Lysosomal phosphoinositide turnover acts upstream of RagGTPase–mTORC1 and controls muscle growth

Picot, Melanie; Hifdi, Nesrine; Vaucourt, Mathilde; Mansat, Melanie; Li, Peng; Singer, Isabel; Chicanne, Gaëtan; Stella, Alexandre; Kuang, Shen; Miceli, Caterina; Henneman, Nathaniel F.; Payrastre, Bernard; Vandromme, Marie; Schiltz, Odile; Nemazanyy, Ivan; de Araujo, Mariana E. G.; Panasyuk, Ganna; Viaud, Julien; Lämmerhofer, Michael; Hnia, Karim

doi:10.1038/s42255-026-01484-1

Download PDF

Article
Open access
Published: 18 March 2026

Lysosomal phosphoinositide turnover acts upstream of RagGTPase–mTORC1 and controls muscle growth

Nature Metabolism volume 8, pages 624–645 (2026) Cite this article

6082 Accesses
13 Altmetric
Metrics details

Subjects

This article has been updated

Abstract

Lysosomes act as metabolic signalling hubs that integrate nutrient availability to coordinate anabolic and catabolic programmes. Mechanistic target of rapamycin complex 1 (mTORC1) is activated at the lysosomal surface by amino acids through RagGTPases recruited by the lysosomal adaptor and MAPK and mTOR activator complex, yet the contribution of lysosomal lipid composition to this pathway remains unclear. Here we identify lysosomal phosphoinositides, PI3P and PI(3,5)P2, as key regulators of lysosomal adaptor and MAPK and mTOR activator complex stability and dynamics at the lysosome. These lipid pools are controlled by the phosphoinositide 3-phosphatase MTM1, mutated in myotubular myopathy, via endoplasmic reticulum–lysosome membrane contact sites. Under endoplasmic reticulum stress, MTM1-dependent phosphoinositide remodelling suppresses RagGTPase–mTORC1 signalling, thereby regulating anabolic–catabolic balance during myogenic differentiation. Restoring mTORC1 activity or lysosomal phosphoinositide homeostasis rescues Rag-dependent signalling and muscle growth in cellular and mouse models of myopathy, uncovering a lysosome-centred metabolic checkpoint with direct disease relevance.

The lysosomal Ragulator complex plays an essential role in leukocyte trafficking by activating myosin II

Article Open access 07 June 2021

Spatial and functional separation of mTORC1 signalling in response to different amino acid sources

Article Open access 09 October 2024

Identification of a resistance-exercise-specific signalling pathway that drives skeletal muscle growth

Article 15 May 2025

Main

Lysosomal dysfunction has been linked to muscle wasting (atrophy) during physiological ageing (sarcopenia) and to genetic related diseases such as lysosomal storage diseases (LSDs) and myopathies¹. How lysosome function is controlled during muscle differentiation and maintenance is not known. The anabolic function of the lysosome relies on the recruitment to and activation of a pool of mechanistic target of rapamycin complex 1 (mTORC1) on the lysosomal surface². Conversely, a lack of nutrients inhibits mTORC1 signalling, promotes lysosome biogenesis and activates catabolic cascades such as lysosomal degradation by autophagy^3,4. Because mTORC1 triggers a fairly resource-intensive anabolic programme, cells have evolved mechanisms to ensure that it only becomes active when sufficient resources are available. As such, lysosomal mTORC1 is activated as a coordinated response to amino acids, cholesterol and glucose availability (through the RagGTPase signalling pathway); growth factors (through inhibition of the tuberous sclerosis complex (TSC) and activation of Rheb GTPase); and cellular energy status (through AMP-activated protein kinase (AMPK) phosphorylation of TSC2 and Raptor)^5,6,7. The Rheb GTPase is activated by the canonical PI3K/AKT/TSC1/TSC2 axis, while several multi-protein complexes have been found to control the RagGTPases. This includes the late endosomal/lysosomal adaptor and MAPK and mTOR activator (LAMTOR)/Ragulator complex (LAMTOR1–5)², GATOR1^8,9, SLC38A9¹⁰ and FLCN–FNIP2 complexes^11,12. Both small GTPases within the heterodimer can associate with either GDP or GTP, and it is their joint status that determines the dimer’s activity. In the presence of anabolic stimuli, RagA/B is bound to guanosine triphosphate (GTP) and RagC/D to guanosine diphosphate (GDP) (GTP-RagA–RagC-GDP), a constellation that allows association with Raptor and thereby mTORC1 recruitment¹³. By contrast, an opposite nucleotide configuration (GDP-RagA–RagC-GTP) weakens the interaction with mTORC1 so that it does not localize to lysosomes and the pathway is suppressed¹⁴. The lysosomal localization of the RagGTPase heterodimer is dictated by the Ragulator complex². Nutrients have been shown to modulate the RagGTPase–Ragulator interaction, ultimately allowing the activated RagGTPase dimer to cycle between lysosome-bound RAGULATOR and the cytoplasm¹⁵. Despite these structural and biochemical evidences, it is still poorly defined how lysosomal membrane composition impacts Ragulator–RagGTPase assembly and function. This regulation might be particularly essential in physiological contexts, such as during cell differentiation when cell fate programmes inflict profound changes on organelle dynamics and homeostasis.

Here, we identify a lysosomal pool of phosphatidylinositol 3-phosphate (PI3P) and phosphatidylinositol 3,5-biphosphate (PI(3,5)P2), that control RagGTPase–mTORC1 signalling. PI3P and PI(3,5)P2 anchors the LAMTOR complex to the lysosome membrane through direct binding to LAMTOR1 and LAMTOR3 subunits. This interaction modulates LAMTOR complex protein turnover and retention of RagGTPase on the lysosome. PI3P and PI(3,5)P2 are controlled by a phosphoinositide conversion cascade involving the PI3P-phosphoinositide phosphatase myotubularin (MTM1), a protein missing in myotubular myopathy (also named X-linked centronuclear myopathy (XLCNM), a deadly untreatable disease in newborns and teenagers). MTM1 controls the lysosomal pool of PI3P and PI(3,5)P2 through endoplasmic reticulum (ER)–lysosome contact sites, enabling the repression of RagGTPase–mTORC1 under physiological stress during muscle differentiation. Thus, our study reveals a novel lysosomal phosphoinositide function in controlling the LAMTOR–RagGTPase axis, with direct impact on future therapies for XLCNM and possibly other diseases where the RagGTPase–mTORC1 axis is overactivated^16,17.

Results

Protein synthesis is affected in congenital myopathies

Dysfunction of cellular catabolism is reported in genetic diseases such as myopathies¹⁸. However, little is known about the status of anabolism, such as translation, in the same pathological context. We used a well-established murine muscle cell line C2C12 to screen using small interfering RNA (siRNA) the impact of 36 congenital myopathy genes on protein synthesis during differentiation (Fig. 1a and Extended Data Fig. 1a). To measure protein synthesis, we employed surface sensing of translation (SUnSET) and amino acid analogues of methionine: Alexa-labelled 488-L-homopropargylglycine (HPG) and/or Alexa 595-L-azidohomoalanine (AHA) (Fig. 1a and Extended Data Fig. 1b). We found that knockdown of five of the tested genes (CryaB, Dnm2, Sepn1, Vcp and Vma21) inhibited the protein synthesis rate, while silencing of four genes (Mtm1, Bag3, Klhl40 and Tia1) activated it (Fig. 1b and Extended Data Fig. 1b). Notably, silencing of Mtm1 resulted in the most overt effect (Fig. 1b). MTM1 is mutated in a severe congenital myopathy with neonatal onset (XLCNM) and encodes the phosphoinositide (PI) 3-phosphatase myotubularin (MTM1)¹⁹. Most patient mutations lead to loss of MTM1 function due to absent/decreased protein levels (Extended Data Fig. 1c). We confirmed the inhibitory effect of MTM1 on protein translation both in XLCNM primary myotubes, as well as in vivo, using a SUnSET assay on muscle sections of the Mtm1-knockout (KO) mouse model (Fig. 1c,d and Extended Data Fig. 1c). These findings prompted us to test the protein synthesis dynamics during a cell differentiation time course. Notably, protein synthesis was increased in Mtm1-KO cells starting from mid-differentiation (5–6 days) and persisted to late stages (9 days) (Fig. 1e and Extended Data Fig. 1d). Unlike mutant cells, control (CTRL) cells showed translational waves with peaks in proliferative cells (myoblasts) at the mid-stage (5–6 days) and the late stage (7–8 days) of myotube differentiation (Fig. 1e and Extended Data Fig. 1d). Importantly, the increased protein synthesis rate observed in XLCNM models was accompanied by altered cell differentiation, manifested by less nuclei per myotube and decreased myotube area (Fig. 1c,e and Extended Data Fig. 1e).

**Fig. 1: Genetic screen identifies MTM1 as a negative regulator of protein synthesis during muscle differentiation.**

MTM1 depletion promotes mTORC1 overactivation in differentiated cells

To identify the molecular cascades involved in the translation abnormalities in Mtm1-KO cells, we treated myotubes at 5-days post-differentiation with compounds known to inhibit protein synthesis (Fig. 2a). We selected compounds that act on key cascades such as the growth factor-AKT/PI3K signalling pathway (wortmanin and LY294002), mTORC1 (torin and AZD8055), its downstream target S6K (PF-4708671) and the translation machinery eukaryotic translation initiation factor 2 α-subunit (eIF2α) (salubrinal). While all tested drugs decreased protein synthesis rate, wortmannin and LY294002 showed deleterious effects on differentiation (Fig. 2b–e and Extended Data Fig. 2a–c). Remarkably, the ATP-competitive inhibitors of mTORC1 (torin 1 and AZD8055) rescued differentiation of Mtm1-KO myotubes (Fig. 2b–e). However, mTORC1 inhibition failed to improve cell differentiation when myotubes were treated at 2 days post-differentiation (Extended Data Fig. 2d), suggesting that the deregulation of mTORC1 in the Mtm1-KO model could be dependent on time-course cell differentiation. Then, we tested the status of the mTORC1 pathway in XLCNM models (mouse and human primary myotubes) (Fig. 2f and Extended Data Fig. 2e). We observed that mTORC1 activity is enhanced in CTRL myoblasts, drops during the first phases of differentiation and then increases after 5–6 days. Interestingly, we found that phosphorylation of downstream targets of mTORC1 (p-S6K (Thr389) and p-4E-BP1(Thr37/46)) were increased in the Mtm1-KO myotubes starting from 5–6 days post-differentiation, while the growth factor/insulin and glucose-sensitive phosphorylation of AKT (S473, T308) was not affected (Fig. 2f). These data suggest that mTORC1 activity is tightly regulated during muscle cell differentiation and that loss of function of MTM1 enhances mTORC1 signalling independently of the PI3K–AKT axis.

**Fig. 2: MTM1 loss enhances mTORC1 signalling during myogenic differentiation.**

The activity of mTORC1 is determined by mTOR recruitment and assembly into a multimeric complex on the lysosome surface¹³. To investigate mTORC1–lysosome crosstalk, we used the C2C12 muscle cell line to deplete MTM1 expression by gene editing (CRISPR/Cas9, MTM1-KO^Cas9). As for the Mtm1-KO mouse, patient primary cells or short hairpin RNA (shRNA)-mediated Mtm1 knockdown, MTM1-KO^Cas9 myotubes showed higher mTORC1 activation (an increase of p-S6K-Thr389 and p4E-BP1-T37/46) and protein synthesis rate in line with defective differentiation capacities manifested by a decreased myotube area and fusion index, as well as the number of nuclei per myotube (Fig. 2g and Extended Data Fig. 2f,g). Consistently, this phenocopied muscle fibres from the Mtm1-KO model, which displayed less myonuclei in basal conditions and following muscle injury/regeneration with notexin (Extended Data Fig. 2h,i). Then, we employed the Lyso-Tag strategy using the TMEM192-3_XHA vector²⁰. CTRL and MTM1-KO^Cas9 cells expressing the Lyso-Tag TMEM192-3_XHA showed perfect colocalization of the Lyso-Tag (3× haemagglutinin (HA) epitope) and the lysosome protein Lamp2 (Extended Data Fig. 3a). Lysosome immunoprecipitation (Lyso-IP) analysis of CTRL myotubes showed specific pull-down of lysosomes free from other organelle markers or cytosolic proteins except for the ER marker calnexin (CANX), owing its close contact with the lysosome through membrane contact sites (MCS) (Extended Data Fig. 3b). The lysosomes in the Lyso-Tag–MTM1-KO^Cas9 cell line did not show any morphological (cellular distribution, number and size) or functional (cathepsin activity on whole-cell or Lyso-IP intact lysosomes) abnormalities (Extended Data Fig. 3c,d). Analysis of the Lyso-Tag–MTM1-KO^Cas9 myotubes and XLCNM myotubes showed higher mTOR enrichment on lysosomes even in the low serum differentiation medium (2% horse serum (HS)) (Fig. 2h and Extended Data Fig. 3e). Finally, we confirmed the increased mTOR association to the lysosome in vivo as manifested by increased mTOR–Lamp2 colocalization in Mtm1-KO skeletal muscle fibres (Fig. 2i). Collectively, these findings support the role of MTM1 in repressing mTORC1 activation during cell differentiation and in adult muscle.

MTM1 represses the Ragulator–RagGTPase axis of mTORC1

Lysosomal mTORC1 activation is coordinated mainly in response to amino acids (through the RagGTPase axis) and to growth factors (through inhibition of TSC and activation of Rheb GTPase). Lyso-IP from CTRL myotubes showed a reduction of mTOR lysosomal association in reaction to nutrient depletion (Hank’s Balanced Salt Solution (HBSS)). Differentiation medium (2% HS) leads to slight increase of mTOR lysosomal association, whereas supplementation with insulin growth factor (IGF) or amino acids enhanced the mTOR lysosomal fraction (Extended Data Fig. 4a). Then, we examined the cellular distribution of mTOR in Lyso-Tag–MTM1-KO^Cas9 (Extended Data Fig. 4b,c) and XLCNM myotubes (Extended Data Fig. 4d,e) under nutritional states. Indeed, nutrient depletion partially decreased lysosomal localization of mTOR as well as mTORC1 activity in MTM1-depleted myotubes (Extended Data Fig. 4b–e). By contrast, CTRL/wild-type (WT) cells showed a depletion of mTOR from the lysosome accompanied by a notable drop in mTORC1 activity (Extended Data Fig. 4a–d). Strikingly, unlike in CTRL cells, the addition of amino acids on MTM1-depleted myotubes did not enhance recruitment of mTOR to the lysosome, while IGF1 addition increased mTOR–lysosome colocalization (Extended Data Fig. 4a–d). This is consistent with the previous analysis of the growth factor-sensitive pathway upstream of mTORC1 showing that MTM1 depletion did not affect the function of the PI3K/AKT axis in XLCNM models (Fig. 2f). Collectively, these observations suggest a sustained activation of the amino acid-sensitive axis in MTM1-depleted myotubes.

To investigate the potential molecular connection between MTM1 and the lysosomal RagGTPase–mTORC1 complex, we used comparative proteomics of purified lysosome (Lyso-IP) from CTRL and Lyso-Tag–MTM1-KO^Cas9 samples. We identified a significant enrichment of the RagGTPase complex (RagA and RagC) and its interactor Raptor (Fig. 3a and Supplementary Table 1). In line with this observation, we performed Gene Ontology (GO) analysis of the dataset and found ‘positive regulation of mTORC1 signaling’ as the most prominent GO term (Supplementary Fig. 1). Importantly, the pentameric Ragulator/LAMTOR complex, which anchors the RagGTPase to the lysosome membrane²¹ was highly enriched in MTM1-KO^Cas9 cells in low serum (2% HS) and HBSS conditions (Fig. 3a–c). Comparative cellular or lysosomal metabolomes in CTRL versus MTM1-KO^Cas9 myotubes did not show any differences in amino acids or other known metabolites that activate mTORC1 (Fig. 3d and Supplementary Fig. 2a,b). Also, the knockdown of SLC38A9 (a lysosomal amino acids transporter that activate RagGTPase) in the Mtm1-KO^Cas9 model did not restore mTORC1 activity or improve cell differentiation (Extended Data Fig. 4e,f). These data suggest that the activation of the RagGTPase axis is probably independent of nutrient/metabolite fluctuations in XLCNM models. To gain mechanistic insights about the sustained RagGTPase activity in MTM1-depleted cells, we first confirmed the increased lysosomal association of RagA and LAMTOR1 (a key member of the LAMTOR complex) in the lysosome in Lyso-Tag–MTM1-KO^Cas9 myotubes as well as in Mtm1-KO skeletal muscle fibres (Extended Data Fig. 5a–c). We then modulated RagA expression/activity in MTM1-KO^Cas9 cells. The knockdown of RagA (shRNA) or ectopic expression of the RagA-GTPase dominant-negative mutants (RagA-T21N and RagA-R37P) in MTM1-KO^Cas9 cells decreased mTORC1 signalling and restored muscle cell differentiation (Fig. 3e,f). By contrast, the expression of the constitutively active mutant (RagA-Q66L) worsened the phenotype (Fig. 3e,f). To gain insights about the activity of the RagGTPase complex in the MTM1-KO^Cas9 model we first employed GTP pull-down assays to pull active RagA (GTP-bound) to confirm the reliability of the assay in HEK393 cells infected with RagA-T21N and RagA-Q66L mutants (Extended Data Fig. 5d). Indeed, the RagA-T21N mutation strongly decreased GTPase binding, whereas the RagA-Q66L showed the reverse effect. Furthermore, in line with inter-subunit exchange of GTP/GDP between RagA/B and RagC/D¹⁴, RagC was found in a GTP-bound state when RagA-T21N was expressed. By contrast, RagA-Q66L expression drastically decreased its GTP-binding capacity (Extended Data Fig. 5d). Then, we tested RagA and RagC GTP binding profiles in muscle cells and found that the RagA-GTP bound form was increased in MTM1-KO^Cas9 myotube lysates, while RagC showed the reverse profile (Extended Data Fig. 5e). Second, we tested TFEB (a transcription factor phosphorylated by mTORC1²²) subcellular localization, which has been suggested to be affected by RagGTPase activity²³. Active RagGTPase heterodimer retains TFEB at the lysosome via a direct binding²³. Our data indicate that in differentiation medium (2% HS), CTRL myotubes showed cytosolic and nuclei TFEB distribution with very few TFEB fractions associated with the lysosome (TFEB–Lamp2 colocalization). However, MTM1-KO^Cas9 myotubes showed fewer TFEB-positive nuclei but higher lysosomal localization (Extended Data Fig. 5f). This cellular distribution was conserved in starved conditions (HBSS) and when treated with torin (or both) (Extended Data Fig. 5f), suggesting that TFEB is retained at the lysosomal surface in MTM1-KO^Cas9 cells, and further demonstrating a sustained activation of RagGTPase in MTM1-depleted myotubes. Finally, as RagGTPases are recruited at the lysosome by the LAMTOR complex²¹ we investigated the involvement of LAMTORs in our disease model. Indeed, the knockdown (shRNAs) of LAMTOR1 decreased mTORC1 signalling and improved MTM1-KO^Cas9 cell differentiation (Fig. 3g,h). Taken together, these data indicate that MTM1 limits the recruitment and activity of the RagGTPase complex, possibly by modulating lysosomal LAMTOR/Ragulator.

**Fig. 3: MTM1 represses the Ragulator–RagGTPase amino acid-sensing arm of mTORC1.**

MTM1 controls a lysosomal phosphoinositide pool that modulates the RagGTPase–mTORC1 axis

MTM1 is a phosphoinositide-3-phosphatase that hydrolyses PI3P and PI(3,5)P2 (Fig. 4a), two phosphoinositides that are present at the lysosome membranes and possibly modulated by PI-kinases²⁴. Thus, we hypothesized that MTM1 PI-phosphatase function and RagGTPase–mTORC1 activity are coupled in muscle cell differentiation. We tested this hypothesis in the MTM1-depleted cell models (Mtm1-KO primary and MTM1-KO^Cas9 myotubes) and in vivo in the Mtm1-KO mice (Fig. 4a and Extended Data Fig. 6a). The latter were subjected to skeletal muscle injury/regeneration, a known approach to trigger de novo myofibre formation (Extended Data Fig. 6a). Notably, ectopic expression of the MTM1 dead-phosphatase mutant (MTM1-C375S) in the MTM1-KO^Cas9 myotubes (Fig. 4b,c), in primary Mtm1-KO myotubes (Extended Data Fig. 6b,c) or in Mtm1-KO skeletal muscle (adeno-associated virus (AAV)-mediated expression) (Extended Data Fig. 6d,e) did not restore global protein translation, mTORC1 overactivation or muscle differentiation. Importantly, the lysosomal levels of the LAMTOR–RagGTPase complex were also dependent of MTM1 lipid phosphatase activity as, in contrast to the MTM1 active enzyme (MTM1-FL), MTM1-C375S expression MTM1-KO^Cas9 cells failed to decrease the lysosomal level of LAMTOR1–5 (LT1–LT5) (Fig. 4d). Then, we tested the hypothesis that MTM1 could control a lysosomal pool of PI3P and PI(3,5)P2 that favour LAMTOR levels and the retention of RagGTPase at the lysosome. We first addressed the localization of these lipids in Lyso-Tag MTM1-KO^Cas9 myotubes using protein domain probes: the 2_XFyve domain of hepatocyte growth factor-regulated tyrosine kinase substrate (HRS) to detect PI3P and SnxA to detect PI(3,5)P2²⁵, as well as the anti-PI(3,5)P2 antibody that we characterized using the PIKFyve inhibitor Apilimod (Apilm) (Extended Data Fig. 6f,g). Of note, PIKFyve is a type III PI5-kinase that produces the major cellular pool of PI(3,5)P2²⁶. These localization data indicated a higher enrichment of PI3P and PI(3,5)P2 on the lysosomal membrane in Lyso-Tag–MTM1-KO^Cas9 myotubes. To validate these findings, we employed mass spectrometry to measure and quantify the entire PI(P)s-ome²⁷: phosphoinositides (PIPx) and phosphatidylInositol (PI) in whole myotube homogenates, as well as for the first time on purified organelles (Lyso-IP) (Fig. 4e,f and Supplementary Fig. 3). Indeed, the total cell cartography of all phosphoinositides showed that the loss of MTM1 impacted the level of its substrates (PI3P and PI(3,5)P2) and products (PI and PI5P) but not the other PIPx (Fig. 4e,f and Supplementary Fig. 3a–c). Particularly, we detected a higher level of PI3P isomer and PI(3,5)P2 isomer with 38:3 fatty acid composition (stearic/dihomolinolenic acid with three unsaturations) accompanied by a drop in PI and PI5P isomers (38:3-PI and 38:3-PI5P), suggesting that MTM1 has a preferential hydrolysis function towards this molecular species (Fig. 4f and Supplementary Fig. 3a–c). Then, we applied the same strategy to measure PIPx on Lyso-IP samples. Except for PI and PI(3,4,5)P3, which were not detected in Lyso-IP samples, we detected the presence of PI(4,5)P2, PI(3,5)P2, PI(3,4)P2, PI3P, PI4P and PI5P (Fig. 4e,f and Supplementary Fig. 3d,e). Consistent with whole-cell analysis, MTM1-KO^Cas9 lysosomes showed a higher level of 38:3 isomers (the major isomer found in lysosomal PIPx) of PI3P and PI(3,5)P2, accompanied by a drop in 38:3 PI5P isomer, while the level of the other PIPx was not affected (Fig. 4f and Supplementary Fig. 3d–f).

Fig. 4: MTM1 regulates RagGTPase–mTORC1 signalling through lysosomal PI3P and PI(3,5)P2 pools. — **Fig. 4: MTM1 regulates RagGTPase–mTORC1 signalling through lysosomal PI3P and PI(3,5)P₂ pools.**

To investigate the link between MTM1-dependent lysosomal PI3P and PI(3,5)P2 production and LAMTOR–RagGTPase, we modulated PI-kinases that produces these phosphoinositides (Fig. 4a). Notably, the Lyso-IP proteomic data showed the presence of the class III PI3-kinase (Vps34/PI3KC3) and the class II PI3-kinase (PI3KC2β) and PIKfyve (Supplementary Table 1). Then, we used the compound IN1 to inhibit Vps34, shRNAs to knockdown PI3KC2β and Apilm to inhibit PIKfyve. Vps34 inhibition did not restore mTORC1 activity or muscle differentiation in MTM1-KO^Cas9 cells (Extended Data Fig. 7a,b). By contrast, genetic inhibition of PI3KC2β decreased mTORC1 signalling and improved muscle cell differentiation (fusion index and myotube area) in MTM1-KO^Cas9 myotubes (Fig. 4g and Extended Data Fig. 7c,d). In comparison to PI3KC2β, PIKfyve inhibition had a mild effect on mTORC1 activity and the myotube phenotype (Fig. 4h and Extended Data Fig. 7c,d). Consistently, mass spectrometry analysis of lysosomal PIPx showed that PI3KC2β inhibition restored 38:3-PI3P and 38:3-PI(3,5)P2 isomer levels in MTM1-KO^Cas9 myotubes, while PIKfyve inhibition only dropped 38:3-PI(3,5)P2 and 38:3-PI5P but increased 38:3-PI3P (Fig. 4i and Extended Data Fig. 7e). Importantly, genetic inhibition of PI3KC2β, and to a lesser degree the inhibition of PIKFyve activity, was able to decrease lysosomal levels of LAMTOR1–5, RagA and Raptor (Fig. 4j). Taken together, our data support that MTM1, PI3KC2β and PIKFyve orchestrate a phosphoinositide conversion cascade to control a lysosomal pool of PI3P and PI(3,5)P2 that are required for fine-tuning RagGTPase–mTORC1 signalling.

MTM1 lipid substrates PI3P and PI(3,5)P2 binds LAMTOR1/3 and modulate complex turnover

A major function of phosphoinositides is to recruit and bind protein effectors to trigger signalling events²⁴. Our data indicate that the LAMTOR complex was highly enriched on MTM1-KO^Cas9 lysosomes in parallel to a retention of RagGTPase activity. Thus, we asked whether there is a physical connection between LAMTOR–RagGTPase and the MTM1 substrates PI3P and PI(3,5)P2 (Fig. 5a). We tested LT1–LT5 and RagA/C binding to phosphoinositides using three independent experimental approaches and detected specific binding of LT1 to PI3P and LT3 to both PI3P and PI(3,5)P2 (Fig. 5b and Extended Data Fig. 8a–c). Phosphoinositides have preferential binding affinity towards positively charged amino acids (Lys/K, Arg/R, His/H), thus, we explored structural models of LAMTOR–RagA/C complex²¹ to identify potential LT1 and LT3 binding sites. We focused on the LAMTOR complex interface with the lysosomal membrane and tested the contribution of binding with PIPx of two conserved lysine residues (K103 and K104) on LT1, as well as two conserved lysine residues (K6 and K34) and the Arg residue (R7) on LT3 (Fig. 5c and Extended Data Fig. 8d). Substitution of these amino acids to alanine/A (single and double) (within the LAMTOR recombinant tetrameric complex) alters LT1 and LT3 binding to PI3P and PI(3,5)P2, respectively (Fig. 5c and Extended Data Fig. 8e). In comparison with wild-type, LT1 and LT3 mutations did not affect LAMTOR complex assembly when overexpressed in HEK293 cells (Extended Data Fig. 8f). However, we noted that LT1 and LT3 mutations had dominant-negative effects towards endogenous LAMTORs, manifested by a decrease of RagA recruitment (Extended Data Fig. 8f). These data suggest that LAMTOR binding to PI3P and/or PI(3,5)P2 contributes to retaining the LAMTOR–RagGTPase complex on the lysosome. Then, we tested the effect of ectopic expression of LT1 and LT3 mutants in MTM1-KO^Cas9 cells with the knockdown of LT1 (LT1-shRNA 3′UTR). Indeed, LT1-K103/K104, LT3-K6/R7/K34 and double mutant constructs showed notable mis-localization from lysosomes, indicating that LT1/LT3–phosphoinositide interactions contribute to maintain the integrity of lysosomal LAMTOR complexes (Fig. 5d). Interestingly, mutations in LT1 and LT3 were able to restore mTORC1 signalling (Fig. 5e) and cell differentiation capacities in MTM1-KO^Cas9 myotubes (Fig. 5f).

Fig. 5: The LAMTOR complex binds PI3P and PI(3,5)P2 through LAMTOR1 and LAMTOR3. — **Fig. 5: The LAMTOR complex binds PI3P and PI(3,5)P₂ through LAMTOR1 and LAMTOR3.**

LAMTOR protein degradation/turnover is controlled by the ubiquitin (Ub) proteasome system which impacts mTORC1 activity²⁸. We therefore tested the hypothesis that LAMTOR retention by PI3P and/or PI(3,5)P2 could affect their protein turnover. Indeed, pulse–chase assays following cycloheximide (CHX) treatment showed accumulation of the LAMTOR complex in MTM1-KO^Cas9 myotubes (Fig. 5g), and this was restored by re-expression of LT1/LT3 phosphoinositide binding mutants (Fig. 5h). However, RagA turnover was slightly affected by LT1/LT3 mutants, possibly through the LAMTOR complex and not a direct binding to phosphoinositides PI3P and/or PI(3,5)P2 (Extended Data Fig. 8b). In line with altered protein turnover, LT1 and LT3 ubiquitination was decreased in MTM1-KO^Cas9 cells, similarly to CTRL/WT myotubes treated with MLN4924, a pan inhibitor of ubiquitination cascades (Fig. 5i). Furthermore, disruption of LAMTOR turnover in CTRL myotubes using MLN4924 promoted LT1 and LT3 lysosomal retention/accumulation similar to the MTM1-KO^Cas9 model (Extended Data Fig. 8g). Collectively, these data demonstrate that PI3P and PI(3,5)P2 function in a lysosomal pathway by anchoring and controlling LAMTOR complex turnover and availability, therefore modulating RagGTPase requirements during cell differentiation.

MTM1 controls PI3P and PI(3,5)P2 levels through ER–lysosome membrane contacts

To address how MTM1 acts on lysosomal PI3P and PI(3,5)P2 to modulate LAMTOR–RagGTPase, we first forced MTM1 to the lysosome (Lyso-Tag fusion protein TMEM192-MTM1-FL and MTM1-C375S inactive mutant) in MTM1-KO^Cas9 cells (Fig. 6a,b). The Lyso-Tag–MTM1-WT expression normalized lysosomal PI3P and PI(3,5)P2 levels (Extended Data Fig. 9a,c), restored mTORC1 signalling and localization and improved the myotube phenotypes, while the Lyso-Tag–MTM1-C375S mutant failed to correct these parameters (Fig. 6b–d). Previous immunodetection and cell fractionation experiments showed an ER/sarcoplasmic reticulum distribution of MTM1 in skeletal muscle²⁹, and given the established role of ER–lysosome MCS in lysosomal homeostasis³⁰, we then tested a possible implication of MTM1 in ER–lysosome MCS to regulate lysosomal PI3P and PI(3,5)P2 and mTORC1 signalling. We tethered MTM1 (MTM1-FL and MTM1-C375S inactive phosphatase) to the ER membrane by fusing it with a MOSPD2 transmembrane domain (Fig. 6e,f). MOSPD2 is a transmembrane ER protein enriched in ER–organelle MCS³¹. Targeting active MTM1 (MTM1-FL) but not MTM1-C375S to ER-MCS returned the level of lysosomal PI3P and PI(3,5)P2 to normal (Extended Data Fig. 9b,c) and restored mTORC1 signalling (Fig. 6g) and the differentiation capacity of MTM1-KO^Cas9 cells (Fig. 6h), suggesting that MTM1 controls the level of lysosomal PI3P and PI(3,5)P2 through ER–lysosome MCS. Consistently, analysis of MCS using the Lyso-Tag antibody (HA) and the ER protein Nogo/RTN4 supported by a proximity ligation assay (PLA) showed a sharp decrease of these membrane contacts in MTM1-KO^Cas9 myotubes (Extended Data Fig. 9d,e). To gain molecular insights in the implication of MTM1-mediated regulation of RagGTPase–mTORC1 through ER–lysosome membrane contacts, we disrupted MCS in CTRL and MTM1-KO^Cas9 myotubes using OSW1. This compound inhibits the oxysterol-binding protein (OSBP) and its paralogue OSBP-related protein 4 (ORP4), two important proteins for MCS maintenance³². In contrast to proliferative cells³³, myotube treatment with OSW1 leads to increased lysosomal levels of mTOR, Raptor, and the LAMTOR–RagA complex concomitantly to increased mTORC1 activity (pS6K/S6K) and lysosomal PI3P and PI(3,5)P2 levels (Fig. 6i,j). By contrast, OSW1 did not have any obvious effect effects on MTM1-KO^Cas9 myotubes (Fig. 6i,j). Consistently, OSW1 led to defective cell differentiation in CTRL myotubes (Fig. 6k). Supporting the role of MTM1 in the regulation of mTORC1 through lysosome–ER MCS, mTOR, OSBP and VAPA/B proteins were co-immunoprecipitated with MTM1 (Extended Data Fig. 9f). This interaction was disrupted by OSW1 or tunicamycin (Tun) treatments (a compound that also disrupts ER–lysosome MCS) (Extended Data Fig. 9f).

**Fig. 6: MTM1 regulates lysosomal phosphoinositides and mTORC1 through ER–lysosome MCS.**

mTORC1 operates both upstream and downstream of ER stress signals, which can either enhance or antagonise the anabolic output of mTORC1³⁴. In addition, Mtm1-KO muscles suffer from sustained ER stress and unfolded protein response activation, manifested by increased phosphorylation of eiF2α (Ser51), IRE1α (Ser720) and PERK (Thr981). Then, we tested if MTM1 could trigger mTORC1 repression in response to ER stress. We confirmed sustained ER stress activation in MTM1-KO^Cas9 myotubes compared with CTRL cells (Extended Data Fig. 9g). In line with the important role of physiological ER stress in myogenesis³⁵, we noted a basal activation of these axes in myotubes starting from day 4 (Extended Data Fig. 9g), which overlapped with higher MTM1 expression and repression of mTORC1 activity (Fig. 2f). Indeed, induction of ER stress with tunicamycin repressed mTORC1 in CTRL cells but not in MTM1-KO^Cas9 or in cells expressing the MTM1-C375S inactive mutant (Extended Data Fig. 9h). Altogether, our results support a role of MTM1 and lysosomal PI3P and PI(3,5)P2 in ER–lysosome crosstalk to modulate RagGTPase–mTORC1 signalling in response to physiological ER stress during muscle cell differentiation.

Inhibition of mTORC1 in the XLCNM mouse model improves disease onset and progression

Mtm1-KO mice develop progressive muscle atrophy beginning at 3–4 weeks post-birth, it progresses to marked weakness at 5–6 weeks, leading to their death at 8–10 weeks of age. Then, we explored the physiological relevance of mTORC1 inhibition in the Mtm1-KO mice model at 2 weeks of age and when mice start to suffer from remarkable muscle atrophy (5 weeks of age). We first analysed mTORC1 signalling in skeletal muscle from 2- and 6-week old Mtm1-KO mice. In line with previous data in another mouse model of Mtm1 loss of function³⁶ and with our cellular data on myotubes, we detected an increase of mTORC1 phospho-targets (pS6K-T389 and p4E-BP1 (T37/36)), while p-AKT (S473 and T308) was not affected (Fig. 7a). This is consistent with our cellular data showing that the growth factor signalling axis of mTORC1 is probably not affected by MTM1 loss of function. Moreover, blood glucose, insulin and IGF1 levels were similar in control and Mtm1-KO mice in fed or fasted conditions (Extended Data Fig. 10a). Then we treated 2- and 5-week-old mice with daily doses (single injection for 20 days) of the ATP-competitive mTORC1 inhibitor AZD8055 (Fig. 7b and Extended Data Fig. 10b). Remarkably, in both experimental setups, treatments extended the lifespan of Mtm1-KO-treated mice with better physiological benefits when treatment was started at 2-weeks of age (Fig. 7c). This was accompanied by improved body weight (Fig. 7d and Extended Data Fig. 10c) as well as muscle mass (tibialis anterior, TA) and function (specific muscle force) (Fig. 7e and Extended Data Fig. 10d). The efficacy of mTORC1 inhibition in TA muscle homogenates was confirmed by the decrease of the mTORC1 downstream target p-S6K-T389 (Fig. 7f). Analyses of muscle fibre morphology and ultrastructure showed increased fibre size in treated Mtm1-KO mice, accompanied by enhanced and uniform oxidative capacity (succinate dehydrogenase (SDH) staining) and improved global myofibre organization (Fig. 7g,h and Extended Data Fig. 10e,f). To address potential side effects of mTORC1 inhibition on healthy muscle, we tested AZD8055 treatment on wild-type mice. Neither muscle mass, nor force and morphology were affected by AZD8055 (Extended Data Fig. 10g–i). However, as expected, inhibition of mTORC1 promoted the formation of some vacuoles reminiscent of autophagosomes (Extended Data Fig. 10h). These findings support the physiological benefit of the inhibition of mTORC1 in the XLCNM mouse model by restoring the anabolic–catabolic balance during muscle growth and maintenance.

Fig. 7: mTORC1 inhibition improves muscle function and extends survival in Mtm1-KO mice. — **Fig. 7: mTORC1 inhibition improves muscle function and extends survival in *Mtm1*-KO mice.**

Discussion

We demonstrate that two phosphoinositides, PI3P and PI(3,5)P2, act upstream of the RagGTPase–mTORC1 complex in post-mitotic cells during muscle differentiation. The turnover of these lipids is controlled by the PI3P and PI(3,5)P2 conversion cascades involving the PI3-phosphtatase MTM1 (an enzyme inactivated in congenital myopathy), the class II PI3KC2β and the PI5-kinase PIKFyve (Fig. 7i). Phosphoinositides can control cell metabolic status at different levels, acting as a stimulator or repressor of mTORC1 signalling depending on their cellular localization. Indeed, PI(3,4,5)P3 or PI(3,4)P2 production at the plasma membrane acts upstream of the growth factor-sensing axis of mTORC1 (PI3K/AKT/TSC1/2), thereby favouring anabolism³⁷, while it has been suggested that specific pools of PI(3,4)P2 or PI(3,5)P2 could be mobilized to repress mTORC1 in proliferative cells under growth factor deprivation^38,39. Moreover, additional stress conditions can mobilize other phosphoinositides such as PI4P to preserve lysosome homeostasis, including channel function, lysosome membrane repair and catabolic status^33,40,41. Our data indicate that in post-mitotic differentiated cells, PI3KC2β controls a lysosomal pool of PI3P that activates mTORC1. This was demonstrated by the decreased lysosomal PI3P level in line with restored mTORC1 signalling in MTM1-depleted cells when PI3KC2β was silenced. In agreement, muscle from mice obtained after a genetic cross between Mtm1-KO and PI3KC2β-D1212A (mice carrying a kinase-dead mutant of PI3KC2β) showed normalization of p-S6K-T389 and PI3P levels (but not PI(3,4)P2) in parallel to improvement of Mtm1-KO mice survival and muscle function⁴². Moreover, primary hepatocyte from PI3KC2β-D1212A mice showed a 60% reduction of PI3P, and PI3KC2β knockdown in human cerebral microvascular endothelial cells dropped the PI3P level without impacting PI(3,4)P2⁴³. As with PI3KC2β-KO mice, PI3KC2β-D1212A mice were as viable and healthy as the CTRL littermates⁴⁴. Notably, these mice were protected from stock injury⁴³ and metabolic stress (enhanced insulin sensitivity, glucose tolerance and protection against high-fat-diet-induced liver steatosis)⁴⁵. Hence, PI3KC2β may trigger mTORC1 signalling in distinct ways depending on cellular context, possibly through acute production of PI(3,4)P2 to repress mTORC1 during growth factor deprivation and through the production of a basal pool of PI3P at the lysosome to activate/sustain mTORC1 during nutrient availability.

In line with the key role of PI3P in mTORC1 signalling, recent evidence in proliferative cells supports that PI3P-positive lysosomes are motile, peripheral and anabolic (with active mTORC1 signalling)³³. This PI3P pool was produced by the class III PI3K (Vps34), under amino acid stimulation, at lysosomes at the cell periphery to recruit the P3P-binding protein FYCO1 and facilitate ER–lysosome contacts⁴⁶. Our data indicate that inhibiting Vps34 in Mtm1-KO myotubes did not restore mTORC1 signalling and worsen the cell phenotype, suggesting that MTM1 and Vps34 did not act on the same lysosomal PI3P pool. However, a possible crosstalk of MTM1 and the Vps15/Vps34 complex in rebalancing a non-lysosomal pool of PI3P could be also emphasized. Our observations favour a model in which MTM1 and PI3KC2β control the balance of the PI3P-activating lysosomal pool of mTORC1 in post-mitotic cells, while the PI3P-dependent pool produced by Vps34 may drive lysosome positioning and other process such as receptor trafficking and macroautophagy. Supporting this, we noted that the loss of function of MTM1 and subsequent PI3P accumulation on the lysosome did not affect lysosome positioning, number or degradative capacity. Notably, the relationship between lysosome localization and mTOR activity is still a matter of contradicting reports. Walton et al. showed that cell acidification drives peripheral redistribution of normally perinuclear lysosomes away from perinuclear Rheb, thereby inhibiting the activity of lysosome-bound mTOR⁴⁷. This is consistent with prior studies showing that perinuclear clustering of lysosomes support mTORC1 signalling⁴⁸. However, later studies supported the idea that the activation of mTORC1 by nutrients correlates with its presence on peripheral lysosomes, whereas starvation causes perinuclear clustering of lysosomes^33,49. Adding a level of complexity in mTORC1–lysosome crosstalk, an ensemble of recent data pointed to the key role of RagGTPase in the retention of mTORC1 at the lysosome. The data also suggested a spatial separation of mTORC1 activities in cells, with lysosomal and non-lysosomal complexes regulating different cellular functions through the phosphorylation of distinct targets at each subcellular location⁵⁰. Moreover, it was suggested that the release of mTORC1 from lysosomes depends on mTOR kinase activity rather that nutrient fluctuations.

Our data indicate that a fraction of the PI3P pool is converted to PI(3,5)P2 by the PI5-Kinase PIKFyve. In comparison with PI3KC2β, PIkFyve inhibition with Apilm showed a partial rescue of RagGTPase–mTORC1 signalling and cell phenotype in MTM1-KO myotubes. Furthermore, targeting PIKFyve and PI3KC2β did not show additive effects, supporting the idea that both enzymes are involved in the same PI3P–PI(3,5)P2 conversion cascade at the lysosome. Notably, we noted that Apilm increased PI3P lysosome levels, which restrained the complete recovery of mTORC1 signalling and cell phenotype in MTM1-KO myotubes. While the role of PIKFyve and its lipid product PI(3,5)P2 in mTORC1 signalling is still a matter of debate⁵¹, our data support that the pool of PI(3,5)P2 controlled by MTM1 and PIKFyve have an effective role (in supporting PI3P) to retain the lysosomal LAMOR–RagGTPase complex and favour mTORC1 activity. Dissecting more precisely the contribution of PI(3,5)P2 in mTORC1 signalling requires more suitable tools and remains a fruitful area for future studies.

Our data support that PI3P and PI(3,5)P2 positively impact RagGTPase activity through LAMTOR/Ragulator. However, the exact mechanism of this activation will need further molecular investigations. Both SLC38A9 and Ragulator were proposed to have GEF-like activity towards RagGTPase⁵². The inhibition of the GEF-like protein SLC38A9 did not show any obvious effect on mTORC1 activity and cell phenotype in Mtm1-KO myotubes, therefore favouring a model where PI3P and PI(3,5)P2 binding to LAMTORs could stimulate Ragulator GEF-like activity towards RagGTPase. Today, there is no available experimental strategy to address Ragulator–GEF activity in cellulo, but future in vitro approaches using reconstituted recombinant LAMTOR–RagGTPase complexes in the presence of phosphoinositides could be used to support our model.

We showed that LAMTOR3 binds mostly PI(3,5)P2, and to a lesser extent PI3P, whereas LAMTOR1 preferentially binds PI3P. This supports a new layer of regulation by phosphoinositides of the LAMTOR complex on the lysosome. Notably, it was proposed that LAMTOR1 anchors the entire LAMTOR complex through on the lysosome membrane through its N-myristoylation and palmitoylation⁵³. The N-myristoyltransferase-1 (NMT1) was proposed to trigger LAMTOR1 myristoylation in cancer cells⁵³, while the cascade and the involved enzymes in the palmitoylation of LAMTOR1 has not been investigated so far. We used recombinant LAMTOR1–5–RagA/C complexes, where LAMTOR1 was depleted of its N-terminal disordered section, suggesting that LAMTOR binding to phosphoinositide is probably independent of these lipid modifications on LAMTOR1. We also showed that LAMTOR retention by lysosomal phosphoinositides impacted their protein turnover by the Ub proteasome system. Indeed, several members of mTORC1, including the LAMTOR complex, are ubiquitinated and degraded to shut down mTORC1 activity²⁸. The output physiological signals are still not fully understood but stressors such as ER stress and the activation of Ub ligases in these conditions could play an important role in modulating mTORC1 components to rebalance anabolic–catabolic signals under stress. Indeed, it was suggested that mTORC1 could operate both upstream and downstream of ER stress signals, which can either enhance or antagonise the anabolic output of mTORC1³⁴. Consistently, we found that during differentiation muscle cells repress mTORC1 in response to ER stress, while the absence of MTM1 interferes with such inhibition. Given the role of ER physiological stress and that mTORC1 activity is modulated during myogenesis, our data support a model in which physiological ER stress triggers mTORC1 activity through MTM1 to ensure the anabolic–catabolic equilibrium required for muscle cell differentiation and maturation. Supporting this model, we demonstrated that MTM1 controls lysosomal PI3P and PI(3,5)P2 through ER–lysosome MCS. Indeed, targeting MTM1 either to the lysosome or to the ER reinstated physiological levels of mTORC1 signalling and restored muscle cell phenotypes in the XLCNM model. This suggests a dynamic phosphoinositide turnover between ER–lysosome membranes. In line with this, our data indicate that MTM1 could interacts with MCS proteins such as the VAPA/B complex and OSBP proteins as well as mTOR. Such associations were abrogated when cells were treated with the MCS inhibitor OSW1. Consistently, loss of function of MTM1 disrupted ER–lysosome MCS, suggesting that MTM1 and/or its lipids are required for functional interconnection between both compartments, at least to repress RagGTPase recruitment and to decrease or turn off mTORC1 signalling. Additional lipids such as cholesterol have been shown to activate mTORC1 through RagGTPase, a mechanism involving ER–lysosome MCS³⁰, thus indicating that the integrity of this compartment is a prerequisite for mTORC1 signalling repression or activation, possibly depending on cell type as well as physiological input. Differentiated muscle cells have an extensive and specialized endo/sarcoplasmic reticulum network, which suggests the presence of specialized ER/SR lysosome MCS to control metabolic fluxes and downstream signalling cascades. Further cellular and molecular investigations are required to better characterize these MCS and their role in organelle metabolic crosstalk in muscle cells.

The efficacy of the mTORC1 inhibitor in the Mtm1-KO mouse model was demonstrated in our study. Although key steps for clinical translation remain yet to be accomplished (for example, tests on the XLCNM dog model and the long-term effects of the drug on muscle and other organ physiology), it is encouraging that lifespan expansion correlated with improvements in muscle physiology and structure even when Mtm1-KO mice were treated at the late stage of disease onset. Thus, our preclinical observations have the potential to expand therapy options for patients with XLCNM. Finally, increasing evidences implicate a decline in lysosomal integrity and function in many diseases, including those involving PI-metabolizing enzymes (neurogenerative, cancer and metabolic diseases)¹. Proteomic and metabolomic profiling of lysosomes coupled with quantitative phosphoinositides, which we employed in this work, provide a framework for future integrative studies that will test the impact of lysosome composition in physiology and human disease mechanisms.

Methods

Muscle primary cells (human and mouse)

XLCNM (1 and 2) primary muscle cells and CTRLs were obtained from the French muscle dystrophy association (Association Contre la Myopathie) myobank after clinical and genetic diagnosis. XLCNM1 and 2 carry EX5del and c.141–144delAGAA, respectively, leading to complete loss of MTM1 expression as confirmed by western blot analysis. XLCNM and CTRL myoblast were maintained in proliferation medium: Dulbecco’s medium supplemented with 20% foetal calf serum and antibiotics (0.1% penicillin–streptomycin) until reaching 60–70% confluency. Medium was replaced every 2 days. The induction of cell differentiation to myotubes was performed by adding the differentiation medium (Dulbecco’s medium, 2% horse serum and 0.1% penicillin–streptomycin). For mouse primary cells, primary muscle cells (satellite cells) were isolated from mouse pup (2–3 weeks old) hindlimbs. Cells were maintained in proliferation medium for 3–4 days to reach 60–70% confluency, and then shifted to a differentiation medium for 9–10 days. For lentiviral expression of MTM1-WT and MTM1-C375S constructs, Mtm1-KO primary cell myoblasts (60–70% confluency) were infected with the indicated lentiviral preparation and maintained in proliferation medium for 48 h. Myoblasts expressing the construct of interest were selected by adding 2 μg ml⁻¹ puromycin (Sigma) for 24 h. Differentiated medium (supplemented with 1 μg ml⁻¹ puromycin) was added to induce myotube formation. For ongoing culture, the medium was replaced every 2 days.

Cell lines

The C2C12 mouse muscle cell line (NCBI Biosample: SAMEA2224966) was cultured in the same proliferation (myoblast) and differentiation (myotube) media. Muscle cells used in this study were analysed at the indicated stage through the manuscript (myoblasts after 3–9 days of differentiation). The MTM1-KO^Cas9 cell line was generated and validated previously⁵⁴. In brief, the CRISPOR webtool was used to design gRNAs targeting Mtm1. The C2C12 line was electroporated with Cas9–sgRNA ribonucleoprotein targeting exon 8 of the Mtm1 gene using the Neon transfection system (Life Technologies). Then, 12.5 pmol of recombinant Cas9 (the encoding plasmid was a gift from Jennifer Doudna, University of California, Berkeley/HHMI) was incubated with 75 pmol of sgRNA synthesized using the EnGen sgRNA Synthesis Kit (New England Biolabs) from the template oligonucleotide (TAATACGACTCACTATAGGTCATCATCTGAGGTCCGATAGTTTTAGAGCTAGAAATAGCAAG) and purified with the RNA Clean & Concentrator-25 Kit (Zymo research). The sgRNA–Cas9 ribonucleoprotein complex was incubated for 10 min at room temperature then mixed and electroporated (1,650 V, 10 ms, three pulses) into 100,000 cells. After cell plating in proliferation medium for 24 h, cells were detached (accutase, BD biosciences) and pelleted. Part of the cell pellet was tested for genomic editing and the remaining part was plated in 96-well plates to obtain a single cell per well (fluorescence-activated cell sorting). Genomic DNA was isolated using the GenElute Mammalian Genomic DNA Miniprep Kit (Sigma–Aldrich) and a 638-bp region surrounding the predicted site of Cas9 activity was amplified. Editing efficiency was verified using the T7 endonuclease assay according to the manufacturer’s instructions (New England Biolabs). The 638-bp region was analysed in the selected clones using the Tracking of Indels by Decomposition webtool to identify clones with frameshift mutations. The isolated positive clones were further characterized by sequencing predictive off-targets (using the CRISPOR prediction tool) and the absence of MTM1 protein in comparison with shRNA-Mtm1 expressing C2C12 cells. For cell starvation, myotubes were cultured with HBSS (without amino acids, glucose and growth factors) for 4 h or stimulated with HBSS medium containing a 2× amino acid cocktail without L-glutamine (50× stock solution, Sigma) or with 40 ng ml⁻¹ rat IGF1 (ThermoFisher scientific) for 4 h.

Finally, for LAMTOR mutation expression, human embryonic kidney HEK-293T cells were used. First, 50–60% confluent HEK-293T cells in 10-cm dishes were transfected with the indicated cDNA mixed with Lipofectamine 2000 reagent (ThermoFisher Scientific) and Opti-pep medium (Sigma) according to the manufacturer’s recommendations. After 24 h, cells were washed twice with phosphate-buffered saline (PBS) and cultured for 36 h with Dulbecco’s medium supplemented with 10% foetal calf serum and antibiotics (0.1% penicillin–streptomycin). When at 80–90% confluence, cells were rinsed twice with PBS and collected for further analysis (IP and/or western blot). All cells were validated to be free of mycoplasma contamination before use.

siRNA- and shRNA-mediated gene knockdown

For congenital myopathy gene silencing, the self-penetrating Accell siRNA (customized SMARTpool library, Dharmacon) was used according to manufacturer’s recommendation. Control siRNA was used (two scrambled siRNAs and one siRNA targeting Gapdh (A-040917-14-0005, Dharmacon)). The efficacy of gene silencing was verified by reverse transcription quantitative PCR using 96-well plates (in triplicate) containing custom primer couples for the 36 analysed genes. Two paris of primers for the housekeeping genes Hprt1 and Rplp0 were used for normalization. Reverse transcription quantitative PCR was performed according to the routine protocol of RNA extraction and reverse transcription with the iScript Reverse Transcription Supermix (Bio-Rad) and used at a 1:100 dilution for qPCR, followed by a qPCR reaction using the iTaq SYBR Green Supermix (Bio-Rad). The analysis was performed using on Light-Cycler 480 instruments (Roche). Experiments were performed in triplicate independent experiments.

For shRNA gene knockdown, pLKO.1-puro vectors encoding shRNA (MISSION shRNA) were obtained from Sigma (validated shRNA). Vector particles were produced and C2C12 cells were transduced by incubation with lentiviral particles for 48 h in proliferation medium in the presence of 2 μg ml⁻¹ puromycin (Sigma) for cell selection. After reaching 60–70% confluency, transduced myoblasts were differentiated as previously described with differentiation medium in the presence of 1 μg ml⁻¹ puromycin. MISSION mouse shRNAs (Sigma) used in this study: Raga (NM_178376, shRNA1-CDS: TRCN0000305245 and shRNA2-CDS: TRCN0000305246), Slc38A9 (NM_178746, shRNA1-CDS: TRCN0000101971 and shRNA2-CDS: TRCN0000101974), Lamtor1 (NM_025605, shRNA1-CDS: TRCN0000121343, shRNA2-CDS: TRCN0000121346 and shRNA-3′UTR: TRCN0000167845), Pik3c2b (NM_001099276, shRNA1-CDS: TRCN0000360942 and shRNA2-CDS: TRCN0000360889) and Mtm1 (NM_019926, shRNA-3′UTR: TRCN0000080723). Control shRNAC003 and shRNAC016 (Sigma–Aldrich) were also used.

Drug treatment of cells

Myotubes were treated with HPG-488 and then with inhibitors targeting nutrient signalling and translational machinery starting from 5 days post-induction of differentiation to avoid interference with the first steps of the differentiation process. Treatment was performed daily for 3 days, and protein synthesis rate (HPG-488 signal) and differentiation capacity (myotube seize and nuclei number per myotube) were addressed. The following drugs were diluted in DMSO to obtain the appropriate concentration: the pan-PI3Ks inhibitors wortmannin (5 mM; W1628, Sigma) and LY294002 (5 mM; L9908, Sigma), the mTORC1-ATP competitors torin 1 (200 nM; 475991, Sigma) and AZD8055 (200 nM; S1555, Selleckchem), the p70 ribosomal S6 kinase (S6K1) inhibitor PF-4708671 (10 mM; PZ0143, Sigma), the eIF2α inhibitor salubrinal (10 mM; SML0951, Sigma), the Vps34 inhibitor IN1 (0.25–1 mM; S7980, Selleck) and the PikFyve inhibitor Apilm (0.05–1 mM; S0785, Selleck). The pulse–chase assay was performed using CHX at 100 μg ml⁻¹ (Sigma–Aldrich). MLN4924 (Tocris) was used at 10 μM and tunicamycin (Tocris) was added at 2.5 μg ml⁻¹.

Protein synthesis measurements using HPG-488/AHA-594 assays and puromycin incorporation

The measurement of protein synthesis rates was addressed by Click-iT HPG-Alexa-Fluor Protein Synthesis Assays (Molecular Probes, Life Technologies) according to the manufacturer’s recommendation. HPG-488 and AHA-594 amino acid analogues were used to measure total protein synthesis or de novo protein synthesis rates, respectively. Of note HPG-488 showed higher incorporation capacity than AHA-594. Both probes were incubated with cells (50 μM in L-serum-free medium) at the indicated stage of differentiation (myoblast or myotubes) for 2–4 h. After PBS washing, cells were fixed with 3.7% formaldehyde in PBS for 15 min at room temperature protected from light and permeabilized with 0.5% Triton X-100 in PBS for 20 min at room temperature. Cells were incubated with nuclei staining (DAPI) solution for 10 min and imaged with a 20× objective using a confocal microscope, with at least nine fields per coverslip. Reconstituted fields were performed in CTRL and XLCNM cells using ImageJ software. For SUnSET analysis in cells, puromycin (2 mg ml⁻¹) diluted in cell medium was incubated for 4 h at 37 °C and puromycin incorporation was detected by western blot using an anti-puromycin antibody (Merck, 12D10 clone). For in vivo SUnSET analysis in skeletal muscle, mice were food deprived for 6 h and then injected (intraperitoneal injection) with puromycin (0.05 μmol g⁻¹ body mass). TA muscles were collected and frozen in 2-methylbutane–liquid nitrogen. Puromycin incorporation was measured on muscle fibres (8 mm transversal sections) using anti-puromycin primary antibodies (Merck, 12D10 clone) and Alexa-Fluor-488 (Invitrogen) secondary antibodies. A total of four WT and four Mtm1-KO mice and at least five sections from each muscle were imaged by confocal microscopy and fluorescent intensity was analysed using ImageJ software.

IP and western blotting

For IP, cells were rinsed with ice-cold PBS and lysed with IP buffer (40 mM HEPES pH 7.4, 5 mM MgCl_2, 10 mM Na₄P₂O₇, 10 mM Na β-glycerol phosphate, 1% Triton and one tablet of protease inhibitor cocktail per 25 ml of lysis buffer). The cell lysates were cleared by centrifugation and protein concentration was measured to equilibrate all supernatant samples using the Bradford assay (Bio-Rad) and bovine serum albumin (BSA) standard ranges. In total, 100 ml of each supernatant was mixed with 2× sodium dodecyl sulfate (SDS) buffer and left for the analysis (input (IN) fraction) and the remaining volumes were incubated with 80 ml of magnetic beads coupled to anti-FLAG, anti-HA (Pierce) or anti-MTM1 (homemade) antibodies pre-washed with IP buffer. Samples were incubated at 4 °C for 2 h. Following IP, the beads were washed four times with IP buffer and two times with IP buffer supplemented with 300 mM NaCl. Immunoprecipitated proteins were denatured by 2× SDS buffer, resolved by SDS–polyacrylamide gel electrophoresis (PAGE) and analysed by immunoblotting. For total lysates, cells or muscles were homogenized with a rotor-stator homogenizer using lysis buffer (50 mM Tris–Cl pH 8.0, 100 mM NaCl and 1 mM EDTA) supplemented with 0.5% Triton X-100, phenylmethylsulfonyl fluoride (PMSF) (1 mM), 10 mM Na β-glycerol phosphate and a complete protease inhibitor tablet (Sigma). Lysates were incubated in ice for 30 min and passed through a 25G syringe (20 times) twice and then cleared by centrifugation at 14,000g to obtain a soluble supernatant fraction. Samples were equilibrated using a Bradford assay (Bio-Rad) and BSA standard ranges. Proteins were denatured by 2× SDS buffer (Laemmli buffer). In total, 30–50 µg of each protein were separated on a 4–12% gradient SDS–polyacrylamide gel (Life Technologies) and transferred on Immobilon-P or nitrocellulose membranes (Millipore). Proteins of interest were revealed by specific primary antibodies (incubation overnight at 4 °C) (see Supplementary Tables 2 and 3 for detailed antibody information), followed by 1 h incubation with horseradish peroxidase-conjugated secondary antibodies. Protein bands were revealed by chemiluminescence (ChemiDoc MP, Bio-Rad Laboratories) with the Clarity Western ECL Substrate Detection System (Bio-Rad Laboratories).

For GTP pull-down, the Active GTPase Pull-Down Assay Kit (Sigma) was used according to manufacturer’s recommendations. A minimum of a 15-cm cell flask for each condition was required to have enough material. Protein extraction was performed using a buffer containing 40 mM HEPES pH 7.4, 150 mM NaCl, 5 mM MgCl₂, 10 mM Na₄P₂O₇, 10 mM Na β-glycerol phosphate, 0.2% CHAPS and one tablet of protease inhibitor cocktail per 25 ml of lysis buffer.

For the Ub trap assay, the Ubiquitin Enrichment Kit (Pierce, Thermo scientific) was used according to manufacturer’s recommendation. A 10-cm culture dish was used for each condition. The same protein extraction buffer was used as for the GTP pull-down assay.

Constructs and cloning

The lentiviral and AAV1-MTM1 (FL and C375S) constructs used in this study have been previously described⁵⁵. For MTM1-targeted expression on the lysosome we used pLV-Puro-UBC-TMEM192-MTM1(FL)-3XHA and pLV-Puro-UBC-TMEM192-MTM1(C375S)-3XHA (generated by Vectorbuilder). For MTM1-targeted expression on the ER the transmembrane domain of the ER protein MOSPD2 (aa493–515) we used pLJC5-TM-MTM1-3xFLAG and pLJC5-TM-MTM1-C375S-3xFLAG (generated by Vectorbuilder). The lentiviral plasmid pLJM1-Flag-RagA (Addgene no. 164357) was used to generate RagA mutants (RagA-Q66L, RagA-R37P and RagA-T21N) using the In-Fusion Mutagenesis Kit (Takara Bio). For phosphoinositide detection the plasmids pGEX-2TK-FYVE (HRS) (a gift from Mark Lemmon, Yale, USA) and pJV0048-6XHis-eGFP-SnxA (a gift from Dr. J. King) were used. For LAMTOR1 and LAMTOR3 constructs, the pRK5-LAMTOR1/p18-FLAG (Addgene no. 42331) and pRK5-HA-LAMTOR3/HBXIP (Addgene no. 42328) plasmids were used to generate mutants (LAMTOR1-K103A/K104A, LAMTOR3-K6A, LAMTOR3-K34 and the LAMTOR3 double mutation K6A/K34A). Primers used for mutagenesis are listed in Supplementary Tables 2 and 3.

Protein purification

The pGEX-4T1-mCherry-FYVE(HRS), pGEX-4T1-eGFP-FYVE(HRS) and pJV0048-6XHis-eGFP-SnxA plasmids were transformed into Escherichia coli BL21(DE3). A single colony was used to prepare a starter culture that was expended and grown at 37 °C until optical density at 600 nm reached 0.6–0.8. The expression of the recombinant proteins was induced with 0.5 mM isopropyl-beta-D-1-thiogalactopyranoside (IPTG) overnight at 18 °C. After centrifugation, bacterial pellets were lysed in a cold buffer containing 50 mM Tris pH 7.5, 150 mM NaCl, 1% Triton X-100, 5 mM dithiothreitol (DTT), 10% glycerol containing protease inhibitor cocktail and 0.5 mg ml⁻¹ lysozyme. Following sonication and centrifugation, proteins were purified with glutathione–sepharose 4B beads, eluted with 20 mM reduced glutathione, dialysed, concentrated and analysed by SDS–PAGE and Coomassie staining.

For RagGTPase dimer expression and purification in bacteria, a pETDuet-1 vector encoding a codon-optimized C-terminally His-tagged RagA and tagless RagC (pSK109) was transformed into BL21(DE3), with purification performed as previously described¹⁴. For LAMTORs, we used two batches of recombinant proteins provided by C. Kuang and M.E.G. de Araujo, respectively^14,21.

In brief, the first batches were purified by co-transforming and expressing, in the BL21(DE3) strain, the five subunits of LAMTOR. Three duet-vectors were transformed sequentially: pETDuet-1:GST-LAMTOR1, pCOLADuet:LAMTOR2|LAMTOR3-His and pACYCDuet:LAMTOR4|LAMTOR5/HBXIP. The resulting strain was grown in Luria broth (LB) until the OD of the culture reached 0.4–0.6 (early log-phase), then the temperature was lowered to 18 °C. Half an hour later, the LAMTOR complex was induced by 0.5 mM IPTG and expressed overnight. The bacterial pellets were resuspended in resuspension buffer (RB; 50 mM NaHEPES pH 7.4, 100 mM NaCl, 2 mM MgCl₂, 2 mM DTT, 0.5 mM PMSF, 0.01% Triton, 100 μg ml⁻¹ lysozyme, 10 μg ml⁻¹ DNase and EDTA-free protease inhibitor). The mixture was stirred at 4 °C for 30 min before passing through the microfluidizer six times at a pressure of 18,000 psi. The cell lysate was cleared by centrifugation at 40,000g for 1 h. The supernatant containing soluble Ragulator was incubated with glutathione resin for an hour, and the resin washed extensively in RB and RB supplemented with 300 mM NaCl. After buffer exchange into buffer NB (50 mM NaHEPES pH 7.4, 100 mM NaCl, 2 mM MgCl₂, 0.5 mM Tris(2-carboxyethyl)phosphine (TCEP), 0.01% Triton and 0.2 mM ATP), on-bead digestion was performed using thrombin at room temperature for an hour, and the flow-through was collected and frozen in liquid nitrogen until use.

For the second batch of the LAMTOR components we employed a similar strategy to that used previously for the LAMTOR–RagA–RagC heptameric complex²¹. In brief, we designed a synthetic gene comprising five open reading frames (ORFs), including individual promoter (polh or p10) and terminator signals (SV40- or HSVtk-polyA signals). Each ORF was codon optimized for expression in Spodoptera frugiperda and flanked by unique restriction sites. The first ORF encoded for LAMTOR1 with an N-terminal deletion of five amino acids. The GB1 domain-tag was used to enhance LAMTOR1 solubility and stability. The second ORF encoded for N-terminal 2× STREP-tagged LAMTOR4 followed by ORFs encoding LAMTOR5, LAMTOR2 and N-terminal 6× histidine-tagged LAMTOR3. To cleave the tags after purification, recognition sites for tobacco etch virus (TEV) protease were inserted N-terminally to the open reading frames. The entire gene was chemically synthesized by GeneArt AG, Life Technologies, and was then inserted into the multiple cloning sites of the acceptor vector pACEBac1, where it is flanked by Tn7 13 transposition elements. Generation of recombinant bacmid DNA was performed by transforming E. coli DH10 Multibac with pACEBac1, allowing the insertion of the expression constructs into the baculoviral DNA via Tn7 transposition. Recombinant bacmid DNA was isolated and used to transfect Spodoptera frugiperda (Sf21) cells, which were cultivated in SF-900 III SFM medium at 100 rpm and 27 °C.

Baculovirus expression was performed as described in ref. ²¹. The pentameric LAMTOR complex was purified by metal chelate affinity purification (Ni-NTA) followed by size exclusion chromatography. The cell pellet was resuspended in 20 mM HEPES pH 8.0, 150 mM NaCl, 2 mM DTT and lysed by flash freezing in liquid nitrogen. Lysates were then supplemented with protease inhibitors and centrifuged at 4 °C, 16,000g for 30 min. The soluble fraction was incubated at 4 °C for 90 min with Ni-NTA-beads equilibrated in lysis buffer. After incubation, the beads were loaded onto a biospin column (7326008 Bio-Rad). The column was subsequently washed with eight column volumes of lysis buffer containing 25 mM imidazole and eluted in 20 mM HEPES pH 8.0, 150 mM NaCl, 2 mM DTT and 250 mM imidazole. After the purification, all tags were cleaved simultaneously by treatment with TEV protease (T4455-1KU, Sigma) overnight. The cleavage was performed during dialysis (10 kDa molecular weight cut-off) against 20 mM HEPES pH 8.0, 150 mM NaCl and 2 mM DTT. To separate the stoichiometric pentameric LAMTOR complex from dimers of LAMTOR2/3 and residual TEV protease, size exclusion chromatography was performed using a Superdex 200 prep column equilibrated in dialysis buffer. The fractions containing the pentameric complex were pooled, and the concentration was adjusted to 1 mg ml⁻¹ using a 10-kDa cut-off concentrator (Z614483-25EA, Sigma).

Finally, for the expression and purification of LAMTOR mutated complex, LAMTOR1–5 were co-transformed and expressed in the BL21(DE3) strain. Three duet-vectors were transformed sequentially: pETDuet-1:GST-LAMTOR1|LAMTOR3-His (ampicillin resistant) (WT and mutants), pCOLADuet:GST-LAMTOR2|LAMTOR4-His (kanamycin resistant) and pACYCDuet:GST-LAMTOR5 (chloramphenicol resistant). The N-terminal GST-Tag on LAMTOR1, LAMTOR2 and LAMTOR5 were followed by TEV cleavage sites. The resulting strain was grown in LB until the OD of the culture reached 0.4–0.6 (early log-phase), then the temperature was lowered to 18 °C. Half an hour later, the LAMTOR complex was induced by 0.5 mM IPTG and expressed overnight. The bacterial pellets were resuspended in suspension buffer 1 (SB1) (50 mM NaHEPES pH 7.4, 100 mM NaCl, 2 mM MgCl₂, 2 mM DTT, 0.5 mM PMSF, 0.01% Triton, 100 μg ml⁻¹ lysozyme, 10 μg ml⁻¹ DNase and EDTA-free protease inhibitor). The mixture was extracted with a Dounce homogenizer at 4 °C before passing through a 25G syringe 20 times. The cell lysate was cleared by centrifugation at 40,000g for 1 h. The supernatant containing the LAMTOR-soluble complex was incubated with glutathione–sepharose beads for 2 h at 4 °C, and the resin washed extensively in SB1 and SB1 supplemented with 300 mM NaCl. After buffer exchange into buffer SB2 (50 mM NaHEPES pH 7.4, 100 mM NaCl, 2 mM MgCl₂, 0.5 mM TCEP, 0.01% Triton and 0.2 mM ATP), on-bead digestion was performed using TEV at 4 °C overnight, and the flow-through was collected and incubated with Ni-NTA resin for half an hour. The resin was washed with SB2 and SB2 supplemented with 30 mM imidazole and 300 mM NaCl, before being eluted in SB2 supplemented with 250 mM imidazole. The elution was collected and concentrated in a 10 kDa molecular weight cut-off concentrator, and ultracentrifuged at 100,000g for 30 min. Finally, fast protein liquid chromatography (FPLC) purification of the LAMTOR complexes (WT and mutants) was applied to a HiLoad 16/60 Superdex 200 gel-filtration column pre-equilibrated in gel-filtration buffer (50 mM NaHEPES pH 7.4, 100 mM KOAc, 2 mM MgCl₂ and 2 mM DTT). The peak corresponding to the LAMTOR complex was collected concentrated in a 10 kDa molecular weight cut-off and analysed in in 3–18% Tris–glycine SDS–PAGE gel. The final yield was ~0.15–0.2 mg LAMTORs per 2 l of initial LB culture.

Lysosome(Lyso)-Tag cell line and Lysosome(Lyso)-IP

C2C12 (WT and Mtm1-KO Cas9) myoblast were infected with a lentiviral preparation to express the Lyso-Tag construct pLJC5-Tmem192-3xHA (Addgene no. 102930) as described previously²⁰. Following selection, cells were differentiated into myotubes and Lyso-Tag expression and specificity were verified by immunoblotting and immunofluorescence. Lysosome integrity (number, size and cell distribution) as well as proteolytic activity (Cathepsin D and B/L Activity Assay Kit, Abcam) were checked in the newly generated cell lines as previously described²⁰. For Lyso-IP, two 10-cm plates were used for each Lyso-IP. Owing to the extensive membrane and cytoskeleton network in myotubes, we adapted the previously described protocol²⁰ to this cell model. In brief, 9-day cultured myotubes were detached from myoblasts using dispase for 30 min at 37 °C to enrich the myotubes. Cells were pelleted and rinsed twice with PBS and 1 ml of potassium buffred PBS (KPBS) (136 mM KCl and 10 mM KH₂PO₄, pH 7.25 (adjusted with KOH)) and centrifuged at 1,000g for 2 min at 4 °C. Pelleted cells were resuspended in 1 ml (50 ml of the total cell lysate was taken and adjusted with 2× SDS for further processing of the whole-cell fraction: ‘inputs’). The remaining cells were gently homogenized with 20 strokes (2 ml Dounce homogenizer) three times, spaced between 2 min incubations on ice. The homogenate was then centrifuged at 1,000g for 5 min at 4 °C and the supernatant containing the cellular organelles, including lysosomes, was incubated with 100 μl of KPBS pre-washed (three times) anti-HA magnetic beads (Sigma). We proceeded to the Lyso-IP assay in our cell model using diffident incubation times (10 min, 20 min, 30 min, 1 h, 2 h and 4 h) and noticed that the 2 h incubation was optimum to capture lysosomes (3XHA-TMEM192 expression) with no contaminant organelles except for very few ER membranes, which are the membrane networks most associated with the lysosome. Lyso-IP beads were then gently washed three times with KPBS and lysosomes were eluted using three rounds of 10 min elution with 2 mg ml⁻¹ HA peptide solution in KPBS (50ul). Eluted samples were equilibrated according to 3XHA-TMEM192 expression and analysed by immunoblotting for organelle markers and the mTORC1 complex or used for proteomic, lipidomic or metabolomic analyses.

Immunofluorescence, PLA and imaging

For imaging analysis, cells were plated on glass slides coated with 0.2 filtered and sterile gelatin (G9391, Sigma) pre-incubated for 1 h at 37 °C. Myosin heavy chain (MHC) labelling was performed on myotubes fixed with 3.7% formaldehyde for 10 min at room temperature, quenched with 50 mM NH₄Cl for 10 min and permeabilized by 0.1% Triton X-100 in Dulbecco’s PBS (DPBS) for 10 min. Following, 1 h of saturation with 10% goat serum in DPBS, cells were incubated for 1 h with antibodies (see Supplementary Tables 2 and 3 for antibody information) and 1 h with Alexa-488-coupled secondary antibodies (ThermoFisher), nuclei were stained with DAPI (Euromedex), and cells were washed and maintained in DPBS. Imaging was performed with the ZOE Fluorescent Cell Imager (Bio-Rad laboratories). Myotube area, percentage of nuclei per myotube and index of fusion were addressed using ImageJ software and homemade plugins. For lysosome and ER labelling, as well as for phosphoinositides, cells were fixed with 3.7% formaldehyde, quenched with NH₄Cl for 10 min and permeabilized with 20 µM digitonin/PIPES-BS (PIPES 20 mM pH 6.8, NaCl 137 mM and KCl 2.7 mM) for 5 min. After saturation with 10% goat serum/PIPES-BS, cells were incubated with 50 µg ml⁻¹ of the GST-mCherry-FYVE-domain (HRS) probe (PI3P detection) and/or primary antibodies at room temperature or overnight at 4 °C. After three washes with PIPES-BS, secondary fluorescent antibodies in 10% goat serum/PIPES-BS were added for 1 h and washed three times with PIPES-BS. Cells were then fixed a second time with 3.7% formaldehyde, nuclei were stained with DAPI, washed three times with DPBS and mounted with FluorSave reagent (Calbiochem).

PLA was performed using the Duolink In Situ Fluorescence Kit (Sigma–Aldrich) according to the manufacturer’s recommendations. In brief, after primary antibody incubation (Nogo and HA), buffer A and B from the kit were prepared and filtered (0.22-mm filter). Cells were washed three times with PBS–3% BSA and once (5 min) with buffer A. A mixed solution of 8 ml of probe Minus and 8 ml of probe Plus diluted in the diluent buffer was prepared at a final volume of 40 ml per sample. Cells were incubated with the mixture for 1 h at 37 °C in a humidified chamber and then washed with buffer 1 twice for 5 min. The ligation was performed for 30 min at 37 °C using 1 ml of the ligase diluted with in 1× ligase buffer (8 ml) and 31 ml of ultrapure water. Following two washes (2 min) with buffer A, the amplification reaction was performed using 0.5 ml polymerase diluted in 1× amplification buffer (8 ml) and 31.5 ml ultrapure water for 100 min at 37 °C. Cells were washed twice with buffer B for 10 min and then for 1 min with the same buffer diluted at 1:100 (v:v ultrapure water). Finally, cells were fixed for 10 min with 4% PFA, rinsed twice with PBS and mounted for microscopy. Imaging was performed with confocal microscopes LSM780 (Zeiss) or LSM900 (Zeiss) or Microscope Elvra PS.1 (Zeiss) for structured-illumination microscopy (SIM) image acquisition. Images were treated with Zen or ImageJ software to quantify fluorescence intensities or colocalization using Mander’s split coefficients (named Mander’s coefficients in the figures) on at least six z-planes for each image. This coefficient is proportional to the amount of fluorescence of the colocalizing pixels or voxels in each colour channel. For morphometric measurements (myotube area, index of fusion and percentage of myotubes with the indicated nuclei number), MHC positive myotubes were isolated (surrounded) using a designed macro on ImageJ or done manually depending on the cell density. The index of fusion was calculated as the ratio of the nuclei number in myotubes with two or more nuclei versus the total number of nuclei. Cell area was recorded on individual myotubes and the percentage of myotubes with <5, 5–10 and >10 nuclei were counted. Data were plotted as individual point for each analysed myotube following at least three independent experiments.

Targeted LC-MS metabolomics analyses

For metabolomic analysis samples (three total cell lysates from independent culture for each cell line or six Lyso-IP from independent experiments for each cell line) were equilibrated according to Lyso-Tag expression (3XHA-TMEM192). The extraction solution (50% methanol, 30% acetonitrile and 20% water) was added to the sample. The volume of the extraction solution was adjusted to lysosome content (1 ml per 100 ml of eluted Lysosomes and 1 ml cell pellet from a 10-cm dish). After the addition of extraction solution, samples were vortexed for 5 min at 4 °C and centrifuged at 16,000g for 15 min at 4 °C. The supernatants were collected and stored at −80 °C until analysis. Liquid chromatography–mass spectrometry (LC-MS) analyses were conducted on a QExactive Plus Orbitrap mass spectrometer equipped with an Ion Max source and a HESI II probe coupled to a Dionex UltiMate 3000 uHPLC system (ThermoFisher). External mass calibrations were performed using a standard calibration mixture every 7 days, as recommended by the manufacturer. The 5 µl samples were injected onto a ZIC-pHILIC column (150 mm × 2.1 mm, inner diameter 5 µm) with a guard column (20 mm × 2.1 mm, inner diameter 5 µm) (Millipore) for LC separation. Buffer A was 20 mM ammonium carbonate, 0.1% ammonium hydroxide (pH 9.2), and buffer B was acetonitrile. The chromatographic gradient was run at a flow rate of 0.200 µl min⁻¹ as follows: 0–20 min, linear gradient from 80% to 20% of buffer B; 20–20.5 min, linear gradient from 20% to 80% of buffer B; 20.5–28 min, 80% buffer B. The mass spectrometer was operated in full-scan polarity-switching mode with the spray voltage set to 2.5 kV and the heated capillary held at 320 °C. The sheath gas flow was set to 20 units, the auxiliary gas flow to 5 units and the sweep gas flow to 0 units. The metabolites were detected across a mass range of 75–1,000 m/z at a resolution of 35,000 (at 200 m/z) with the automatic gain control target at 106 and the maximum injection time at 250 ms. Lock masses were used to ensure mass accuracy below 5 ppm. Data were acquired with Thermo Xcalibur software (ThermoFisher). The peak areas of metabolites were determined using Thermo TraceFinder software (ThermoFisher), identified by the exact mass of each singly charged ion and by the known retention time on the HPLC column. Analysis was performed using the online software Metaboanalyst (McGill university). Heat maps were generated using Prism 9 software.

Nano-liquid chromatography-tandem mass spectrometry analysis of lysosomal proteomes

Analysis was performed on four independent Lyso-IP samples for each cell line. For each sample, dried protein extracts were solubilized with 25 µl of 5% SDS. Proteins were submitted to reduction and alkylation of cysteine residues by addition of TCEP and chloroacetamide to a final concentration, respectively, of 10 mM and 40 mM. Protein samples were then processed for trypsin digestion on S-trap Micro devices (Protifi) according to manufacturer’s protocol, with the following modifications: precipitation was performed using 211 µl S-Trap buffer, and 1 µg trypsin was added per sample for digestion in 25 µl ammonium bicarbonate 50 mM pH 7. Tryptic peptides were resuspended in 17 µl of 2% acetonitrile and 0.05% trifluoroacetic acid and analysed by nano-liquid chromatography tandem mass spectrometry using an UltiMate 3000 system (NCS-3500RS Nano/Cap System; ThermoFisher Scientific) coupled to an Orbitrap QEx-Plus mass spectrometer (ThermoFisher Scientific). Each sample was loaded onto a C18 pre-column (300 µm inner diameter × 5 mm; ThermoFisher Scientific) in a solvent made of 2% acetonitrile and 0.05% trifluoroacetic acid, at a flow rate of 20 µl min⁻¹. After 5 min of desalting, the pre-column was switched online with the analytical C18 column (75 µm inner diameter × 50 cm, in-house packed with Reprosil C18) equilibrated in 95% solvent A (5% acetonitrile and 0.2% formic acid) and 5% solvent B (80% acetonitrile and 0.2% formic acid). Peptides were eluted using a 5–50% gradient of solvent B over 105 min at a flow rate of 300 nl min⁻¹. The mass spectrometer was operated in a data-dependent acquisition mode with the Xcalibur software. Mass spectrometry survey scans were acquired with a resolution of 70,000 and an AGC target of 3 × 10⁶. The ten most intense ions were selected for fragmentation by high-energy collision-induced dissociation, and the resulting fragments were analysed at a resolution of 17,500, using an AGC target of 1 × 10⁵ and a maximum fill time of 50 ms. Dynamic exclusion was used within 30 s to prevent repetitive selection of the same peptide.

For bioinformatics analysis, raw mass spectrometry files were processed with the Mascot software (version 2.7.0) for database searches and Proline (version 2.1.2) for label-free quantitative analysis. Data were searched against Mus musculus entries of the SwissProt protein database (release SwissProt 20201022; Mus musculus 17,060 entries). Carbamidomethylation of cysteines was set as a fixed modification, whereas oxidation of methionine was set as a variable modification. Specificity of trypsin/P digestion was set for cleavage after K or R, and two missed trypsin cleavage sites were allowed. The mass tolerance was set to 10 ppm for the precursor and to 20 mmu in tandem mass spectrometry mode. The minimum peptide length was set to seven amino acids, and identification results were further validated in Proline by the target decoy approach using a reverse database at both a peptide spectrum match (PSM) and protein false-discovery rate of 1%. For label-free relative quantification of the proteins across biological replicates and conditions, cross-assignment of peptide ion peaks was enabled inside groups with a match time window of 1 min, after alignment of the runs with a tolerance of ±600 s.

Median ratio fitting computes a matrix of abundance ratios calculated between any two runs from ion abundances for each protein. For each pair-wise ratio, the median of the ion ratios is then calculated and used to represent the protein ratio between these two runs. A least-squares regression is then performed to approximate the relative abundance of the protein in each run in the dataset. This abundance is finally rescaled to the sum of the ion abundances across runs. A Student’s t-test (two-tailed t-test, equal variances) was then performed on log₂-transformed values to analyse differences in protein abundance in all biologic group comparisons. The significance level was set at P ≤ 0.05, and ratios were considered relevant if higher than ±4.

Whole-cell and lysosome lipidomics

The protocol was adapted from Li et al.²⁷. Myotubes (10⁶ in 10-cm culture dishes in duplicate) were used. Samples were sequentially extracted using the methods described below (methods 1 and 2). Internal standards containing EquiSPLASH LIPIDOMIX (Avanti Polar Lipids), PI(3)P 32:0-d62 (Cayman Chemical), PI(4,5)P2 32:0-d62 (Cayman Chemical) and PI(3,4,5)P3 32:0-d62 (Cayman Chemical) were added to samples with 1 ml 10:3 (v:v) tert-butyl methyl ether (MTBE):methanol (MeOH) and ~15 mg zirconia beads (CarlRoth). After homogenization in Precellys for ten cycles (6,800 rpm, cycle time 10 s, pulse 30 s, 4 °C), samples were centrifuged for 5 min (16,000g at 4 °C). The supernatant was dried under nitrogen protection and reconstituted in 100 µl methanol for the analysis of PIs by method 1. The residue was mixed with 1.0 ml 1000:300:65 (v:v:v) MTBE:MeOH:2 N hydrochloric acid (HCl). After homogenization in Precellys for two cycles (4,500 rpm, cycle time 10 s, pulse 30 s, 4 °C), 250 µl 0.1 N HCl was added followed by vortexing for 5 min at 4 °C. After centrifugation (5 min, 16,000g, 4 °C), the upper layer was taken for PIPx (method 2). For PIPx samples, 50 µl trimethylsilyl diazomethane solution (2 M in hexane) was added and the reaction was allowed for 10 min at room temperature. The methylation was quenched with 5 µl acetic acid. After washing with 250 µl 1:1 (v:v) MeOH:water, the upper layer was dried and reconstituted in 50 µl methanol for the analysis of PIPx by method 2.

Lysosome samples (starting material of 10¹⁰ in 15-cm culture dishes in duplicate followed by Lyso-IP) were mixed with 1 ml 1000:300:65 (v:v:v) MTBE:MeOH:2 N HCl. After homogenization in Precellys for two cycles (4,500 rpm, cycle time 10 s, pulse 30 s, 4 °C), 250 µl 0.1 N HCl was added followed by vortexing for 5 min at 4 °C. After centrifugation (5 min, 16,000g, 4 °C), the upper layer was taken for methylation. A 50 µl trimethylsilyl diazomethane solution (2 M in hexane) was added and the reaction was allowed for 10 min at room temperature. The methylation was quenched with 5 µl acetic acid. After washing with 250 µl 1:1 (v:v) MeOH:water, the upper layer was dried under nitrogen and reconstituted in 50 µl MeOH for PIPx analysis.

Ultra-high performance liquid chromatography coupled to mass spectrometry analysis of PI was carried out as reported previously²⁷. In brief, an ACQUITY UPLC CSH C18 Column (100 mm × 2.1 mm, 1.7 μm, 130 Å) with an ACQUITY UPLC CSH C18 Van-Guard Pre-column (5 mm × 2.1 mm, 1.7 μm, 130 Å) (Waters, Milford, MA, USA) was used for liquid chromatography separation. Method 2: the analysis of PIPx was performed on a Daicel Chiralpak IB-U column (100 mm × 3.0 mm, 1.6 μm) with a KrudKatcher Ultra HPLC in-line filter (0.5 μm) (Phenomenex). Sequential window acquisition of all theoretical fragment ion mass spectra experiments was used in both methods to obtain comprehensive tandem mass spectrometry data. Selected reaction monitoring was applied to acquire targeted mass spectrometry data. Analyst v1.7 was used in all analytical systems for data acquisition and system control. PeakView v2.2 was used for manual data evaluation and LC-MS performance checks by comparing periodically with quality control samples. MasterView v1.1 was used for identification. MultiQuant v3.0 was used for peak integration, linear regression and concentration calculations. Statistical analysis and visualization were performed with R.

Phosphoinositide pull-down and flotation assay

Recombinant proteins and phosphoinositide interactions were first addressed by lipid–protein pull-down assay. The phosphoinositide beads sample pack contained control beads, PI beads, PI(3)P beads, PI(4)P beads, PI(5)P beads, PI(3,4)P2 beads, PI(3,5)P2 beads, PI(4,5)P2 beads and PI(3,4,5)P3 beads (Echelon Bioscience). Purified proteins of interest were diluted (2 µg ml⁻¹ final concentration) in binding/wash buffer (10 mM HEPES pH 7.4, 150 mM NaCl, 0.25% NP-40 and supplemented with a protease inhibitor cocktail). PI-beads were washed four times with binding/wash buffer at 4 °C by gentle rotation for 5 min and centrifugation at 300g for 5 min. A volume of 50 µl of beads was added to protein solution and incubated for 2 h at 4 °C. After five washing cycles, bead–protein complexes were dissolved in 2× SDS buffer (Laemmli buffer) and binding proteins were revealed by immunoblotting.

For the flotation assay, 200 µl of 1 mM CTRL liposome suspension composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC; Avanti Polar Lipids) and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (POPE; Avanti Polar Lipids) at a molar ratio of 70:30. The phosphoinositide liposomes were composed of POPC:POPE:PI (molar ratio of 60:20:10). The lipid mixtures in glass tubes were adjusted to a 200 µl final volume with chloroform and homogenized for 10 min at 37 °C. The solvent was evaporated for 30 min under nitrogen to form a dry film followed by overnight incubation in a vacuum chamber to eliminate traces of solvent. Dry lipid mixtures were resuspended in preheated (37 °C) PBS to obtain 1 mM lipid suspensions and then incubated for 1 h in a 37 °C water bath with vigorous vortexing cycles every 10 min. Lipid mixtures were subjected to five cycles of freezing and thawing (in liquid nitrogen and a 37 °C bath, respectively). Following centrifugation (1,000g for 5 min at room temperature), liposomes were transferred into 1.5-ml microcentrifuge tubes and sonicated to obtain liposomes that could be stored at −80 °C until use. Before incubation with the proteins of interest, thawed liposomes were sonicated for 60 min in a water bath sonicator filled with ice to obtain a clear suspension. Liposome diameters were around 50 nm (measured by dynamic light scattering). In ultracentrifuge polycarbonate tubes, proteins of interest (1 µg) were mixed with 20 µl of each liposome (CTRL/POPC:POPE and POPC:POPE:PI liposomes) in a final volume of 100 µl PBS (without Ca/Mg). The protein–liposome mixtures were incubated for 20 min at room temperature, and 100 μl of 60% sucrose in PBS was added to yield a 30% final sucrose concentration. The mixture was overlayed with 250 μl of 25% sucrose and 50 μl of PBS (without Ca/Mg) on top. Tubes were centrifuged at 200,000g for 1 h at 20 °C (TLA100 rotor, Beckman). The liposome–protein complexes migrate to the top buffer layer. The bottom, middle and top fractions were collected with a Hamilton syringe and submitted to trichloroacetic acid precipitation followed by analysis of the protein of interest by immunoblotting.

Mice experiments

Male Mtm1-KO mice (genetic background 129/Sv) and littermate WT mice were used (X-linked disease). Care and manipulation of mice was approved by the ethical committee for animal experimentation, University of Strasbourg, and the Centre Régional d’Exploration Fonctionnelle et Ressources Expérimentales, University of Toulouse, according to national and European legislations on animal experimentation. Mice for each experimental protocol had the same age and similar weights. Animals showing signs of suffering during treatment were prematurely scarified. Animals were housed in a temperature-controlled room (19–22 °C) with a 12 h:12 h light:dark cycle. Mice had ad libitum access to food and water. Standard chow (SAFE R04-10) was provided, and mice displaying a myopathic phenotype received DietGel Recovery (Bio-Serv, 72-06-5022) as a dietary supplement. Blood glucose, insulin and IGF1 were measured by glucometer (Roche), a mouse insulin ELISA kit and a mouse IGF1 ELISA Kit (Sigma), respectively. For treatment with AZD8055, 2-week and 5-week-old mice were injected intraperitoneally with AZD8055 (10 mg day⁻¹ kg⁻¹) for 20 days. For toxin-induced muscle degeneration/regeneration followed by AAV injection of MTM1 constructs, 10 μl of Notexin (Lotaxan Laboratory) at 25 μg ml⁻¹ into the TA muscle after 7 days (the beginning of muscle regeneration phase), AAV1-MTM1-FL (CTRL) or AAV1-MTM1-C375S (inactive mutant) particles were injected in TA muscle at 2 × 10¹⁰ viral genomes per ml. Mice were weighed daily and food intake was measured every day for all groups. At the end of the experimental protocols, muscle force was measured on TA muscle, otherwise mice were killed by CO₂ inhalation followed by cervical dislocation, and TA muscles were rapidly dissected. For protein expression analysis, part of the TA was homogenized with a Dounce homogenizer in lysis buffer (50 mM Tris–Cl pH 8.0, 100 mM NaCl and 1 mM EDTA) supplemented with 1% Triton X-100, PMSF (1 mM) and a complete protease inhibitor tablet. Lysates were incubated on ice for 30 min, passed through a 25G syringe ten times and then cleared by centrifugation at 14,000g to obtain soluble fractions. The last part of the TA muscle was conserved in nitrogen-cooled isopentane and liquid nitrogen for muscle transversal sectioning and histology analysis as previously described⁵⁵.

Muscle histology, morphometry and physiology

For isolated fibre bundles, muscles (TAs) were dissected and washed with PBS. Muscles were incubated for 15 min with PBS–4%PFA before a second wash with PBS and mechanical dissection under a microscope until obtaining fibre bundles of one to four fibres. Following saturation with PBS–10% foetal bovine serum and 3% BSA for 2 h at room temperature, fibres were permeabilized in PBS 1% Triton X-100 for 30 min, washed with PBS (10 min) and incubated with the primary antibodies overnight at 4 °C. After five washes with PBS–0.1% Triton X-100, a second saturation step was performed for 30 min and fibres were incubated with the secondary antibody for 2 h at room temperature (in the dark). After five washes with PBS–0.1% Triton X-100 and two washes with PBS, fibres were incubated with DAPI for 10 min, washed once with PBS and mounted on slides for microscopic observation.

Transverse TA cryosections (10 μm) were fixed and stained with hematoxylin and eosin (H&E) or used for SDH activity assays. Sections were imaged with a slide scanner NanoZoomer 2HT equipped with the fluorescence module L11600-21 (Hamamatsu Photonics). Fibre diameter and nuclei position were analysed in H&E sections from TA mouse skeletal muscle using FIJI image analysis software. For transmission electron microscopy (TEM) analysis, TA muscle was minced into small pieces for organelle suspension and were then fixed in 2.5% gluteraldehyde with 0.1 M sodium cacodylate buffer (pH 7.2) for 24 h at 4 °C, washed in 0.1 M cacodylate buffer for 30 min and post-fixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 1 hr at 4 °C. Following stepwise dehydration with increasing concentrations of ethanol and embedding in Epon 812, the sections were observed with a Morgagni 268D electron microscope.

Muscle force was monitored in in situ TA muscle in response to sciatic nerve stimulation. Mice were anaesthetized and immobilized on a platform and hindlimbs were fixed with pins and clamps. The distal muscle tendon was attached to the servomotor system (Aurora Scientific). A surgical opening was performed to access to the sciatic nerve and stimulated distally with a bipolar silver electrode using 0.1-ms-duration supramaximal square-wave pulses. The absolute maximal force generated during isometric contractions in response to electrical stimulation (500-ms-duration stimulation train, 100 Hz frequency) was recorded. Absolute maximal force was determined at optimal length (length at which maximal tension was obtained during the tetanus). Absolute maximal force was normalized to muscle weight as an estimate of specific maximal force. After contractile measurements, mice were euthanized with an overdose of pentobarbital, and muscles were dissected and weighed.

Statistics and reproducibility

Statistical analyses are presented as mean ± s.d. or ± s.e.m. and were performed on data obtained from at least three independent experiments using biologically independent samples (biological replicates), unless otherwise indicated. For some in vitro confirmation experiments, two biological replicates were performed using pooled material from three to four independent cell preparations; these experiments are identified in the figure legends, and no statistical analysis was performed for these cases. For cell-based experiments, biological replicates correspond to independent muscle cell batches (CTRL or genetically modified), derived from different passages and cultured on separate days for each experiment. For in vitro experiments using recombinant proteins and lipid preparations, at least three independent batches were prepared separately, and experiments were performed independently. No a priori power calculations were performed. Sample sizes were determined on the basis of previously published studies in which statistically significant differences were observed^29,55. Sample sizes (cells and mice) are indicated in the corresponding figure legends. Experiments were not randomized, and no data were excluded from the analysis, except for Mtm1-KO mice that did not survive until the end of the treatment period. Data collection and analysis were not performed blind to the conditions of the experiments (XLCNM cells and mouse models showed visual phenotypical differences). GraphPad Prism (version 9.4.0) was used for data visualization and statistical analyses. Comparisons between two experimental groups were performed using a two-sided Student’s t-test for assumed normally distributed data (not formally tested) or a two-sided Mann–Whitney U test for non-normally distributed data. Comparisons involving more than two groups were analysed using one-way analysis of variance (ANOVA), followed by appropriate multiple-comparisons post hoc tests (Dunnett’s, Tukey’s or Šidák’s), as specified in the figure legends. Statistical significance was defined as P < 0.05. Exact P values are reported in the figure captions, except when P < 0.0001.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium available via the PRIDE partner repository at https://www.ebi.ac.uk/pride/archive/projects/PXD065512. All data supporting the findings of this study are available within the Article and its Supplementary Information and available via Figshare at https://doi.org/10.6084/m9.figshare.31168189 (ref. ⁵⁶). Requests for additional data can be referred to the corresponding authors; these will be reviewed in consultation with the institutional R&D (INSERM Transfer), which will determine the terms of a data transfer agreement between the recipient institution and the INSERM. Source data are provided with this paper.

Change history

01 April 2026
In the version of this article initially published, Nesrine Hifdi and Mathilde Vaucourt (UMR1297 Inserm/Université Toulouse, Institut des Maladies Métaboliques et Cardiovasculaires (I2MC), Toulouse, France) were listed incorrectly with additional affiliations, which are now removed in the HTML and PDF versions of the article.

References

Zoncu, R. & Perera, R. M. Built to last: lysosome remodeling and repair in health and disease. Trends Cell Biol. 32, 597–610 (2022).
Article CAS PubMed PubMed Central Google Scholar
Sancak, Y. et al. Ragulator-Rag complex targets mTORC1 to the lysosomal surface and is necessary for its activation by amino acids. Cell 141, 290–303 (2010).
Article CAS PubMed PubMed Central Google Scholar
Liu, G. Y. & Sabatini, D. M. mTOR at the nexus of nutrition, growth, ageing and disease. Nat. Rev. Mol. Cell Biol. 21, 183–203 (2020).
Article CAS PubMed PubMed Central Google Scholar
Napolitano, G., Di Malta, C. & Ballabio, A. Non-canonical mTORC1 signaling at the lysosome. Trends Cell Biol. 32, 920–931 (2022).
Article CAS PubMed Google Scholar
Gwinn, D. M. et al. AMPK phosphorylation of raptor mediates a metabolic checkpoint. Mol. Cell 30, 214–226 (2008).
Article CAS PubMed PubMed Central Google Scholar
Inoki, K., Li, Y., Xu, T. & Guan, K. L. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes Dev. 17, 1829–1834 (2003).
Article CAS PubMed PubMed Central Google Scholar
Menon, S. et al. Spatial control of the TSC complex integrates insulin and nutrient regulation of mTORC1 at the lysosome. Cell 156, 771–785 (2014).
Article CAS PubMed PubMed Central Google Scholar
Bar-Peled, L. et al. A tumor suppressor complex with GAP activity for the Rag GTPases that signal amino acid sufficiency to mTORC1. Science 340, 1100–1106 (2013).
Article CAS PubMed PubMed Central Google Scholar
Shen, K. et al. Architecture of the human GATOR1 and GATOR1–Rag GTPases complexes. Nature 556, 64–69 (2018).
Article CAS PubMed PubMed Central Google Scholar
Fromm, S. A., Lawrence, R. E. & Hurley, J. H. Structural mechanism for amino acid-dependent Rag GTPase nucleotide state switching by SLC38A9. Nat. Struct. Mol. Biol. 27, 1017–1023 (2020).
Article CAS PubMed PubMed Central Google Scholar
Petit, C. S., Roczniak-Ferguson, A. & Ferguson, S. M. Recruitment of folliculin to lysosomes supports the amino acid-dependent activation of Rag GTPases. J. Cell Biol. 202, 1107–1122 (2013).
Article CAS PubMed PubMed Central Google Scholar
Tsun, Z. Y. et al. The folliculin tumor suppressor is a GAP for the RagC/D GTPases that signal amino acid levels to mTORC1. Mol. Cell 52, 495–505 (2013).
Article CAS PubMed PubMed Central Google Scholar
Rogala, K. B. et al. Structural basis for the docking of mTORC1 on the lysosomal surface. Science 366, 468–475 (2019).
Article CAS PubMed PubMed Central Google Scholar
Shen, K., Choe, A. & Sabatini, D. M. Intersubunit crosstalk in the Rag GTPase heterodimer enables mTORC1 to respond rapidly to amino acid availability. Mol. Cell 68, 821 (2017).
Article CAS PubMed PubMed Central Google Scholar
Lawrence, R. E. et al. A nutrient-induced affinity switch controls mTORC1 activation by its Rag GTPase-Ragulator lysosomal scaffold. Nat. Cell Biol. 20, 1052–1063 (2018).
Article CAS PubMed PubMed Central Google Scholar
Kim, Y. C. et al. Rag GTPases are cardioprotective by regulating lysosomal function. Nat. Commun. 5, 4241 (2014).
Article CAS PubMed PubMed Central Google Scholar
Okosun, J. et al. Recurrent mTORC1-activating RRAGC mutations in follicular lymphoma. Nat. Genet. 48, 183–188 (2016).
Article CAS PubMed Google Scholar
Hnia, K., Clausen, T. & Moog-Lutz, C. Shaping striated muscles with ubiquitin proteasome system in health and disease. Trends Mol. Med. 25, 760–774 (2019).
Article CAS PubMed Google Scholar
Hnia, K., Vaccari, I., Bolino, A. & Laporte, J. Myotubularin phosphoinositide phosphatases: cellular functions and disease pathophysiology. Trends Mol. Med. 18, 317–327 (2012).
Article CAS PubMed Google Scholar
Abu-Remaileh, M. et al. Lysosomal metabolomics reveals V-ATPase- and mTOR-dependent regulation of amino acid efflux from lysosomes. Science 358, 807–813 (2017).
Article CAS PubMed PubMed Central Google Scholar
de Araujo, M. E. G. et al. Crystal structure of the human lysosomal mTORC1 scaffold complex and its impact on signaling. Science 358, 377–381 (2017).
Article PubMed Google Scholar
Settembre, C. et al. A lysosome-to-nucleus signalling mechanism senses and regulates the lysosome via mTOR and TFEB. EMBO J. 31, 1095–1108 (2012).
Article CAS PubMed PubMed Central Google Scholar
Martina, J. A. & Puertollano, R. Rag GTPases mediate amino acid-dependent recruitment of TFEB and MITF to lysosomes. J. Cell Biol. 200, 475–491 (2013).
Article CAS PubMed PubMed Central Google Scholar
Posor, Y., Jang, W. & Haucke, V. Phosphoinositides as membrane organizers. Nat. Rev. Mol. Cell Biol. 23, 797–816 (2022).
Vines, J. H. et al. A PI(3,5)P2 reporter reveals PIKfyve activity and dynamics on macropinosomes and phagosomes. J. Cell Biol. 222, e202209077 (2023).
Article CAS PubMed PubMed Central Google Scholar
Rivero-Rios, P. & Weisman, L. S. Roles of PIKfyve in multiple cellular pathways. Curr. Opin. Cell Biol. 76, 102086 (2022).
Article CAS PubMed PubMed Central Google Scholar
Li, P. & Lammerhofer, M. Isomer selective comprehensive lipidomics analysis of phosphoinositides in biological samples by liquid chromatography with data independent acquisition tandem mass spectrometry. Anal. Chem. 93, 9583–9592 (2021).
Article CAS PubMed Google Scholar
de Araujo, M. E. et al. Stability of the endosomal scaffold protein LAMTOR3 depends on heterodimer assembly and proteasomal degradation. J. Biol. Chem. 288, 18228–18242 (2013).
Article PubMed PubMed Central Google Scholar
Amoasii, L. et al. Myotubularin and PtdIns3P remodel the sarcoplasmic reticulum in muscle in vivo. J. Cell Sci. 126, 1806–1819 (2013).
CAS PubMed Google Scholar
Lim, C. Y. et al. ER–lysosome contacts enable cholesterol sensing by mTORC1 and drive aberrant growth signalling in Niemann–Pick type C. Nat. Cell Biol. 21, 1206–1218 (2019).
Article CAS PubMed PubMed Central Google Scholar
Di Mattia, T. et al. Identification of MOSPD2, a novel scaffold for endoplasmic reticulum membrane contact sites. EMBO Rep. 19, e45453 (2018).
Article PubMed PubMed Central Google Scholar
Radulovic, M. et al. Cholesterol transfer via endoplasmic reticulum contacts mediates lysosome damage repair. EMBO J. 41, e112677 (2022).
Article CAS PubMed PubMed Central Google Scholar
Ebner, M. et al. Nutrient-regulated control of lysosome function by signaling lipid conversion. Cell 186, 5328–5346 (2023).
Article CAS PubMed Google Scholar
Appenzeller-Herzog, C. & Hall, M. N. Bidirectional crosstalk between endoplasmic reticulum stress and mTOR signaling. Trends Cell Biol. 22, 274–282 (2012).
Article CAS PubMed Google Scholar
Xiong, G. et al. The PERK arm of the unfolded protein response regulates satellite cell-mediated skeletal muscle regeneration. Elife 6, e22871 (2017).
Article PubMed PubMed Central Google Scholar
Fetalvero, K. M. et al. Defective autophagy and mTORC1 signaling in myotubularin null mice. Mol. Cell. Biol. 33, 98–110 (2013).
Article CAS PubMed Google Scholar
Manning, B. D., Tee, A. R., Logsdon, M. N., Blenis, J. & Cantley, L. C. Identification of the tuberous sclerosis complex-2 tumor suppressor gene product tuberin as a target of the phosphoinositide 3-kinase/AKT pathway. Mol. Cell 10, 151–162 (2002).
Article CAS PubMed Google Scholar
Fitzian, K. et al. TSC1 binding to lysosomal PIPs is required for TSC complex translocation and mTORC1 regulation. Mol. Cell 81, 2705–2721 (2021).
Article CAS PubMed Google Scholar
Marat, A. L. et al. mTORC1 activity repression by late endosomal phosphatidylinositol 3,4-bisphosphate. Science 356, 968–972 (2017).
Article CAS PubMed Google Scholar
Rodgers, S. J., Jones, E. I., Mitchell, C. A. & McGrath, M. J. Sequential conversion of PtdIns3P to PtdIns(3,5)P2 via endosome maturation couples nutrient signaling to lysosome reformation and basal autophagy. Autophagy 19, 1365–1367 (2022).
Tan, J. X. & Finkel, T. A phosphoinositide signalling pathway mediates rapid lysosomal repair. Nature 609, 815–821 (2022).
Article CAS PubMed PubMed Central Google Scholar
Massana-Munoz, X. et al. Inactivating the lipid kinase activity of PI3KC2beta is sufficient to rescue myotubular myopathy in mice. JCI Insight 8, e151933 (2023).
Article PubMed PubMed Central Google Scholar
Anquetil, T. et al. PI3KC2beta inactivation stabilizes VE-cadherin junctions and preserves vascular integrity. EMBO Rep. 22, e51299 (2021).
Article CAS PubMed PubMed Central Google Scholar
Sabha, N. et al. PIK3C2B inhibition improves function and prolongs survival in myotubular myopathy animal models. J. Clin. Invest. 126, 3613–3625 (2016).
Article PubMed PubMed Central Google Scholar
Alliouachene, S. et al. Inactivation of the class II PI3K-C2beta potentiates insulin signaling and sensitivity. Cell Rep. 13, 1881–1894 (2015).
Article CAS PubMed PubMed Central Google Scholar
Hong, Z. et al. PtdIns3P controls mTORC1 signaling through lysosomal positioning. J. Cell Biol. 216, 4217–4233 (2017).
Article CAS PubMed PubMed Central Google Scholar
Walton, Z. E. et al. Acid suspends the circadian clock in hypoxia through inhibition of mTOR. Cell 174, 72–87 (2018).
Article CAS PubMed PubMed Central Google Scholar
Clippinger, A. J. & Alwine, J. C. Dynein mediates the localization and activation of mTOR in normal and human cytomegalovirus-infected cells. Genes Dev. 26, 2015–2026 (2012).
Article CAS PubMed PubMed Central Google Scholar
Korolchuk, V. I. et al. Lysosomal positioning coordinates cellular nutrient responses. Nat. Cell Biol. 13, 453–460 (2011).
Article CAS PubMed PubMed Central Google Scholar
Fernandes, S. A. et al. Spatial and functional separation of mTORC1 signalling in response to different amino acid sources. Nat. Cell Biol. 26, 1918–1933 (2024).
Article CAS PubMed PubMed Central Google Scholar
Jin, N. et al. Roles for PI(3,5)P2 in nutrient sensing through TORC1. Mol. Biol. Cell 25, 1171–1185 (2014).
Article PubMed PubMed Central Google Scholar
Shen, K. & Sabatini, D. M. Ragulator and SLC38A9 activate the Rag GTPases through noncanonical GEF mechanisms. Proc. Natl Acad. Sci. USA 115, 9545–9550 (2018).
Article CAS PubMed PubMed Central Google Scholar
Chen, Y. C. et al. N-myristoyltransferase-1 is necessary for lysosomal degradation and mTORC1 activation in cancer cells. Sci. Rep. 10, 11952 (2020).
Article CAS PubMed PubMed Central Google Scholar
Mansat, M. et al. MTM1-mediated production of phosphatidylinositol 5-phosphate fuels the formation of podosome-like protrusions regulating myoblast fusion. Proc. Natl Acad. Sci. USA 121, e2217971121 (2024).
Article CAS PubMed PubMed Central Google Scholar
Gavriilidis, C. et al. The MTM1–UBQLN2–HSP complex mediates degradation of misfolded intermediate filaments in skeletal muscle. Nat. Cell Biol. 20, 198–210 (2018).
Article CAS PubMed Google Scholar
Picot, M. et al. Biological replicates for immunoblots. Figshare https://doi.org/10.6084/m9.figshare.31168189 (2026).

Download references

Acknowledgements

We are grateful to the Institut National de la Santé et de la Recherche Médicale and the University of Paul Sabatier, Toulouse. We thank all the technical platforms of the Institute of Metabolic and Cardiovascular Diseases. We thank R. Flores from the Imaging Platform for ImageJ-Macro microscopy image analysis and D. Ramel for his helpful guidance in microscopy investigations. We thank M. Torjmane for her help in the revision of the manuscript. We thank the J. Laporte Lab and C. Kretz (IGBMC, Strasbourg) for their support and for animal care during the first proof-of-principle experiments. We gratefully acknowledge the Association Française Contre les Myopathies for financial support to K.H. (grant no. 23101) and M.P. PhD fellowship (grant no. 22115), as well as the Myotubular Trust for support to K.H.and M.P. (grant no. 20_INSERM_HNIAK), the ANR-PRC (grant no. MuscLY 2021–2024 to K.H. and G.P.), the ANR France Relance Programme (grant no. ANR-21-PRRD-0002-01 to K.H.), the Proteomics French Infrastructure ANR-10-INBS-08 and the Fonds Européens de Développement Régional Toulouse Métropole and the Région Midi-Pyrénées (to O.S. and A.S.), the E-Rare JTC 2017 TREAT-MTMs (to M.M. and B.P.), The China Scholarship Council (grant no. 201807060010 to P. L.), the German Research Foundation (grant no. 374031971-TRR 240 to M.L), the European Research Council (grant no. ERC-CoG-MetaboSENS-819543) and the ANR-PRC (grant no. MuscLY 2021–2024 to G.P.).

Author information

These authors contributed equally: Nesrine Hifdi, Mathilde Vaucourt.

Authors and Affiliations

UMR1297 Inserm/Université Toulouse, Institut des Maladies Métaboliques et Cardiovasculaires (I2MC), Toulouse, France
Melanie Picot, Nesrine Hifdi, Mathilde Vaucourt, Melanie Mansat, Gaëtan Chicanne, Caterina Miceli, Bernard Payrastre, Marie Vandromme, Julien Viaud & Karim Hnia
Institute of Pharmaceutical Sciences, Pharmaceutical (Bio-)Analysis, University of Tübingen, Tübingen, Germany
Peng Li & Michael Lämmerhofer
Division Cell Biology Biocenter Innsbruck, Center for Chemistry and Biomedicine, Innsbruck, Austria
Isabel Singer & Mariana E. G. de Araujo
IPBS, CNRS, UMR 5089, Proteomics and Mass Spectrometry of Biomolecules, Toulouse, France
Alexandre Stella & Odile Schiltz
Program in Molecular Medicine, University of Massachusetts Chan Medical School, Worcester, MA, USA
Shen Kuang
Institut Necker-Enfants Malades, INSERM U1151/CNRS UMR 8253, Université de Paris Cité, Paris, France
Nathaniel F. Henneman & Ganna Panasyuk
Platform for Metabolic Analyses, Structure Fédérative de Recherche Necker, INSERM US24/CNRS UAR 3633, Paris, France
Ivan Nemazanyy

Authors

Melanie Picot
View author publications
Search author on:PubMed Google Scholar
Nesrine Hifdi
View author publications
Search author on:PubMed Google Scholar
Mathilde Vaucourt
View author publications
Search author on:PubMed Google Scholar
Melanie Mansat
View author publications
Search author on:PubMed Google Scholar
Peng Li
View author publications
Search author on:PubMed Google Scholar
Isabel Singer
View author publications
Search author on:PubMed Google Scholar
Gaëtan Chicanne
View author publications
Search author on:PubMed Google Scholar
Alexandre Stella
View author publications
Search author on:PubMed Google Scholar
Shen Kuang
View author publications
Search author on:PubMed Google Scholar
Caterina Miceli
View author publications
Search author on:PubMed Google Scholar
Nathaniel F. Henneman
View author publications
Search author on:PubMed Google Scholar
Bernard Payrastre
View author publications
Search author on:PubMed Google Scholar
Marie Vandromme
View author publications
Search author on:PubMed Google Scholar
Odile Schiltz
View author publications
Search author on:PubMed Google Scholar
Ivan Nemazanyy
View author publications
Search author on:PubMed Google Scholar
Mariana E. G. de Araujo
View author publications
Search author on:PubMed Google Scholar
Ganna Panasyuk
View author publications
Search author on:PubMed Google Scholar
Julien Viaud
View author publications
Search author on:PubMed Google Scholar
Michael Lämmerhofer
View author publications
Search author on:PubMed Google Scholar
Karim Hnia
View author publications
Search author on:PubMed Google Scholar

Contributions

Conceptualization: K.H. and M.P. Methodology: K.H., M.P., M.M., N.F.H., M. Vaucourt., C.M., P.L., I.S., G.C., A.S., S.K., J.V., M. Vandromme., O.S., I.N. and M.L. Investigation: K.H., M.P., M.M., N.F.H., C.M., M. Vaucourt., P.L., I.S., G.C., A.S., M.E.G.A. and J.V. Visualization: M.P., N.F.H., M.V., B.P. and M. Vandromme. Funding acquisition: K.H. and G.P. Project administration: K.H. and G.P. Supervision: K.H. Writing the original draft: K.H. Corrections and suggestions: K.H., G.P., M.E.G.A. and M.L.

Corresponding author

Correspondence to Karim Hnia.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Source data

Extended Data Fig. 10 Physiological benefits of mTORC1 inhibition in the Mtm1-KO mouse model.

a, Serum levels of glucose (n = 8 mice per group), insulin (n = 10 mice per group), and IGF1 (n = 8 mice per group) measured in 3–4-week-old Mtm1-KO mice under fed and fasted conditions. Data are shown as mean ± s.e.m., One-way ANOVA followed by Šídák’s multiple-comparisons test. b, Food intake of CTRL (vehicle-treated) and Mtm1-KO mice during AZD8055 treatment. Data are shown as mean ± s.e.m., n = 8 mice per group. Two-sided Mann–Whitney test at the indicated time points. c,d, mTORC1 inhibition improves body weight, tibialis anterior (TA) muscle weight (n = 8 mice per group), and specific muscle force (n = 6 mice per group) in 5-week-old Mtm1-KO mice. Data are shown as mean ± s.e.m., One-way ANOVA followed by Tukey’s multiple-comparisons test. e, Representative images of muscle morphology from 5-week-old Mtm1-KO mice treated with AZD8055, assessed by hematoxylin and eosin (H&E), succinate dehydrogenase (SDH) staining, and transmission electron microscopy (TEM). White scale bars, 500 µm; black scale bars, 2 µm. f, Muscle fiber diameter distribution following AZD8055 treatment in Mtm1-KO mice. Data are shown as mean ± s.d., n = 6 mice per group. One-way ANOVA followed by Tukey’s multiple-comparisons test. g, Impact of mTORC1 inhibition in wild-type (WT) mice. Body weight was monitored during and after AZD8055 treatment. TA muscle weight and specific force were assessed at the experimental endpoint. Data are shown as mean ± s.e.m., n = 8 mice per group. h, Representative muscle morphology images (H&E and TEM) from WT mice treated with AZD8055, showing no overt alterations in muscle architecture, except for the expected induction of autophagy, evidenced by the presence of vesicular structures resembling autophagosomes. White scale bars, 500 µm; black scale bars, 2 µm. i, Muscle fiber diameter distribution following AZD8055 treatment in WT mice. Data are shown as mean ± s.e.m., n = 6 mice per group.

Source data

Supplementary information

Supplementary Information (download PDF )

Supplementary Figs. 1–3.

Reporting Summary (download PDF )

Supplementary Table 1 (download XLSX )

Lysosome proteomics of myotubes from Mtm1-KO versus CTRL.

Supplementary Tables 2 and 3 (download XLSX )

Antibody information and primers used for mutagenesis.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Picot, M., Hifdi, N., Vaucourt, M. et al. Lysosomal phosphoinositide turnover acts upstream of RagGTPase–mTORC1 and controls muscle growth. Nat Metab 8, 624–645 (2026). https://doi.org/10.1038/s42255-026-01484-1

Download citation

Received: 14 January 2025
Accepted: 09 February 2026
Published: 18 March 2026
Version of record: 18 March 2026
Issue date: March 2026
DOI: https://doi.org/10.1038/s42255-026-01484-1

Subjects

Abstract

Similar content being viewed by others

Main

Results

Protein synthesis is affected in congenital myopathies

MTM1 depletion promotes mTORC1 overactivation in differentiated cells

MTM1 represses the Ragulator–RagGTPase axis of mTORC1

MTM1 controls a lysosomal phosphoinositide pool that modulates the RagGTPase–mTORC1 axis

MTM1 lipid substrates PI3P and PI(3,5)P2 binds LAMTOR1/3 and modulate complex turnover

MTM1 controls PI3P and PI(3,5)P2 levels through ER–lysosome membrane contacts

Inhibition of mTORC1 in the XLCNM mouse model improves disease onset and progression

Discussion

Methods

Muscle primary cells (human and mouse)

Cell lines

siRNA- and shRNA-mediated gene knockdown

Drug treatment of cells

Protein synthesis measurements using HPG-488/AHA-594 assays and puromycin incorporation

IP and western blotting

Constructs and cloning

Protein purification

Lysosome(Lyso)-Tag cell line and Lysosome(Lyso)-IP

Immunofluorescence, PLA and imaging

Targeted LC-MS metabolomics analyses

Nano-liquid chromatography-tandem mass spectrometry analysis of lysosomal proteomes

Whole-cell and lysosome lipidomics

Phosphoinositide pull-down and flotation assay

Mice experiments

Muscle histology, morphometry and physiology

Statistics and reproducibility

Reporting summary

Data availability

Change history

01 April 2026

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Extended data

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

Search

Quick links