Acetylation of PAX7 controls muscle stem cell self-renewal and differentiation potential in mice

Abstract

Muscle stem cell function has been suggested to be regulated by Acetyl-CoA and NAD+ availability, but the mechanisms remain unclear. Here we report the identification of two acetylation sites on PAX7 that positively regulate its transcriptional activity. Lack of PAX7 acetylation reduces DNA binding, specifically to the homeobox motif. The acetyltransferase MYST1 stimulated by Acetyl-CoA, and the deacetylase SIRT2 stimulated by NAD +, are identified as direct regulators of PAX7 acetylation and asymmetric division in muscle stem cells. Abolishing PAX7 acetylation in mice using CRISPR/Cas9 mutagenesis leads to an expansion of the satellite stem cell pool, reduced numbers of asymmetric stem cell divisions, and increased numbers of oxidative IIA myofibers. Gene expression analysis confirms that lack of PAX7 acetylation preferentially affects the expression of target genes regulated by homeodomain binding motifs. Therefore, PAX7 acetylation status regulates muscle stem cell function and differentiation potential to facilitate metabolic adaptation of muscle tissue.

Fig. 1. PAX7 acetylation regulates its transcriptional activity.

a Schematic representation of PAX7 protein including the paired, octapeptide, and homeodomain, as well as the location of the two lysine residues that are acetylated. b, c Representative MS/MS spectrum of (b) the acetyl-lysine K105 peptide YQETGSIRPGAIGGSK(Ac)PR and (c) the acetyl-lysine K193 peptide HSIDGILGDK(Ac)GNR from mouse PAX7. Fragment assignments were determined using Mascot software. PAX7-FLAG was ectopically expressed in primary myoblasts, purified using anti-FLAG agarose and resolved on SDS-PAGE. The band corresponding to PAX7 protein was excised and subjected to LC-MS/MS spectrometry. d, e Amino acid sequence alignment of PAX7 K105 and K193-containing regions among the indicated species. f Acetylation of PAX7 controls Myf5 expression. Cells were transfected with plasmids encoding WT or mutant PAX7-FKHR fusion as indicated, together with a luciferase reporter plasmid containing Myf5 −111 kb enhancer and renilla internal control reporter plasmid DNA. Luciferase activity was normalized to control in which Myf5 reporter plasmid was not transfected. Data are presented as mean values ± SEM (n = 5 independent experiments) (One-way ANOVA uncorrected Fisher’s LSD test: **p = 0.002; ***p = 0.0009). g PAX7 binding to chromatin is regulated by acetylation. C2C12 myoblasts were transfected with plasmids encoding FLAG-tagged WT or K105/193R PAX7 as indicated, and PAX7 recruitment at the Myf5 −111 kb and Myf5 −57.5 kb enhancers was determined by chromatin immunoprecipitation using anti-FLAG antibodies. Data are presented as mean values ± SEM (n = 3 independent experiments) (Two-way ANOVA uncorrected Fisher’s LSD test: ***p < 0.0001 −111 kb IgG vs WT; **p = 0.0043 −111 kb WT vs KR; ***p < 0.0001 −57,5 kb IgG vs WT; **p = 0.0043 −57,5 kb WT vs KR). h Electromobility shift assay (EMSA) showing WT and mutant PAX7 binding to full-length, homeobox motif or paired motif of the Myf5 −111 kb enhancer. Binding is abrogated when a cold probe is used as a competitor. Specificity of the PAX7 binding is controlled by supershift using an anti-PAX7 antibody. Empty arrows indicate shift, while black arrows indicate supershift.

Fig. 2. PAX7 protein interacts with MYST1, SIRT1, and SIRT2.

a PAX7 interacts with MYST1. Cells were transfected with plasmids expressing HA-MYST1 and FLAG-PAX7, as indicated. The interaction between PAX7 and MYST1 was determined by immunoprecipitation using anti-FLAG antibodies (top) or anti-HA antibodies (bottom). Immunoblotting was performed with the indicated antibodies. b PAX7 interacts with SIRT1. Cells were transfected with plasmids expressing HA-SIRT1 and FLAG-PAX7, as indicated. The interaction between PAX7 and SIRT1 was determined by immunoprecipitation using anti-FLAG antibodies (top) or anti-SIRT1 antibodies (bottom). Immunoblotting was performed with the indicated antibodies. c PAX7 interacts with SIRT2. Cells were transfected with plasmids expressing MYC-SIRT2-FLAG and HA-PAX7, as indicated. The interaction between PAX7 and SIRT2 was determined by immunoprecipitation using anti-HA antibodies (top) or anti-FLAG antibodies (bottom). Immunoblotting was performed with the indicated antibodies. d, e Representative confocal images displaying the co-localization between PAX7 and MYST1 (d) or SIRT2 (e). Primary myoblasts were immunostained with PAX7 (red) and MYST1 or SIRT2 (green), and nuclei were counterstained with DAPI (blue). Scale bar represents 10 μm. Images are representative of ≥80% of the cells examined (n = 3 experiments from biologically independent samples). f PAX7 acetylation levels are regulated by MYST1 and SIRT2. Cells were transfected with plasmids encoding PAX7 (WT or K105/193R), MYST1, SIRT2, and SIRT1, and lysates were subjected to immunoprecipitation using anti-acetyl-lysine antibodies. Immunoblotting was performed with the indicated antibodies. g Quantification of the immunoblotting signals from (f). The densitometric analysis of the level of acetylated PAX7, relative to total PAX7 signal, is shown. Data are presented as mean values ± SEM (n = 3 independent experiments) (One-way ANOVA uncorrected Fisher’s LSD test: *p = 0.049).

Fig. 3. MYST1 and SIRT2 control the expression of PAX7 target genes.

a, b Expression level of Pax7 and Myf5 in primary myoblasts after treatment for 24 h with a the SIRT2 inhibitor AGK2 (10 and 20 μM) or b the MYST1 inhibitor MG149 (20 and 40 μM), as indicated. Gene expression was determined by RT-qPCR (normalized to Rps18 expression). Data are presented as mean values ± SEM (n = 6 biologically independent samples) (Two-way ANOVA uncorrected Fisher’s LSD test: *p = 0.021 in (a); **p = 0.0052 in (a) *p = 0.018 in (b)). c Expression level of Myst1, Sirt2, Pax7, and Myf5 in primary myoblasts after transfection with siRNA against Myst1 or Sirt2, as indicated. Gene expression was determined by RT-qPCR (normalized to Rps18 expression). Data are presented as mean values ± SEM (n = 6 biologically independent samples) (Two-way ANOVA uncorrected Fisher’s LSD test: ***p < 0.0001 siMyst1 (Myst1); **p = 0.0015 siSirt2 (Myst1); ***p < 0.0001 siMyst1/siSirt2 (Myst1); **p = 0.0023 siMyst1 (Sirt2); ***p < 0.0001 siSirt2 (Sirt2); ***p < 0.0001 siMyst1/siSirt2 (Sirt2); **p = 0.0053 siMyst1 (Myf5); ***p = 0.0006 siSirt2 (Myf5)).

Fig. 4. MYST1 and SIRT2 control the level of PAX7 acetylation in satellite cells.

a Representative PAX7:acetylated-lysine PLA (red) performed on satellite cells on cultured myofibers following siRNA treatments against Myst1 or Sirt2, as indicated. Satellite cells are labeled with SYNDECAN-4 (SDC4, green) and nuclei are counterstained with DAPI (blue). Scale bar represents 10 μm. Images are representative of ≥75% of the cells examined (n = 3 mice). b, c Quantification of the PLA signals from (a), represented as the mean ± SEM. The PLA was quantified by counting the number of nuclear PLA puncta for each satellite cell. b (n = 69 siCtrl, n = 61 siMyst1 cells from 3 mice). c (n = 86 siCtrl, n = 77 siSirt2 cells from 3 mice) (Two-tailed paired t test: **p = 0.0082 in (b); **p = 0.0096 in (c)). d Representative PAX7:acetylated-lysine PLA (red) performed on satellite cells cultured on myofibers in presence of DMSO, AGK2 or MG149, as indicated. SDC4 marks the satellite cells (green) and nuclei are counterstained with DAPI (blue). Scale bar represents 10 μm. Images are representative of ≥75% of the cells examined (n = 3 mice). e Quantification of the PLA signals from (d), represented as the mean ± SEM. The PLA was quantified by counting the number of nuclear PLA puncta for each satellite cell (n = 107 DMSO, n = 107 MG149, n = 90 AGK2 cells from 3 mice).

Fig. 5. MYST1 and SIRT2 control satellite stem cell self-renewal.

a Representative immunostaining of an asymmetric (YFP⁺/YFP⁻) and a symmetric (YFP⁻/YFP⁻) satellite stem cell division on myofibers cultured for 42 h. Myofibers were immunostained with YFP (green) and PAX7 (red), and nuclei were counterstained with DAPI (blue). Scale bar represents 10 μm. Images are representative of ≥90% of the divisions examined (n = 8 mice). b Proportion of asymmetric satellite stem cell division on myofibers cultured for 42 h. Myofibers were transfected with siRNA against Sirt2 or Myst1 as indicated. Data are presented as mean values ± SEM (n = 8 samples from independent mice) (One-way ANOVA uncorrected Fisher’s LSD test: *p = 0.0124 siMyst1; *p = 0.0193 siSirt2). c Number of satellite cells on myofibers cultured for 42 h and transfected as in (b). Data are presented as mean values ± SEM (n = 8 samples from independent mice) (One-way ANOVA uncorrected Fisher’s LSD test: *p = 0.0204). d Representative immunostaining of satellite cells on myofibers cultured for 72 h. Myofibers were transfected with siRNA against Myst1 or control and were stained for PAX7 (red). Nuclei were counterstained with DAPI (blue). Scale bar represents 50 μm. Images are representative of ≥75% of the myofibers examined (n = 3 samples from independent mice). e Number of satellite cells on myofibers cultured for 72 h. Myofibers were transfected with siRNA against Myst1 or control. Data are presented as mean values ± SEM (n = 3 samples from independent mice) (Two-tailed paired t test: **p = 0.007).

Fig. 6. PAX7 acetylation controls satellite cell self-renewal and regenerative potential.

a Representative immunostaining of TA muscle cross-section in uninjured muscles and after 3 rounds of CTX injury. Tissues were immunostained with DYSTROPHIN (DYS, red) and PAX7 (green), and nuclei were counterstained with DAPI (blue). Scale bar represents 100 μm. Images are representative of ≥80% of the tissues examined (n = 3 for Pax7^WT mice, n = 4 for Pax7^KR mice). b Quantification of the number of satellite cells in uninjured muscles and injured muscles following one CTX injection in adult and aged mice (geriatric, >21 months), and triple CTX injection in adult mice. Data are presented as mean values ± SEM (n = 3 for Pax7^WT mice, n = 4 for Pax7^KR mice) (Two-tailed unpaired t test: *p = 0.0139 in 1× CTX aged; *p = 0.0188 in 3× CTX). c Representative immunostaining of tibialis anterior (TA) muscle cross-section after 3 rounds of CTX injury. Top: tissues were immunostained with LAMININ (white), MHCIIA (green) and MHCIIB (red), and nuclei were counterstained with DAPI (blue). Bottom: Representative Matlab SMASH analysis of individual muscle fibers based on laminin staining, and enlarged images of cropped area of MHCIIA (green) and MHCIIB (red) immunostaining. Scale bar represents 200 μm. Images are representative of ≥80% of the tissues examined (n = 6 mice per group). d Injured TA muscle weight relative to body weight after 3 rounds of CTX injury. Data are presented as mean values ± SEM (n = 6 for Pax7^WT mice, n = 10 for Pax7^KR mice) (One-way ANOVA uncorrected Fisher’s LSD test: **p = 0.0011). e Minimal fiber Feret’s diameter from Pax7^WT and Pax7^KR mice after 3 consecutive rounds of CTX injury represented as absolute number of muscle fibers. Data are presented as mean values ± SEM (n = 6 mice per group) (Multiple unpaired t tests: *p = 0.0174). f Proportion of type IIa and type IIb fibers in Pax7^WT and Pax7^KR TA muscles after three consecutive rounds of CTX injury. Data are presented as mean values ± SEM (n = 5 mice per group) (Two-way ANOVA uncorrected Fisher’s LSD test: *p = 0.018).

Fig. 7. PAX7 acetylation regulates asymmetric stem cell division.

a Representative immunostaining of satellite cells on myofibers from Pax7^WT and Pax7^KR mice, immediately after isolation (0 h). Myofibers were immunostained with YFP (green) and PAX7 (red), and nuclei were counterstained with DAPI (blue). Scale bar represents 50 μm. Images are representative of ≥75% of the myofibers examined (n = 7 mice per genotype). b Absolute numbers of YFP⁻ satellite cells from Pax7^WT:Myf5-Cre:R26R-EYFP and Pax7^KR:Myf5-Cre:R26R-EYFP mice immediately after isolation. Data are presented as mean values ± SEM (n = 7 mice per genotype) (Two-tailed unpaired t test: *p = 0.0368). c Total number of satellite cells on myofibers from Pax7^WT:Myf5-Cre:R26R-EYFP and Pax7^KR:Myf5-Cre:R26R-EYFP mice immediately after isolation. Data are presented as mean values ± SEM (n = 7 mice per genotype). d Representative immunostaining of satellite cells on myofibers from Pax7^WT and Pax7^KR mice, after 42 h of culture. Myofibers were immunostained with YFP (green) and PAX7 (red), and nuclei were counterstained with DAPI (blue). Scale bar represents 50 μm. Images are representative of ≥75% of the myofibers examined (n = 7 mice per genotype). e, f Absolute numbers per fiber of (e) symmetric (YFP−/YFP−) and (f) asymmetric (YFP+/YFP−) satellite cell doublets from Pax7^WT:Myf5-Cre:R26R-EYFP and Pax7^KR:Myf5-Cre:R26R-EYFP mice at 42 h post-isolation. Data are presented as mean values ± SEM (n = 7 mice per genotype) (Two-tailed unpaired t test: *p = 0.0276). g Proportion of asymmetric stem cell divisions after 42 h of culture. Data are presented as mean values ± SEM (n = 7 mice per genotype). h Total satellite cell numbers after 42 h of culture. Data are presented as mean values ± SEM (n = 7 mice per genotype).

Fig. 8. PAX7 acetylation regulates target gene expression in satellite cells.

a Heatmap from normalized expression matrix of differentially expressed genes between Pax7^WT and Pax7^KR satellite cells. The number of PAX7 ChIP-seq peaks associated with each gene is indicated. Genes in bold with an asterisk are significantly different (Wald test followed by multiple testing correction using the Benjamini–Hochberg method to control the false discovery rate, with application of a cut-off of 0.1 (approximating to a p-value of 3.5 × 10⁻⁴) to corrected values). Heatmap coloring indicates row Z-score; color bars to the left indicate log2 fold change, and log2 mean expression. b Plot illustrating the distance from the closest PAX7 ChIP-seq peak and transcriptional start site (TSS) to the closest PAX7 ChIP-seq peak vs the negative log p-value for all PAX7 target genes. Differentially and significantly expressed genes between Pax7^WT and Pax7^KR satellite cells are highlighted in red (Wald test followed by multiple testing correction using the Benjamini–Hochberg method to control the false discovery rate, with application of a cut-off of 0.1 (approximating to a p-value of 3.5 × 10⁻⁴) to corrected values). c Motifs identified de novo in Pax7 peaks are similar to motifs described in Soleimani et al.. The homeobox motif is enriched in PAX7 target genes (all genes), as well as in the subset that is differentially expressed in Pax7^KR satellite cells. In contrast, the paired motif is found only when analyzing peaks associated with all PAX7 target genes, but not when considering only the subset of genes differentially expressed in Pax7^KR satellite cells. Numbers indicate the percentage of peaks containing the different motifs, as well as the p value calculated by HOMER. d Gene ontology (GO) term enrichment analysis (biological process) for PAX7 target genes that are differentially expressed between Pax7^WT and Pax7^KR satellite cells.