Skeletal muscle MACF1 maintains myonuclei and mitochondria localization through microtubules to control muscle functionalities

Skeletal muscle is made from multinuclear myofiber, where myonuclei are positioned at the periphery or clustered below neuromuscular junctions (NMJs). While mispositioned myonuclei are the hallmark of numerous muscular diseases, the molecular machinery maintaining myonuclei positioning in mature muscle is still unknown. Here, we identified microtubule-associated protein MACF1 as an evolutionary conserved regulator of myonuclei positioning, in vitro and in vivo, controlling the “microtubule code” and stabilizing the microtubule dynamics during myofibers maturation, preferentially at NMJs. Specifically, MACF1 governs myonuclei motion, mitochondria positioning and structure and acetylcholine receptors (AChRs) clustering. Macf1-KO in young and adult mice decreases muscle excitability and causes evolutionary myonuclei positioning alterations in adult mice, paralleled with high mitochondria content and improved resistance to fatigue. We present MACF1 as a primary actor of the maintenance of synaptic myonuclei and AChRs clustering, peripheral myonuclei positioning and mitochondria organization through the control of microtubule network dynamics in muscle fibers.


23
Skeletal muscle is made from multinuclear myofiber, where myonuclei are positioned at the 24 periphery or clustered below neuromuscular junctions (NMJs). While mispositioned myonuclei 25 are the hallmark of numerous muscular diseases, the molecular machinery maintaining 26 myonuclei positioning in mature muscle is still unknown. Here, we identified microtubule-27 associated protein MACF1 as an evolutionary conserved regulator of myonuclei positioning, in 28 vitro and in vivo, controlling the "microtubule code" and stabilizing the microtubule dynamics 29 during myofibers maturation, preferentially at NMJs. Specifically, MACF1 governs myonuclei 30 motion, mitochondria positioning and structure and acetylcholine receptors (AChRs) 31 clustering. Macf1-KO in young and adult mice decreases muscle excitability and causes 32 evolutionary myonuclei positioning alterations in adult mice, paralleled with high mitochondria 33 content and improved resistance to fatigue. We present MACF1 as a primary actor of the 34 maintenance of synaptic myonuclei and AChRs clustering, peripheral myonuclei positioning 35 and mitochondria organization through the control of microtubule network dynamics in muscle 36 fibers. 37 Folker, 2018). This precise organization of myonuclei is correlated with particular myonuclei 50 shape in muscle fibers, the alteration of which has recently emerged as potentially contributing 51 to several muscular diseases ( Proteins localized inside myonuclei such as Lamin A/C, Emerin or Nuclear Envelope Proteins 54 (NEPs), that couple the nuclear lamina to cytoskeleton through the "LInker of Nucleoskeleton 55 and Cytoskeleton" (LINC) complex, play critical roles in the maintenance of myonuclei shape 56 and localization (Chang et  MACF1 gene (inducing a decrease in the total amount of MACF1 protein) has been linked to a 77 new neuromuscular disease (Jørgensen et al., 2014). Also important in this respect, it was 78 recently proposed that Shot/MACF1 participates in the formation and maintenance of a 79 perinuclear shield in the muscle of drosophila larvae, contributing once again in myonuclei 80 localization (Wang et al., 2015). 81 Here, we demonstrate in vitro and in vivo that MACF1 is implicated in the maintenance of 82 myonuclei patterning in myofibers, starting at the NMJs, and controlling both myonuclei and 83 mitochondria dynamics through the regulation of microtubules. 84 85 86

88
MACF1 is not required in precocious myonuclei positioning but is essential during 89 myofibers maturation in the maintenance of myonuclei localization. 90 91 To identify new factors that contribute specifically to myonuclei spreading in myofibers, we 92 purified proteins able to bind to microtubules, stabilized with Taxol ® , from 3 days mouse 93 primary myotubes (Fig1A). The major Microtubule Associated Protein (MAP) identified by 94 mass-spectrometry using this protocol revealed the significant presence of MACF1/ACF7 95 protein, a member of the plakin family ( Supplementary Fig1A-B). Although MACF1 is initially 96 described to be predominantly expressed in neurons and muscle (Bernier et al., 2000), its role 97 during muscle fibers formation and behavior is poorly understood. Published work suggests 98 that plakin family members play specific roles through the regulation of placement and 99 function of specific organelles such as nucleus, mitochondria, Golgi apparatus, and 100 sarcoplasmic reticulum (Boyer et al., 2010). Mouse Macf1 mRNA was previously show to 101 increase steadily during myogenesis (Sun et al., 1999) and MACF1 was proposed to contribute 102 in muscle fiber to form a flexible perinuclear shield in collaboration with two other MAPs 103 named EB1 and Nesprin (Wang et al., 2015). We confirmed the increase of Macf1 mRNA and 104 MACF1 protein during early steps of myotubes formation (3 to 5 days of differentiation) using 105 RT-qPCR and Western blotting on mouse myoblast C2C12 cells (Supplementary Fig1C-D). 106 We first addressed the role of MACF1 in early steps of myotubes formation (Fig1A) Myonuclei localization in different type of muscles was studied using a histological staining 244 approach in muscle cross-section. Laminin staining was used to determine the limit of each 245 myofiber and Dapi to detect myonuclei (Fig4A). We observed a dramatic delocalization of 246 myonuclei in 12-month-adult muscle fibers revealed by the presence of myonuclei in the center 247 or dispatched in the myofibers rather than at the periphery in the Tibialis Anterior, Soleus and 248 Gastrocnemius with a respective increase of 275, 153 and 230% of the delocalization index 249 (Fig4B, Supplementary Fig3F). Interestingly, no significant delocalization of myonuclei was process as evidenced by the absence of staining for embryonic myosin heavy chain in 257 myofibers with mispositioned myonuclei (data not shown). We confirmed the accumulation of 258 mislocalized myonuclei in myofibers from 12-month-adult mice, using single isolated 259 myofibers from the Tibialis Anterior or Extensor Digitorum Longus muscles (Fig4F-G, green 260 arrows). Altogether, these results demonstrate that MACF1 is implicated in the maintenance of 261 myonuclei localization in muscle myofibers of adult mice that is not associated to a 262 regeneration process. This role for MACF1 seems to be a long-term process since no 263 significant myonuclei mis-localization is observed in younger mice. 264 265 MACF1 muscle-KO mice exhibit early neuromuscular junction alteration with a reset of 266 the related "muscle-microtubule code".

268
Our in vitro results suggested that MACF1 is implicated in myonuclei spreading in an 269 Agrin/NMJ dependent manner (Fig1-2). Thus, we wondered if in our in vivo conditional mutant 270 model, synaptic AChRs clustering size and density were impacted by the absence of MACF1.

271
To address this question, we isolated individual's myofibers from the Tibialis Anterior and the 272 Extensor Digitorum Longus muscles and visualized AChRs clustering using bungarotoxin 273 staining in 12-months-adult mice (Fig5A). In this condition, we found that the size of synaptic 274 AChRs clusters was highly diminished in both muscles. To determine if the fragmentation of 275 NMJ-related AChRs was an early process, we isolated individual's myofibers from the Rectus 276 Lateralis muscle, known to be multi-innervated, allowing an access to multiple NMJs, from 4-277 months-young mice ( Fig 5B). In this muscle, NMJs were much more fragmented in PTMs, Tubulin de-tyrosination is associated with longer-lived microtubules, whereas more 294 dynamic microtubules are found to be mainly tyrosinated (Bulinski and Gundersen, 1991). To 295 see if MACF1 plays a role in the correct formation of tubulin code, we addressed changes in 296 microtubule tyrosination/de-tyrosination status in conditionally mutant mice compared to 297 control mice (Fig5E). In Macf1 f/f Cre-muscle, tyrosinated-tubulin was found to locate 298 preferentially at the vicinity of myonuclei along myofibers. No particular enrichment was 299 found close to the AChRs clusters at the NMJ (Fig5E, top). Nonetheless, de-tyrosinated-tubulin 300 was found at the vicinity of myonuclei with a noticeable enrichment at the NMJ, close to the 301 AChRs clusters (Fig5E, top). In conditionally mutant mice (Macf1 f/f Cre+), although myonuclei 302 shape was changed, tyrosinated-tubulin was still found at the vicinity of myonuclei associated 303 with few aggregates along myofibers, but this time, the presence of tyrosinated-tubulin was 304 enriched at the vicinity of AChRs clusters. In the same perspective, de-tyrosinated-tubulin was 305 highly reduced both around myonuclei and near AChRs clusters (Fig5E, bottom). These data 306 show that MACF1 plays a role in the maintenance of the pool of de-tyrosinated-tubulin which 307 consequently stabilizes the microtubule network at both the vicinity of myonuclei and at the 308 NMJs.

310
Adult MACF1 muscle-KO mice exhibit increased myofibers with high mitochondria 311 content.

313
Our in vitro results suggested that MACF1 is implicated in mitochondria spreading and 314 fragmentation in muscle fibers (Fig2A, D-E). Since skeletal muscles are composed of a 315 functional and metabolic continuum of slow (type I) and fast fibers (types IIa and IIx), we first 316 questioned whether there were changes in the mitochondrial pool in adult conditional mice. We 317 measured the intensity of succinate dehydrogenase staining, indicative of mitochondrial activity 318 and found a significant increase in the number of fibers with positive succinate dehydrogenase 319 staining in Tibialis Anterior muscle in conditionally mutant mice (Macf1 f/f Cre+) compared to 320 control mice, while no effect was observed in Soleus muscle (Fig6A-B). Quantification of 321 proteins comprising the electron transport chain was then performed on Gastrocnemius muscles 322 and our results confirmed an increase in total amount of mitochondria (Tom20 relative to actin) 323 and no changes in protein contents from the electron transport chain (CI, II, III, IV and V 324 relative to Tom20) (Fig6C). The intensity of succinate dehydrogenase staining is commonly 325 used to discriminate slow and fast fibers. Since our quantification suggests a change in the 326 repartition of slow/fast fibers in Macf1 f/f Cre+ muscle, we next investigated if the proportion of 327 slow fibers was changed in conditional mutant mice (Macf1 f/f Cre+). Cross-sections of 12-328 month-adult muscles from the Tibialis Anterior, Soleus and Gastrocnemius were analyzed for 329 slow myosin content and no alteration of the proportion in each muscle type was observed 330 compared to control mice (Fig6D-E). We then used electron microscopy to visualize the 331 ultrastructure of mitochondria in Tibialis Anterior muscle of Macf1 f/f Cre+ mice (Fig6F). This 332 approach confirmed an apparent increase in mitochondria content in muscle fibers associated 333 with the presence of spherical mitochondria in-between myofibrils and close to myonuclei 334 (Fig6F). All together, these data show that MACF1 is involved in the maintenance of 335 mitochondria architecture in muscle fibers and that the loss of MACF1 is associated with a 336 redistribution of mitochondrial content dependent on muscle type. 337 338 there was also no change in the voltage-dependent Ca 2+ channel activity of the dihydropyridine 385

Muscle fibers from MACF1-KO mice show muscle function defects and increased fatigue
receptor (also referred to as Ca V 1.1, the voltage-sensor of excitation-contraction coupling) in 386 fibers from Macf1 f/f Cre+ mice (not shown). All in all, these data show that muscle from Macf1 387 conditional mutant mice exhibit delay in excitability in response to nerve stimulation and 388 although the T-tubule network organization seems affected in some myofibers, the global 389 efficiency of SR-Ca 2+ release seems not engaged. Interestingly, muscles from Macf1 390 conditional mice display a higher resistance to fatigue compared to muscles of control mice. However, processes involved in maintenance of myonuclei at the periphery of muscle fibers are 404 still unknown.

406
The present study is the first to demonstrate that a protein related to the microtubule network 407 controls and maintains peripheral myonuclei localization in mature myofibers. Here, we show 408 that MACF1 is progressively accumulated at the onset of myotubes formation, with no evident 409 implication in early steps of myonuclei spreading or myotubes elongation but remarkable 410 actions in maturation steps to set the distance between adjacent myonuclei (Fig1-2). 411 Accordingly, conditionally mutant mice (Macf1 f/f Cre+) exhibit a strong decrease in peripheral 412 myonuclei, but surprisingly, this alteration becomes significant only at 12-months of age 413 (Fig4). This alteration comes after a slight reduction in myonuclear domains in myofibers 414 reflected by a decrease in the mean diameter of myofibers in 8-month-adult mice. These data 415 suggest a complex interplay between factors that maintain peripheral myonuclei localization 416 and demonstrate that the absence of MACF1 contributes to a long-term modification of this 417 equilibrium that breaks in 12 month-adult Macf1 f/f Cre+ mice.

419
It has already been shown that atrophy can result from a progressive alteration of NMJs 420 (Bonaldo and Sandri, 2013 Table  634 showing mass spectrometry identification of MACF1 and Kif4b in 3 days-old primary Movie 3. Nuclear movements in primary myotubes transfected with a scrambled shRNA 681 (green) and lamin-chromobody® (red). Picture were recorded every 15 min. Movie correspond 682 to Figure 3C sh-Scramble. 683 684 Movie 4. Nuclear movements in primary myotubes transfected with a pool of 4 individual 685 shRNAs targeting Macf1 (green) and lamin-chromobody® (red). Picture were recorded every 686 15 min. Movie correspond to Figure 3C sh-MACF1. 687 688

Production of wild type Agrin recombinant proteins 709
For production of recombinant proteins, the stably transfected HEK293-EBNA cell lines were 710 grown to about 80 % confluence and were transferred to expression medium without FBS. 711 Conditioned medium containing secreted proteins was collected every 3 days and replaced with 712 fresh expression media for 12 days. Conditioned medium was centrifuged at 2,000xg for 10 713 min to pellet the cells before storing at -20°C. After thawing, recombinant proteins were 714 purified from conditioned media by HPLC, on a HiTRAP Imac HP column (GE Healthcare, For primaries cells, siRNA were transfected using Lipofectamine 2000 (ThermoFisher 728 Scientifics, 11668-019) at the final concentration of 2nM. shRNA (Geneocopia), Eb1 or RFP-729 Lamin-chromobody (Chromotek) cDNA were transfected in cells using Llipofectamine 3000 730 (ThermoFisher Scientifics, L3000-008). 731 732 siRNA Sense oligonucleotide sequence Anti-Sense oligonucleotide sequence Protein sample preparation 734 735 For primary cultured cells or C2C12 cell lines, cells were harvested, using 1X Trypsin for 5min 736 at 37°C and centrifuged at 1500RPM for 5min at 4°C. Cell pellets were diluted and incubated 737 in the optimal volume of RIPA lysis buffer containing phosphatases inhibitors (Sigma, P5726-738 5mL) and proteases inhibitors (Sigma, P8340) for 10min at 4°C. Following a sonication and a 739 centrifugation at 12000RPM for 10min at 4°C, protein samples were collected for further uses. 740 The concentration of proteins was determined using BCA protein assay kit (Thermo Fisher 741 Scientifics, 23225) as described by the manufacturer.

743
Western blot and dot blot analysis 744 To carry out western blots, the same amount of sample were loaded in 6% acrylamide gels and 745 were migrated at 130V for 10min followed by 160V for 90min. iBlot 2 mini slacks (Thermo 746 Fisher Scientifics, IB23002) semi-dry system was used to transfer the proteins to nitrocellulose 747 membranes. Membranes were then saturated in 5% milk in TBS for 1H at RT and were 748 incubated in primary antibodies in 5% milk in TBS over night at 4°C. Following washes by 749 0.1% Tween-20-1X TBS, the membranes were incubated in HRP conjugated secondary 750 antibodies in 5% milk in TBST for 1H at RT. Following washes by 0.1% Tween-20-1X TBS 751 the detection of the target proteins was carried out using Super Signal West Femto (Thermo 752 Fisher Scientifics, 34095) and ChemiDoc imaging system (BioRad).

754
To carry out dot blots, Bio-Dot SF Microfiltration Apparatus plates (BioRad) were used to 755 transfer the protein samples onto the nitrocellulose membranes. Membranes were then saturated 756 in 5% milk in TBS for 1H at RT and were incubated in primary antibodies over night at 4°C. 757 Following washes by 0.1% Tween-20-1X TBS, the membranes were incubated in HRP 758 conjugated secondary antibodies in 5% milk in TBST for 1H at RT. Following washes by 0.1% 759 Tween-20 in TBS the detection of the target proteins was carried out using Super Signal West 760 Femto (Thermo Fisher Scientifics, 34095) and ChemiDoc imaging system (BioRad).

762
Primary cells immunofluorescence staining 763 764 Cells were fixed in 4%PFA in PBS for 20min at 37°C followed by washes with PBS and 765 permeabilization with 0.5% Triton-X100 in PBS for 5min at RT. Following washes with PBS, 766 cells were saturated with 1% BSA in PBS for 30min at 37°C and incubated in primary 767 antibodies over night at 4°C. Following washes with 0.05% Triton-X100-1X PBS, cells were 768 incubated in secondary antibodies or dyes for 2H at RT followed by washes with 0.05% Triton-769 X100 in PBS and a last wash in PBS. Cultured myofibers were imaged using either Z1-770 AxioObserver (Zeiss) or confocal SP5 microscope (Leica).

772
Isolation of mono-myofibers and immunofluorescence staining 773 774 Following the dissection of the whole muscle from the mice, TA or EDL muscle blocks were 775 fixed in 4%PFA in PBS for 2H at RT. After washes, 30 to 50 mono-myofibers were isolated 776 per staining from each muscle. Myofibers were then permeabilized using 0.5% Triton-X100 in 777 PBS for 5min at RT and saturated in 1% BSA in PBS for 30min at RT. Based on the 778 experiments, myofibers were incubated in desired primary antibodies at 4°C over night.

779
Following washes with 0.05% Triton-X100 in PBS, myofibers were incubated in secondary 780 antibodies or dyes for 2H at RT followed by washes with 0.05% Triton-X100 in PBS and a last 781 wash in PBS. Myofibers were mounted on slides using fluromount Aqueous mounting (Sigma, 782 F4680-25mL) and kept at 4°C or -20°C. Slides were analyzed using confocal SP5 microscope 783 (Leica) or TI-Eclipse (Nikon Time-lapse images were acquired using Z1-AxioObserver (Zeiss) with intervals of 15minutes. 791 Final videos were analyzed using Metamorph (Zeiss) and SkyPad plugin as described before 792 (Cadot et al., 2014).

794
Electronic-Microscopy 795 796 Tissues were cut into small pieces and fixed in 2% glutaraldehyde for 2H at 4°C. Shot GD9507 UAS-RNAi line from VDRC collection crossed to Mef2-GAL4 driver has been 810 used to attenuate shot gene expression specifically in muscles. Third instar larvae were 811 dissected in physiological salt with 25mM EDTA. Body wall muscles were fixed with 4% 812 formaldehyde in PBS for 15min and then rinsed three times for 5min each in PBS with 0.5% 813 Tween 20 (PBT). Muscles were blocked for 30min with 20% horse serum in PBT at RT.

814
Staining was performed by using primary antibodies applied overnight at 4°C and after washing 815 3 times in PBT secondary antibodies were applied at RT for 1H. The following primary 816 antibodies were used: anti-Brp1 (1:100; DSHB, Nc82-s), anti-Shot (1:100; DSHB, mAbRod1). 817 818 Mouse model 819 The following mice have been described previously, mice that carry a loxP-flanked allele of 820 Macf1 (Goryunov et al., 2010) and Hsa-Cre transgenic mice (Miniou et al., 1999 solution. Estimation of the T-tubule density from the di-8-anepps fluorescence was carried out 880 from a largest possible region of interest excluding the plasma membrane, within each fiber. 881 For each fiber, two images taken at distinct locations were used. Analysis was carried out with 882 the ImageJ software (National Institute of Health). Automatic threshold with the Otsu method 883 was used to create a binary image of the surface area occupied by T-tubules. The "skeletonize" 884 function was then used to delineate the T-tubule network. T-tubule density was expressed as 885 the percent of positive pixels within the region. Sarcomere length was estimated from half the 886 number of fluorescence peaks (T-tubules) along the length of the main axis of a given fiber. To 887 assess variability in T-tubule orientation, objects within two T-tubule binary images of each 888 fiber were outlined and particle analysis was performed to determine the angle of all objects 889 yielding a perimeter larger than an arbitrary value of 10µm. For each fiber, the standard 890 deviation of angle values was then calculated. This analysis was performed on 10 muscle fibers 891 from 3 Macf1f/f Cre-and from 3 Macf1f/f Cre+ mice, respectively.

893
Intracellular Ca 2+ in voltage-clamped fibers 894 895 Single fibers were isolated from FDB muscles as described previously (Jacquemond, 1997). In 896 brief, muscles were incubated for 60min at 37 °C in the presence of external Tyrode containing 897 2 mg.mL collagenase (Sigma, type 1). Single fibers were obtained by triturating the 898 collagenase-treated muscles within the experimental chamber. 899 Isolated muscle fibers were handled with the silicone voltage-clamp technique (Lefebvre et al., 900 2014). Briefly, fibers were partly insulated with silicone grease so that only a short portion (50-901 100µm long) of the fiber extremity remained out of the silicone. Fibers were bathed in a 902 standard voltage-clamp extracellular solution containing (in mM) 140 TEA-methanesulfonate, 903 2.5 CaCl2, 2 MgCl2, 1 4-aminopyridine, 10 HEPES and 0.002 tetrodotoxin. An RK-400 patch-904 clamp amplifier (Bio-Logic, Claix) was used in whole-cell configuration in combination with 905 an analog-digital converter (Axon Instruments, Digidata 1440A) controlled by pClamp 9 906 software (Axon Instruments). Voltage-clamp was performed with a micropipette filled with a 907 solution containing (in mM) 120 K-glutamate, 5 Na2-ATP, 5 Na2-phosphocreatine, 5.5 MgCl2, 908 15 EGTA, 6 CaCl2, 0.1 rhod-2, 5 glucose, 5 HEPES. The tip of the micropipette was inserted 909 through the silicone within the insulated part of the fiber and was gently crushed against the 910 bottom of the chamber to ease intracellular equilibration and decrease the series resistance. 911 Intracellular equilibration of the solution was allowed for 30min before initiating 912 measurements. Membrane depolarizing steps of 0.5s duration were applied from -80mV. 913 Confocal imaging was conducted with a Zeiss LSM 5 Exciter microscope equipped with a 63x 914 oil immersion objective (numerical aperture 1.4). Rhod-2 fluorescence was detected in line-915 scan mode (x,t, 1.15 ms per line) above 560nm, upon excitation from the 543nm line of a HeNe 916 laser. Rhod-2 fluorescence transients were expressed as F/F0 where F0 is the baseline 917 fluorescence. The Ca2+ release flux (rate of SR Ca2+ release) was estimated from the time 918 derivative of the total myoplasmic Ca2+ ([Catot]) calculated from the occupancy of 919 intracellular calcium binding sites following a previously described procedure (Kutchukian et 920 al., 2017).

922
In vivo force measurements 923 Mice were initially anesthetized in an induction chamber using 4% isoflurane. The right 924 Hindlimb was shaved before an electrode cream was applied at the knee and heel regions to 925 optimize electrical stimulation. Each anesthetized mouse was placed supine in a cradle 926 allowing for a strict standardization of the animal positioning. Throughout a typical experiment, 927 anesthesia was maintained by air inhalation through a facemask continuously supplied with 928 1.5% isoflurane. The cradle also includes an electrical heating blanket in order to maintain the 929 animal at a physiological temperature during anesthesia. Electrical stimuli were delivered 930 through two electrodes located below the knee and the Achille's tendon. The right foot was 931 positioned and firmly immobilized through a rigid slipper on a pedal of an ergometer 932 (NIMPHEA_Research, AII Biomedical SAS) allowing for the measurement of the force 933 produced by the Hindlimb muscles (i.e., mainly the Gastrocnemius muscle). The right knee 934 was also firmly maintained using a rigid fixation in order to optimize isometric force 935 recordings. Monophasic rectangular pulses of 0.2ms were delivered using a constant-current 936 stimulator (Digitimer DS7AH, maximal voltage: 400V). The force-frequency curves were 937 determined by stepwise increasing stimulation frequency, with resting periods > 30s between 938 stimuli in order to avoid effects due to fatigue. For each stimulation train, isometric peak force 939 was calculated. After a 3-min recovery period, force was assessed during a fatigue protocol 940 consisting of 30Hz stimulation trains of 0.3s delivered once every second for 180s. The peak 941 force of each contraction was measured and averaged every 5 contractions. A fatigue index 942 corresponding to the ratio between the last five and the first five contractions was determined. 943 Force signal was sampled at 1000Hz using a Powerlab system and Labchart software 944 (ADinstruments). 945

Quantification methods for myonuclei spreading in myotubes 946
Quantifications in immature myotubes were assessed using an analysis tool developed in our 947 team. An image analysis performed in ImageJ® software is combined with a statistical analysis 948 in RStudio® software. This provides quantifications of parameters, ranked by myonuclei 949 content per myotubes, regarding phenotype of myotubes (area, length) and their respective 950 myonuclei positioning compare to centroid of myotubes (DMcM). 951 MSG diagrams were obtained through the normalization of lengths of all analyzed myotubes 952 (independently to their myonuclei content) to 100%. White lines represent myonuclei density 953 curves assessing the statistical frequency for myonuclei positioning along myotubes. Each color 954 group reflects statistical estimation of myonuclei clustering along myotubes.

956
Quantification methods for EB1 Comets 957 To determine EB1 comets speed, four continuous frame of a time-lapse movie of EB1-GFP 958 were overlapped, the first two are color-coded in green and the last two are color-coded in red.

959
Comets length was counted if green and red length were equal, traducing a comets growing in 960 the same focal plan. 961

Quantification of mitochondria fragmentation 962
In order to analyze mitochondria repartition, Cytochrome-C staining was used as representative 963 of the mitochondrial network in long differentiated myofibers. Representative images were 964 taken from myofibers and the analysis of images was done using ImageJ® software. One or 965 several Regions Of Interest (ROI) were selected per image, these regions were set just next-to 966 or close to myonuclei. Following a threshold determination, the whole information of each 967 stained particle (representing a single or a group of mitochondria) was extracted from ImageJ 968 software. The particles with circularity equivalent to 1 were eliminated in order to purify our 969 results from any unwanted background errors. Particles with feret less than 0.75µm (mean size 970 of mitochondria) were eliminated to purify our results. Finally, the ratio of number of particles 971 per area of ROI was calculated.

973
Quantification of AChRs fragmentation 974 Bungarotoxin staining was used to analyze formation of acetylcholine receptors clusters in 975 mature myofibers. Representative images were taken from myofibers of each condition as 976 described before and the analysis of images was done using ImageJ® software. Briefly, a 977 threshold of 8µm was set as the minimal size for AChRs clusters. Myofibers with at least one 978 cluster of 8µm or bigger were considered as positive. 979 980 981 Statistical analysis 982 The statistical analyses were performed using Student's t test or the Mann-Whitney test 983 according to samples amount and distribution 984