress, and decrease protein synthesis, eventually leading to defective muscle mass and force (Fig. 8j). Collectively, we suggest that FABP3 would be a valuable therapeutic target for intervention of sarcopenia.

Methods
Animal models. All animal experiments were performed according to protocols approved by the Animal Care and Use Committee of the Korea Research Institute of Bioscience and Biotechnology (KRIBB). Young naive (3-month-old) and aged naive (22-24-month-old) C57BL/6 mice were purchased from the Laboratory Animal Resource Center (KRIBB). The mice were fed standard chow (Teklad F6 Rodent Diet 8664, Harlan Teklad, Indianapolis, IN), and housed under controlled temperature at 22-24°C and a 12 h light/12 h dark cycle with a humidity between 40 and 60%. Plasmid DNA was transfected into mouse TA muscle using an electric pulse mediated gene transfer technique. Young mice were anesthetized with isoflurane, and 50 μg of plasmid DNA containing either HA-cherry-FABP3 or HAcherry in 50 μl with saline was injected into TA muscle with a 27-gauge needle. An electric field was then applied to the muscle using an electric pulse generator (Electro Square Parator ECM830; BTX). TA muscle overexpressing FABP3 was collected 5 days after transfection. For FABP3 knockdown, TA muscle of aged mouse was injected with 50 μl Ad-shFABP3 adenovirus (10 8 CFU, shADV-258510; Vector Biolabs) or shControl (1782; Vector Biolabs). FABP3-knockdown TA muscles of aged mice were collected 5 days after infection. The contralateral TA was used as an intrasubject control. To induce atrophy and recovery, the hindlimb was immobilized for 5 days by stapling the foot, exploiting normal dorsotibial flexion using an Autosuture Royal 35 W skin stapler 42 . One tine was inserted close to the toe on the plantar portion of the foot, while the other was inserted in the distal portion of the gastrocnemius. The hindlimb was remobilized for 5 days by removing the tine. Mice were sacrificed, and isolated muscle tissues were used for further analysis.
Immunofluorescence. C2C12 myoblasts were seeded on 6-well plates with coverslips and differentiated into myotubes. Myotubes were rinsed in PBS, fixed in 4% paraformaldehyde for 15 min, and then washed three times with PBS and permeabilized in 0.1% Triton X-100 in PBS for 15 min. The myotubes blocked with 0.2% BSA for 30 min and then incubated with anti-MYH (sc-376157, Santa Cruz , and morphology (c). FABP3 expression was induced with Cre recombinase-carrying adenovirus (Ad-Cre) infected into fully differentiated C2C12 myotubes harboring Cre-inducible FABP3 constructs. FABP3-overexpressing myotubes or palmitate (500 μM)-treated myotubes were treated with DHA (100 μM) or vehicle. a Membrane fluidity was measured by FRAP analysis. Note the t 1/2 values in FRAP analyses. Black, control, n = 6; gray, DHA, n = 6; deep blue, palmitate, n = 5; light blue, palmitate+DHA, n = 9; red, FABP3, n = 5; light red, FABP3+DHA, n = 9 each myotube. b Immunoblot analysis (left) and quantification (right) of the indicated proteins and puromycin incorporation. Myotubes were incubated with puromycin (1 μM) for 30 min (n = 3 independent experiments). c the effect of DHA supplementation on FABP3-induced loss of atrophy recovery. FABP3-overexpressing or palmitate-treated myotubes were treated with dexamethasone (10 μM). Representative images (left) of MyHC (green) and DAPI (blue) staining in myotubes following DHA supplementation (100 μM) for 24 h (n = 3 independent experiments). Scale bar, 100 μm. Quantification (right) of myotubes diameter treated with dexamethasone. Data are presented as means ± S.E.M. Two-tailed unpaired Student's t-test was used. Source data are provided as a Source Data file. TA muscles of aged mice were infected with Ad-shFABP3 or Ad-shControl virus, immobilized with a surgical staple 5 days after infection, and allowed to recover (remobilization) after 5-day immobilization. b TA muscle mass was measured at 5 days after immobilization (n = 5) and 5 days after remobilization (n = 7). Changes in muscle weight for immobilization and remobilization were expressed as the percentage of the contralateral non-immobilized muscle weight. Biotechnology) overnight at 4°C, followed by Alexa Fluor 488 (#A-21121, Thermo-Fisher) for 60 min. Coverslips were mounted with Vectashield with DAPI mounting medium (VECTOR Laboratories). Images were captured using a Nikon Eclipse Ti-U inverted microscope and Nikon DS-Ri2 camera using NIS-Elements software.
Puromycin incorporation. To label newly-synthesized proteins with puromycin, C2C12 myotubes were incubated with puromycin (1 μM) for 30 min and harvested. For in vivo labeling, mice were intraperitoneally injected with puromycin (0.04 mmol/kg). TA muscles were harvested after 30 min. Cell or muscle lysates were analyzed using SDS-PAGE followed by immunoblotting. Puromycin-labeled proteins were detected by antipuromycin antibodies (PMY-2A4, DSHB).
Immunoblot analysis. Muscle tissues and myotubes were homogenized in lysis buffer containing 20 mM HEPES (pH 7.2), 150 mM NaCl, 0.5% Triton X-100, 0.1 mM Na 3 VO 4 , 1 mM NaF, 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride (AEBSF), and 5 mg/ml aprotinin (Sigma-Aldrich). The lysates were centrifuged at 15,000×g for 20 min at 4°C, and the supernatants were subjected to SDS-PAGE followed by immunoblot analysis. Antibodies used are as follows: Ex vivo isometric force and fatigue measurements. Intact TA muscles were dissected from the hindlimb of euthanized mice and mounted vertically between a force transducer (Model FT03, Glass Instruments, USA) in an organ bath with platinum electrodes and continuous perfusion with 95% O 2 + 5% CO 2 -saturated Krebs-Ringer solution (118 mM NaCl, 4.75 mM KCl, 24.8 mM NaHCO 3 , 1.18 mM KH 2 PO 4 , 2.5 mM CaCl 2 •2H 2 0, 1.18 mM MgSO 4 , and 10 mM glucose). Optimal muscle stretch was determined by applying a single twitch at supramaximal voltage (100 V for 1 ms) using a previously described protocol with slight modification 43,44 , and set at the length that generated maximal twitch force. After 10 min of equilibration, TA muscles were subjected to different force frequencies (tetani with increasing stimulation frequencies at  Hz every 500 ms with 2-min recovery intervals). The fatigue properties of TA muscles were assessed through repeated stimulation for 10 min at frequency of 1 Hz and 100 V. All experiments were performed at 25°C. Data acquisition and analysis were performed using LabChart Pro Software (Version 8; AD instruments, Pty Ltd.). Muscle length, diameter, and wet weight were measured at the end of each experiment.
Lipidomic analysis. For lipid extraction, 50 mg muscle tissue and 2.8 mm zirconium oxide beads were placed in a homogenizing tube with 500 μl 70% methanol solution. The sample was homogenized twice for 20 s at a speed of 5000 rpm. Next, 300 μl 70% methanol and 400 μl CHCl 3 were added and the mixture was vortexed for 1 min. The mixture was incubated at room temperature for 10 min and centrifuged at 13,000 rpm for 10 min. The lower organic layer was collected and dried under N 2 gas. The dried samples were reconstituted into 250 μl isopropanol:acetonitrile:water (2:1:1, v/v/v). Each sample was filtered, and 5 μl was used for lipidomic analysis. Lipids were measured using a triple TOF 5600 mass spectrometry (MS) system (SCIEX, Concord, Canada) coupled with ultra-performance liquid chromatography (UPLC) system (Waters, Milford, USA). An acquity UPLC BEH C18 column (1.7 μm × 2.1 × 100 mm) was used for lipid separation. The mobile phase was 10 mM ammonium acetate in acetonitrile:water (4:6, v/v; solvent A) and 10 mM ammonium acetate in acetonitrile:isopropanol (1:9, v/v; solvent B). Samples were eluted at 0.35 ml/min for 20 min. The eluate was analyzed with electrospray ionization (ESI) in positive and negative modes. Mass range was 50-1000 m/z. An automated calibrant delivery system was used to maintain mass accuracy. Raw UPLC/QTOF MS spectral data were preprocessed using Markerview software (SCIEX, Concord, Canada) and normalized to the total spectral area. Lipids were identified using Lipid Maps (www.lipidmaps.org), Human Metabolome (www. hmdb.ca), and Metlin (metlin.scripps.edu) databases. Identification was confirmed using MS/MS pattern and retention time of lipid standards (Avanti Polar Lipids, Alabaster, USA and Sigma-Aldrich, St Louis, USA). For lipidomic data analysis, each lipid species is indicated as a percentage of the total identified lipids. Differential-abundance analysis was performed using a twotailed Student's t-test and presented as bar graphs or volcano plots. To determine fold-changes, each dataset was normalized, and the ratios of young vs. aged muscles, FABP3-overexpressing vs. young muscles, and FABP3-knockdown vs. aged muscles were log2-transformed and presented as volcano plots. For correlation analysis, the log-ratio of the indicated condition for each lipid species was computed using RStudio. For principal component plots, the log-ratio of the indicated condition for each lipid species was computed using principal component analysis in Cluster 3.0 45 .
Fluorescence recovery after photobleaching (FRAP) assay. For FRAP assay in myotubes, C2C12 myoblasts were differentiated into myotubes, and then were stained with BODIPY 500/510 C1, C12 (4,4-Difluoro-5-Methyl-4-Bora-3a,4a-Diaza-s-Indacene-3-Dodecanoic Acid) (Invitrogen) at 2 μg/ml in PBS for 10 min at 37°C. FRAP images were acquired with a Zeiss LSM800 confocal microscope equipped with a live cell chamber (set at 37°C or 32°C and 5% CO 2 ) and Zen software (Zeiss). Myotubes were excited with a 488 nm laser and emission between 493 and 589 nm was recorded. Images were acquired with 16 bits image depth and 256 × 256 resolution using~1.34 μs pixel dwell settings. Four prebleaching images were collected and then the region of interest was bleached with 70% laser power. Fluorescence intensity was measured for 5 min every 10 s. t 1/2 was defined as the time the fluorescence intensity reached 50% of maximum fluorescence recovery 46 .