Generation of a MyoD knock-in reporter mouse line to study muscle stem cell dynamics and heterogeneity

Summary Myoblast determination protein 1 (MyoD) dynamics define the activation status of muscle stem cells (MuSCs), aiding in muscle tissue regeneration after injury. However, the lack of experimental platforms to monitor MyoD dynamics in vitro and in vivo has hampered the investigation of fate determination and heterogeneity of MuSCs. Herein, we report a MyoD knock-in (MyoD-KI) reporter mouse expressing tdTomato at the endogenous MyoD locus. Expression of tdTomato in MyoD-KI mice recapitulated the endogenous MyoD expression dynamics in vitro and during the early phase of regeneration in vivo. Additionally, we showed that tdTomato fluorescence intensity defines MuSC activation status without immunostaining. Based on these features, we developed a high-throughput screening system to assess the effects of drugs on the behavior of MuSCs in vitro. Thus, MyoD-KI mice are an invaluable resource for studying the dynamics of MuSCs, including their fate decisions and heterogeneity, and for drug screening in stem cell therapy.


INTRODUCTION
Skeletal muscles display a remarkable regenerative capacity, which depends on skeletal muscle stem cells (MuSCs), also known as satellite cells, found around the myofibers and underneath the basal lamina. 1 MuSCs represent myogenic progenitors that supply nuclei to growing myofibers during development and regeneration. 2 In adult muscles, MuSCs enter a mitotically quiescent state. However, in response to injury, they are activated and enter the cell cycle to rapidly expand myogenic progenitors that either differentiate into new myofibers or undergo self-renewal for lifelong maintenance of the MuSC pool. 3,4 MuSC dynamics are regulated by multiple transcription factors sequentially expressed during myogenesis. Quiescent MuSCs express members of the paired homeodomain family of transcriptional factors paired box protein Pax7, 5 and a small subset of quiescent MuSCs express Pax3, the paralog of Pax7. [6][7][8] Myogenic regulatory factors (MRFs) comprising myoblast determination protein 1 (Myod1, MyoD) and myogenic factor 5 (Myf5) are fundamental regulators of skeletal muscle lineage determination and development and are also induced in activated MuSCs during regeneration in adult mice. Interestingly, quiescent MuSCs are primed to rapidly activate the myogenic program while remaining quiescent. [9][10][11] These unique paradoxical features are mainly achieved by the translational suppression of MyoD and Myf5 mRNAs with microRNAs and RNA-binding proteins (RBPs). Myf5 transcripts are repressed by miR-31 and fragile X mental retardation protein associated with cytoplasmic RNA granules. 9,12,13 MyoD transcripts are repressed by the activity of RBPs, such as Zfp36 and Staufen1 (Stau1), through their 3 0 -UTR. 14,15 Upon MuSC activation, MyoD and Myf5 transcripts are translated to initiate the myogenic program by relieving this translational block.
The functions of MyoD and Myf5 as established transcriptional factors for myogenetic lineage specification are partially redundant and contribute to establishment of the myogenic progenitors during embryogenesis. [16][17][18][19][20][21] In addition, MyoD plays a unique role in adult muscle regeneration, as reported by the delayed onset of myogenic differentiation and increased propensity for self-renewal in MyoD knockout (KO) mice, suggesting that MYOD regulates the balance between self-renewal and differentiation during regeneration. [22][23][24][25] Furthermore, MyoD KO mice on a mdx background, a model of Duchenne muscular dystrophy, shows poor survival and die around the age of 12 months 22 MYOD has also been used as an activation marker for MuSCs during myogenic development and regeneration. The combination of MYOD and PAX7 immunostaining is frequently used to define the divergent fates of MuSCs during myogenesis and culture. Quiescent MuSCs are PAX7 positive (+) but MYOD-negative (À), whereas activated MuSCs co-express PAX7 and MYOD. Most activated MuSCs then proliferate, downregulate PAX7, and differentiate by expressing MYOGENIN, a member of the MRFs. In contrast, a small number of PAX7 (+), MYOD (+) activated MuSCs downregulate MYOD while maintaining PAX7 to return to the state resembling quiescence, which represents ''reserve stem cell''. [26][27][28] The mechanism by which activated MuSCs determine cell fate to proceed or return to the myogenic program remains unclear. Particularly, although MuSCs are maintained in an undifferentiated state by manipulating substrate elasticity or p38 Mitogen-activated Protein Kinase (MAPK) activity, [29][30][31] stimulating and maintaining the reserve stem cell population, a defining feature of stem cells in vitro and in vivo, remains a major challenge in muscle stem cell research.
Since Pax7 and MyoD are transcription factors, a major challenge is the lack of an efficient tool to isolate and analyze immunolabeled MuSC populations at different myogenic stages in vitro and in vivo. Recently, two Pax7-knock-in (KI) reporter mouse lines have been generated. 32,33 The Pax7-KI mouse line is a valuable tool for isolating quiescent MuSCs and studying a heterogeneous population of quiescent MuSCs based on PAX7 expression levels. However, several limitations of the Pax7-KI mouse line hinder our understanding of the transition from quiescence to activation and vice versa. MYOD expression pattern reflects the myogenic stage of MuSCs more accurately during myogenic progression. Moreover, accumulating evidence indicates that MuSCs are functionally heterogeneous populations, including rapidly dividing majorities and slowly dividing minorities. 34,35 Therefore, recapitulating endogenous MYOD expression in MuSCs with a fluorescent reporter is urgently required to distinguish and study the divergent phenotypes of MuSCs during myogenesis.
CRISPR-Cas9 is a potent tool for precise genome editing, and we used it to generate a MyoD-KI mouse line expressing tdTomato from the MyoD locus, without affecting the endogenous MYOD functions. Here, we assessed the MyoD-KI tdTomato reporter mirrors endogenous expression patterns in vitro and in vivo. Furthermore, we analyzed the tdTomato fluorescence intensity of MuSCs isolated from MyoD-KI mice in correlation with the activation status of MuSCs in culture, as evidenced by protein and RNA-sequencing analyses. Finally, using the features of MyoD-KI tdTomato fluorescence in cultured MuSCs, we aimed to establish an unbiased and high-throughput screening system to analyze the effect of compounds on MuSCs based on tdTomato fluorescence intensity. Overall, our MyoD-KI mouse line would open up an opportunity to elucidate the molecular mechanisms of MuSC dynamics, including their heterogeneity and fate determination mechanism, and to develop a potential drug target for MuSCs for regenerative medicine.

Generation of a MyoD-KI mouse line by CRISPR-Cas9
To generate MyoD-KI (MyoD KI/+ ) mice using the CRISPR-Cas9 system, a single guide RNA (sgRNA) was designed to target the region of the stop codon of MyoD ( Figure 1A). sgRNA and Cas9 were inserted into the px330-mC plasmid driven by the U6 and CBh promoter, respectively. For the KI donor DNA vector, the tdTomato sequence, preceded by a P2A peptide sequence, was cloned into the px330-mC plasmid (Figure 1A). The KI of the donor DNA vector consisted of two homology arms corresponding to the end of exon 3 and the start of the 3 0 -UTR of MyoD sequence ( Figures 1A and 1B). Microinjection of these vectors into the pronucleus of the C57BL/6J zygote resulted in the generation of the MyoD-KI locus via homologous recombination ( Figure 1B). The binding of Stau1 to the 3 0 -UTR of MyoD represses the protein translation of MyoD. 15 Therefore, before generating the F0 MyoD KI/+ mice, we confirmed if the insertion of the tdTomato sequence immediately before the 3 0 -UTR of MyoD affects its post-transcriptional regulation by Staufen1 (E-I) Genomic DNA analysis of F0 MyoD KI/+ or WT mice tails using PCR. The primer sets used for step 1, 4, and 5 are depicted in (B). Image showing WT (lane 3) and 3 different lines of F0 male MyoD KI/+ mice (lane 4-6). Lane 7 is Cas9-amp control to assess the random integration. Each expected PCR amplicon size is indicated in parentheses. M1, 1 kb DNA ladder; M2, 100 bp DNA ladder. All data are represented as the mean G standard error of the mean (s.e.m.) **p < 0.01, ***p < 0.001. iScience Article in vitro. To determine Staufen1 activity on the tdTomato_3 0 -UTR of MyoD, an expression vector expressing tdTomato_3 0 -UTR of the MyoD sequence under the Cytomegalovirus (CMV) promoter was created, and the plasmid was co-transfected with the Stau1_IRES_EGFP expression vector into the HEK293T cell line (Figure 1C). The tdTomato fluorescence decreased markedly as the expression of Stau1-EGFP increased ( Figures 1C and 1D), suggesting that the 3 0 -UTR-mediated translational regulation of Staufen1 was not disrupted by the insertion of the tdTomato sequence into the MyoD 3 0 -UTR in vitro.

ll
We initially screened for the presence of tdTomato (399bp) in the genomic DNA of 74 newborn mice of the F0 generation by using PCR with tdTomato-specific primer pairs ( Figure 1E). Three male F0 founders (#10, 25, and 27) carrying the tdTomato sequence on their genome were further subjected to PCR screening with Cas9-(959 bp) and ampicillin-specific primer pairs (560 bp) to confirm that it was not the result of random integrations ( Figures 1F and 1G). Furthermore, precise KI of the tdTomato sequence at the MyoD locus was verified by two additional PCR screenings using the primers listed in Figure 1B. The expected amplicons were confirmed in all three F0 founders ( Figures 1H and 1I). Male F0 founders (MyoD KI/+ ) were mated with female C57BL/6J mice to generate heterozygous KI offspring for subsequent analyses. PCR screening was repeated on the F1 offspring for further confirmation ( Figure S1).

MyoD-tdTomato fluorescence patterns partially mirror endogenous MYOD protein expression in vivo
First, we evaluated the endogenous tdTomato fluorescence in MyoD KI/+ embryos at E10.5 and E13.5 (Figures 2A-2D). Consistent with a previous result of whole-mount in situ hybridization of MyoD mRNA at E10.5, 36 tdTomato fluorescence was observed in the hypaxial myotome of interlimb somites, forelimb bud, and brachial arches of E10.5 MyoD KI/+ embryos ( Figure 2B). Additionally, tdTomato fluorescence was detected in the trunk, limbs, and craniofacial muscles at E13.5 of whole MyoD KI/+ embryos ( Figure 2C), which also represents a pattern similar to the endogenous MyoD mRNA expression, as described previously. 21 The sagittal section at E13.5 of MyoD KI/+ embryo showed that tdTomato (+) cells were clearly localized at the level of tongue, tail, and trunk muscles ( Figure 2D).
Next, to determine whether MuSCs expressing tdTomato can be quantified by flow cytometry, we isolated the mononuclear population from the hindlimb muscles of juvenile and adult MyoD KI/+ mice. Consistent with the analysis of TA muscle cross-sections with antibodies against PAX7 and tdTomato ( Figures  To further investigate whether tdTomato followed the expression dynamics of MYOD in the context of acute muscle injury, we injected barium chloride (BaCl 2 ) into the TA muscles of adult MyoD KI/+ mice. 40 Upon acute muscle injury, strong tdTomato signals were observed in the regenerating myofibers with central nuclei at day 4 post-injury ( Figure S3A). However, at day 14 post-injury, the tdTomato fluorescence in the regenerating myofibers disappeared and reached a level similar to the wild-type (WT) background (Figure S3A), except for the mononuclear cells associated with the myofibers ( Figure S3B) and small myofibers ( Figure S3C).
Next, we analyzed the tdTomato (+) mononuclear fraction by flow cytometry at days 4 and 14 after injury ( Figures 2J and 2K). We observed that the tdTomato (+) mononuclear fraction rapidly increased at day 4 post-injury but decreased by day 14 after injury ( Figures 2J and 2K). Next, to determine whether the expanded tdTomato (+) cells induced by the injury also recapitulate endogenous MYOD expression dynamics, the tdTomato (+) cells from the damaged TA muscles of adult MyoD KI/+ mice at days 4 and 14 post-injury were sorted and immediately stained with antibodies against PAX7 or MYOD, and tdTomato ex vivo (Figures 2L and S4A-S4C). First, we found that, an average of 56% tdTonato (+) cells were negative for PAX7 at day 4 and the fraction significantly reduced at day 14 ( Figure S4D). More importantly, we confirmed that an average of 99% tdTomato (+) cells were MYOD (+) at day 4 post-injury (Figures 2L-2M). However, only an average of 49% tdTomato (+) were immunolabelled with MYOD antibody ( Figures 2L-2M), suggesting that the tdTomato protein remains longer than endogenous MYOD protein.

MyoD-tdTomato fluorescence intensity detected in activated MuSCs recapitulates endogenous MYOD protein levels in vitro
To determine whether tdTomato fluorescence also recapitulated MYOD dynamics in in vitro and ex vivo cultures of MuSCs, we first used magnetic cell sorting (MACS) with a7-integrin microbeads to enrich MuSCs and non-myogenic progenitors from limb muscles of adult MyoD KI/+ mice ( Figure 3A). After 4-day of culture, strong tdTomato fluorescence was exclusively observed in a7-integrin (+) MuSCs, but not in fibroblast-like non-myogenic cells ( Figure 3A), suggesting that tdTomato is only expressed in myogenic progenitors. To further confirm tdTomato expression pattern, single extensor digitorum longus (EDL) myofibers were analyzed by immunohistochemistry. Single EDL myofiber from adult MyoD KI/+ mice stained with MYOD antibodies revealed the absence of MYOD signals in the myonuclei, and the tdTomato fluorescence was observed only faintly when compared to those from adult MyoD +/+ mice ( Figure S5A).  Figure S6A), and a fraction of tdTomato (+) MuSCs also expressed MYOG ( Figure S6B). To further assess the fidelity of the tdTomato fluorescence in MuSCs of MyoD KI/+ mice, we cultured MACS-isolated MuSCs for 4 days and verified that an average of 98% of tdTomato (+) cells were positive for MYOD (+) ( Figures 3F and 3G). To analyze the tdTomato fluorescence during myogenic differentiation in vitro, we isolated MuSCs from adult MyoD KI/+ mice and cultured them in a differentiation medium (DM) for either 3 or 8 days. Immunofluorescence analysis of MYOD and tdTomato shows that the tdTomato-expressing myotubes accumulate MYOD in the myonuclei at day 3 ( Figure S5B). However, by day 8, while the accumulation of MYOD protein in the myonuclei became fuzzy, the tdTomato signal weakly remained in those myotubes as depicted by the yellow box in Figure S5C.
Activated MuSCs are heterogeneous populations that associate with distinct myogenic states. On day 4 in culture, a majority of MuSCs expressing MYOD undergo proliferation or differentiation, while a small number show MYOD downregulation, thereby maintaining an undifferentiated state representing ''reserve cell'' population. 11,26,28 We observed weak or no MYOD-expressing MuSCs from MyoD KI/+ mice ( Figure 3F). Notably, cells with low MYOD expression appeared to have low tdTomato expression ( Figure 3F, white arrows). Further, flow cytometry analysis demonstrated that cultured MuSCs expressing low and high levels of tdTomato were distinguishable ( Figure 3H). To verify that tdTomato fluorescence intensity can recapitulate endogenous MYOD protein expression level, MuSCs expressing tdTomato at low and high levels were isolated using flow cytometer and immunostained for MYOD and tdTomato ( Figure 3I). MuSCs expressing low tdTomato fluorescence exhibited low endogenous MYOD protein levels and vice versa ( Figures 3J and 3K). Finally, we found a positive correlation between tdTomato fluorescence and MYOD protein levels (R 2 =0.4154, p < 0.0001) when each cell was plotted ( Figure 3L). Taken together, the MyoD KI/+ reporter mouse is a useful tool for monitoring endogenous MYOD dynamics in vitro, particularly during the early activation phase of myogenesis.

Homozygous MyoD-KI mice (MyoD KI/KI ) show no abnormality in growth and myogenic differentiation
Previously, MyoD-null MuSCs exhibited increased basal numbers, delayed entry into the cell cycle, and defects in proliferation and differentiation. Therefore, we investigated whether function was maintained in MyoD-KI homozygous mice (MyoD KI/KI ) ( Figure 4A). Both MyoD KI/KI and MyoD KI/+ mice grew normally, similar to the wild type (WT, MyoD +/+ ) mice with respect to body and muscle tissue weight ( Figures 4B-4D). No abnormal structure of myofibers with central nucleation or reduction in myofiber cross-sectional area was observed in mice as assessed using hematoxylin and eosin staining ( Figures 4E and 4F). Furthermore, the number of MuSCs associated with single EDL myofibers remained unchanged in MyoD +/+ , My-oD KI/+ , and MyoD KI/KI mice ( Figure 4G). Subsequently, comparison of the endogenous MYOD expression iScience Article levels in cultured MuSCs from all three genotypes using immunofluorescence indicated similar levels of MYOD expression (Figures 4H and 4I). Conversely, tdTomato fluorescence intensity was higher in My-oD KI/KI mice than that in MyoD KI/+ mice, suggesting that the MYOD protein was translated from transcripts originating from both alleles ( Figures 4J and 4K).
Next, we analyzed the regenerative ability in MyoD +/+ , MyoD KI/+ , and MyoD KI/KI mice. We injected BaCl 2 into TA muscles of MyoD +/+ , MyoD KI/+ , and MyoD KI/KI mice. At day 14 after injury, CSA and the number of PAX7 (+) cells were examined. Our results show that both CSA and the number of PAX7 (+) were almost identical among the three genotypes, suggesting that the insertion of tdTomato at the MyoD locus does not impair MuSCs function even under strong regenerative stress ( Figures 4L-4N). Together, these data demonstrated that tdTomato KI at the MyoD locus does not inhibit endogenous MYOD expression and muscle development and regeneration.

MyoD-tdTomato fluorescence intensity defines the activation status of MuSCs
MyoD KI/+ reporter mice recapitulated MYOD expression without interrupting its endogenous function; therefore, we investigated whether tdTomato fluorescence defines the state of MuSCs during myogenesis in vitro.
First, MACS-isolated MuSCs were cultured for five days and assessed using immunofluorescence for PAX7 and tdTomato expression ( Figure 5A). Heterogeneous tdTomato fluorescence patterns were observed in MuSCs at this stage. Of note, MuSCs expressing low levels of tdTomato expressed high levels of PAX7 (Figure 5A, white arrows) and vice versa ( Figure 5A, yellow arrows). To confirm that tdTomato fluorescence intensity correlates with the myogenic state of MuSCs, cultured MuSCs expressing low and high tdTomato (MyoD-tdTomato low and MyoD-tdTomato high ) were sorted using flow cytometry ( Figure 5B). Immunofluorescence labeling for PAX7 and tdTomato showed that around 80% of MuSCs in the MyoD-tdTomato low were positive for PAX7. This rate was significantly reduced to an average of 35% in the MyoD-tdTomato high (Figures 5C and  5D). We simultaneously performed immunofluorescence analysis for MYOG. MYOG-expressing MuSCs in the MyoD-tdTomato low population were almost undetectable ( Figures 5E and 5F). Conversely, an average of 47% of MuSCs in the MyoD-tdTomato high population expressed MyoG (Figures 5E and 5F). Additionally, the distinct characteristics between the two populations were further confirmed using real-time quantitative PCR (RT-qPCR) for Pax7, MyoD, tdTomato, and Myogenin ( Figure 5G), thereby indicating that the MyoD-tdTomato low population was enriched with undifferentiated stem-like cells, and the MyoD-tdTomato high population contained committed proliferating and MYOG (+) cells.

RNA-seq analysis of the MyoD-tdTomato low and MyoD-tdTomato high populations revealed unique gene expression profiles
MyoD KI/+ mice can distinguish undifferentiated (MyoD-tdTomato low ) and committed (MyoD-tdTomato high ) MuSCs from heterogeneous populations in culture, which has been challenging to explore this far. Therefore, obtaining whole-genome expression profiles from these two populations would greatly benefit the identification of new regulatory genes in MuSCs. We performed RNA sequencing of these two populations isolated using FACS after 5-day of culture. Further, we analyzed the whole gene expression in freshly isolated MuSCs from Pax7-YFP KI mice. 32 Principal component analysis revealed a clear sample separation among groups ( Figure 6A). As expected, the transcriptome profile of freshly isolated MuSCs (QSCs) was  iScience Article distinct from that of MyoD-tdTomato low and MyoD-tdTomato high populations ( Figure 6A). We detected 1479 genes that were differentially expressed between the MyoD-tdTomato low and MyoD-tdTomato high populations, visualized using a heatmap ( Figure 6B). Additionally, we visualized selected genes related to MuSC quiescence (Pax7, CD34, Calcr, Col5a1, Spry1, Adgrf5, Arrb1, Notch1-3, Hey1, and Heyl), activation (MyoD, Myf5, and Hes1), and differentiation (Myog, Myf6, Hes6, Myh3, Myh7, and Myh8) among the groups ( Figure 6C). The quiescence marker genes were abundantly expressed in QSCs, while some genes overlapped with the MyoD-tdTomato low population. When we compared the quiescence marker genes expression level between the QSCs and MyoD-tdTomato Low populations, these markers were significantly higher in QSCs than the MyoD-tdTomato low population, except for Pax7 and Myf5 ( Figures 6C, 6D, 6F). The activation and differentiation genes were more accumulated in the MyoD-tdTomato high population than the MyoD-tdTomato low population ( Figures 6C-6G). Together, whole gene expression profiles also support the notions that MyoD-tdTomato low population represent undifferentiated stem-like cells with intermediate characteristics between quiescent and committed/activated MuSCs, and the MyoD-tdTomato high population represents committed proliferating and differentiating cells.
To further characterize MyoD-tdTomato low and MyoD-tdTomato high populations, we performed gene ontology (GO) analysis on differentially expressed genes (DEGs). Comparison of MyoD-tdTomato low with MyoD-tdTomato high revealed 590 upregulated and 871 downregulated genes in the MyoD-tdTomato low population. Of the 590 upregulated genes, 426 genes were commonly upregulated between MyoD-tdTomato low vs. MyoD-tdTomato high and QSCs vs. MyoD-tdTomato high ( Figure 6H). Of the 871 downregulated genes, 535 genes were commonly downregulated between MyoD-tdTomato low vs. MyoD-tdTomato high and QSCs vs. MyoD-tdTomato high ( Figure 6I). Among the 426 upregulated genes, there was a strong enrichment for terms related to blood vessel development, extracellular matrix (ECM) organization, and tissue morphogenesis ( Figure 6J, Tables S1 and S2). In contrast, among the 535 downregulated genes, there was a strong enrichment for terms related to muscle structure/tissue development and myotube differentiation ( Figure 6K, Tables S1 and S2).
Since our MyoD KI/+ model allowed us to obtain a unique intermediate stage of MuSCs (Myod-tdTomato low ) wherein they express high levels of Pax7 and Myf5, and low MyoD in vitro, next, we considered the genes that were exclusively upregulated or downregulated in the MyoD-tdTomato low population without corresponding changes reported in QSCs vs. MyoD-tdTomato high (Figures 6H-6I). Interestingly, within 164 genes upregulated in Myod-tdTomato low population, those associated with tube morphogenesis, including Angpt1, Foxc2, Cxcl12, Csf1, Mmp14, and Smad7 ( Figure 6L, Tables S1 and S2) was determined to be the most significantly represented class of genes, consistent with previous observations. 41 -45 The second most significant class of genes was those involved in ECM organization, which includes Col4a2, Has2, Fn1, lamb1, and Nid1 ( Figure 6L, Tables S1 and S2). [46][47][48][49] In contrast, among the 336 genes downregulated exclusively in Myod-tdTomato low population, there was a strong enrichment for terms related to muscle structure development and muscle cell differentiation ( Figure 6M, Tables S1 and S2). These transcriptome data of cultured MuSCs from our MyoD KI/+ mice will be a unique resource for elucidating the undefined molecules associated with stemness properties in MuSCs for future research. Our data suggest that the MyoD-tdTomato low population displays an undifferentiated state and expresses defined stem cell-like markers, whereas the MyoD-tdTomato high population exhibits committed/proliferative myogenic progenitor features ( Figures 5 and 6). Therefore, we hypothesized that monitoring the fluorescence pattern/ratio of MyoD-tdTomato would enable us to screen the global activation state of cultured MuSCs, which has been difficult to analyze so far without immunostaining. The laminin 511-E8 fragment (LM511-E8) is an important ECM component in the maintenance of stemness in MuSCs. 50 Thus, we cultured MACS-isolated MuSCs on either gelatin or LM511-E8 for 5 days ( Figure 7A), and tdTomato fluorescence intensity and ratio in the MuSCs were analyzed by microscopically ( Figures 7B and 7C) or by flow cytometry (Figures 7D-7F). MuSCs cultured on LM511-E8 exhibited low levels of tdTomato fluorescence and a higher percentage of MyoD-tdTomato low population than those cultured on gelatin ( Figures 7B-7F).
In addition to the substrate for MuSC culture, we tested whether the MyoD KI/+ mice would be a valuable tool for assessing the effect of compounds on MuSC status in a high-throughput manner. We cultured MACS-isolated MuSCs in the presence of SB203580, a p38 MAPK inhibitor that is an established target for maintaining stemness in MuSCs ( Figure 7G). 29,30,51 FACS analysis of the MuSCs demonstrated that the p38 MAPK inhibitor significantly maintained the MyoD-tdTomato low population in a dose-dependent manner, when compared with the dimethyl sulfoxide (DMSO) control ( Figures 7H-7J), thereby indicating that the MyoD KI/+ mouse line is useful for screening a potential compound/substrate on the effect of MuSC fate decision.
Furthermore, to develop a more efficient screening platform for validating the effect of drugs and substrates on the effects of the myogenic state using a MyoD KI/+ model, we established a 96-well-based fluorescence image quantification screening system using an image cytometer equipped with KEYENCE's microscope. After 5-day of MuSCs culture with DMSO or SB203580 in a 96-well-plate, the tdTomato fluorescence intensity of each well was captured, and the data were visualized using a heatmap ( Figure 7K) and a plot graph ( Figure 7L). Consistent with the results of the FACS analysis, we detected that SB203580 significantly suppressed MyoD-tdTomato fluorescence in a dose-dependent manner when compared to the DMSO control ( Figures 7K and 7L). Therefore, these data suggest that this screening system using My-oD KI/+ mice will be beneficial for identifying new substances that regulate the fate of MuSCs in vitro.

DISCUSSION
Herein, using CRISPR-Cas9, we generated a KI mouse line that recapitulates endogenous MYOD expression using a tdTomato fluorescent reporter in vivo and in vitro. Paradoxically, MyoD transcripts are present but absent at the protein level in quiescent MuSCs, which is controlled by translational suppression by RBPs, such as Stau1, through the MyoD 3 0 -UTR. 15 Therefore, we first evaluated whether CRISPR-Cas9-mediated gene editing on MyoD 3 0 -UTR with the addition of the tdTomato sequence would not disrupt the translational suppression of MyoD in vitro. If the addition of a tdTomato sequence changes the accessibility of Stau1 to the MyoD 3 0 -UTR, aberrant translation of MYOD is induced. We confirmed that Stau1-mediated translational suppression is preserved in the tdTomato-MyoD 3 0 -UTR sequence. We successfully obtained the F0 founder and established the MyoD-KI mouse line, with no inappropriate mutations in the endogenous MyoD sequence, including the stop codon ( Figures S8 and S9). We confirmed the restricted expression of MyoD-tdTomato in myogenic progenitors and activated MuSCs but not in quiescent MuSCs. Additionally, homozygous MyoD KI/KI mice are viable with no morphological abnormalities in skeletal muscle and expression levels similar to endogenous MYOD in WT or MyoD KI/+ mice. iScience Article Importantly, we found that MyoD-tdTomato fluorescence intensity mirrors the endogenous MYOD protein expression in cultured MuSCs. While it is clear that not all quiescent MuSCs are functionally homogeneous, 38 their progeny are diverse, especially when activated. Previous studies using a label-retaining assay have revealed distinct populations in MuSCs. 34,35 MuSCs with limited proliferative output are enriched for self-renewal, whereas fast-dividing MuSCs proceed to differentiation. However, though MYOD is most frequently used to quantify heterogeneity in the MuSC population, no direct evidence of heterogeneity based on MYOD expression has been reported due to the lack of tools to distinguish between MYODlabeled and -unlabeled MuSCs. To the best of our knowledge, our study is the first to isolate and analyze heterogeneous MuSC populations based on MYOD expression.
We performed RNA-seq analysis to compare whole transcriptome expression profiles in MyoD-tdTomato low-and high-expressing MuSCs and freshly isolated MuSCs. RNA-seq and protein analysis confirmed that the MyoD-tdTomato low population was enriched in undifferentiated MuSC features, and the MyoD-tdTomato high population showed committed and differentiating cell features. Previously reported quiescent MuSC-associated genes, such as CD34, Calcr, Spry1, and Notch1-3, are highly expressed in freshly isolated MuSCs. Adgrf5, an adhesion G protein-coupled receptor 116, and its downstream effecter Arrb1, important for the quiescence and long-term maintenance of the MuSC pool, as we previously reported, 52 are also enriched in this population. Notably, the expression of these quiescence signatures is downregulated in the MyoD-tdTomato low population but remains highly maintained when compared to the MyoD-tdTomato high population. Intriguingly, Pax7 expression was not significantly different between freshly isolated MuSCs and MyoD-tdTomato low population. Furthermore, we found that the Myf5 transcript was highly enriched in the MyoD-tdTomato low population when compared to the freshly isolated MuSCs and the MyoD-tdTomato high population. Myf5 and MyoD functionally compensate for each other in myogenic lineage determination during embryonic development but play distinct roles in adult regeneration. 22,25,53 Moreover, the absence of MyoD leads to upregulation of Myf5 and an increase in the number of MuSCs as a consequence of increased self-renewal rather than myogenic differentiation. 22,25 Therefore, we interpret these data to suggest that the MyoD-tdTomato low population expressing high levels of Pax7 and Myf5 contains a rare intermediate state between quiescent and activated MuSCs. 23,25,54 Our whole transcriptome data will be a great database to decipher the new molecular mechanism for a small unique subset of self-renewing MuSCs and to determine their fate in future studies. We found that the MyoD-tdTomato low population expresses genes associated with endothelial cell development. Interestingly, the endothelial and skeletal muscle lineages in the trunk and limbs arise from common embryonic progenitors. 42,55 Furthermore, the angiopoietin 1/Tie-2 axis, which is highly enriched in the MyoD-tdTomato low population, regulates self-renewal of MuSCs. 41 It is tempting to hypothesize that genes enriched in the MyoD-tdTomato low population recapitulate, at least in part, myogenesis by common multipotent progenitors. Using our database as a stepping stone, the functions of these undefined genes can be verified to unveil the new paradigm of MuSC research.
MYOD expression has been used as an activation marker for MuSCs for several decades. However, there is a lack of tools to efficiently quantify MYOD (+) or (À) cells. Usually, the fraction of immunolabeled cells positive for PAX7 and/or MYOD is quantified to indicate self-renewed, activated, and committed MuSCs in cultured MuSCs or on the muscle cross-sections. However, such visual assessment of images by individuals can lead to bias in their perception of patterns in the images and in subsequent interpretation of data. Additionally, the activated MuSCs also exhibit heterogeneous MYOD expression patterns, with some showing bright and others showing weak MYOD; thus, the visual evaluation of immunolabeled cells by individuals is limited in reproducibility. Here, we propose a model in which MyoD-tdTomato fluorescent pattern/intensity efficiently identifies undifferentiated stem-like or proliferative/committed MuSCs using a non-biased approach, such as flow cytometry, without immunolabeling. Therefore, this unbiased method using MyoD-KI mice is expected to contribute to the replicability and reproducibility in skeletal muscle research.
Furthermore, a high-throughput system for comparing the effects of compounds in regulating MuSC selfrenewal and expansion has not yet been established, although many studies have attempted to expand MuSCs without losing stem cell regenerative capacity. 13,30,31, 50 We developed a platform to screen compounds for manipulating MuSC fates based on monitoring MyoD-tdTomato fluorescence intensity using an image cytometer. Since MuSCs are rare, a 96-well-plate system could be a beneficial screening system for validating many candidates from the library for their ex vivo expansion toward regenerative medicine. iScience Article In summary, we successfully developed a MyoD-KI mouse line to monitor MYOD expression using a tdTomato fluorescent reporter, without interfering with endogenous MYOD protein expression during myogenesis. Furthermore, MyoD-tdTomato fluorescence intensity partially recapitulates the myogenic progression of MuSCs in vitro and in vivo. Therefore, our MyoD-KI mice can be used as a screening tool to easily and efficiently distinguish between undifferentiated and committed MuSCs under various conditions, including knockout of a gene of interest or in the presence of an undefined compound of interest. Moreover, our MyoD-KI mice and the RNA-seq database in the present study will be valuable resources for identifying a key molecule regulating a small subset of the reserve MuSC population during myogenesis.

Limitations of the study
Our study reveals that the tdTomato sequence immediately before the MyoD 3 0 UTR sequence does not interfere the post-translational regulation by Stafune1 in HEK293T cells, but does not investigate the context of myogenic progenitors. Although, we confirmed that the adult MuSCs from MyoD KI/+ mice specifically express tdTomato upon activation, further investigation is required to reveal whether the posttranslational regulators of MyoD 3 0 UTR, such as Staufen1 15 and Tristetraprolin (TPP), 14 can be functional in the MuSCs from MyoD KI/+ mice.
We have shown that MyoD-tdTomato fluorescence successfully recapitulates the endogenous MYOD protein levels under several conditions. However, we must point out that MyoD-tdTomato fluorescence is not perfect in some conditions as MyoD expression dynamically changes during myogenesis. For instance, at day 14 after BaCl 2 injury, only 49% of tdTomato signals are successfully overlapped with endogenous MYOD signals detected by the antibody. This could be attributed to the fact that the tdTomato protein has a longer half-life than MYOD in MuSCs, wherein MyoD expression has already been downregulated or tuned off. Future research using these MyoD KI/+ mice should consider this limitation.
Moreover, our current evidence supports that tdTomato signals in the cryosections of E10.5, E13.5 embryos, and injured TA muscles, are primarily restricted to the dermomyotome area, muscle-forming regions, and regenerating myofibers, respectively. However, it remains unclear whether the tdTomato signals detected in the cryosections of E10.5 and E13.5 embryos, and in the regenerating myofibers, are positive for endogenous MYOD. This is because the MYOD antibody used in the present study has limitations in detecting MYOD on the PFA-prefixed mouse cross-sections in our protocol.
We also note that the molecular features of the MyoD-tdTomato-negative population remain unknown. We purified quiescent MuSCs from MyoD KI/+ mice by using MACS and culture expansion. However, MACSbased isolation showed a slight contamination of non-myogenic cells, which are completely negative populations for MyoD-tdTomato ( Figures S7A-S7C). Thus, we were unable to perform whole transcriptome analysis of the MyoD-tdTomato-negative population in the present study. Nonetheless, there would be a very minor MyoD-tdTomato negative MuSC population, which may represent a rare subset of stress-resistant MuSCs expressing Pax3, as reported previously. 6,8 In the future, the combination of our MyoD-KI mice with single-cell RNA-seq will help identify the population and its molecular signatures. This experiment is currently underway in our laboratory.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following: