A defined N6-methyladenosine (m6A) profile conferred by METTL3 regulates muscle stem cell/myoblast state transitions

Muscle-specific adult stem cells (MuSCs) are required for skeletal muscle regeneration. To ensure efficient skeletal muscle regeneration after injury, MuSCs must undergo state transitions as they are activated from quiescence, give rise to a population of proliferating myoblasts, and continue either to terminal differentiation, to repair or replace damaged myofibers, or self-renewal to repopulate the quiescent population. Changes in MuSC/myoblast state are accompanied by dramatic shifts in their transcriptional profile. Previous reports in other adult stem cell systems have identified alterations in the most abundant internal mRNA modification, N6-methyladenosine (m6A), conferred by its active writer, METTL3, to regulate cell state transitions through alterations in the transcriptional profile of these cells. Our objective was to determine if m6A-modification deposition via METTL3 is a regulator of MuSC/myoblast state transitions in vitro and in vivo. Using liquid chromatography/mass spectrometry we identified that global m6A levels increase during the early stages of skeletal muscle regeneration, in vivo, and decline when C2C12 myoblasts transition from proliferation to differentiation, in vitro. Using m6A-specific RNA-sequencing (MeRIP-seq), a distinct profile of m6A-modification was identified, distinguishing proliferating from differentiating C2C12 myoblasts. RNAi studies show that reducing levels of METTL3, the active m6A methyltransferase, reduced global m6A levels and forced C2C12 myoblasts to prematurely differentiate. Reducing levels of METTL3 in primary mouse MuSCs prior to transplantation enhanced their engraftment capacity upon primary transplantation, however their capacity for serial transplantation was lost. In conclusion, METTL3 regulates m6A levels in MuSCs/myoblasts and controls the transition of MuSCs/myoblasts to different cell states. Furthermore, the first transcriptome wide map of m6A-modifications in proliferating and differentiating C2C12 myoblasts is provided and reveals a number of genes that may regulate MuSC/myoblast state transitions which had not been previously identified.


Introduction
Skeletal muscle regeneration is reliant upon a population of skeletal muscle-specific adult stem cells (MuSCs) and their ability to employ a myogenic program to form new skeletal muscle. In mature skeletal muscle, MuSCs remain in a quiescent state until activated by injury associated ques, at which point they enter the cell cycle, give rise to a transient myoblast population, and divide 1,2 . After a series of cell divisions, a subset of MuSCs will return to the quiescent state to renew the MuSC population that is able to respond to subsequent injuries while another subset of MuSCs will commit to the myogenic lineage and proceed to terminal differentiation 3 . MuSCs that differentiate, will restore damaged skeletal muscle through formation of new myofibers or fusion with existing myofibers. The process of MuSC differentiation and fusion is known to include the choreographed expression of several transcription factors termed the myogenic regulatory factors (MRFs) including myoblast determination protein 1 (MyoD), myogenic factor 5 (Myf5), and myogenin (Myog) 2 . The temporal sequence of the expression of each MRF is well documented however, less is known about how modification of these mRNAs influences their translation and regulatory functions.
Posttranscriptional modifications, particularly the most abundant mRNA modification, N6-methyladenosine (m 6 A), have rapid, pronounced effects on mRNA stability, splicing, and translation, and therefore a cell's transcriptional profile. m 6 A modification requires a defined set of "writer" proteins including the active methyltransferase, methyltransferase-like 3 (METTL3), to deposit methyl groups on mRNA 4 . Homozygous knockout of Mettl3 is embryonic lethal in mice 5 . Mettl3 knockout in embryonic stem cells, however, is viable and has been shown to increase cell proliferation and reduce differentiation capacity when cells are in the naive state 6 . Intriguingly, the opposite effects of Mettl3 knockout on cell state were observed in primed, pluripotent stem cells 7 . The effect of METTL3 depletion on stem cell state is also dependent on stem cell lineage. For example, reducing m 6 A levels in hematopoietic stem cells reduces proliferation and promotes differentiation 8 whereas increasing m 6 A levels in adult neural stem cells reduces differentiation 9 . Therefore, m 6 A modifications are central in regulating stem cell state, however, their role is lineage dependent.
To date, only two studies have evaluated the effects of m 6 A modification on myoblasts 10,11 . The first study evaluated the effect of depletion of the m 6 A "eraser," mRNA demethylase, Fat mass and obesity-associated protein (FTO), in the immortalized myoblast cell line, C2C12 myoblasts. Depletion of FTO led to impaired differentiation and fusion in vitro and impaired postnatal skeletal muscle development in vivo 10 . It is difficult to attribute these effects solely to reductions in m 6 A as FTO has recently been shown to act on m 6 A as well as N6,2-Odimethyladenosine (m 6 A m ), a distinct mRNA modification which occurs exclusively near the mRNA cap 12 . The second study used siRNA to knockdown the m 6 A writer, METTL3, which would be expected to have inverse results from knockdown of the m 6 A eraser, FTO, however a reduction in MyoD mRNA levels and protein expression of the terminal differentiation marker embryonic myosin heavy chain (eMHC) was observed 11 . Consequently, m 6 A modification appears to be implicated in C2C12 myoblast state transitions, but the directionality of the effect cannot be inferred from the available literature.
Our objective was to determine if m 6 A deposition on mRNA via Mettl3 is a regulator of MuSC/myoblast state in vitro and in vivo. This study defines the temporal dynamics of m 6 A levels in whole skeletal muscle tissue during the regeneration process, post injury, and in C2C12 myoblasts during in vitro proliferation and differentiation. Using liquid chromatography/mass spectrometry (LC-MS), we identified increased m 6 A levels in skeletal muscle tissue in the early stages of tissue regeneration compared to pre-injury. Furthermore, LC-MS analysis and m 6 A-specific RNA-sequencing (MeRIP-seq) demonstrated increased m 6 A in C2C12 myoblasts during proliferation compared to differentiation. MeRIP-seq of proliferating and differentiating C2C12 myoblasts reveals the first comprehensive profile of mRNAs that are preferentially m 6 A-modified in different myoblast states and reveals that m 6 A modification is closely linked to genes implicated in transcriptional regulation during proliferation. Finally, we show that Mettl3 knockdown slows proliferation in primary mouse MuSCs and enhances their primary transplant capacity.

Results
Global m 6 A levels decline during the MuSC transition from proliferation to differentiation Global m 6 A levels were assessed in skeletal muscle tissue after BaCl 2 induced injury using LC-MS. In whole skeletal muscle tissue, relative m 6 A levels increased at 3 days post injury (dpi) compared to baseline (Fig. 1a). 3 dpi corresponds to a time of substantial MuSC proliferation but also infiltration of a number of other cell types that support the skeletal muscle regenerative process [13][14][15] . Global m 6 A levels at later times points when MuSC differentiation is initiated (i.e., 5 and 10 dpi) were not different from baseline suggesting this increase in m 6 A levels was specific to MuSC proliferation. To understand the dynamics of global m 6 A levels at the MuSC/myoblast level, during this period of rapid proliferation, C2C12 myoblasts ( Fig. 1b) and primary mouse myoblasts ( Supplementary Fig. 1A) were used. m 6 A levels were greater when cells were cultured in proliferation favoring media for 24 h (GM) compared to when cells were switched from GM to serum restricted, differentiation inducing media for 3 days (3d DM) in both myoblast models. Because both myoblast models behaved similarly and due to the large amount of input required for downstream applications, C2C12 cells were used for subsequent cell culture experiments, unless otherwise specified.
MeRIP-sequencing reveals dynamic changes in transcriptspecific m 6 A-modification abundance but not the frequency of m 6 A-modification location on transcripts To map the changes in the transcriptome marked by m 6 A as myoblasts transition from proliferation to differentiation, MeRIP-seq was performed in GM and 3d DM samples. Confirming our LC-MS data, MeRIP-seq revealed an enrichment of m 6 A-modified transcripts in GM (vs. 3d DM) samples (Fig. 1c, Supplementary Tables 1, 2). The distribution of m 6 A across transcripts was similar in GM and 3d DM samples with modifications being most abundant at the boundary of the coding sequence and 3′ regions (Fig. 1d). Therefore, we concluded that a global-and transcript-specific program of m 6 A enrichment is associated with maintaining myoblasts in a proliferative state and that this effect is not driven by differences in the location of m 6 Amodification on transcripts.
Characterization of the proliferation-specific m 6 A profile at the individual transcript level MeRIP-seq analysis revealed 1607 m 6 A peaks representing 862 unique genes enriched in GM compared to 3d DM (Fig. 1c). The top m 6 A-enriched transcripts, based on fold change (GM to 3d DM), were Dmp1, Ccdc73, GM5141, Zfp316, Zfp157, Cebpa, Gprin1, Chst7, Ncam1, and Nlrx1 (Supplementary Table 1). Comparing the MeRIP-seq analysis with results from concurrently performed bulk RNA-seq (Supplementary Table 3) it was noted that only 14 transcripts were both m 6 A-enriched in GM compared to 3d DM (Supplementary Table 4) and differentially expressed in GM (vs. 3d DM). Of the 14 common transcripts, 7 were zinc finger proteins. To evaluate the effect of m 6 A-modification on individual gene expression, we compared whether having at least one m 6 A-modification predicted relative gene expression (Fig. 2a). Overall, transcripts with at least one enriched m 6 A-modification site in myoblasts during proliferation (i.e., GM samples) were moderately more likely to have increased gene expression; increased expression was determined by fold change >1.0 (GM/3d DM) for transcripts in the RNA-seq dataset (Fig. 2a).
To identify Molecular Function Pathways associated with the m 6 A-modified mRNAs that were overrepresented in GM, we performed PANTHER GO-slim Molecular Function analysis (Fig. 2b, Supplementary Table 5) 16 . An emerging theme from the PANTHER analysis was that m 6 A-modified transcripts that clustered into related Molecular Function Pathways were related to transcriptional regulation (e.g., transcription factor binding, transcription coregulator activity, and chromatin binding). We next performed PANTHER GO-slim analysis on differentially regulated transcripts during GM compared to 3d DM, based on bulk RNA-seq results (Supplementary Table 6); we identified nine overlapping Molecular Function Pathways between the RNA-seq and MeRIP-seq analyses ( Fig. 2c and Supplementary Table 7). Similar to our analysis at the single transcript level, the incomplete overlap of affected Molecular Function Pathways between the MeRIP-seq and RNA-seq datasets indicates that some of the functional outcomes of m 6 Amodifications during GM are beyond enhanced transcript stabilization and/or decay.

Characterization of the differentiation-specific m 6 A profile at the individual transcript level
We repeated the analysis for transcripts that emerged as m 6 A-enriched in the 3d DM MeRIP-seq dataset; 573 m 6 A peaks were identified as associated with 340 unique genes that were significantly enriched in 3d DM vs. GM samples (Fig. 1c). Using fold change, the top 10 m 6 A-enriched transcripts in 3d DM were Flcn, Zfp712, Zfp59, Zfp964, Atxn712, Gsdme, Taf12, Col23a1, Ankfy1, and Cep851 (Supplementary Table 2). Only six transcripts were both differently regulated in the RNA-seq dataset and m 6 Aenriched in 3d DM based on the MeRIP-seq dataset (Supplementary Table 8). Similar to GM, in 3d DM having at least one m 6 A-modification enrichment site present on transcript tended to increase the likelihood that the corresponding transcript was expressed to a greater degree in 3d DM vs. GM (Fig. 3a).
PANTHER GO-slim Molecular Function Pathway analysis revealed fewer Molecular Function Pathways related to transcriptional regulation than appeared in the GM dataset (Figs. 2b, 3b). The most prominent pathway enriched in 3d DM was microtubule binding (Fig. 3b and  Supplementary Table 8). Furthermore, unlike in GM, all Molecular Function Pathways identified in the MeRIP-seq dataset emerged as enriched in the RNA-seq dataset suggesting a closer link to m 6 A-modification and transcript abundance in 3d DM than in GM ( Fig. 3c and Supplementary Tables 5,9,10).

Mettl3 regulates the myoblast transition from proliferation to differentiation
We hypothesized that m 6 A levels were enriched in GM samples due to increased expression of the active m 6 A writer Mettl3. To model MuSC activity during regeneration, C2C12 myoblast and primary mouse myoblasts were cultured then prompted to differentiate in vitro by serum restriction; Mettl3 expression levels markedly declined in both C2C12 myoblast (Fig. 4a) and primary mouse myoblasts ( Fig. 4b) after differentiation initiation.
To determine if prevention of m 6 A-modification deposition alone was adequate to initiate differentiation, C2C12 myoblasts were treated with siRNA targeting Mettl3 (Fig. 4c), which was successful in reducing the frequency of m 6 A-modification (Fig. 4d). Knockdown of Mettl3 reduced cell count (Fig. 4e). We next evaluated the effect of Mettl3 knockdown on MRFs associated with proliferation (Pax7), early differentiation (Myod1), and late differentiation (Myog). Gene expression of Myod1 and Myog were significantly increased after Mettl3 knockdown (Fig. 4f). Finally, Mettl3 knockdown significantly increased the number of C2C12 myoblasts expressing eMHC, a marker of terminal differentiation (Fig. 4g), indicating a transition from proliferation to differentiation despite cells being cultured in proliferation favoring media. Altogether these data support a model whereby Mettl3 knockdown in proliferating C2C12 myoblasts reduces global levels of m 6 A-modification and causes premature differentiation of myoblasts.

Mettl3 knockdown affects primary MuSC engraftment
To test the effect of Mettl3 knockdown on primary mouse MuSCs, MuSCs were isolated from CMV-luc mice, treated with one of four different short-hairpin constructs targeting Mettl3 (shMettl3) or a nonmammalian control (shCON, Fig. 5a); the two most effective constructs at reducing Mettl3 levels were used for transplantation experiments. Knockdown of Mettl3 in primary mouse MuSCs effectively reduced luciferase expression (a surrogate for nuclei number, Fig. 5b) and percent confluence (Supplementary Fig. 2A) after 7 days of culture in a proliferation favoring media. When Mettl3 was knocked down prior to transplantation the likelihood for engraftment and bioluminescent expression of engrafted MuSCs improved (Fig. 5c, d). Additionally, we performed serial transplants to identify if MuSCs maintained engraftment potential; however, none of the shMettl3 or shCON transplants were capable of engraftment upon secondary transplantation, indicating no MuSCs were retained in a stem like state after primary transplantation (data not shown).

Discussion
In the experiments described here, we provide the first evidence that global m 6 A levels increase after skeletal muscle injury at a time point corresponding with rapid MuSC proliferation. Additionally, the use of LC-MS m 6 A measurement and MeRIP-seq build upon an earlier report that showed, using the RNA dot blot method, global m 6 A levels in primary mouse myoblasts or C2C12 myoblasts (see figure on previous page) Fig. 4 Mettl3 knockdown forces premature myoblast differentiation. a Mettl3 gene expression levels were assessed via qPCR in C2C12 myoblasts in GM as well 3d and 5d DM (n = 3). ***P < 0.001. b Gene expression levels of Mettl3 were measured via qPCR in GM as well 3d and 5d DM in primary mouse myoblasts (n = 3). ***P < 0.001. c siRNA targeting Mettl3 was applied to C2C12 myoblasts for 2 days in proliferation favoring media and protein levels (immunoblotting) and gene expression (qPCR) levels of Mettl3 demonstrated an efficient knockdown (n = 3). d mRNA was isolated from C2C12 myoblasts after 2 days of siMettl3 or siCON and global m 6 A-modification levels were determined via by LC-MS (n = 6). **P < 0.01. e Total C2C12 myoblast number was counted at the time points indicated after treatment with siMettl3 or siCON (n = 3). *P < 0.05. f Gene expression levels of Pax7, Myod1, and Myog were measured via qPCR after 2 days of siMettl3 or siCON (n = 3). *P < 0.05; ***P < 0.001. g Immunocytochemistry for eMHC in C2C12 cells (left) and quantification (right) after 2 days of siMettl3 or siCON (n = 3). ***P < 0.001.

Fig. 5
Mettl3 knockdown enhances MuSC engraftment after primary transplantation. a Primary mouse MuSCs were isolated from CMV-luc mice, treated with shRNAs targeting Mettl3 or a nonmammalian control sequence, and cultured for 7 days prior to measuring Mettl3 gene expression levels via qPCR (n = 4). ***P < 0.001; ****P < 0.0001. b Luminescence was measured in primary mouse MuSCs after 7 days of culture and treatment with shMettl3 or shCON (n = 4). *P < 0.05. c Schematic of primary mouse MuSC transplantation strategy; Primary mouse MuSCs were isolated from CMV-luc mice and transplanted into the TA of NSG knockout mice (n = 8 transplants per condition) and followed for 40 days. d Bioluminescence was measured at the time points indicated (left) and representative images are shown with the number of transplants with successful engraftment indicated (right). *P < 0.05; **P < 0.01 (main effect of shRNA, one-way ANOVA).
are increased during proliferation and decline during differentiation in vitro 10 . Our MeRIP-seq analysis extended previous findings showing that not only are global m 6 A levels greater in proliferation, but that there are also unique subsets of transcripts which are preferentially methylated in different cell states (i.e., proliferation and differentiation).
The transition from myoblast proliferation to differentiation is accompanied by drastic alterations in the transcriptome [17][18][19] . Transcriptional changes influencing MuSC/myoblast state transitions have been shown to be affected by a number of processes including changes in DNA and histone methylation of the MRFs 20 . This study provides the first, comprehensive identification of transcripts that are m 6 A-modified in myoblasts during proliferation and in early differentiation. We found no differences in methylation of transcription factors known to drive MuSC state changes such as Pax7, Myod1, or Myog suggesting that the MRFs are not direct targets of dynamic m 6 A-modification but rather, changes in their expression resulting in altered cell state is a secondary effect. Intriguingly, of the top m 6 A-enriched transcripts in GM and in 3d DM, many have not been previously linked to MuSC/myoblast function but have been demonstrated to regulate transcription and even cell state in other cell types. For example, a number of zinc finger proteins including Zfp316 and Zfp157 (in GM) as well as Zfp712, Zfp59, and Zfp964 (in 3d DM) were among the top ten transcripts differentially m 6 A-enriched during proliferation and differentiation, respectively. While many of the zinc finger proteins that emerged from the MeRIP-seq dataset have not been previously implicated in MuSC/ myoblast state transitions, the regulation of myogenic differentiation by other zinc finger proteins is well established 21 . Notably, Zfp637 was identified as an m 6 Aenriched transcript in GM, compared to 3d DM. Zfp637 repression is reported to suppress C2C12 myoblast proliferation and its over expression prevents C2C12 myoblast differentiation 22 . Furthermore, the PANTHER GOslim Molecular Function analysis revealed that 11/17 of the Molecular Function pathways that were overrepresented by m 6 A-enriched transcripts in GM (vs. 3d DM) were related to transcriptional regulation. Therefore, it appears likely that during proliferation, m 6 Amodification is used, in part, to affect the abundance of transcription regulating proteins, such as zinc finger proteins, which in turn define a transcriptional profile favoring cell cycle progression.
A dynamic pattern of Mettl3 expression during myoblast state changes was observed however, the upstream regulation of this expression remains elusive. m 6 A has been shown to be responsive to a number of external cues such as heat shock 23 or inflammatory challenges 24 at the cellular level or restraint stress at the organismal level 25 .
During skeletal muscle injury a number of changes to the MuSC microenvironment occur, including cytokine release from invading immune cell populations 26,27 and release of growth factors from the extracellular matrix 28 . It is likely that one or a combination of these factors is able to repress Mettl3 expression and drive the reduction in m 6 A levels observed, thereby facilitating the transition from proliferation to differentiation in MuSCs at the correct time.
Transplantation of Mettl3-knockdown MuSCs showed improved engraftment capacity compared to controls. Engraftment potential determined by bioluminescence imaging has been identified to be an approximate representation of maintenance of MuSC self-renewal, and therefore stem potential 29 , but is also influenced by the MuSCs' niche colonization capability 30 . Culturing of MuSCs in vitro prior to transplantation, as was done in this study, is documented to prompt the transition from a stem to a progenitor like state and impair engraftment potential after transplantation 29,31,32 . We hypothesize that this enhanced engraftment capacity of Mettl3-knockdown MuSCs is due to increased resistance to activation, and delayed progression toward a progenitor like state, prior to transplantation, resulting in enhanced niche colonization capability. The failure of secondary transplants, from any condition, to engraft supports a loss of stem potential and underscores that the increase in bioluminescent signal observed in Mettl3-knockdown MuSCs was likely due to enhanced niche colonization and not maintenance of a stem like state. A delay to activation of MuSCs due to Mettl3 knockdown supports a framework where the function of m 6 A-modification in MuSCs is to regulate transitions between cell states including quiescence to activation as well as proliferation to differentiation.
In conclusion, we elucidate METTL3-mediated m 6 A regulation as an important regulator of MuSC/myoblast state transitions and provide the first comprehensive map of m 6 A-modified transcripts during myoblast proliferation and differentiation. Future investigation into the biological function of the genes from the datasets developed here may glean new insight into the earliest stages of terminal differentiation commitment in MuSC/myoblasts. Furthermore, an understanding of the upstream signal that leads to the decline in Mettl3 transcription upon differentiation initiation would be invaluable in understanding how MuSC-microenvironment communication underlies transcriptional changes that regulate the initial steps in myogenic differentiation.

Mice and animal care
All mice were maintained according to the Guide for the Care and Use of Laboratory Animals and all procedures were approved by The Cornell University Institutional Animal Care and Use Committee. For skeletal muscle injury experiments, previously published B6. Cg-Pax7 tm1(Cre/ERT2)Gaka /J mice 33 (JAX: 017763) were crossed with B6N.129S6-Gt(ROSA)26Sor tm1(CAG-tdTomato*,−EGFP)/Ees /J mice 34 (JAX: 005304) to produce the following mouse: Pax7 CreERT2 ;Rosa26 nTnG (hereafter referred to as Pax7 nTnG ). Pax7 nTnG mice were used for all muscle injury experiments. Mice ubiquitously expressing luciferase, FVB-TG(CAG-luc,-GFP)L2G85Chco/J (JAX: 008450, CMV-luc mice) 35,36 , were used for the isolation of donor cells for transplantation experiments and were transplanted into immunodeficient NSG mice (NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ, JAX: 005557) 37,38 . Animals were randomly assigned to treatment groups and the treatment group of samples derived from each animal were blinded to the investigator performing downstream analyses. The number of animals per experimental condition are reported in the figure legends and sample sizes were based on convention in the field.

Skeletal muscle injury
To induce skeletal muscle injury, 50 µL of 1.2% BaCl 2 (Sigma, St. Louis, MO, USA) was injected into the tibialis anterior (TA) muscle of Pax7 nTnG mice under isoflurane anesthesia. Mice were sacrificed at 0 (baseline), 1, 3, 5, and 10 dpi and the TA muscle was snap frozen in liquid nitrogen and stored at −80°C until mRNA isolation. Cultures were not tested for mycoplasma contamination. When cultures were between 80-100% confluent they were switched to a media containing high glucose DMEM supplemented with 2% heat-inactivated horse serum to promote differentiation. GM samples were taken after 24 h of culture in growth medium when cultures werẽ 50% confluent and DM samples were taken at various days after the switch to differentiation media, as indicated. All C2C12 myoblasts were used between passage 5 and 10 and were cultured in a 5% CO 2 atmosphere at 37°C on collagen (Type I, Rat Tail, Corning, Pittston, PA, USA) coated plates. As per convention in the field, all experiments were repeated at least three times for all cell culture experiments (excluding MeRIP-seq experiments) with specific values noted in the figure legends.

Primary mouse myoblast culture
Passage 17-20 primary mouse myoblasts, obtained as previously described, were used 39 . Primary mouse myoblasts were cultured in a growth medium of 43% low glucose DMEM (Gibco, Gaithers MD, USA), 40% Ham's F10 (Gibco, Gaithers MD, USA), 15% fetal bovine serum, 1% penicillin-streptomycin, 1% Glutamax, and 2.5 ng/mL basic fibroblast growth factor (Promega, Madison, WI, USA). When cultures were between 80 and 100% confluent, they were switched to a media containing 2% heatinactivated horse serum. Similar to C2C12 myoblast experiments, GM samples were obtained at~50% confluency (24 h after seeding) and DM samples at various days after the switch to differentiation media. Primary mouse myoblasts were seeded on collagen (Type I, Rat Tail) coated plates and maintained in a 5% CO 2 atmosphere at 37°C. and m/z 268-136 for adenosine. The m 6 A to total adenosine (m 6 A + total adenosine) peak area ratio was calculated and differences in ratios reported.

MeRIP-sequencing
RNA for MeRIP-seq was isolated from C2C12 myoblasts using the Omega E.Z.N.A. ® Total RNA Kit I (Omega Bio-Tek Inc, Norcross, GA, USA). RNA quality was determined by denaturing agarose gel electrophoresis. mRNA was isolated from >120 µg total RNA using Arraystar Seq-Star TM poly(A) mRNA Isolation Kit (ArrayStar Inc, Rockville, MD, USA) as per the manufacturer's directions. Isolated mRNA was fragmented to a median size of 10 nt in 20 µL of a buffer containing 10 mM Zn 2+ and 10 mM Tris-HCl (pH 7.0) for 7 min at 94°C. Fragment size was confirmed by agarose gel electrophoresis. Ten percent of fragmented mRNA was saved for RNA-sequencing to serve as input control for MeRIPseq. The remaining fragmented mRNA was used for m 6 A immunoprecipitation. m 6 A immunoprecipitation was performed by combining fragmented mRNA with 2 µg anti-m 6 A antibody (Cat: 202 003, Synaptic Systems, Goettingen, Germany) and rotated head over tail for 120 min at 4°C in 500 µL total volume. Twenty microliters of Dynabeads TM M-280 sheep anti-rabbit IgG (clone 11204D, Thermo Fisher Scientific Waltham, MA, USA) were blocked with 0.5 mg/ mL bovine serum albumin at 4°C for 120 min. The blocked Dynabeads and fragmented mRNA/antibody mixture were combined for 120 min at 4°C. The solution was then washed 3× with a buffer of 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.1% NP-40 followed by 2× washes with a buffer comprised of 10 mM Tris-HCl (pH 7.4), 50 mM NaCl, and 0.1% NP-40. mRNA immunoprecipitated to the Dynabeads TM was separated by incubation with 200 µL of an elution buffer [10 mM Tris HCl (pH 7.4) 1 mM EDTA, 0.05% SDS, 40 U proteinase K (Qiagen, Hilden, Germany)] for 30 min at 50°C. mRNA was then extracted by phenol-chloroform and ethanol precipitated before sequencing.
The KAPA Stranded mRNA-seq Kit (Illumina, San Diego, CA, USA) was used to prepare the RNA-seq libraries for m 6 A immunoprecipitated mRNA and input control mRNA libraries. Quality of the libraries was determined using an Agilent 2100 Bioanalyzer. The HiSeq 3000/4000 PE Cluster Kit (Illumina, San Diego, CA, USA) was used to generate clusters from the prepared libraries following the manufacturer's instructions. The clustered libraries were sequenced on an Illumina HiSeq 4000 system.

MeRIP-sequencing analysis
The 5′, 3′-adaptor trimmed reads were aligned to the mm10 reference genome using HISAT2 (v2.1.0) 41 . ExomePeak was used for MeRIP peak calling and statistically significant MeRIP enriched regions (MeRIP peaks) were identified at the sample level with a significance threshold of P ≤ 0.05 42 . ExomePeak was also used to compare transcripts which were differentially m 6 Amodified between the GM and 3d DM samples with significance set at P ≤ 0.05 42 .

RNA-sequencing analysis
The input control RNA-sequencing results were subjected to differential gene expression analysis using edgeR 43 in RStudio (Version 3.6.1). Genes with at least two counts per million in at least three of the samples were kept for analysis.

Pathway analysis
Pathway analysis was conducted using the PANTHER (version 14.1) GO-Slim Molecular Function Overrepresentation Test with a false discovery rate (FDR) correction 16 . PANTHER analysis was performed using lists of m 6 A-modified transcripts that were enriched (P < 0.05) in GM and separately on lists of transcripts that were enriched in 3d DM based on the MeRIP-seq dataset. PANTHER analysis was also performed using the list of transcripts that were differentially regulated between GM and 3d DM according to the RNA-seq data (FDR < 0.05, log fold change >1.5 or <−1.5).

Immunoblotting
Cell lysates were collected using RIPA buffer containing protease (cOmplete, Roche, Basel, Switzerland) and phosphatase (PhosSTOP, Roche, Basel, Switzerland) inhibitors. Following cell lysis, the protein fraction was cleared of other cellular debris via centrifugation (12,000 × g, 15 min, 4°C). Ten micrograms of protein, as determined by the bicinchoninic acid assay (Thermo Fisher Scientific Waltham, MA, USA), was loaded on 10% SDS gels and transferred to PVDF membranes. Membranes were blocked with a chemiluminescent blocking buffer (bl∅k TM

C2C12 myoblast population expansion assay
To determine the effect of Mettl3 knockdown of cell number, C2C12 myoblasts that were cultured in 6-well cell culture dishes were counted using the Moxi Z Mini Automated Cell Counter (ORFLO Technologies, Ketchum, ID, USA) 12, 24, 48, and 72 h after seeding and siRNA transfection in growth medium.

Immunocytochemistry
Immunocytochemistry was used to determine expression of eMHC. C2C12 myoblasts were fixed for 20 min in 4% paraformaldehyde at RT before 3 × 5 min washes in PBS. Fixed, C2C12 myoblasts were blocked for 60 min at RT in 5% goat serum then incubated with primary eMHC antibody supernatant (DSHB Hybridoma Product F1.652, F1.652 was deposited to the DSHB by Blau, Helen M.) 44 overnight at 4°C. C2C12 myoblasts were then washed 3 × 5 min in PBS and incubated in secondary antibody (Alex Fluor ® 488 goat anti-mouse, Cat: A11029, Thermo Fisher Scientific, Waltham, MA, USA) diluted 1:1000 in 1% goat serum. Hoechst 33342 was used as a counterstain to identify nuclei and stained cells were imaged on the Celigo S Imaging Cytometer (Nexcelom Bioscience, Lawrence, MA, USA).

Primary mouse MuSC transplantation
For primary transplantation experiments primary MuSCs isolated from CMV-luc mice from each shRNA condition (shCON, sh2, sh3, and sh2/3) were individually resuspended in PBS at a concentration of 175 MuSCs per 20 µL PBS 24 h after transfection in culture. NSG knockout mice were anesthetized with isoflurane and 20 µL of the MuSC cell suspension was injected into their TA muscles. Each TA muscle received MuSCs from one shRNA condition (i.e., shCON, sh2, sh3, and sh2/3) in a random fashion for a total of 8 replicates per condition across 16 recipient mice.
For secondary transplantation, the TA muscles of primary transplant recipient mice were harvested and digested as described in the primary mouse MuSC isolation section. The cell suspension resulting from each TA muscle was evenly divided and injected into two recipient TA muscles for a total of 8 replicates per condition across 16 recipient mice.

In vitro luciferase assay
To determine the effect of shRNA conditions on MuSC proliferation after 7 days in culture, primary MuSCs from CMV-luc mice were incubated in fresh growth medium containing 0.15 mg/mL D-Luciferin (Gold Biotechnology, St. Louis, MO, USA) for 15 min at 37°C prior to measurement on a SpectraMax M3 plate reader (Molecular Devices, San Jose, CA, USA).

In vivo bioluminescent imaging
In vivo bioluminescent imaging was used to track MuSC behavior after transplantation. Transplant recipient mice were anesthetized using isoflurane and injected with 125 µL of 30 mg/mL D-Luciferin prior to imaging (Gold Biotechnology, St. Louis, MO, USA). Bioluminescent imaging was performed using the IVIS ® Spectrum in vivo imaging system (PerkinElmer, Waltham, MA, USA). Digital images were obtained 12 min after injections and analyzed using Living Image Software (PerkinElmer, Waltham, MA, US). The engraftment threshold was set at 50,000 photons per second. Bioluminescent images were taken at 3, 11, 17, 25, 31, and 40 days after primary transplantation. For secondary transplantation experiments, bioluminescent images were taken 35 days after transplantation.

Confluence analysis
Confluence measurements were performed as previously described 45 . Briefly, confluence was determined as the area of the cell culture surface covered by cells as measured by the Celigo S Imaging Cytometer (Nexcelom Bioscience, Lawrence, MA, USA).

Statistics
Statistical analyses for experimental data were performed in GraphPad (Version 1.0.136) with either twotailed unpaired t tests or one-way analysis of variance (ANOVA). Data are displayed as means ± standard deviation. Statistical significance was determined at P < 0.05. When a main effect of a one-way ANOVA was found to be statistically significant, a Tukey-Kramer post hoc test was conducted.