Limiting cap-dependent translation increases 20S proteasomal degradation and protects the proteomic integrity in autophagy-deficient skeletal muscle

ABSTRACT

Postmitotic skeletal muscle critically depends on tightly regulated protein degradation to maintain proteomic stability. Impaired macroautophagy/autophagy-lysosomal or ubiquitin-proteasomal protein degradation causes the accumulation of damaged proteins, ultimately accelerating muscle dysfunction with age. While in vitro studies have demonstrated the complementary nature of these systems, their interplay at the organism levels remains poorly understood. Here, our study reveals novel insights into this complex relationship in autophagy-deficient skeletal muscle. We demonstrated that despite a compensatory increase in proteasome level in response to autophagy impairment, 26S proteasome activity was not proportionally enhanced in autophagy-deficient skeletal muscle. This functional deficit was partly attributed to reduced ATP levels to fuel the 26S proteasome. Remarkably, we found that activation of EIF4EBP1, a crucial inhibitor of cap-dependent translation, restored and even augmented proteasomal function through dual mechanisms. First, genetically activating EIF4EBP1 enhanced both ATP-dependent 26S proteasome and ATP-independent 20S proteasome activities, thereby expanding overall protein degradation capacity. Second, EIF4EBP1 activation caused muscle fiber transformation and increased mitochondrial biogenesis, thus replenishing ATP levels for 26S proteasome activation. Notably, the improved performance of the 20S proteasome in EIF4EBP1-activated skeletal muscle was attributed to an increased abundance of the immunoproteasome, a subtype specially adapted to function under oxidative stress conditions. This dual action of EIF4EBP1 activation preserved proteomic integrity in autophagy-deficient skeletal muscle. Our findings uncover a novel role of EIF4EBP1 in improving protein quality control, presenting a promising therapeutic strategy for autophagy-related muscular disorders and potentially other conditions characterized by proteostatic imbalance.

Abbreviations: 3-MA: 3-methyladenine; ACAC/ACC: acetyl-Coenzyme A carboxylase; AMPK: AMP-activated protein kinase; ATG5: autophagy related 5; ATG7: autophagy related 7; ATP: adenosine triphosphate; ATP5F1A/ATP5A: ATP synthase F1 subunit alpha; CKM-Cre: creatine kinase, muscle-Cre; CMA: chaperone‐mediated autophagy; CTSB: cathepsin B; CTSK: cathepsin K; CTSL: cathepsin L; CUL3: cullin 3; EDL: extensor digitorum longus; EIF4E: eukaryotic translation initiation factor 4E; EIF4EBP1: eukaryotic translation initiation factor 4E binding protein 1; EIF4F: eukaryotic translation initiation factor 4F complex; FBXO32/ATROGIN1/MAFbx: F-box protein 32; GFP: green fluorescent protein; IFNG/IFN-γ: interferon gamma; KEAP1: kelch-like ECH-associated protein 1; LAMP1: lysosomal-associated membrane protein 1; LAMP2: lysosomal-associated membrane protein 2; MAP1LC3/LC3: microtubule-associated protein 1 light chain 3; MEF: mouse embryonic fibroblast; Myl1/Mlc1f-Cre: myosin, light polypeptide 1 (promoter driving Cre recombinase); mRFP: monomeric red fluorescent protein; MTOR: mechanistic target of rapamycin kinase; MTORC1: MTOR complex 1; NFE2L1/NRF1: nuclear factor, erythroid derived 2, like 1; NFE2L2/NRF2: nuclear factor, erythroid derived 2, like 2; NFKB1/NFκB1: nuclear factor of kappa light polypeptide gene enhancer in B cells 1, p105; OXPHOS: oxidative phosphorylation; PPARGC1A/PGC1α: peroxisome proliferator activated receptor, gamma, coactivator 1 alpha; PSMB5: proteasome (prosome, macropain) subunit, beta type 5; PSMB6: proteasome (prosome, macropain) subunit, beta type 6; PSMB7: proteasome (prosome, macropain) subunit, beta type 7; PSMB8: proteasome (prosome, macropain) subunit, beta type 8 (large multifunctional peptidase 7); PSMB9: proteasome (prosome, macropain) subunit, beta type 9 (large multifunctional peptidase 2); PSMB10: proteasome (prosome, macropain) subunit, beta type 10; PSME1: proteasome (prosome, macropain) activator subunit 1 (PA28 alpha); PSME2: proteasome (prosome, macropain) activator subunit 2 (PA28 beta); RBX1: ring-box 1; SQSTM1/p62: sequestosome 1; SREBF1/SREBP1: sterol regulatory element binding transcription factor 1; STAT3: signal transducer and activator of transcription 3; TRIM63/MURF1: tripartite motif-containing 63; ULK1: unc-51 like kinase 1; UPS: ubiquitin-proteasome system

Introduction

The autophagy-lysosomal system is an evolutionarily conserved cellular process for intracellular degradation, which is essential for cellular renovation during aging and cellular survival under extreme conditions such as nutrient depletion or stress. This process is crucial for long-lasting post-mitotic tissues like skeletal muscle and neurons due to their low turnover rate. Dysfunctions in this system are linked to various muscular and neurological diseases [1]. A recent clinical study reported twelve individuals from five families carrying harmful, recessive mutations in ATG7 (autophagy related 7), an E1-like ubiquitin-activating enzyme homolog essential for autophagosome formation. The primary function of ATG7 is to facilitate the lipidation of Atg8-family proteins including the MAP1LC3/LC3 and GABARAP subfamilies in autophagosome formation. This can be achieved either directly or by promoting the assembly of the E3 ubiquitin ligase complex, comprising ATG12 (autophagy related 12), ATG5, and ATG16L1 (autophagy related 16 like 1) denoted as ATG12–ATG5-ATG16L1, which is responsible for the second autophagy conjugation reaction that enables Atg8-family protein lipidation [2]. These individuals exhibited complex muscular and neurological conditions beginning in childhood, including brain abnormalities, low skeletal muscle tone, and skeletal muscle weakness. Skeletal muscle biopsies of these patients displayed cytoplasmic inclusions containing ubiquitinated proteins and SQSTM1/p62 (sequestosome 1), indicating a deficiency in autophagy and an increase in protein aggregation [3].

In postmitotic skeletal muscle, protein synthesis and degradation are closely controlled by the MTOR (mechanistic target of rapamycin kinase) complex 1 (MTORC1) to uphold proteomic stability. MTORC1 activation is essential for growth hormone or exercise-mediated protein synthesis, promoting muscle growth. Conversely, MTORC1 inhibition is necessary for autophagy reactivation to clean and recycle long-lived proteins and damaged organelles during starvation or injury, maintaining cellular homeostasis and facilitating muscle repair. However, with aging, MTORC1 becomes constitutively active, rendering it less responsive to various internal and external stimuli [4,5]. This unregulated activity can potentially disrupt the delicate balance between protein synthesis and degradation, leading to impaired skeletal muscle function and age-related skeletal muscle weakness, as demonstrated in the mouse model with hyperactivated MTORC1 in the skeletal muscle [6].

Our previous studies demonstrated that restricting cap-dependent translation can restore autophagy deficits and rescue skeletal muscle functions in adult mice with hyperactivated MTORC1 in skeletal muscle [7,8]. This was achieved by overexpressing a mutated version of EIF4EBP1 (eukaryotic translation initiation factor 4E binding protein 1) specifically in myofibers. This mutation in EIF4EBP1 prevents the formation of the initiation complex necessary for cap-dependent translation by sequestering EIF4E (eukaryotic translation initiation factor 4E), thereby inhibiting its function [9]. Although the activation of this EIF4EBP1 mutant does not alleviate the MTORC1-mediated inhibition of autophagy initiation factor ULK1 (unc-51 like kinase 1), it restores autophagy flux and lysosomal activities [7]. Thus, EIF4EBP1 activation preserves proteomic integrity by targeting the cap-dependent translation, presenting a potential therapeutic strategy for conditions associated with autophagy deficiencies and consequent skeletal muscle pathologies.

Disrupting autophagy by selectively deleting Atg7 in differentiated myofibers leads to the accumulation of defective proteins and mitochondria, ultimately accelerating skeletal muscle atrophy and weakness in adult mice [10], similar to individuals with recessive ATG7 mutations, as described above. In this study, we aimed to determine whether restricting cap-dependent translation could also protect proteomic integrity in Atg7-deficient skeletal muscle. We found that EIF4EBP1 activation preserved proteomic stability primarily through the up-regulated activities of the ubiquitin-proteasome system in skeletal muscle with autophagy deficiencies. Mechanistically, EIF4EBP1 activation not only restored ATP level to maintain 26S proteasome activity in Atg7-deficient skeletal muscle but also enhanced ATP-independent 20S proteasome activity by inducing immunoproteasomal biogenesis through the transcription factor NFKB/NF-κB (nuclear factor of kappa light polypeptide gene enhancer in B cells), previously shown to regulate the transcription of immunoproteasome subunits [11]. Overall, these results suggest that EIF4EBP1 serves as an internal node coordinating anabolic protein synthesis and catabolic protein degradation systems, thereby upholding proteomic stability in skeletal muscle.

Results

Activation of EIF4EBP1 mitigates the accumulation of protein aggregates in Atg7-deficient skeletal muscle

To explore whether limiting cap-dependent translation could have a broad impact on alleviating muscle pathologies associated with autophagy deficiencies, we generated mice with a deletion of Atg7 and transgenic overexpression of a human EIF4EBP1 mutant (referred to as EIF4EBP1mt) specific to myofibers. The expression of the EIF4EBP1mt transgene is repressed by a loxP-flanked STOP codon cassette. When Cre recombinase is expressed in a specific tissue, it excises this repressive element, allowing for the expression of EIF4EBP1mt to sequester EIF4E from forming the eukaryotic translation initiation factor 4F (EIF4F) complex, thereby suppressing cap-dependent translation in a tissue-specific manner [9]. Conditional Atg7^f/f and EIF4EBP1mt mice were bred with CKM-Cre transgenic mice, where Cre recombinase expression is controlled by the CKM (creatine kinase, muscle) promoter (referred to as EIF4EBP1mt-atg7mKO mice). Consequently, the deletion of Atg7 and the overexpression of EIF4EBP1mt were restricted to fully differentiated skeletal muscles (Fig. S1A).

A subfamily of Atg8-family proteins in mammals, MAP1LC3/LC3 (microtubule-associated protein 1 light chain 3), is highly expressed in mouse skeletal muscle and frequently employed to examinee autophagy activity. Deletion of Atg7 led to the accumulation of unprocessed LC3 (referred to as LC3-I) and the autophagy receptor SQSTM1/p62 in the skeletal muscle of young mice (4-months-old), whether the mice were in a fed state (Fig. S1B) or undergoing prolonged starvation (Figure 1A and D). Note that the increase in lipidated LC3 (referred to as LC3-II) in the skeletal muscle of atg7mKO mice after 48 h of starvation was comparable to control mice (Figure 1A and D). This could be attributed to either the incomplete deletion of Atg7 or the contamination of other cell types with intact Atg7, whose activities got induced by prolonged fasting (Fig. S1A). To further investigate the direct role of atg7 deletion in regulating autophagy in muscle cells, we isolated the primary myoblasts from conditional Atg7^f/f mouse muscles and employed lentiviral delivery of Cre recombinase to excise the floxed Atg7 alleles (Atg7^f/f). In line with the findings from atg7mKO mouse muscles, myoblasts that lack Atg7 exhibited a marked rise in the buildup of and SQSTM1/p62 and LC3-I (Fig. S5A, S5B). When bafilomycin A₁ was used to prevent the degradation of autolysosomes, in conjunction with torin, an inhibitor of MTOR that derepressed the formation of autophagosomes, lipidated LC3-II remained detectable in Atg7-deficient myoblasts, albeit at much lower levels than in the control myoblasts (Fig. S5B). Another possibility is that LC3 lipidation can occur without ATG7 under certain conditions, as seen in mouse macrophages elicited by thioglycolate [12] and mouse embryonic fibroblasts (MEFs) infected with vaccinia [13], which may also be relevant in this context of skeletal muscle during prolonged fasting. Despite the presence of LC3-II, the efficiency of LC3 lipidation in Atg7-deficient skeletal muscles was still lower when the substantial level of LC3-I was taken into account (Figure 1A and D). Typically, once cytoplasmic cargo is taken into autophagosomes, autophagosomes will ultimately fuse with acidic endolysosomal compartments for degradation. The buildup of LC3-I and SQSTM1/p62 in the soluble fraction from atg7mKO mouse skeletal muscle implied impaired autophagosome formation, leading to an excess of ubiquitinated proteins in the insoluble fraction (Figure 1B and E, and Fig. S1C), an indication of protein aggregation. In addition, the insoluble fraction showed increased levels of SQSTM1/p62 and LC3-I (Figure 1B and E, and Fig. S1D), further providing evidence of defective autophagy machinery in atg7mKO mouse skeletal muscle. Notably, activation of EIF4EBP1 led to reduced protein aggregation in the Atg7-deficient skeletal muscles (Figure 1B and E, and Fig. S1C, S1D). The decline of aggregated ubiquitinated proteins in the insoluble fraction was not caused by the ubiquitin expression level in EIF4EBP1mt-atg7mKO mouse skeletal muscle, as evidenced by comparable levels of mono-ubiquitin expression and polyubiquitinated proteins in the soluble fraction among the various groups (Fig. S1E). While the LC3-II levels did not change, the protein levels of SQSTM1/p62 and LC3-I were significantly lower in the skeletal muscle of EIF4EBP1mt-atg7mKO mice compared to atg7mKO mice (Figure 1A, 1D and Fig. S1B).

Figure 1.

EIF4EBP1 activation relieves the protein aggregation in autophagy-deficient muscle. (A) Immunoblotting of autophagy markers SQSTM1/p62 and LC3 was performed in the soluble fraction extracted from TA of 4-month-old male mice following 48-h fasting (n = 3). (B) Immunoblotting of aggregated ubiquitinated protein, SQSTM1/p62, and LC3 was performed in the insoluble fraction extracted from the tibialis anterior (TA) muscle of 4‐month‐old male mice following 48-h fasting (n = 3). (C) Immunoblotting of lysosome markers LAMP1, LAMP2, CTSL, and CTSB was performed in the soluble fraction extracted from the TA of 4‐month‐old male mice following 48-h fasting (n = 3). (D) the quantification of (A). (E) the quantification of (B). (F) the quantification of (C). For all the quantification of immunoblotting, the relative protein expression levels were normalized to control mouse group, and data were shown as mean. Individual points corresponded to one mouse. Ponceau S was used as a loading control. Statistical significance was determined by student t-test for insoluble SQSTM1 and LC3 in (E) and One-Way ANOVA with Tukey’s multiple comparison test for other proteins. Only significant differences (p < 0.05) were shown. The quantification of (A) and (C) were also shown in Table S2, and (B) was shown in Table S3. (G) Confocal imaging of EGFP and RFP-LC3 signal in the cross-section of the TA muscle from 4-month-old male transgenic mice carried CAG-RFP-EGFP-LC3 and another indicated genotype. A wheat germ agglutinin (WGA) conjugate of Alexa Fluor 647 was used as a marker of the plasma membrane. Scale bar: 30 µm. Images were representative for n = 3 mice/genotype.

Autophagy in myofiber was further monitored using transgenic mice expressing LC3 fused to green fluorescent protein (GFP) and monomeric red fluorescent protein (mRFP). The fluorescence of GFP is quenched in the acidic and protease-rich environment of the lysosomal lumen, whereas mRFP fluorescence remains relatively stable. During fasting, activated autophagic flux promotes the fusion of the autophagosomes with lysosomes to form autolysosomes. This process was indicated by the presence of only mRFP without GFP-labeled puncta structures in the skeletal muscle of control mice after 24 h (Figure 1G) or 48 h of fasting (Fig. S1F). Conversely, impaired fusion or lysosomal acidity resulted in increased GFP and RFP signal colocalization in LC3 puncta structures, as demonstrated in the skeletal muscle of control mice treated with colchicine (Figure 1G). Thus, we employed the LC3 transgenic system to track the various phases of autophagosome and autolysosome development in skeletal muscle. In the skeletal muscles of atg7mKO mice, a noticeable increase in both RFP and GFP signals was observed, with the signals appearing large and irregular in shape, indicating LC3 accumulation in the cytosol rather than its conjugation into autophagosomes. LC3 puncta structures were sparsely observed in fasted atg7mKO mouse skeletal muscle and often exhibited double-positive for mRFP and GFP, in contrast to solely mRFP labeling seen in fasted control mice (Figure 1G and Fig. S1F). However, LC3 puncta formation was recovered in fasted EIF4EBP1mt-atg7mKO mouse skeletal muscle (Figure 1G), and the LC3 puncta structures that were exclusively mRFP-positive, indicative of autolysosome structure, were observed following prolonged fasting (Fig. S1F).

Research has demonstrated that chaperone-mediated autophagy (CMA) can be activated as a compensatory mechanism when autophagosome formation is inhibited. This was observed in mouse fibroblast cells (NIH3T3) exposed to 3-methyladenine (3-MA), a known class III phosphatidylinositol 3-kinase inhibitor that prevents autophagosome formation, as well as in immortalized MEFs obtained from embryos lacking Atg5 (autophagy related 5), which is crucial for autophagosome formation and expansion [14]. Although the levels of LAMP2A (lysosomal-associated membrane protein 2A), which is essential for CMA [15], were three times higher in EIF4EBP1-activated skeletal muscle, the amount of K63-linked polyubiquitin substrates designated for autophagic degradation [16,17] remained similarly elevated in both atg7mKO and EIF4EBP1mt-atg7mKO mouse muscle (Fig. S1H). This further supports the notion that the autophagy impairment was not alleviated by the activation of EIF4EBP1.

Alternatively, intracellular degradation can occur through direct lysosomal engulfment without the involvement of autophagosomes. A previous study showed that activating EIF4EBP1 to suppress cap-dependent translation can alleviate protein aggregation in MTORC1 hyperactivated mouse skeletal muscle by expanding lysosomal degradation capacity [7]. In this study, we found slightly elevated protein levels of LAMP1 (lysosomal-associated membrane protein 1) and LAMP2 in the skeletal muscle of both atg7mKO and EIF4EBP1mt-atg7mKO mice (Figure 1C and F, and Fig. S1I), with no noticeable change in their transcriptions (Fig. S1J). Despite this increase in lysosomal structural proteins, there was no corresponding rise in the expression and activities of lysosomal proteases such as CTSL (cathepsin L) or CTSB (cathepsin B) (Figure 1C and F, and Fig. S1G, S1I, S1J). The lack of difference in lysosomal biogenesis and activity between the atg7mKO and EIF4EBP1mt-atg7mKO mouse skeletal muscles suggests that activating EIF4EBP1 to inhibit cap-dependent translation alleviates the protein aggregation in Atg7-deficient skeletal muscle without involving lysosomes.

Activation of EIF4EBP1 ameliorates proteostasis in Atg7-deficient muscle by engaging the ubiquitin-proteasomal system

Previous research has shown that SQSTM1/p62 plays a role in increasing proteasomal biogenesis as a compensatory mechanism when autophagy is suppressed [18]. SQSTM1/p62 sequesters KEAP1 (kelch-like ECH-associated protein 1), an adaptor protein for the CUL3 (cullin 3)-RBX1 (ring-box 1)-dependent E3 ubiquitin ligase complex. This sequestration protects NFE2L2/NRF2 (nuclear factor, erythroid derived 2, like 2) from degradation. NFE2L2 subsequently regulates various cellular processes through its target genes, including those encoding multiple subunits of the 20S and 19S proteasomes [19]. Deletion of NFE2L2 suppresses growth and induces apoptosis in ATG7-deficient cancer cell lines, implying that proteasome activation is necessary for compensating impaired autophagy and maintaining cell survival [20]. Given the interplay between autophagy-lysosomal or ubiquitin-proteasomal systems, the two principal degradation pathways in eukaryotes, we investigated whether EIF4EBP1 activation mitigated protein aggregation by enhancing the ubiquitin-proteasomal system.

The 26S proteasome is an ATP-dependent proteolytic complex composed of a catalytic 20S core particle and one or two 19S regulatory particle caps. The 19S component is responsible for identifying ubiquitinated proteins and employing ATP to unfold and transport them to the 20S core component for degradation [21]. Our study revealed that skeletal muscle from both atg7mKO and EIF4EBP1-atg7mKO mice exhibited increased protein levels of 19S and 20S proteasome subunits (Figure 2A, 2B, and Fig. S2A), as well as enhanced assembly of the 26S proteasome (Figure 2C, 2D, and Fig. S2C, S2D). However, these elevated protein levels of proteasome subunits were not a result of increased transcription (Fig. S2B) induced by NFE2L2 stabilization (Fig. S2A) as a compensatory response to autophagy inhibition in either mouse model. Notably, EIF4EBP1mt-atg7mKO mouse skeletal muscle demonstrated the most pronounced elevation in 26S proteasome assembly (Figure 2C), and ex vivo ATP-dependent proteolytic activities (Figure 2E and Fig. S2G), surpassing the levels observed in atg7mKO mice. The increase in proteasome abundance and assembly was also evidenced in EIF4EBP1mt-muscle mouse muscle, suggesting that EIF4EBP1 is a key factor driving proteasome activation (Fig. S2E, S2F).

Figure 2.

EIF4EBP1 activation enhances the 26S proteasomal degradation capacity in skeletal muscle. (A) Immunoblotting of 19S (PSMD1) and constitutive 20S proteasome subunit proteins (PSMA2, PSMB5, PSMB6, and PSMB7) was performed in the soluble fraction extracted from the TA of 4‐month‐old male mice following 48-h fasting (n = 4). (B) the quantification of (A). (C) the quantification of (D). (D) Active proteasome complexes were resolved by native gel electrophoresis using proteasome lysates extracted from a pool of muscle of 4-month-old male mice under fed conditions. Native gel immunoblotting of 19S (PSMD1) and constitutive 20S proteasome subunit proteins (PSMB5, PSMB7, and PSMA2) were showed (n = 4). (E) Proteasome activity analysis in the presence of ATP was performed in lysate extracted from 4-month-old male mouse muscle under fed conditions (n = 4 for 4EPB1mt-muscle group and EIF4EBP1mt-atg7mKO group; n = 3 for control group and atg7mKO group). (F) Immunoblotting of K48 linked ubiquitinated protein and muscle-specific E3 ubiquitin ligases (FBXO32, and TRIM63) was performed in the soluble fraction extracted from TA of 4-month-old male mice following 48-h fasting (n = 3). (G) the quantification of (F). For all the quantification of immunoblotting shown in (B), (C), and (G), the relative protein expression levels were normalized to control group, and data were shown as mean. Individual points corresponded to one mouse. Ponceau S was used as a loading control. Statistical significance was determined by One-Way ANOVA with Tukey’s multiple comparison test for other proteins. Only significant differences (p < 0.05) were shown. The quantification of (A) and (F) were also shown in Table S4 and (D) was shown in Table S5.

The master transcription factor that regulates the proteasome complex, NFE2L1/NRF1 (nuclear factor, erythroid derived 2, like 1), whose protein stability is regulated by proteasome activities, is involved in the proteasome recovery pathway. In cases of proteasome inhibition, NFE2L1 avoids degradation and triggers the transcription of proteasome complexes [11]. In mouse skeletal muscles, we detected several forms of NFE2L1 in different sizes (Fig. S2A), similar to what has been observed in mouse retina [22] and adipose tissues [23]. The level of the long isoform (referred to as NFE2L1a, which migrates near the ladder of 130 kDa) serves as a secondary indicator of proteasome function. We found that the protein level of NFE2L1a was exclusively down-regulated in EIF4EBP1mt-atg7mKO and EIF4EBP1mt-muscle mouse skeletal muscles (Fig. S2A, S2E), providing additional evidence of elevated 26S proteasome activities in these mice (Figure 2E and Fig. S2G). Increases in transcription of shorter isoforms (referred to as NFE2L1b, which migrates near the ladder of 100 kDa) [24] were also observed in EIF4EBP1mt-atg7mKO mouse skeletal muscle (Fig. S2B), but this did not lead to an upregulated protein level (Fig. S2A).

The presence of lysine 48 (K48) polyubiquitin chains denotes proteins destined for degradation through the proteasome [25]. We observed an extensive accumulation of K48 ubiquitinated proteins in Atg7-deficient skeletal muscles without a corresponding increase in the skeletal muscle-specific E3 ligases, FBXO32/ATROGIN1/MAFbx (F-box protein 32) and TRIM63/MURF1 (tripartite motif-containing 63) (Figure 2F, 2G, and Fig. S2B), suggesting compromised proteasome activity. This accumulation of K48 polyubiquitinated proteins might be explained by excess SQSTM1/p62 causing compensatory proteasomal biogenesis while also compromising proteasomal degradation by delaying the transport of ubiquitinated proteins to the proteasome [26]. Additionally, autophagy deficiency could lead to the accumulation of dysfunctional proteasomes over time, dampening their function as they are also targeted by autophagy, a process known as proteaphagy [27]. We confirmed this by observing higher levels of aggregated proteasome subunit in Atg7-deficient skeletal muscle (Fig. S2H, S2I). Nonetheless, consistent with a twofold increase in the 26S proteasomal degradation activities (Figure 2E and Fig. S2G), we observed a decrease in K48-linked polyubiquitinated proteins in the skeletal muscle of EIF4EBP1mt-atg7mKO mice compared to atg7mKO mice, reaching levels similar to control mice (Figure 2F). This finding confirms higher functional proteasome activities in EIF4EBP1-activated skeletal muscle.

Taken together, our results indicate that even though the protein levels of 26S proteasome are increased in atg7mKO mouse muscle, the proteasome activities decrease over time in atg7mKO mouse muscle as shown by the increased protein aggregation and inefficient clearance of K48-labeled protein substrates. EIF4EBP1 activation enhances the proteasomal degradation capacity in atg7mKO mice by increasing the 26S proteasome assembly in the absence of increased transcription of proteasome subunit.

Activation of EIF4EBP1 rescues ATP depletion and type-IIb myofiber atrophy in Atg7-deficient muscle

Fully assembled 20S core and 19S regulatory complexes are essential for proteasomal degradation, as free proteasome subunits lack catalytic ability. While the total protein amount of 26S proteasome subunits remained unchanged, EIF4EBP1mt-atg7mKO mice exhibited a notable increase in 26S proteasome assembly compared to atg7mKO mice. This led us to investigate the underlying mechanism for this change in assembly without a significant difference in total protein levels.

Adenosine triphosphate (ATP) plays a crucial role in multiple steps of proteasomal degradation, including 26S proteasome assembly and substrate unfolding by 19S regulatory complex [28–33]. A previous study has shown that endurance exercise training, known to induce mitochondrial biogenesis and shifts from glycolytic to oxidative fiber types, can increase the 26S proteasome activity in skeletal muscle [34]. Additionally, EIF4EBP1 activation in skeletal muscle has been shown to increase the proportion of oxidative fibers [9]. Thus, we hypothesized that EIF4EBP1 activation enhances 26S proteasome complex assembly by increasing ATP content in skeletal muscle.

ATG7-dependent autophagy is vital for maintaining cellular health in post-mitotic tissues like skeletal muscle through continuous breakdown and recycling of cellular components and organelles. The deficiency of Atg7 leads to the accumulation of damaged mitochondria, which results in elevated oxidative stress, muscle weakness, and atrophy [10,35,36]. A decrease in ATP levels was noted in skeletal muscle lacking Atg7, utilizing the Myl1 (mosin, light polypeptide 1) fast promoter to drive CRE recombinase (Myl1-Cre) [37], but there was no sign of AMP-activated protein kinase (AMPK) activation [10]. Although the presence of dysfunctional mitochondria was not yet apparent in the skeletal muscle of 4-month-old atg7mKO mice (Fig. S3A), their ATP production was significantly lower than the control (Figure 3A). In agreement with earlier findings, our atg7mKO mice also did not show increased phosphorylation of AMPK at tyrosine 172 nor its downstream substrate ACAC/ACC (acetyl-Coenzyme A carboxylase) at serine 79 (Fig. S3F). This reduction could be attributed to increased aggregation of ATP5F1A/ATP5A (ATP synthase F1 subunit alpha), a key subunit of the mitochondrial ATP synthase complex involved in ATP synthesis and hydrolysis, in the insoluble fraction of atg7mKO mouse skeletal muscle (Figure 3C, 3E).

Figure 3.

EIF4EBP1 activation restores the ATP levels with a shift toward oxidative fiber composition in autophagy-deficient muscles. (A) Muscle ATP measurement was performed in extensor digitorum longus muscle (EDL) of 4-month-old male mice following 48-h fasting (n = 3). Data were shown as mean and individual points corresponded to one mouse. (B) Immunoblotting of mitochondrial proteins was performed in the soluble fraction extracted from TA of 4-month male mice following 48-h fasting. (C) Immunoblotting of mitochondrial proteins was performed in the insoluble fraction extracted from TA of 4-month-old male mice following 48-h fasting. (D) the quantification of (B). (E) the quantification of (C). For all the quantification of immunoblotting, individual points corresponded to one mouse and data were shown as mean. The relative protein expression levels were normalized to control group. Statistical significance was determined by One-Way ANOVA with Tukey’s multiple comparison test for other proteins. Only significant differences (p < 0.05) were shown. Ponceau S was used as a loading control. Quantification of (B) and (C) are also shown in Table S7. (F) Muscle fiber type percentage analysis was performed in 4-month-old male mouse soleus muscle (SM) (n = 6 for control and EIF4EBP1mt-muscle group; n = 3 for atg7mKO group and n = 4 for EIF4EBP1mt-atg7mKO group) (G) Muscle fiber type percentage analysis was performed in 4-month-old male mouse EDL (n = 8 for control group, n = 6 for EIF4EBP1mt-muscle group, n = 3 for atg7mKO, and EIF4EBP1mt-atg7mKO group). (H) Muscle fiber mean size analysis was performed in 4-month-old male mouse EDL under fed conditions (n = 5 for control group; n = 3 for EIF4EBP1mt-muscle, atg7mKO, and EIF4EBP1mt-atg7mKO groups). Data in (F), (G), and (H) were shown as mean and individual points corresponded to one mouse. Statistical significance was determined by One-Way ANOVA with Tukey’s multiple comparison test. Only significant differences (p < 0.05) were shown. (I) Quantification of Type IIb myofiber size was analyzed in 4-month-old male mouse EDL under fed conditions (n = 5 for control group; n = 3 for EIF4EBP1mt-muscle, atg7mKO, and EIF4EBP1mt-atg7mKO groups). Data were shown as mean ± SEM and statistical significance was determined by Two-Way ANOVA with Tukey’s multiple comparison test compared within the size class. *p<.05; **p<.01, ***p<.001, ****p<.0001 indicated statistical significance in comparison to control group; #P<.05; ##P<.01, ###P<.001, ####P<.0001 indicated significance in comparison to atg7mKO group. Color indicates statistical significance between atg7mKO (red), EIF4EBP1mt-atg7mKO (purple), and EIF4EBP1mt-muscle (green). (J) MyHC staining of EDL was performed in 4-month-old male mouse on fed conditions. Scale bar: 200 μm. Images were representative for n = 5 mice for control group and n = 3 mice for other groups. (K) MyHC staining was performed in SM of 4-month-old male mouse on fed conditions. Scale bar: 200 μm. Images were representative for n = 4 mice for atg7mKO group and n = 3 mice for other groups.

Activation of EIF4EBP1 in atg7mKO mouse skeletal muscles restored ATP production to control levels (Figure 3A), likely by promoting mitochondrial biogenesis [9]. We observed a substantial increase in oxidative phosphorylation (OXPHOS) system complex levels (Figure 3B, 3D), accompanied by a trend toward increased transcription (Fig. S3B) in EIF4EBP1mt-atg7mKO mice compared to both control and atg7mKO mice. This increase in OXPHOS gene transcription is previously linked to enhanced translation of PPARGC1A/PGC1-$\mathit{\alpha }$ (peroxisome proliferator activated receptor, gamma, coactivator 1 alpha), a key regulator of mitochondrial biogenesis, induced by EIF4EBP1 activation [9]. The EIF4EBP1mt-atg7mKO mouse skeletal muscle exhibited a higher level of soluble ATP5F1A protein compared to atg7mKO (Figure 3B, 3D), while its level in the insoluble fraction remained consistent with the control and lower than atg7mKO (Figure 3C, 3E), indicating a rescue effect.

The increased OXPHOS expression resulting from EIF4EBP1 activation was largely influenced by a shift toward a more oxidative muscle phenotype in EIF4EBP1mt-atg7mKO mice (Figure 3F, 3G, 3J, 3K), consistent with the previous study on EIF4EBP1 activation [9]. Furthermore, EIF4EBP1 activation counteracted the atrophy of glycolytic type fibers in atg7mKO mouse skeletal muscles (Figure 3H, 3I, 3J), without significantly affecting oxidative type fibers (Figure 3K and Fig. S3C, S3D, S3E). The increase in AMPK phosphorylation, paired with higher total levels of AMPK proteins (Fig. S3F), may simply signify greater oxidative characteristics in EIF4EBP1-activated skeletal muscle. Thus, our results demonstrate that EIF4EBP1 activation in Atg7-deficient muscle rescues ATP depletion and type-IIb myofiber atrophy. The replenished ATP level powers the 26S proteasome to counteract the lack of autophagy and preserve the proteostasis in the skeletal muscle of EIF4EBP1mt-atg7mKO mice.

20S proteasomal degradation capacity is expanded by EIF4EBP1 activation

It should be noted that, besides 26S proteasome, the assembly of 20S proteasomes was notably higher in the muscle of EIF4EBP1mt-atg7mKO mice than in control and atg7mKO mice (Figures 2C, 2D). Correlating with the increased formation of 20S proteasome, the ATP-independent degradation activities of 20S proteasome were significantly higher in myofibers with EIF4EBP1 activation (Figure 4A). The 20S proteasome can eliminate damaged proteins without the use of ATP and is responsible for a large portion of oxidized protein degradation [38]. A recent study conducted in vitro has shown that 20S proteasomes can degrade both ubiquitin-conjugated and misfolded proteins, similar to 26S proteasomes. In ATP-limited conditions, elevated levels of 20S proteasomes are vital for removing damaged proteins, as 26S proteasome activity is ATP-dependent. This offers significant advantages for cell survival during prolonged hypoxia, especially in case of ischemia-induced heart failure [39]. Further, proteasomes can adapt their subunit composition in response to various stresses. For instance, exposure to inflammatory cytokines like IFNG/IFN-γ (interferon gamma) leads to the replacement of standard catalytic β-subunits, including PSMB5 (proteasome (prosome, macropain) subunit, beta type 5), PSMB6, and PSMB7 with inducible immunoproteasome subunits, including PSMB8 (proteasome (prosome, macropain) subunit, beta type 8 (large multifunctional peptidase 7)), PSMB9 (proteasome (prosome, macropain) subunit, beta type 9 (large multifunctional peptidase 2)), and PSMB10 (proteasome (prosome, macropain) subunit, beta type 10). Additionally, the 19S proteasome is substituted by the 11S proteasome, composed of PA28α/β (encoded by PSME1 [proteasome (prosome, macropain) activator subunit 1 (PA28 alpha)] and PSME2 [proteasome (prosome, macropain) activator subunit 2 (PA28 beta)]). Beyond its role in immune responses, the immunoproteasome has been found to possess nonimmune functions that help alleviate oxidative stress in neurodegenerative and cardiovascular disorders [40]. Notably, the elevated 20S proteasome activity in EIF4EBP1-activated skeletal muscle was more susceptible to inhibition by ONX-0914 (Fig. S4A), a compound that targets the immunoproteasome chymotrypsin-like catalytic subunit, PSMB8 [41].

Figure 4.

EIF4EBP1 activation promotes the atp-independent 20S proteasomal degradation via increasing immunoproteasome subunit expression and assembly. (A) 20S Proteasome activity analysis was performed with lysates without ATP extracted from 4-month-old male mouse muscle under fed conditions (n = 4 for 4EPB1mt-muscle group and EIF4EBP1mt-atg7mKO group; n = 3 for control group and atg7mKO group). (B) Immunoblotting of immunoproteasome subunit proteins immunoproteasome (PSMB8, PSMB9, and PSMB10) and 11S regulatory proteasome (PSME1, and PSME2) was performed in the soluble fraction extracted from TA of 4-month-old male mouse following 48-h fasting (n = 4). (C) the quantification of (B). (D) Active proteasome complexes were resolved by native gel electrophoresis using proteasome lysates extracted from a pool of muscle of 4-month-old male mice under fed conditions (n = 4). Native gel immunoblotting of immunoproteasome subunit proteins immunoproteasome (PSMB9) and 11S regulatory proteasome (PSME1/2) was presented. (E) the quantification of (D). (F) the quantification of (G). (G) Immunoblotting of proteasomal transcription factor proteins STAT3, and NFKB1 was performed in the soluble fraction extracted from TA of 4-month-old male mouse following 48-h fasting (n = 4). (H) Representative immunoblotting of STAT3, NFKB1, GAPDH (used as cytoplasmic fraction marker), and Histone H3 (used as nuclei fraction marker) was performed in nuclear fraction, cytosolic fraction, and total lysate extracted from muscles of 4-month-old male mouse following 48-h fasting. The other three accompanied immunoblotting images are presented in Fig. S4G. (I) the quantification of (H) and Fig. S4G (n = 4). For all the quantification of immunoblotting, the relative protein expression levels were normalized to control group, and data were shown as mean. Individual points corresponded to one mouse. Ponceau S was used as a loading control. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparison test for other proteins. Only significant differences (p < 0.05) were shown. The quantification of (B) and (G) were shown in Table S8, (D) was shown in Table S9, and (H) was shown in Table S10.

We found that the skeletal muscle of EIF4EBP1mt-muscle and EIF4EBP1-atg7mKO mice displayed higher expression of immunoproteasome subunit (Figure 4B, 4C, and S4B, S4C) and their integration into functional proteasomes (Figure 4D, 4E, and Fig. S4D, S4E) compared to control or atg7mKO mice. This increase is partially attributed to elevated transcription of these subunits (Fig. S4F). Further analysis of transcription factors, STAT3 (signal transducer and activator of transcription 3) and NFKB1, known regulators of immunoproteasome subunits, revealed increased protein levels of both in EIF4EBP1-activated skeletal muscle (Figure 4F, 4G, and S4C). However, only NFKB1 was enriched in the nuclei fraction (Figure 4H, 4I, and Fig. S4G). Notably, NFKB1 function depends on proteasome activation, as the maturation of its precursor, p105, into p50 requires 20S proteasome activity [42]. The mature form of NFKB1-p50 significantly increased in both total lysate (Figure 4G, 4F) and nuclei fraction (Figure 4H, 4I, and Fig. S4G) from EIF4EBP1-activated skeletal muscle. This effect appeared to be primarily driven by EIF4EBP1 activation, as enrichment of NFKB1-p50 in the nuclei fraction was also observed in EIF4EBP1-activated skeletal muscle (Figure 4H, 4I, and Fig. S4G). These results indicate that EIF4EBP1 activation expands proteasomal degradation capacity, at least partially, by inducing immunoproteasomal biogenesis via NFKB1 activation.

Our findings elucidate the role of EIF4EBP1 activation in expanding proteasomal degradation capacity to mitigate proteolytic stress in autophagy-deficient skeletal muscle (Figure 5). This adaptive response not only enhances our understanding of proteostasis mechanisms but also suggests potential therapeutic strategies for conditions associated with autophagy deficiency. Our research demonstrates that EIF4EBP1 activation influences myofiber physiology through at least two pathways: first, by reprogramming myofibers toward an oxidative phenotype, thereby increasing ATP production and supporting 26S proteasome activities; and secondly, by promoting proteasomal biogenesis, particularly the immunoproteasome, via NFKB1-mediated transcriptional activation. These proteasomal adaptations appear capable of functioning under stress conditions. Collectively, these results underscore the significance of EIF4EBP1 in maintaining proteostasis through proteasomal compensation, offering insights that may inform future therapeutic approaches.

Figure 5.

Illustration depicting the mechanism by which EIF4EBP1 activation restores proteostasis in autophagy-deficient skeletal muscle through dual mechanisms of proteasome enhancement and mitochondrial biogenesis. In healthy skeletal muscle, the autophagy-lysosome pathway and ubiquitin-proteasome system collaboratively maintain proteostasis by clearing damaged organelles and misfolded proteins (top panel). atg7mKO mouse skeletal muscle exhibits impaired autophagy, leading to accumulation of damaged mitochondria, dysfunctional proteasomes, and protein aggregates. Despite increased 26S proteasome protein levels, insufficient ATP production renders this compensatory mechanism ineffective (left bottom panel). Our study reveals a novel therapeutic approach: EIF4EBP1 activation-induced restriction of cap-dependent translation significantly alleviates protein aggregation in Atg7-deficient muscle through two key mechanisms (right bottom panel): 1. Enhanced immunoproteasomal biogenesis and activity via NFKB1 activation, expanding overall proteolytic capacity. 2. Increased mitochondrial biogenesis through PPARGC1A/PGC1-α upregulation, restoring ATP levels crucial for powering 26S proteasome function. This dual action of EIF4EBP1 activation effectively compensates for autophagy deficiency, presenting a promising strategy for maintaining proteome integrity in muscle disorders characterized by impaired autophagy.

Discussion

Our study reveals a potential therapeutic approach for preserving proteomic stability in tissues with impaired autophagy through the restriction of cap-dependent translation, ultimately leading to enhanced 20S proteasome activation. The 20S proteasome, with its diverse compositional variants, demonstrates remarkable adaptability to various cellular conditions, including energy depletion and inflammation [40]. Recent research has expanded our understanding of the 20S proteasome’s capabilities, revealing its ability to degrade not only oxidized proteins [38], but also ubiquitinated proteins, a function previously attributed solely to the 26S proteasome [39]. Furthermore, the proteasome’s role in mitochondria-associated degradation has been established, serving as a crucial surveillance system for the mitochondrial proteome [43].

Given the observed impairment of the autophagy-lysosomal system in numerous age-related ailments and situations, particularly in post-mitotic tissues such as neurons and muscles [1], the 20S proteasome emerges as an appealing target for restoring proteomic stability in autophagy deficiency-related disorders. Our research aligns with and extends previous studies that noted increased proteasome activity in response to genetic hyperactivation of MTORC1 signaling in MEFs [44] and skeletal muscle [45], interpreted as a compensatory reaction for impaired autophagy. While previous investigations have primarily focused on the role of SREBF1/SREBP1 (sterol regulatory element binding transcription factor 1) and NFE2L1 in proteasomal biogenesis downstream of MTORC1 [44,45], our study provides novel insights into the regulatory mechanisms at play. We demonstrate that EIF4EBP1 activation in skeletal muscle leads to a selective downregulation of NFE2L1 protein level without concomitant transcriptional changes. We showed that activation of EIF4EBP1 led to an enhancement of proteasomal degradation capacity, potentially contributing to the increased instability of NFE2L1 in EIF4EBP1-activated skeletal muscle. A recent study employing the Targets of RNA-binding proteins Identified By Editing/TRIBE assay in a human prostate cancer cell line, which revealed physical interactions between active EIF4EBP1 and NFE2L1 mRNA [46]. This finding suggests an alternative hypothesis where EIF4EBP1 might directly suppress NFE2L1 mRNA translation through physical binding. However, this mechanism requires further investigation to confirm its relevance in skeletal muscle specifically.

Our research further highlights the complex interplay between MTORC1 signaling and proteasome function. While acute MTORC1 inhibition has been demonstrated to enhance proteasome expression and activities in various cell types and yeast [47,48], potentially as a mechanism to replenish free amino acid reserves [49], our findings demonstrate that EIF4EBP1 activation leads to increased proteasomal biogenesis at the protein level and enhanced proteasomal activities in skeletal muscle. The absence of transcriptional changes in proteasome subunits suggests that this increase may be attributed to upregulated translation of proteasomal genes, possibly through mechanisms similar to those proposed for 19S regulatory-particle assembly chaperones/RPACs [50].

In the context of aging and muscle physiology, our study offers new perspectives on the regulation of proteasome activity. Previous research has reported decreased 20S proteasome activity in aged skeletal muscle [51–55]. Our current findings suggest that EIF4EBP1 activation can counteract this decline by boosting both 20S and 26S proteasome activities, thus promoting muscle health as demonstrated in our earlier publication [9]. Notably, we observe increased expression and association of 11S regulatory proteins and 20S catalytic core subunits, as well as enhanced mitochondrial biogenesis and ATP level, collectively expanding the total proteasomal degradation capacity in autophagy-deficient muscle. In particular, immunoproteasome expression and activities were significantly up-regulated in EIF4EBP1-activated skeletal muscle. Previous studies have reported an increase in the levels and functions of immunoproteasomes in fibroblasts from primate species with longer lifespans when compared to those with shorter lifespans. Enhanced biogenesis and activity of the immunoproteasome were noted in the livers of mouse strains with longer lifespans and in those treated with longevity-promoting drugs, implying that a more active immunoproteasome may be associated with longevity [56]. In conclusion, our study provides compelling evidence for the therapeutic potential of enhancing immunoproteasome function as a strategy to maintain proteome integrity in autophagy-deficient tissues. These findings have broad implications for a range of muscular and neurological disorders associated with autophagy dysfunction [1] and open new avenues of targeted interventions in age-related pathologies characterized by proteostatic imbalance.

Materials and methods

Generation of genetic mice models

All animal studies are reviewed and approved by the NUS Institutional Animal Care and Use Committee (IACUC; R17–0195, R20–1535). Muscle-specific knockout or transgenic mice are generated by crossing conditional flox mice with CKM-Cre mice. Control mice included in this study are littermates of the mutant mice but do not express Cre-recombinase. Multiple mouse models are included in this study. CKM-Cre mice are obtained from The Jackson Lab (006475). Tg-EIF4EBP1mt mice are generated from our previous study [9]. Atg7 flox mice (Atg7^f/f) are generated and provided by the RIKEN BRC through the National Bio-Resource Project of the MEXT, Japan [57]. atg7mKO mice are bred from Atg7 flox mice and CKM-Cre mice. EIF4EBP1mt-muscle mice are bred from Tg-EIF4EBP1mt mice and CKM-Cre mice. EIF4EBP1mt-atg7mKO mice are bred from atg7mKO mice and EIF4EBP1mt-muscle mice. All mice are back-bred 5 generations to the C57BL/6J background and housed in the National University of Singapore in a temperature-controlled condition with a 12-h light/dark cycle and have free access to food and water unless specified.

Mice harvesting and tissue collection

Mice are harvested at 4 months of age. All the harvesting is done at 1 pm-2 pm to avoid the effect of circadian rhythm. Mutant mice and their littermate control mice are always harvested together and used in the study. Muscle samples for histology studies are snap-frozen in liquid nitrogen cold isopentane (2-methylbutane) (Sigma-Aldrich, M32631) and kept at −80°C for long-term storage. All samples for biochemistry studies are frozen with liquid nitrogen and kept at −80°C for long-term storage.

RNA extraction

All consumptions used for RNA extraction are RNAase-free. Frozen samples are taken out from a −80°C refrigerator in dry ice. Muscle samples are crushed into powder in liquid nitrogen cold metal mortar and collected into 2-mL microcentrifuge tubes. Muscle powders are then homogenized with an electric homogenizer in 300 µL of ice-cold TRIzol Reagent (Invitrogen, 15,596,018) until fully homogenized. Ice-cold TRIzol Reagent (700 µL) is then added into the tube to make the final volume equal to 1 mL. An aliquot (200 µL) of chloroform is then added to the lysates and mixed properly by shaking the tube vigorously. The lysates are placed on ice for 5 min to allow the proper extraction of RNA. The lysates are then centrifuged at 4°C for 20 min at 21,160 ×g. The clear layer of lysates is then transferred into new 1.5-mL microcentrifuge tubes with 500 µL of isopropanol, mixed properly by shaking the tube vigorously. The lysates are then centrifuged at 4°C for 10 min at 21,160 ×g. Supernatants are discarded. Pellets are then washed with 500 µL of 70% Methanol by inverting the tube up and down and then centrifuged at 4°C for 10 min at 21,160 ×g. Supernatants are discarded. Pellets are air-dried for 5 min to eliminate the remaining ethanol dissolved in DEPC-treated water (ThermoFisher, AM9915G) and kept at −80°C for long-term storage.

Reverse transcription PCR and real-time PCR analysis

The RNA concentration of the tissue lysates is measured by IMPLEN Nanophotometer N60. The same amount of RNA is reverse transcript to cDNA by High-Capacity cDNA Reverse Transcription Kits (Applied Biosystem, 4,368,814) following the protocol provided by the manufacturer. Real-time PCR is performed with GoTaq® qPCR Master Mix (Promega, A6102) followed by the protocol provided by the manufacturer in the Roche LightCycler® 480 Instrument II. Reactions are performed in triplicate, and relative amounts of cDNA are normalized to Ppia/cyclophilin A (peptidylprolyl isomerase A) [58]. The primers used for real-time PCR are listed in Table S1.

Protein isolation

Frozen samples are taken out from the −80°C refrigerator in dry ice. Muscle samples are crushed into powder in liquid nitrogen cold metal mortar and collected into 2-mL microcentrifuge tubes. Muscle powders are then homogenized with an electric homogenizer (Omni International, TH-02) 3 times x 5 s in 300 µL ice-cold 0.1% SDS RIPA buffer (50 mm Tris, pH 8, 150 mm NaCl, 0.5% sodium deoxycholate [Sigma-Aldrich, D6750-10 G], 0.1% SDS, 1% Triton X-100 [Sigma-Aldrich, T8787]) with Protease Inhibitor Cocktail (Roche, 04,693,124,001) and Phosphatase Inhibitor Cocktail II and III (Sigma-Aldrich, P5726 and P0044). Lysates are then sonicated for 1 time x 5 s under high power for tissue or 1 time x 15 s under high power for cells (Bioruptor Plus, Diagenode) and centrifuged at 4°C for 30 min at 21,160 ×g. Supernatants are collected as the SDS soluble fraction. Pellets are further homogenized by pipetting up and down in ice-cold 2% SDS RIPA buffer (50 mm Tris, pH 8, 150 mm NaCl, 0.5% sodium deoxycholate, 2% SDS, 1% Triton X-100) with Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail II and III. Lysates are then sonicated 10 times x 1 minute under high power. Place the lysates on ice for 30 min to allow the proper extraction of proteins. Centrifuge the lysates at 4°C for 10 min at 21,160 ×g. Discard the pellets, and the supernatant is collected as a SDS insoluble fraction. As for cells, cells are incubated in 100 µL/well of the same lysis buffer for 5 min at 4°C and collected into 1.5-mL microcentrifuge tubes with cell scrapers. Cell lysates are homogenized by pipetting up and down. Lysates are then sonicated for 5 s under high power for tissue or 15 s under high power for cells (Bioruptor Plus, Diagenode). The lysates are placed on ice for 30 min to allow the proper extraction of proteins and then are collected as total fractions for western blot analysis.

Western blot analysis

The protein concentration of the protein lysate extract from the tissue is measured with Pierce BCA Protein Assay Kit (ThermoFisher, 23,225) following the protocol provided by the manufacturer. The same amount of total protein (~30 µg for tissue, ~5 µg for cell) for each sample is loaded into freshly made 8%-15% SDS-PAGE gel. Gels are run for 30 min under 70 V and increased to 120 V till the end. Gels are then transferred to the PVDF membrane (0.2 µm; Bio-Rad, 1,620,177) for blotting LC3 antibody, or nitrocellulose membranes (0.2 µm; Bio-Rad, 1,620,112) for blotting the other antibodies. The gels are transferred in 1X transfer buffer with 20% methanol for 90 min under 100 V for blotting small proteins or transferred in 1X transfer buffer with 10% methanol and 0.05% SDS overnight under 30 V for blotting big proteins. After transfer, membranes are stained with Ponceau S (Sigma, P3504) and imaged with iBright FL1500 as a loading control. Membranes are then washed with TBST (0.1% Tween 20 [Sinopharm Chemical, T2008687] in Tris-Buffered Saline [137 mm NaCl, 2.7 mm KCl and 24.8 mm Tris]) and blocked for 1 hour in 5% milk in TBST at room temperature on an orbital shaker (60 rpm). After blocking, incubate the membrane with primary antibody (1:1000 in 1% BSA [Sigma-Aldrich, A3803]) overnight in a cold room (4). The membranes are then washed with TBST for 3 times x 10 min. Membranes are then incubated with secondary antibodies (1:5000 in 5% milk) for flexible times depending on the abundance of the proteins. After incubating with secondary antibody, membranes are washed with TBST 3 times x 10 min and incubated with ECL reagents (Pierce, 32,106, for strong signals; Amersham, RPN2232, for weak signals) and image with iBright FL1500. The intensity of the bands is quantified by Fiji ImageJ and normalized to the intensity of Ponceau S. The primary antibody used for western blot analysis is listed in Table 1.

Table 1.

List of primary antibodies.

Antibody	Origin	Company	Cat. No.	Experiment
EIF4EBP1	Rabbit	Cell Signaling Technology	9452S	WB
ACAC/Acetyl-CoA Carboxylase	Rabbit	Cell Signaling Technology	3662S	WB
ACAC/Acetyl-CoA Carboxylase p-S79	Rabbit	Cell Signaling Technology	3661	WB
PRKAA/AMPKα (23A3)	Rabbit	Cell Signaling Technology	2603S	WB
PRKAA/AMPKalpha p-T172 (40H9)	Rabbit	Cell Signaling Technology	2535S	WB
ATG7	Rabbit	Cell Signaling Technology	8558	WB
FBXO32/Atrogin-1/MAFbx (F-9)	Mouse	Santa Cruz Biotechnology	sc -166,806	WB
CTSB/cathepsin B (D1C7Y) XP	Rabbit	Cell Signaling Technology	31718S	WB
CTSL/cathepsin L	Goat	R&D	AF1515	WB
GAPDH (14C10)	Rabbit	Cell Signaling Technology	2118S	WB
Histone H3	Rabbit	Genetex	GTX122148	WB
HSPA8/HSC70	Rabbit	Proteintech	10654–1-AP	WB
HSP90 (C45G5)	Rabbit	Cell Signaling Technology	4877S	WB
K48-linkage-Specific Polyubiquitin (D9D5)	Rabbit	Cell Signaling Technology	8081	WB
K63-linkage-Specific Polyubiquitin	Rabbit	abcam	ab179434	WB
LAMP1 (1D4B)	Rat	Santa Cruz Biotechnology	sc -19,992	WB
LAMP2 (ABL-93)	Rat	DSHB	ABL-93	WB
LAMP2A	Rabbit	abcam	ab18528	WB
LC3B	Rabbit	Sigma-Aldrich	L7543	WB
TRIM63/MuRF1 (C-11)	Mouse	Santa Cruz Biotechnology	sc -398,608	WB
NFE2L1/TCF11/NRF1 (D5B10)	Rabbit	Cell Signaling Technology	8052S	WB
OXPHOS Cocktail	Mouse	Abcam	ab110411	WB
SQSTM1/p62	Mouse	Abnova	H00008878-M01	WB
PSMA2	Rabbit	Cell Signaling Technology	2455S	WB
PSMB10	Rabbit	Cell Signaling Technology	17579	WB
PSMB5	Rabbit	Abcam	ab3330	WB
PSMB6 (E1K9O)	Rabbit	Cell Signaling Technology	13267	WB
PSMB7 (E1L5H)	Rabbit	Cell Signaling Technology	13207S	WB
PSMB8 (LMP7)	Rabbit	Abcam	ab3329	WB
PSMB9 (LMP2)	Rabbit	Abcam	ab3328	WB
PSMD1	Rabbit	Abcam	ab2941	WB
PSME1 (PA28A)	Rabbit	Cell Signaling Technology	9643	WB
PSME2 (PA28B)	Rabbit	Cell Signaling Technology	2409	WB
SDHA (F-2)	Mouse	Santa Cruz Biotechnology	sc -390,381	WB
STAT3 (D3Z2G)	Rabbit	Cell Signaling Technology	12640	WB
NFKB1 p105/p50 (D4P4D)	Rabbit	Cell Signaling Technology	13586S	WB
Ubiquitin (E4I2J)	Rabbit	Cell Signaling Technology	43124	WB
MYH2	Mouse-IgG1	DSHB	SC-71	IF
MYH4	Mouse-IgM	DSHB	BF-F3	IF
MYH7	Mouse-IgG2b	DSHB	BA-F8	IF

Proteasome activity assay

Frozen muscle samples are taken out from a −80°C refrigerator in dry ice crushed into powder in liquid nitrogen cold metal mortar and collected into 2-mL microcentrifuge tubes. Muscle powders are then homogenized with an electric homogenizer 3 times x 5 s in ice-cold lysis buffer containing 50 mm Tris – HCl, pH 7.5, 250 mm sucrose (Sigma-Aldrich, 84,097), 5 mm MgCl₂, 0.5 mm EDTA, and 1 mm DTT for measuring the 20S proteasome activity and immunoproteasome activity. For measuring the 26S proteasome activity, 2 mm ATP (Sigma-Aldrich, A7699) is added freshly into the lysis buffer. Proteasome lysates are then centrifuged at 4°C for 30 min at 12,000 ×g. Supernatants are collected for measuring proteasomal degradation capacity and native gel immunoblotting. The protein concentration of the proteasome lysate is measured by Bradford’s essay (Bio-Rad 5,000,006). Proteasome lysates are diluted with lysis buffer to 1 µg/µL according to the protein concentration. As for the 20S and 26S proteasome activity measurement, the reaction mix for the experiment group is 20 µL diluted proteasome lysate +20 µL lysis buffer +10 µL DMSO +50 µL 200 µM Suc-LLVY-AMC (Bachem, I-1395), reaction mix for the negative control group is 20 µL diluted proteasome lysate +20 µL lysis buffer +10 µL 200 mm MG132 (Sigma-Aldrich, M7449-1 ML) + 50 µL 200 µM Suc-LLVY-AMC. As for the immunoproteasome activity measurement, the reaction mix for the experiment group is 20 µL diluted proteasome lysate +10 µL lysis buffer +2.5 µL DMSO/10 µM ONX-0914 (MedChemExpress, HY-13207) + 17.5 µL DMSO +50 µL 200 µM Suc-LLVY-AMC, reaction mix for the negative control group is 20 µL diluted proteasome lysate +10 µL lysis buffer +10 µL 10 µM ONX-0914 + 10 µL 200 mm MG132 + 50 µL 200 µM Suc-LLVY-AMC. Reaction mixes are loaded onto a black 96-well plate with a flat bottom and incubated at 37°C for 60 min. Fluorescence intensity is read by Tecan Infinite M200 plate reader at 390 nm excitation/460 nm emission wavelength.

Subunit-specific proteasome activity assay

Subunit specific proteasome activity is measured by Proteasome-Glo™ Assays Kit (Promega, G8531) followed by the protocol provided by the manufacturer. Briefly, frozen muscle samples are taken out from a −80°C refrigerator in dry ice crushed into powder in liquid nitrogen cold metal mortar and collected into 2-mL microcentrifuge tubes. Muscle powders are then homogenized with an electric homogenizer 3 times x 5 s in ice-cold lysis buffer containing 5 mm EDTA (in phosphate-buffered saline [PBS; Gibco, 18,912–014], pH 7.2). Proteasome lysates are then centrifuged at 4°C for 10 min at 12,000 ×g twice. Supernatants are collected for measuring proteasomal degradation capacity. The protein concentration of the proteasome lysate is measured by Bradford’s assay (Bio-Rad, 5,000,006). Proteasome lysates are diluted with lysis buffer to 1 µg/µL according to the protein concentration. The reaction mix for the experiment group is 20 µL diluted proteasome lysate +20 µL lysis buffer +10 µL DMSO +50 µL Proteasome Glo detection cocktail, the reaction mix for the negative control group is 20 µL diluted proteasome lysate +20 µL lysis buffer +10 µL 200 mm MG132 + 50 µL Proteasome Glo detection cocktail. Reaction mixes are loaded onto a black 96-well plate with a flat bottom and incubated at 37°C for 60 min. Luminescence intensity is read by Tecan Infinite M200 plate reader.

Native gel immunoblotting

The protein concentration of proteasome lysate (with ATP) is measured by Bradford’s essay. Proteasome lysate is diluted to 1 µg/µL with lysis buffer and mixed with 0.25 times volume of 5X native gel loading dye (250 mm Tris, pH 7.4, 50% glycerol, 60 ng/ml xylene cyanol). Mixed lysate (10 µL) is loaded for each sample into 3–12% SDS-free native gel (Invitrogen, BN1003BOX). Gels are run in ice-cold running buffer (90 mm Tris base-boric acid, pH 8, 5 mm MgCl₂, 0.5 mm EDTA, 1 mm ATP-MgCl₂ [250 mm ATP, and 250 mm MgCl₂ diluted in 500 mm Tris, pH 7.0]) for 4 h at 150 V in a cold room. After running, gels are calibrated in SDS-PAGE running buffer for 10 min at room temperature and overnight transferred to PVDF membrane (0.2 µm) at 15 V in 1X transfer buffer with 1% methanol in a cold room. The following steps are the same as the western blot analysis mentioned above.

Nuclear fraction isolation

Muscle samples (~5 mg) are taken out from −80°C refrigerator in dry ice and minced into small pieces (2–3 mm³) in ice-cold lysis buffer containing 10 mm HEPES-KOH, pH 7.3, 10 mm KCl, 5 mm MgCl₂, 0.1% NP-40 (BioChemika, 74,385), 0.1 mm PMSF (Sigma-Aldrich, P7621) with Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail II and III. Lysates are then centrifuged at 4°C for 3 min at 1,000 ×g. Supernatants are discarded, and pellets are resuspended with 1.5 mL of ice-cold lysis buffer and transferred into a small Dounce homogenizer (Kimble Kontes, 886,000–0021) and homogenized by 20–30 strokes within 3 min. Lysates are transferred back to the original tubes. Lysates (100 µL) are aliquots into a new microcentrifuge tube and frozen at −80°C as the total lysates. The rest of the lysates are centrifuged at 4°C for 10 min at 1,000 ×g, and supernatants are transferred into a new microcentrifuge tube and frozen at −80°C as the cytosolic fraction. Pellets are resuspended with 10 mL ice-cold lysis buffer and filtered through a 40-µm filter (Fisher Scientific, 22-363-547). Lysates are centrifuged at 4°C for 10 min at 1,000 ×g, and supernatants are discarded, and pellets are washed with 1 mL ice-cold lysis buffer and centrifuged at 4°C for 10 min at 1,000 ×g. Pellets are collected as a nuclear fraction and lysed with 0.1% SDS RIPA buffer and processed in the same way as muscle protein extraction discussed above.

Adenosine triphosphate (ATP) assay

ATP content in the muscle lysates is measured by ATP Assay Kit (Colorimetric/Fluorometric) (Abcam, ab83355) followed by the protocol provided by the manufacturer. Briefly, extensor digitorum longus (EDL) muscle is homogenized with a Dounce homogenizer in the lysis buffer provided by the kit. The lysates are then deproteinized by treatment with trichloroacetic acid (TCA; Sigma-Aldrich, T6399-100 G). Extra TCA is then neutralized by KOH. The lysates are then mixed with the detecting reagent provided by the kit and loaded into a black 96-well plate with a flat bottom. A fluorometric essay is used for the measurement included in this study.

MYH (myosin, heavy polypeptide) staining

Mouse skeletal muscles are sectioned into 15-um cross-sections at −25°C in a cryostat station and then air-dried at room temperature for around 30 min. Sections are then washed with PBS for 10 min to remove the remaining O.C.T. Compound (Sakura, 4583). Sections are then blocked with 10% goat serum (Gibco, 16,210,072) in PBS for 120 min and incubated with primary antibody dissolved in blocking solution overnight at 4°C. After incubating with primary antibody, sections are washed with PBS for 3 times x 5 min and incubated with secondary antibody dissolved in blocking solution for 120 min at room temperature. Sections are then washed with PBS 3 times x 5 min and mounted with ProLong™ Gold Antifade Mountant (Invitrogen, P36930). Sections are imaged with Tissue Fax at 20X magnification and quantified by Fiji ImageJ for fiber proportion and Feret diameter. Primary antibodies used for MYH staining is listed in Table 1.

Nicotinamide adenine dinucleotide (NADH) staining

Mice skeletal muscles are sectioned into 15-µm cross-sections at −25°C in a cryostat station and air-dried at room temperature for around 30 min until the sections are fully dried. Sections are then washed with PBS for 10 min to remove the remaining O.C.T. Compound. The staining solution is prepared by mixing the NBT buffer containing 2 g/L Nitroblue tetrazolium (Sigma-Aldrich, 11,585,029,001) and 0.05 M Tris buffer, pH 7.2 with NADH solution containing 1.6 g/L of NADH (Roche, 10,107,735,001) and 0.05 M Tris buffer in 1:1 ratio. The staining solution is then warmed up for 10 min at 37°C. The slides are incubated in the mixture for 5 ~ 30 min at 37°C. Sections are then washed with PBS 3 times x 5 min and mounted with Prolong gold mounting medium (Invitrogen, P36930). Sections are imaged with Tissue Fax at 20X magnification.

Wheat germ agglutinin (WGA) staining

Mice skeletal muscles are sectioned into 15-µm cross-sections at −25°C in a cryostat station and air-dried at room temperature for around 30 min until the sections are fully dried. Sections are then washed with PBS for 10 min to remove the remaining O.C.T. Compound. Sections are then incubated in Wheat Germ Agglutinin (WGA) Alexa Fluor™ 647 Conjugate (1:200; Invitrogen, W32466) for 10 min at room temperature. Sections are then washed with PBS 3 times x 5 min and mounted with Prolong gold mounting medium. Sections are imaged with confocal microscopy at 60X magnification.

Magic red staining

Mice skeletal muscles are sectioned into 15-µm cross-sections at −25°C in a cryostat station and air-dried at room temperature for around 30 min until the sections are fully dried. Sections are then washed with PBS for 10 min to remove the remaining O.C.T. Compound. Magic Red substates for CTSB (Bio-Rad, ICT-938), CTSK (Bio-Rad, ICT940), CTSL (Bio-Rad, ICT-942) are reconstituted with 200 µL DMSO. Sections are then incubated in a mix of Magic Red substrates for CTSB, CTSK, and CTSL (1 µL for each substrate), Wheat Germ Agglutinin (WGA) Alexa FluorTM 488 Conjugated (1:200; Invitrogen, W11261), and DAPI (1:1000; Sigma-Aldrich, D9542) diluted in 1X PBS for 30 min at room temperature. Sections are then washed with PBS 3 times x 5 min and mounted with Prolong gold mounting medium. Sections are imaged with confocal microscopy at 60X magnification.

Cell culture and treatment

Primary myoblasts are isolated from three one-month old female mice, carrying Atg7 flox allele. Myoblasts are individually cultured with DMEM (Hyclone, SH30243.01) and Ham’s F10 (Pan Biotech, P04-12500) containing 20% FBS (Hyclone, SV30160.03) according to their original identity. Atg7 is deleted through infection with lentivirus expressing Cre recombinase. Lentivirus is produced by transfecting LentiX-293T cells with lentivirus packaging plasmids (pMDLg/pRRE [Addgene, 12,251; deposited by Didier Trono lab], pRSVRev [Addgene, 12,253; deposited by Didier Trono lab] and pMD2.G [Addgene 12,259; deposited by Didier Trono lab]). pLV-EF1-Cre-PGK-puro (Addgene, 108,543; deposited by Javier Alcudia lab) is also co-transfected with the packaging plasmid to create Cre lentivirus. As the control, pLenti X1 Puro empty (w610–1; Addgene, 20,953; deposited by Eric Campeau lab) is co-transfected with the packaging plasmids to produce the control virus. The control myoblasts are those infected with control lentivirus. Medium is replaced 16 h post transfection. The condition medium with lentivirus is collected for infecting myoblasts. After infection, myoblasts are selected with 1ug/mL of puromycin (Sigma, P9620) for 6 days and are allowed to recover for another 6 days. To test autophagy flux in these primary myoblasts, control cells and Atg7 knockdown myoblasts are treated with 1 µM torin (Tocris, 4247) or 100 nM bafilomycin A₁ (Sigma-Aldrich, B1793) or both for 2 h. DMSO are used as a vehicle control.

Statistical analysis and illustration

The statistical analysis done for Figure 1B (LC3–1 and SQSTM1/p62), Fig. S1D, S2E, S2F, S4B, S4C, S4E, and S5A are analyzed by two-tailed unpaired t-test and for Fig. S2G, S4A, and S5B are analyzed by Two-Way ANOVA followed by Šídák’s multiple comparison test. The statistical analysis done for quantification of myofiber size in Figure 3I, Fig. S3D, and S3E are determined by Two-Way ANOVA with Tukey’s multiple comparison test compared within the size class between genetic groups. The rest of the statistical analyses done in this study are analyzed by One-Way ANOVA followed by Tukey’s multiple comparison test. All the quantification images included in this study are drawn by GraphPad Prism, the schematic diagram in Figure 5 is drawn by BioRender (https://www.biorender.com/).

Funding Statement

The work was supported by the National Medical Research Council [OFIRG18nov-0093, OFIRG24jan-0020], and the Ministry of Education Tier1 NUHSRO/2022/070/T1/Seed-Mar/06 and Tier 2 MOE-T2EP30223-0005 (to Shih-Yin, Tsai).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2025.2457925