Calorie restriction and rapamycin distinctly mitigate aging-associated protein phosphorylation changes in mouse muscles

Aging is a major risk factor for chronic, inflammatory and malignant diseases, with sarcopenia (age-related loss of muscle mass and strength) as a key consequence. Perturbations contributing to sarcopenia include reduced anabolic signaling, insulin resistance and chronic inflammation. The mTOR pathway, particularly mTORC1, regulates muscle proteostasis by promoting anabolic processes and inhibiting autophagy. mTORC1 activity is elevated in aged mouse muscle; sustained activation causes late-onset myopathy with impaired autophagy, whereas long-term inhibition with rapamycin preserves muscle mass and function. CR and RM are among the most robust interventions to delay aging, but their efficacy varies across muscles (soleus, tibialis anterior, triceps brachii, gastrocnemius). Prior studies mainly analyzed mRNA responses, leaving signaling via protein phosphorylation insufficiently understood. This study addresses that gap by profiling phosphoproteomes in four distinct mouse muscles across adult, aged, and aged animals subjected to long-term CR or RM to determine how these interventions modulate aging-associated signaling changes.

Previous work established increased mTORC1 activity as a hallmark of aged skeletal muscle and linked sustained mTORC1 activation to myopathy with defective autophagy, while rapamycin mitigates age-related muscle decline. CR and RM extend healthspan/lifespan across species and improve muscle function in aged mice, though responses differ by muscle type. Earlier investigations in these muscles largely focused on transcriptome-level changes, lacking pathway-level phosphorylation data. The mTOR pathway has substrates differentially sensitive to RM; prolonged RM can also inhibit mTORC2. Limited phosphosite annotations in repositories like PhosphoSitePlus constrain kinase attribution, highlighting a need for broader phosphoproteomic mapping to understand aging and interventions at the signaling level.

Experimental design: Male C57BL/6JRj mice. Adult control group at 10 months (10M_WT). From 15 to 30 months, mice received: control diet (30M_WT), calorie restriction (30M_CR; 65% of standard diet without restricting vitamins/minerals), or rapamycin diet (30M_RM; 42 mg/kg chow encapsulated RM, ~4.2 mg/kg/day based on intake). All mice single-caged; food removed morning prior to sacrifice; tissues collected within a 3–4 h window late morning. Four muscles analyzed: soleus (slow-twitch), tibialis anterior, triceps brachii, gastrocnemius (primarily fast-twitch). Sample preparation and phosphoproteomics: Approximately 3 mg muscle tissue per sample was snap-frozen, mechanically ground, lysed in 8 M urea/0.1 M ammonium bicarbonate with phosphatase inhibitors, reduced (TCEP), alkylated (chloroacetamide), diluted, and digested with trypsin. Peptides were desalted and enriched for phosphopeptides using Fe(III)-IMAC on an AssayMAP Bravo platform. LC-MS/MS: EASY-nLC 1000 coupled to Q-Exactive HF, 75 µm × 37 cm in-house packed C18 column, 90-min gradient to 30% B, DDA with top-10 HCD, dynamic exclusion 20 s; MS1 at 120k resolution (200 m/z), MS2 at 30k. Global proteomics: Aliquots (0.25 µg) analyzed on Orbitrap Fusion Lumos with EASY-nLC 1200, 75 µm × 36 cm C18 column, multistep gradient to 95% B; DDA with 3 s cycle, MS1 at 120k, MS2 in ion trap (Rapid), HCD 35%, dynamic exclusion 30 s. Data processing: Progenesis QI (v2.0) for precursor intensity extraction; mgf exported and searched with MASCOT against Swiss-Prot Mus musculus plus contaminants (forward/reverse) with full tryptic specificity, up to 3 missed cleavages; fixed carbamidomethyl (C), variable oxidation (M) and phosphorylation (S/T/Y for phospho runs); tolerances 10 ppm precursor, 0.02/0.6 Da fragments for phospho/proteome. FDR controlled at 1% at protein level. Normalization and statistics with limma (v3.64.0). Custom phosphoproteomics scripts deposited (Zenodo DOI: 10.5281/zenodo.10635552). Analyses: Principal component analysis of phosphopeptide intensities; comparison with total protein abundance to assess contribution of expression to phosphosite intensity. Differential phosphosite analysis computed as log2 fold changes between 30M_WT vs 10M_WT (AGE), 30M_CR vs 30M_WT (CR), 30M_RM vs 30M_WT (RM) separately per muscle. Curated lists of direct mTORC1, Rps6kb, upstream mTORC1 regulators, and mTORC2 substrates were interrogated. Kinase Set Enrichment Analysis (KSEA) used PhosphoSitePlus kinase–substrate sets with weighted Kolmogorov–Smirnov tests to infer kinase activity changes. Gene Ontology overrepresentation with clusterProfiler; heatmaps with ComplexHeatmap and circlize. Tests for phosphopeptide changes independent of protein levels used two-tailed t-tests comparing protein intensities in a target muscle versus others; discordant directionality or non-significant protein changes indicated phosphosite-specific regulation.

• Broad phosphoproteome coverage: 6960 non-redundant phosphosites detected (6526 in ≥48/95 samples); 1415 sites not present in PhosphoSitePlus (novel). Novel sites span 227 proteins, including 19 proteins not previously known to be phosphorylated in mouse.
• Muscle specificity: 6780 sites detected in all four muscles; PCA separated soleus (PC1) and tibialis anterior (PC2) from triceps/gastrocnemius, indicating muscle-specific phosphorylation patterns. Phosphosite intensities partially correlate with protein abundance (strongest in soleus) but many sites show regulation beyond protein levels (e.g., Myom1 S863 elevated in soleus despite lower protein; Actn2 S109 elevated in tibialis anterior without protein correlation).
• mTOR pathway modulation: Long-term RM broadly reduced phosphorylation of RM-sensitive mTORC1 substrates and most consistently reduced Rps6kb substrates (including Rps6), validating expected target inhibition. CR effects were more variable across muscles and sometimes opposite to RM (e.g., Eif4ebp1 T69 increased with CR, decreased with RM).
• Upstream mTORC1 regulators: Akt1s1 (PRAS40) S213 phosphorylation consistently reduced by RM; Akt1s1 S204 increased with aging in 3/4 muscles, mitigated by CR and especially RM. Rptor S722 (AMPK site) phosphorylation increased by CR and RM, consistent with mTORC1 inhibition.
• mTORC2 substrates: Five sites in five substrates detected; all upregulated with age, suggesting increased mTORC2 activity in aging muscle. CR and RM reduced phosphorylation at TM sites Akt2 S450 and Prkca T638, consistent with mTORC2 inhibition upon prolonged RM.
• Global regulated sites across muscles: Commonly regulated phosphopeptides across all four muscles: AGE 227, CR 291, RM 172. AGE-associated changes involved stress granules (G3bp1 S149), actin stress fibers (Cap1 S307), autophagy (Sqstm1 T269), fatty acid metabolism (Acaca), with many sites annotated to AMPK; CR did not consistently mitigate AMPK-target increases. CR affected cell adhesion/motility/differentiation and reduced Rps6 S236, consistent with mTORC1 suppression. RM effects concentrated on mTORC1 pathway and increased Camk2b-associated phosphorylation, including Camk2b itself.
• Unannotated sites: 204 (AGE), 255 (CR), 150 (RM) consistent kinase-unannotated sites across muscles. Treatments generally mitigated AGE effects. CR countered AGE-related changes in neurofilaments (Nefh, Nefm), ER stress/trafficking proteins (Ubxn6, Pdia6, Hspa5, Naca, Tom1, Syt2). RM countered changes in glucose metabolism (Tbc1d4), mitochondria (Atp2a1/SERCA1), Tardbp, and muscle structure (Dmd), and increased Synm S1502 phosphorylation. Some sites responded to treatments but not age (e.g., increased phosphorylation in Flnc, Lmod2, Mybph, Nefm; Bnip3 S79).
• Kinase activity inference (KSEA): Aging reduced Akt2 and increased Mapk14/p38α activity in soleus and tibialis anterior; both CR and RM counteracted these changes (CR also in gastrocnemius). RM suppressed age-related increases in Stk39 phosphorylation; CR increased Pfkfb2 and Gsk3a phosphorylation. RM increased multiple Camk2b sites including autophosphorylation at T287 (persistent activity), paralleled by increased Hdac4 S245, consistent with Camk2b–Hdac4 signaling and potential NMJ involvement.
• Atlas expansion and tissue specificity: Known and novel phosphoproteins showed comparable abundance, arguing against detection bias. Novel phosphoprotein genes exhibit muscle-restricted expression patterns in Human Protein Atlas data, underscoring the value of muscle-focused phosphoproteomics.

The data show that both CR and RM counteract aging-associated increases in mTOR pathway activity in skeletal muscle, with RM exerting a more uniform and robust suppression across muscles than CR. Aging increased phosphorylation in diverse signaling proteins, while phosphorylation in structural muscle components (e.g., neurofilaments, myofibrillar proteins) decreased and was restored by the interventions, suggesting improved muscle structure/function signaling. Differential responses across muscles reflect intrinsic fiber-type and functional differences and likely underlie previously observed disparities in treatment efficacy. KSEA highlighted consistent aging-associated declines in Akt2 and increases in p38α activity that were mitigated by CR and RM, while also revealing muscle-specific modulation of kinases such as ROCK2 and CaMKIIβ. Prolonged RM reduced phosphorylation of both mTORC1 substrates and certain mTORC2-dependent sites, including TM sites, aligning with known long-term RM effects. The expansion of the phosphosite atlas, especially muscle-specific and previously unannotated sites, provides new opportunities to map kinase–substrate relationships and to understand the signaling basis of sarcopenia and its amelioration by CR and RM. Potential NMJ-related mechanisms (Camk2b autophosphorylation and Hdac4 regulation) emerge as targets for further investigation.

This study delivers a comprehensive phosphoproteomic resource across four mouse skeletal muscles, revealing how aging and long-term CR or RM remodel signaling. RM consistently reduces phosphorylation of mTORC1 and Rps6kb substrates and, with CR, mitigates increased activity of mTOR complexes and age-associated kinase activity shifts (notably Akt2 and p38α). Both interventions restore phosphorylation signatures of structural muscle proteins linked to contraction and cytoskeleton. The dataset adds 1415 novel phosphosites and identifies 19 novel mouse phosphoproteins, many with muscle-restricted expression, substantially expanding the muscle phosphoproteome. Future directions include elucidating the cognate kinases for unannotated sites, dissecting the differential sensitivity and feedback within mTORC1 substrate networks under long-term intervention, validating the functional role of Camk2b–Hdac4 signaling at the neuromuscular junction during aging, and exploring muscle-specific kinase pathways (e.g., ROCK2) to tailor interventions.

Phosphosite coverage remains incomplete and biased by detection limits; many regulated sites lack annotated cognate kinases, complicating mechanistic attribution. Phosphorylation states are transient and tissue/condition specific, challenging comprehensive capture. Muscle-specific variability limits generalization from any single muscle. Functional validation of many phosphorylation changes was not performed. The study used only male mice and focused on two ages (10 and 30 months) with interventions from 15 to 30 months, which may not capture sex- or age-window-specific effects. Prolonged RM can impact mTORC2, potentially confounding interpretation of mTORC1-specific effects. Proteomic quantification is label-free and subject to inherent quantitative uncertainties.