DAXX-dependent H3.3 deposition maintains myoblast cell identity independently of other histone chaperone complexes

Summary

H3.3 histone chaperone DAXX regulates heterochromatin silencing; however, its function in transcription regulation remains understudied. Here, we show that Daxx knockout (KO) myoblasts have impaired differentiation and fusion. Transcriptomic analysis revealed a loss in myogenic gene expression and broad transcription dysregulation in Daxx KO myoblasts. Chromatin immunoprecipitation followed by sequencing demonstrated a marked reduction in H3.3 deposition at myogenic loci in Daxx KO myoblasts, which was further linked to decreased H3K27ac. Intriguingly, the double KO of Daxx and Hira resulted in distinct transcriptomic alterations than those of single KOs, demonstrating that DAXX and HIRA have both overlapping and unique roles in H3.3 incorporation. Our findings establish DAXX as a critical regulator of myogenic gene expression and muscle cell identity through a distinct mechanism from that of HIRA and highlighted an unanticipated plasticity in the deposition loci for DAXX and HIRA in myoblasts.

Graphical abstract

Highlights

• •DAXX maintenance of myoblast identity is linked to H3.3 deposition at myogenic loci
• •DAXX and HIRA have overlapping but non-compensatory functions in preserving myoblast identity
• •Both DAXX and HIRA regulate H3.3 deposition in euchromatin in myoblasts

Cell biology; Transcriptomics

Introduction

Chromatin architecture and nucleosome composition regulate transcriptional expression to maintain cell state and identity. Fine-tuned modifications in DNA methylation, nucleosome histone variant composition, and histone posttranslational modifications (PTMs), among other processes, contribute to this regulation. In eukaryotic cells, nucleosomes are composed of a tetramer of H3-H4 histones and two dimers of H2A-H2B histones that are wrapped by 147 base pairs of DNA linked by the histone H1.¹ Nucleosomes are dynamic structures, and their remodeling is associated with cellular processes such as DNA replication, DNA repair, transcription, and heterochromatin silencing. The turnover of nucleosomes may underlie the exchange of histone variants and alter the proportion of parental versus newly synthesized histones and the ratio of PTMs associated with each histone variant.^1,2 Moreover, specific PTMs associated with newly synthesized histones in pre-assembly complexes were identified, such as acetylation of lysine 5 and 12 of histone H4 and methylation of lysine 9 of histone H3, in which deposition occurs after the S phase of the cell cycle.^3,4 Therefore, the maintenance of the chromatin landscape upon cell division is crucial for the preservation of cell fate and identity. Histone variant H3.3 is linked to the maintenance of gene status upon cell division and consequently sustains epigenetic memory, in particular through the acquisition of the transcriptionally permissive histone mark H3K4me3.^5,6 In the mouse, except for histone H4, histones have several variants, which are defined as canonical and non-canonical regarding their deposition in cell cycle-dependent and -independent ways, respectively.⁴ Histone H3 has two canonical variants, H3.1 and H3.2, the non-canonical variant H3.3, and the centromeric-specific variant CENPA.^1,4 In addition, 14 new H3.3 sub-variants with tissue-specific expression were identified.⁷ H3.3 is incorporated into the genome by two specific histone chaperones, HIRA and DAXX, which belong to two independent pathways for histone deposition depending on the genomic loci. The deposition of H3.3 in promoters, active gene bodies, transcription start sites (TSSs), and transcription end sites (TESs) is regulated by HIRA and is associated with transcription regulation.^8,9 Conversely, DAXX deposits H3.3 in constitutive heterochromatin such as telomeres and pericentromeric regions, where it presents the transcriptional repressive PTM H3K9me3.^10,11,12,13

The link between H3.3 deposition and cell lineage progression has been recognized in several contexts, such as fibroblast reprogramming and transdifferentiation, hematopoietic stem cell (HSC) differentiation, transition of hemogenic endothelial to hematopoietic stage, and neurogenesis, among others.^{14,15,16,17,18,19} Studies performed in skeletal muscle cell lines, in resting muscles, or during regeneration showed that myogenic cell identity and myoblast differentiation rely on the HIRA-H3.3 incorporation pathway.^{20,21,22,23,24} Skeletal muscle progenitor cells are characterized by the expression of the paired-box transcription factor PAX7. Upon lineage commitment into the myoblast state, the cells express the myogenic regulatory factors MYF5, MRF4, MYOD, and MYOG, the latter being associated with terminal differentiation.^25,26,27 To initiate differentiation and fusion, myoblasts express Myod1. The transcription of Myod1 is regulated by HIRA-mediated H3.3 incorporation in the regulatory regions of the Myod1 locus.^20,21,23 Moreover, skeletal muscle stem cells require HIRA-H3.3 deposition at myogenic loci to maintain the transcriptional permissive histone mark H3K27ac and sustain gene transcription.^23,28 In the absence of HIRA, muscle stem cells lose identity and are unable to replenish the pool or regenerate the tissue.²³

While H3.3 involvement in transcriptional regulation is HIRA mediated, the function of the histone chaperone activity of DAXX in this process remains understudied. DAXX is commonly associated with ATRX, which recognizes repetitive regions on DNA, to sustain heterochromatin silencing via H3.3 incorporation, required for the maintenance of the H3K9me3 repressive mark.^29,30,31,32 Recently, it was shown that the H3K9me3 PTM mark is present in H3.3 in a pre-deposition complex with DAXX and that it is independent of ATRX.³⁰ DAXX histone chaperone-independent function is linked to retrotransposon silencing by physical interaction with SETDB1, KAP1, and HDAC1.²⁹ Although, in this context, H3.3 is not deposited, DAXX-H3.3 physical interaction stabilizes the DAXX protein.^29,33 The recruitment of histone deacetylase (HDAC) by DAXX to inhibit gene expression was further reported in several cell lines, including myogenic cells.^34,35 While DAXX histone chaperone activity has been mostly evaluated in the context of heterochromatin silencing, a few studies demonstrated a role for DAXX-mediated H3.3 deposition and transcriptional regulation. In developing neurons, DAXX-H3.3 deposition regulates early neuronal activation pathways.³⁶ Moreover, DAXX-H3.3 contributes to the regulation of adult HSC homeostasis.³⁷ Furthermore, we observed that around 10% of H3.3-detected peaks were still present in regulatory regions, promoters, and gene bodies in Hira knockout (KO) myoblasts,²³ which possibly highlights the presence of other H3.3 histone chaperones. This suggested that DAXX-H3.3 deposition may not be restricted to heterochromatin as previously reported.^11,38

In this report, we demonstrate that DAXX-mediated H3.3 deposition in myoblasts directly regulates skeletal muscle cell identity, myogenic gene expression, and myoblast differentiation. We performed a comprehensive comparative analysis of H3.3 incorporation by HIRA and DAXX genome-wide in myoblasts, where HIRA and DAXX gene expression is abolished. Our study reveals a unique genomic deposition plasticity of H3.3 by these two chaperone complexes in myoblasts, which differs from the patterns observed in pluripotent stem cells.

Results

Daxx deletion in myoblasts results in myogenic impairment

We previously demonstrated that deleting HIRA in myoblasts leads to a loss of skeletal muscle cellular identity while not completely eliminating H3.3-detected peaks in gene bodies and regulatory regions.²³ Therefore, we investigated whether DAXX could regulate gene expression via H3.3 deposition in the myogenic lineage. We mutated Daxx in the myoblast C2C12 cell line using CRISPR-Cas9 (Daxx KO) (Table S1), at the level of exon 3, which partially codes for the protein histone-binding domain.³⁹ We validated the loss of DAXX expression in several clones by immunostaining and western blot (Figures 1A and 1B). We observed that HIRA expression was maintained in the Daxx KO cells (Figure S1A). Next, we quantified the number of myoblasts that expressed the myogenic marker PAX7 when cultured under low-density and high-serum conditions, representing a proliferative environment. We identified a significantly reduced number of PAX7⁺ cells in Daxx KO cells (Figures 1C and 1D). The expression of the myogenic commitment marker MYOD was also decreased as analyzed by immunostaining and western blot, both of which showed reduced protein levels (Figures 1C–1E). Consistently, expression levels for Pax7, Myod1, and, as expected, Daxx were reduced in the absence of DAXX (Figure 1F), while Hira expression levels were not changed. Consistent with the loss of several myogenic markers, when cultured under low-serum conditions, Daxx mutant myoblasts displayed impaired differentiation capacity, as observed by myosin staining (MF20 antibody) (Figures 1G and 1H). The control and Daxx mutant cells showed a similar proliferation rate (EdU incorporation assay, Figures S1B and S1C). Since DAXX has also been shown to play an essential role in maintaining genomic integrity, we analyzed cell survival and the levels of γH2A.X to detect DNA breaks.⁴⁰ Immunostaining using a phosphorylated form of caspase-3 did not detect significant apoptotic signals (Figures S1D and S1E). Moreover, we observed that Daxx KO cells did not show significant changes in the number of γH2A.X foci per nucleus compared to control myoblasts (Figures S1F and S1G). To confirm that the robust differentiation phenotype observed was due to the lack of DAXX, we performed rescue experiments by transfecting control and Daxx KO myoblasts with plasmids containing the full-length Daxx cDNA or an empty plasmid. Overexpression of DAXX in Daxx KO myoblasts rescued the differentiation potential of myoblasts (Figures 1I, 1J, and S1H). Taken together, these results show that like HIRA,²³ DAXX is required to maintain myogenic gene expression and differentiation and that these two histone chaperone pathways cannot compensate for the loss of each other.

Figure 1.

Daxx KO impairs myogenic lineage progression

(A) Immunostaining with DAXX (green) antibody and the nuclear marker DAPI (blue) in C2C12 (top) and Daxx KO (bottom) cell lines (n = 3 C2C12 and n = 3 Daxx KO independent culture experiments). Scale bars, 40 μm.

(B) Western blot for DAXX and GAPDH in C2C12 (n = 2 independent protein samples) and Daxx KO (in 4 independent clones) (n = 2 independent protein samples).

(C) Co-immunostaining with PAX7 (red) and MYOD (green) antibodies and the nuclear marker DAPI (blue) in C2C12 (top) and Daxx KO (bottom) cell lines (n = 5 C2C12 and n = 5 Daxx KO independent culture experiments). Scale bars, 100 μm.

(D) Quantification of the number of PAX7-positive cells among DAPI-positive cells in (C) (n = 5 C2C12 and n = 5 Daxx KO independent culture experiments). Error bars, mean ± SD, Mann-Whitney test.

(E) Western blot or MYOD and H3 in C2C12 (n = 2 independent protein samples) and Daxx KO (in 4 independent clones) (n = 2 independent protein samples) cell lines.

(F) RT-qPCR analyses of the mRNA levels of Pax7, Myod1, Daxx, and Hira in control (gray) (n = 3 independent RNA samples) and Daxx KO (black) cells (n = 3 independent RNA samples). For each gene, the mRNA levels of the control cells were normalized to 1. Error bars, mean ± SD, two-tailed unpaired t test.

(G) Immunostaining with MF20 antibody (green) to visualize myosins and the nuclear marker DAPI (blue) in C2C12 (top) and Daxx KO (bottom) cell lines. Scale bars, 40 μm.

(H) Quantification of the number of MF20-positive nuclei among DAPI-positive nuclei in (G) (C2C12 n = 6, Daxx KO n = 5, independent differentiation assays). Error bars, mean ± SD, Mann-Whitney test.

(I) Immunostaining with MF20 antibody (green) to visualize myosins and the nuclear marker DAPI (blue) in C2C12 transfected with empty plasmid (left), and Daxx KO cell lines transfected with empty (middle) or Daxx cds-containing plasmid (right). Scale bars, 100 μm.

(J) Quantification of the number of MF20-positive nuclei among DAPI-positive nuclei in (I) (n = 3 independent differentiation assays for each condition). Error bars, mean ± SD, two-tailed unpaired t test.

p values below 0.05 were considered significant.

See also Figures S1 and S6.

DAXX histone chaperone activity regulates gene transcription

To explore changes in muscle gene expression that might explain the observed cell phenotype and to assess genome-wide effects on the transcriptome of Daxx KO cells, we performed RNA sequencing (RNA-seq) on cells cultured under proliferative conditions (low-density and high-serum environment) (Figure S2A). We observed significant transcriptome alteration in Daxx KO compared to C2C12 controls, with 14.8% of expressed genes being dysregulated (4,975 out of 33,686 expressed genes) (Figures 2A and 2B). This suggests a broad impact of impaired DAXX-mediated H3.3 deposition on gene expression. Consistent with the myoblast phenotype, Gene Ontology (GO) analysis of down-regulated genes in Daxx KO cells revealed terms associated with muscle cell development, myofibril organization, and skeletal muscle contraction and fusion (Figure 2C). Interestingly, GO terms associated with genes up-regulated in Daxx KO myoblasts were linked to telomere maintenance, DNA replication, ribosome assembly, and DNA metabolic processes, among other processes (Figure 2D). Selected down-regulated myogenic genes (Pax7, Myod1, Myog, Mymk, and Mymx) and genes from pathways linked to myogenesis such as bone morphogenetic protein (BMP) signaling pathway-associated (Bmp4, Id1, and Id2) and Notch signaling pathway-associated genes (Notch1, Hey1, Hey2, and Heyl) are represented in the heatmap with the normalized reads per gene and in individual triplicates (Figure 2E). Selected up-regulated genes are linked to alternative lineages (Tbx20, Isl1, Nefl, and Efna5), Wnt signaling pathway (Wnt5b, Wnt6, Dkk2, and Lgr5), and genome and cell maintenance (Igtp, Muc1, Rad50, and Dach1) (Figure 2E). We validated the RNA-seq results by performing quantitative reverse-transcription PCR (RT-qPCR) to a group of selected genes (Figure 2F). To determine whether the myoblast phenotype and transcriptomic changes were linked to changes in H3.3 incorporation at myogenic gene loci, we performed H3.3 chromatin immunoprecipitation followed by sequencing (ChIP-seq) in Daxx KO and control cells. Genome-wide analysis showed a 66% reduction in the number of H3.3 peaks in the absence of DAXX compared with control C2C12 myoblasts (12,779 vs. 38,339 detected peaks) (Figure 2G). Additionally, unlike what we observed in Hira KO myoblasts, where the number of H3.3 peaks was evenly decreased throughout the genome,²³ the loss of H3.3 peaks in Daxx KO myoblasts was predominantly associated with intronic and intergenic regions (Figure 2H). Remarkably, these genomic locations are associated with regulatory elements,^41,42 suggesting that DAXX-mediated H3.3 deposition may be associated with enhancer elements and play a role in transcriptional regulation. Consistently, since chromatin accessibility can be associated with gene transcription, we further investigated the opening status of chromatin in control and Daxx KO myoblasts by assay for transposase-accessible chromatin followed by sequencing (ATAC-seq) and identified an average of 65,284 vs. 46,941 detected peaks (Figure S2B) but without major changes in gene bodies and TSSs (Figures S2C and S2D). These findings demonstrate that DAXX function, in a direct or indirect manner, is required for H3.3 deposition across the myoblast genome, with a pronounced affinity for intergenic and intronic regions.

Figure 2.

Daxx KO prevents myogenic gene expression

(A) MA plot of C2C12 over Daxx KO RNA-seq data. Significantly dysregulated genes are highlighted in blue (false discovery rate [FDR] < 0.05).

(B) Number of up-regulated (2,364 genes, blue) and down-regulated (2,611 genes, red) genes in Daxx KO compared to C2C12.

(C and D) Gene Ontology analysis for biological processes of the down-regulated (C) and up-regulated (D) genes in Daxx KO cells. Selected enriched terms are presented according to the fold enrichment.

(E) Heatmap with the number of normalized reads of C2C12 and Daxx KO RNA-seq data for a selection of genes from the indicated cell lineages, in individual triplicates.

(F) RT-qPCR analyses of the mRNA levels of selected genes down-regulated and up-regulated in Daxx KO cell lines. Control (gray) (n = 3 independent RNA samples) and Daxx KO cells (black) (n = 3 independent RNA samples). For each gene, the mRNA levels of the control cells were normalized to 1. Error bars, mean ± SD, two-tailed unpaired t test.

(G) Total number of H3.3 called peaks (tag number) (q value = 5e−2) in the ChIP-seq of C2C12 (gray, 38,439 peaks) (n = 1) and Daxx KO cells (black, 12,779 peaks) (n = 1).

(H) Percentage of the total number of H3.3 peaks distributed in distinct genomic regions as indicated in C2C12 or Daxx KO cells.

p values below 0.05 were considered significant.

See also Figure S2.

Loss of myogenic gene expression in Daxx KO myoblasts is linked to decreased H3.3 deposition

To further analyze changes in H3.3 incorporation in myogenic cells, we plotted the H3.3 average intensity signal in gene bodies of Daxx KO and control cells. We observed that H3.3 signal is significantly decreased in Daxx KO cells when plotting all genes (Figure 3A). Strikingly, the loss of H3.3 deposition in Daxx KO cells was specifically associated with the loci of genes whose expression was down-regulated (Figure 3A), mirroring the pattern observed in Hira KO myoblasts.²³ To address changes in the histone PTM landscape in gene bodies, we performed ChIP-seq for marks linked to active (H3K4me3 and H3K27ac) and repressive (H3K9me3 and H3K27me3) transcription. The H3K4me3 PTM mark is associated with active promoters, and the deposition of this mark does not show major changes when plotting all genes (Figure 3B) or splitting between up- and down-regulated genes (Figure S3A). However, H3K27ac, which marks active promoters and enhancers,⁴³ is decreased in Daxx KO myoblast gene loci compared to control (Figure 3C) independently of their expression (Figure S3B). When analyzing repressive histone marks, we detected decreased H3K9me3 concomitantly with an increased average signal of H3K27me3 at TSSs (Figures 3D and 3E). Consistent with the previously described presence of H3K27me3 mark at promoters and H3K9me3 mark in gene bodies,⁴⁴ we observed that the H3K9me3 decrease was maintained throughout the gene body, while H3K27me3 increase was restricted to TSSs (Figures 3D and 3E). The alteration of the repressive histone PTM marks was observed independently of gene expression changes (Figures S3C and S3D). This suggests the presence of a compensation mechanism in the cell to overcome the loss of the repressive mark H3K9me3 by the addition of H3K27me3 marks. The absence of H3.3 in myogenic gene loci in Daxx KO myoblasts is linked to decreased H3K27ac mark and impacts gene expression, despite the maintenance of H3K4me3, leading to the lack of myogenic capacity of the cells, as observed in Hira KO cells.²³ This intricate profile is highlighted in specific myogenic loci such as Pax7, Myod1, and Mymk (Figures 3F, 3G, and 3H). Conversely, genes that are up-regulated do not show changes in H3.3 incorporation (Figures 3A–3I, 3J, and 3K). Strikingly, ATAC-seq-detected peaks were reduced in both gene loci of down-regulated genes, linked to the loss of H3.3 (Figures 3F, 3G, and 3H) and gene loci of up-regulated genes (Figures 3I, 3J, and 3K), showing a broad reduction of chromatin accessibility in the absence of DAXX, similar to what we observed in Hira KO myoblasts.²³ This up-regulation is linked with the loss of the repressive mark H3K9me3 at the TSS. Strikingly, in the same gene loci, the H3K27me3 mark is acquired at the TSS in Daxx KO cells, but this is not sufficient to rescue the repressive status. This can be observed in the loci of genes with up-regulated expression, such as Dkk2, Igtp and Wnt5b (Figures 3I, 3J, and 3K). These findings demonstrate that DAXX regulates myogenic gene expression through the deposition of H3.3 at myogenic gene loci. Furthermore, they suggest that in myoblasts, the histone chaperone complexes HIRA²³ and DAXX are required for correct deposition of H3.3 at loci of genes critical for maintaining cellular identity and ensuring their sustained expression.

Figure 3.

DAXX deposits H3.3 in myogenic gene loci

(A) ChIP-seq average signal profiles (ratio to input) in the promoter region (±3 kb around the TSS), TSS, gene body, and TES for H3.3 in C2C12 (blue) and Daxx KO (red) cells are shown for all genes (left), up-regulated (middle) genes in Daxx KO, and down-regulated genes in Daxx KO (right).

(B–E) ChIP-seq average signal profiles (ratio to input) in the promoter region (±3 kb around the TSS), TSS, gene body, and TES for H3K4me3 (B), H3K27ac (C), H3K9me3 (D), and H3K27me3 (E) in C2C12 (blue) and Daxx KO (red) cells shown for all genes.

(F–K) ATAC-seq (green), ChIP-seq profiles for H3.3 (orange), H3K27ac (dark blue), H3K4me3 (red), H3K27me3 (purple), and H3K9me3 (light blue) in the genomic loci of Pax7 (F), Myod1 (G), Mymk (H), Dkk2 (I), Igtp (J), and Wnt5b (K) in control (WT, top lines) and Daxx KO (KO, bottom lines).

See also Figures S2 and S3.

Myoblasts deleted for HIRA and DAXX lose myogenic identity, but the up-regulated gene set is distinct from individual mutated lines

Since single mutants for Daxx or Hira²³ did not lead to a complete loss of H3.3 in euchromatin, we generated a Hira and Daxx double KO (dKO) C2C12 line to investigate the resulting changes in gene expression and myogenic identity. We validated the absence of both DAXX and HIRA proteins by immunostaining and western blot (Figures 4A and 4B). As expected, the loss of both Daxx and Hira led to a reduced number of myogenic cells (Figures 4C–4G) and reduced MYOD protein levels (Figure 4B). As anticipated, this resulted in the absence of differentiated myosin+ (MF20 antibody) cells (Figures 4H and 4I). Consistently, the mRNA levels for Pax7, Myf5, Myod1, and Myog were reduced compared to the controls, while the housekeeping gene Tbp was not changed (Figure 4J). The dKO cells showed a proliferation rate comparable to that of C2C12 control cells (Figures S4A and S4B) but showed increased DNA damage as observed by γH2A.X immunostaining (Figures S4C and S4D). To identify the transcriptomic changes in the dKO cell line, we performed RNA-seq from cells cultured in the proliferative environment (low-density and high-serum conditions) (Figure S4E). We observed striking changes in the transcriptome of dKO samples compared to C2C12 controls, with 39.3% of expressed genes being dysregulated (13,255 out of 33,686 total number of expressed genes) (Figures 5A and 5B). This indicates a major genome-wide alteration in transcriptional activity. Strikingly, no significant changes were observed in the expression of other chromatin modifiers such as Dek, Chd1, and Chd2 in dKO cells (Figure S4F). Consistent with the single Hira or Daxx KO lines, GO analysis of genes down-regulated in dKO cells revealed terms associated with skeletal muscle adaptation, contraction, and myofibril assembly (Figure 5C). However, up-regulated GO terms refer to calcium transport, retinoic acid, and collagen biosynthesis (Figure 5D), showing specific and distinct pathways compared to the single Daxx KO data (Figure 2D, GO: telomere maintenance, DNA repair) or the single Hira KO data²³ (GO: alternative lineage development). Selected dysregulated genes belonging to these GO terms were used for heatmap visualization and RT-qPCR validation (Figures 5E and 5F). To cross-compare the genes whose expression was dysregulated between each single KO and the dKO lines, we plotted the genes into a Venn diagram and observed that while only 15% and 27% were specific to Daxx KO and Hira KO, respectively, 38% of the dysregulated genes in the dKO were specific to this dataset (Figure 5G). Moreover, we performed GO analysis on genes dysregulated specifically in each individual KO and dKO. While Daxx KO did not present significant GO terms, Hira KO dysregulated genes were linked to peroxisomal transport and mitochondrial translation and dKO linked to protein PTMs (Figures S4G and S4H). These results show that in the absence of DAXX and HIRA in myoblasts, the myogenic identity is lost, but the outcome for gene expression is distinct from each of the individual mutants.

Figure 4.

Daxx and Hira dKO cell line has impaired myogenic identity

Co-immunostaining with DAXX (green) and HIRA (red) antibodies and the nuclear marker DAPI (blue) in C2C12 (top) and dKO (bottom) cell lines (n = 3 C2C12 and n = 3 dKO independent culture experiments). Scale bars, 40 μm.

(B) Western blot for DAXX, HIRA, MYOD, and H3 in C2C12 (n = 2 independent protein samples) and dKO (2 independent clones) (n = 2 independent protein samples).

(C) Co-immunostaining with PAX7 (green) and MYOD (red) antibodies and the nuclear marker DAPI (blue) in C2C12 (top) and dKO (bottom) cell lines (n = 3 C2C12 and n = 3 dKO independent culture experiments). Scale bars, 40 μm.

(D) Quantification of the number of PAX7-positive cells among DAPI-positive cells in (C) (n = 3 C2C12 and n = 3 dKO independent culture experiments). Error bars, mean ± SD, two-tailed unpaired t test.

(E) Quantification of the number of MYOD-positive cells among DAPI-positive cells in (C) (n = 3 C2C12 and n = 3 dKO independent culture experiments). Error bars, mean ± SD, two-tailed unpaired t test.

(F) Co-immunostaining with MYOG (red) and the nuclear marker DAPI (blue) in C2C12 (top) and dKO (bottom) cell lines (n = 3 C2C12 and n = 3 dKO independent culture experiments). Scale bars, 40 μm.

(G) Quantification of the number of MYOG-positive cells among DAPI-positive cells in (F) (n = 3 C2C12 and n = 3 dKO independent culture experiments). Error bars, mean ± SD, two-tailed unpaired t test.

(H) Immunostaining with MF20 antibody (green) to visualize myosins and the nuclear marker DAPI (blue) in C2C12 (top) and dKO (bottom) cell lines. Scale bars, 40 μm.

(I) Quantification of the number of MF20-positive nuclei among DAPI-positive cells in (H) (n = 3 C2C12 and n = 3 dKO independent culture experiments). Error bars, mean ± SD, two-tailed unpaired t test.

(J) RT-qPCR analyses of the mRNA levels of Pax7, Myf5, Myod1, Myog, and Tbp in the dKO cells normalized to control = 1 (n = 3 independent RNA samples for control and dKO). For each gene, the mRNA levels of the control cells were normalized to 1. Error bars, mean ± SD, two-tailed unpaired t test.

p values below 0.05 were considered significant.

See also Figures S4 and S6.

Figure 5.

Transcriptomic analysis of dKO cells shows a distinct phenotype compared to single mutation myoblast lines

(A) MA plot of C2C12 over dKO RNA-seq data. Significantly dysregulated genes are highlighted in blue (FDR < 0.05).

(B) Number of up-regulated (6,816 genes, blue) and down-regulated (6,439 genes, red) genes in dKO compared to C2C12.

(C and D) Gene Ontology analysis for biological processes of the down-regulated (C) and up-regulated (D) genes in dKO cells. Selected enriched terms are presented according to the fold enrichment.

(E) Heatmap with the number of normalized reads of C2C12 and dKO RNA-seq data for a selection of genes from the indicated cell lineages, in individual triplicates.

(F) RT-qPCR analyses of the mRNA levels of selected genes down-regulated and up-regulated in dKO cell lines. Control (n = 3–6 independent RNA samples) and dKO cells (n = 3–6 independent RNA samples). For each gene, the mRNA levels of the control cells were normalized to 1. Error bars, mean ± SD, two-tailed unpaired t test.

(G) Venn diagram depicting the number of dysregulated genes in each single KO and in the dKO mutant lines.

p values below 0.05 were considered significant.

See also Figure S4.

DAXX and HIRA incorporate H3.3 in regulatory regions

In pluripotent stem cells, H3.3 genomic deposition mediated by DAXX and HIRA occurs in heterochromatin and euchromatin, respectively, but how these pathways regulate H3.3 incorporation in somatic cell lines remains underinvestigated.^10,11 Our results reveal that in myoblasts, DAXX facilitates H3.3 deposition in euchromatin—gene bodies and regulatory regions in myoblasts. To further address these differences, we compared HIRA- and DAXX-mediated H3.3 deposition in distinct genomic locations in myoblasts. We plotted the H3.3 average intensity signal in enhancers (p300 ChIP-seq data from Asp et al.⁴⁵ and at ATAC-seq peaks). Our analysis revealed that H3.3 deposition is decreased in both Daxx KO and Hira KO myoblasts but with an enhanced effect in Hira KO cells (Figures 6A and 6B). This demonstrates that while HIRA is primarily responsible for the incorporation of H3.3 at regulatory regions, DAXX also contributes to H3.3 deposition at these genomic locations, albeit to a lesser extent.

Figure 6.

DAXX and HIRA deposit H3.3 in both euchromatin and heterochromatin in C2C12 myoblasts

(A) ChIP-seq average signal profiles (ratio to input) in enhancer regions for H3.3 in C2C12 (blue), Daxx KO (red), and Hira KO (green) cells.

(B) ChIP-seq average signal profiles (ratio to input) for H3.3 plotted into the ATAC-seq peaks in C2C12 (blue), Daxx KO (red), and Hira KO (green) cells.

(C and D) ChIP-seq average signal profiles (ratio to input) for H3.3 (C) and H3K9me3 (D) plotted into genomic repetitive regions in C2C12 (blue), Daxx KO (red), and Hira KO (green) cells.

(E) UpSet plot depicting the H3.3 peak intersections between the different datasets, C2C12, Daxx KO, and Hira KO.

(F) Percentage of the number of H3.3 shared peaks between Daxx KO and Hira KO datasets, distributed in distinct genomic loci.

(G) Gene Ontology analysis for biological processes for the genes associated with peaks shared between Daxx KO and Hira KO datasets. Selected enriched terms are presented according to the fold enrichment.

(H–J) ChIP-seq profiles for H3.3 (orange), in the genomic loci of Dlx4 (H), Cbx5 (I), Jmjd6 (J), in control (WT, top lines), Hira KO (middle lines), and Daxx KO (bottom lines).

(K) ChIP-RT-qPCR for H3.3 on the regions selected from the ChIP-seq data for Myod1, Tbx4, Dll1, and Notch1 genes in C2C12 (gray) and dKO (black) cell lines (n = 3 independent biological samples per condition). Error bars, mean ± SD, two-tailed paired t test.

p values below 0.05 were considered significant.

See also Figure S5.

To identify the shared and specific H3.3-detected peaks in control, Daxx KO, and Hira KO C2C12 myoblasts, we plotted the annotated peaks of each dataset in an upset plot. We observed that while the majority of the H3.3-detected peaks were specific to control myoblasts (29,952 peaks), shared peaks between the three datasets or between Hira and Daxx KO represented 890 peaks (Figure 6C). This indicates that only a limited number of H3.3 peaks are retained in both Hira and Daxx KO cell lines. Remarkably, the majority of the H3.3 peaks maintained in both Hira and Daxx KO (shared peaks) are present in promoters (Figure 6D) and could be associated with gene maintenance of gene expression, since genes losing H3.3 in myoblasts are down-regulated (Figure 3A). Consistently, GO analysis for genes associated with this set of peaks highlighted terms related to cell process maintenance: proliferation, chromatin organization, nuclear transport, and protein PTMs (Figure 6E). Maintenance of H3.3 at promoters in control, Daxx KO, and Hira KO can be observed at the selected gene loci for Dlx4, Cbx5, and Jmjd6 (Figures 6F–6H). Next, we performed a peak intersection analysis between the datasets to examine H3.3 peaks specific to Daxx KO and Hira KO samples. We found that in Daxx KO, H3.3 peaks are more frequently retained in promoter regions, whereas in Hira KO cells, they are more commonly maintained in intergenic regions (Figures S5A and S5B). Our findings are consistent with the observation that DAXX predominantly deposits H3.3 in intronic and intergenic regions (Figure 2H). Remarkably, GO terms linked to H3.3 peaks specifically maintained in Daxx KO and Hira KO samples are linked to nuclear transport and organization, cell membrane maintenance, and endocytosis/exocytosis, respectively (Figures S5C and S5D). These data suggest that DAXX could be mostly involved in regulating the expression of genes related to cellular membrane and adhesion processes, while HIRA plays a more prominent role in the regulation of transcription of genes linked to nuclear organization. Since DAXX was shown to regulate transposable element (TE) silencing, we performed TE expression analysis using the RNA-seq data but did not detect major transcriptomic changes with only 0.03% of TEs being dysregulated (Figures S5E and S5F). To determine whether Daxx KO or Hira KO, specifically H3.3 peaks, was maintained in the dKO cells, we performed ChIP-RT-qPCR for some of the candidate regions. We observed that Dll1 (Daxx KO-specific peak), Tbx4 (Hira KO-specific peak), and Notch1 (shared peak between Daxx KO and Hira KO) genomic loci were enriched for H3.3 in dKO cells but not in control cells, while H3.3 is detected in Myod1 loci in control but not in dKO cells (Figure 6I) (of note, given the variability of the samples, ChIP-RT-qPCR for H3.3 at Myod1 and Tbx4 loci do not meet statistical significance). These data suggest that dKO myoblasts retain the H3.3 peaks detected in the single KO lines.

Taken together, our findings reveal a flexible genomic pattern of H3.3 deposition by the DAXX and HIRA histone chaperone complexes in myoblasts, highlighting shared and unique H3.3 peaks in the absence of either DAXX or HIRA.

Discussion

DAXX and HIRA are H3.3 histone chaperones with distinct regions for genomic deposition. In pluripotent stem cells, HIRA is associated with transcription regulation, and DAXX is linked to heterochromatin deposition.^10,11,38 Here, we show that in myoblasts, both DAXX and HIRA are required to regulate myogenic gene expression and identity via H3.3 histone deposition in myogenic gene loci.

Regulation of transcription by DAXX via H3.3 histone chaperone activity is understudied. Hence, the DAXX function as a co-repressor has been linked to transcriptional regulation through its binding to transcription factors and recruitment of HDACs.³⁵ DAXX was shown to inhibit myogenic differentiation by physically interacting with the E2A transcription factor and repressing E2A-mediated transcription.³⁴ DAXX transcriptional regulation is in part mediated by its sub-nuclear localization. During myogenic differentiation, before myoblast fusion, promyelocytic leukemia nuclear bodies, matrix-associated domains that serve as reservoirs for protein recruitment and PTMs, are disassembled. This disassembly leads to re-localization of DAXX to chromatin, where it contributes to the regulation of transcription.^46,47,48,49 Several studies revealed the recruitment of DAXX to regulatory elements, particularly in cancer cell lines. While these studies did not show whether this was associated with genome-wide recruitment of H3.3,^50,51,52 DAXX-mediated H3.3 incorporation was documented at CDKN1A and GADD45A promoter regions as well as sub-telomeric regions lost in DAXX KO human cell lines.⁵³ In addition, in vivo studies in mice showed that DAXX H3.3 histone chaperone activity is required for early immediate gene expression during neuronal activation, particularly at the Bdnf and Fos loci.³⁶ Together with our data, this establishes that DAXX can regulate transcription as a co-repressor but also as an H3.3 chaperone.

In this study, we observed that both DAXX and HIRA are required for the maintenance of H3.3 in myogenic gene loci and display global genomic H3.3 incorporation. The decreased H3.3 detected peaks in Hira KO cells are evenly distributed across the genome.²³ However, in the Daxx KO cell line, we observed a particularly significant reduction in H3.3 within introns and intergenic regions. This aligns with studies in pluripotent stem cells where Hira KO leads to decreased H3.3 incorporation at promoters and TSSs, while transcription factor-binding sites largely maintain H3.3 peaks,¹⁰ potentially due to the involvement of another chaperone complex. Additionally, H3.3 has been shown to be incorporated at both promoters and enhancers to activate gene expression,⁵⁴ a process that requires H3K27ac PTM.²⁸ Our study reveals that DAXX and HIRA play distinct yet overlapping roles in regulating myogenic gene expression through the deposition of H3.3 histones in myoblasts. These results expand our understanding of the functional versatility of DAXX by demonstrating its involvement in H3.3-mediated chromatin dynamics.

The DAXX/ATRX complex has been shown to regulate H3.3 deposition at repetitive regions/heterochromatin and is required for establishing H3K9me3 PTM necessary for silencing these repeats.^11,38,55 Notably, we observed that single KO for either Hira or Daxx resulted in decreased H3.3 incorporation at heterochromatin. However, only in Daxx KO cells, H3K9me3 PTM was lost at these regions, consistent with the role of DAXX, but not HIRA, in heterochromatin silencing.⁵⁵ This finding aligns with previously identified pre-deposition complexes of H3.3K9me3, deposited by DAXX in heterochromatin,³⁰ and suggests that H3.3 can be incorporated by both HIRA and DAXX at heterochromatin.

We identified H3.3 peaks that were shared or specific to each of the single mutants (Daxx or Hira KO) and absent in the control. This led us to hypothesize that other chaperone activity-containing proteins could be implicated in H3.3 incorporation in myoblasts. Indeed, in dKO cells, we detected the presence of some of these peaks and a distinct transcriptome signature compared to the single KO lines. This observation suggests a distinct chromatin state compared with the single KO lines and suggests the presence of an additional chaperone for H3.3 deposition in the genome of myogenic cells. Other proteins shown to incorporate H3.3 into chromatin are part of the chromodomain helicase DNA-binding protein family, such as CHD1 and CHD2.^56,57,58 Particularly, CHD2 physically interacts with MYOD and locates to Myog loci to mediate H3.3 incorporation and activation of Myog expression.⁵⁹ Future studies should aim to identify and characterize these potential chaperones, which may work in concert with DAXX and HIRA to regulate H3.3 incorporation. Understanding the interplay between these factors could provide deeper insights into chromatin remodeling processes and their impact on gene expression during myogenic differentiation. Additionally, further exploration of the sub-nuclear localization dynamics of DAXX during differentiation could uncover new mechanisms by which chromatin architecture and gene expression are coordinated during muscle cell development. These studies will not only enhance our understanding of muscle biology but may also inform therapeutic strategies for muscle-related diseases where these pathways are dysregulated.

Limitations of the study

This study primarily investigated the role of histone chaperone DAXX in C2C12 myoblasts, a well-established in vitro model. While powerful, this model lacks additional in vivo context, myogenic signals, and microenvironmental factors. We show that myogenic gene expression and cell identity are regulated by DAXX, but one cannot conclude if this is through a direct or indirect effect from the H3.3 deposition function of DAXX. Our work revealed residual H3.3 peaks in Daxx and Hira dKO cells by ChIP-RT-qPCR, which could be linked to the involvement of other deposition complexes that were not investigated. Finally, in Daxx KO and Hira KO myoblasts, H3.3 deposition is abrogated, and this is further linked to the loss of H3K27ac. The precise connection between these chaperones and the recruitment of histone-modifying complexes remains to be elucidated.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Joana Esteves de Lima (joana.esteves-de-lima@inserm.fr).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •Data: data are available within the article and supplementary files, and accession codes in key resources table. The RNA-seq, ChIP-seq, and ATAC-seq sequence data that support the findings of this study have been previously published and deposited in GEO NCBI with the accession code GSE161056²³ or newly deposited in GEO NCBI with the accession codes GSE277345 (ATAC-seq), GSE277346 (ChIP-seq), and GSE277347 (RNA-seq).
• •Code: no code was generated in this study.
• •Other items: any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We thank lab members for reading and commenting on the manuscript. We thank Frédéric Auradé (UPEC, Créteil, France) for providing the pMITomaR1 plasmid. We thank Jeffrey Dilworth (University of Madison) for scientific input and comments on the manuscript. We thank Stéphane Kerbrat from the IMRB genomic platform. We thank Odile Neyret from the IRCM molecular biology platform. We thank Sylvie Manin for lab management support. We thank the funding agencies Association Française contre les Myopathies (AFM) grant Translamuscle II - n° 22946 and Agence Nationale pour la Recherche (ANR), grant MyoID - ANR-22-CE13-0003-01.

Author contributions

V.Z., S.C., and F.I. performed and analyzed experiments contributing equally to this work. F.R. designed, supervised experiments and acquired funding. J.E.d.L. designed, performed, analyzed, and supervised experiments, wrote the manuscript, and acquired funding. All authors read, edited and commented on the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-DAXX	Santa Cruz	Cat#sc-7152; RRID: AB_2088784
Mouse monoclonal IgG anti-HIRA	Active Motif	Cat#39557; RRID: AB_2793256
Mouse monoclonal IgG1 anti-PAX7	Santa Cruz	Cat#sc-81648; RRID: AB_2159836
Rat monoclonal IgG2a anti-MYOD	Active Motif	Cat#39991; RRID: AB_2793421
Mouse monoclonal IgG2b anti-MF20	DSHB	MF 20; RRID: AB_2147781
Rabbit polyclonal anti-Cleaved Caspase-3	Cell Signaling	Cat#9661; RRID: AB_2341188
Rabbit polyclonal anti-gammaH2A.X	Abcam	Cat#ab11174; RRID: AB_297813
Mouse monoclonal anti-MYOD	Dako	Cat#M3512; RRID: AB_2148874
Rabbit monoclonal anti-GAPDH	Cell Signaling	Cat#2118; RRID: AB_561053
Rabbit polyclonal anti-H3	Abcam	Cat#ab1791; RRID: AB_302613
Peroxidase Goat polyclonal anti-rabbit	Vector Laboratories	Cat#PI-1000; RRID: AB_2336198
Peroxidase Goat polyclonal anti-mouse	Vector Laboratories	Cat#PI-2000; RRID: AB_2336177
Rat monoclonal IgG2a anti-H3.3	Cosmo Bio	Cat#CE-040B
Rabbit polyclonal anti-H3K27ac	Abcam	Cat#ab4729; RRID: AB_2118291
Rabbit polyclonal anti-H3K4me3	Merck Millipore	Cat#07-473; RRID: AB_1977252
Rabbit monoclonal anti-H3K27me3	Cell Signaling	Cat#9733; RRID: AB_2616029
Rabbit polyclonal anti-H3K9me3	Active Motif	Cat#39161; RRID: AB_2532132
Chemicals, peptides, and recombinant proteins
DMEM	Gibco	Cat#41966029
Fetal calf serum (FCS)	Eurobio	Cat#CVFSVF00-01
Pen/Strep	Gibco	Cat#15070063
Gelatin from pork skin	Sigma	Cat#G1890
Lipofectamine LTX PLUS reagent	Invitrogen	Cat#15338030
IgG-free BSA	Jackson	Cat#001-000-162
DAPI	Sigma	Cat#D9542
Protease Inhibitor Cocktail	Roche	Cat#04-693-116-001
Gelatin from cold water fish	Sigma	Cat#G7765
SuperSignal West Pico	ThermoScientific	Cat#35060
SuperScript III Reverse Transcriptase	Invitrogen	Cat#18080-093
Micrococcal Nuclease (MNase)	Sigma	Cat#N5386
Hydroxyapatite	BioRad	Cat#158-2000
Protein-A-coated magnetic beads	Diagenode	Cat#C03010020
Protein-G-coated magnetic beads	Diagenode	Cat#C03010021
DMSO	ThermoScientific	Cat#20688
Critical commercial assays
EdU Click-iT PLUS Kit	Invitrogen	Cat#C10640
NucleoSpin RNA extraction kit	Macherey Nagel	Cat#740955
KAPA Hyper Prep Kits	Roche	Cat#07962363001
NEBNext Library Quant Kit for Illumina	NEB	Cat#E7640AA
Deposited data
RNA-seq, ChIP-seq and ATAC-seq (Hira KO)	Esteves de Lima et al.23	GEO: GSE161056
ATAC-seq (Daxx KO)	This paper	GEO: GSE277345
ChIP-seq (Daxx KO)	This paper	GEO: GSE277346
RNA-seq (dKO)	This paper	GEO: GSE277347
Experimental models: Cell lines
Mouse: C2C12 cell line	ATCC	CRL1772
Mouse: Daxx KO C2C12 line	This paper	N/A
Mouse: Daxx and Hira double KO C2C12 line	This paper	N/A
Oligonucleotides
RT-qPCR primers, cDNA cloning primers, sgRNA primers, ChIP-RT-qPCR primers	This paper	See Table S1
Recombinant DNA
pU6-(BbsI)_CBh-Cas9-T2A-mCherry	Chu et al.60	Addgene plasmid #64324
pMITomaR1	Frédéric Auradé (UPEC, Créteil, France)	N/A
pMITomaR1-Daxx	This paper	N/A
Software and algorithms
Galaxy	The Galaxy Community61	https://usegalaxy.org
BOWTIE2	Langmead and Salzberg62	http://bowtie-bio.sourceforge.net/bowtie2/
featureCounts	Liao et al.63	N/A
DESeq2	Love et al.64	N/A
MACS2	Zhang et al.65	https://github.com/macs3-project/MACS
Integrated Genome Browser (IGB)	Freese et al.66	https://bioviz.org/igb/
ChIPseeker	Yu et al.67	N/A
deepTools2	Ramírez et al.68	https://deeptools.readthedocs.io/
ImageJ	Schneider et al.69	https://imagej.nih.gov/ij/
GraphPad Prism version 8	GraphPad Software	https://www.graphpad.com/
BioRender	BioRender.com	https://www.biorender.com/
displayR software	Displayr	https://www.displayr.com/
Zeiss Zen Lite v2.3 software	Zeiss	https://www.zeiss.com/

Experimental model and study participant details

C2C12 (ATCC: CRL1772) cell line was obtained from DSMZ Germany. Authentication as stated by DSMZ Germany ”was confirmed as mouse by PCR species and recently as Mus musculus by COI DNA Barcoding“. We performed the Daxx KO and the Hira:Daxx double KO by Crispr/Cas9 in this study. Hira KO was previously described.²³ C2C12 KO cell line mutations were confirmed by sequencing. The absence of RNA and protein was confirmed by RT-qPCR and immunohistochemistry, respectively. Cells were grown in DMEM (Gibco, 41966029) supplemented with 10% fetal calf serum (FCS) (Eurobio, CVFSVF00-01) and 1% Pen/Strep (Gibco, 15070063) at 37°C. Cell lines were tested negative for mycoplasma contamination.

Method details

Culture of C2C12 cells and transfection

C2C12 cell line (ATCC: CRL1772), obtained from DSMZ Germany, were grown in DMEM (Gibco, 41966029) supplemented with 10% fetal calf serum (FCS) (Eurobio, CVFSVF00-01) and 1% Pen/Strep (Gibco, 15070063). For differentiation assays, C2C12 were seeded at a confluence of 80% with 2% FCS for 3 days. For EdU proliferation analysis, cells were incubated with 10mM EdU (Invitrogen, A10044) for 2h (Invitrogen, EdU Click-iT PLUS Kit, C10640). Daxx full-length cDNA was obtained from C2C12 cDNA with the primers described in Table S1, which included BamHI and MfeI restriction sites, respectively. The cDNA was ligated into the pMITomaR1 plasmid, a modified version of pMIGR1 with IRES-TdTomato instead of IRES-nlsGFP, kindly provided by Frédéric Auradé (UPEC, Créteil, France), previously linearized with XhoI (ThermoScientific, ER0692) and EcoRI (ThermoScientific, ER0271), using the T4 DNA-ligase (ThermoScientific, EL0011). 1.5 x 10⁵ C2C12 cells were transfected with 7μg of plasmid with Lipofectamine LTX PLUS reagent (Invitrogen, 15338030) and 3 days post-transfection the cells were collected for analysis.

Generation of the Daxx KO and Daxx+Hira dKO C2C12 line using CRISPR-Cas9

Daxx knockout and Daxx and Hira double knockout lines were generated by CRISPR-Cas9 using the pU6-(BbsI)_CBh-Cas9-T2A-mCherry plasmid, a gift from Ralf Kuehn (Addgene plasmid #64324; http://n2t.net/addgene:64324; RRID: Addgene_64324).⁶⁰ The single guide RNAs (sgRNAs) were obtained from the sgRNA optimized library.⁷⁰ The primers were designed with added BbsI (ThermoScientific, FD1014) restriction sites and an extra G/C for increased hU6 promoter efficiency (Table S1 and²³) annealed and ligated to BbsI-linearized pU6-CBh-Cas9-T2A-mCherry plasmid using the T4-ligase (ThermoScientific, EL0011). 1.5 x 10⁵ C2C12 cells were transfected with 7μg of plasmid with Lipofectamine LTX PLUS reagent (Invitrogen, 15338030) and 3 days post-transfection the cells were FACS-isolated for mCherry-positive. The sorted cells were seeded at a low confluence (500 cells per 10cm dish) in dishes previously coated with 0.1% gelatin from pork skin (Sigma, G1890). Individual clones were isolated, expanded, genotyped by PCR and sequenced to confirm the presence of the mutation.

Immunohistochemistry

Cells were fixed in 4% PFA for 8min at room temperature (RT), permeabilized in 0.5% Triton in PBS 1X for 10min at RT and incubated 1h at RT in blocking solution consisting of 5% immunoglobin G (IgG)-free BSA (Jackson, 001-000-162) in PBS 1X. Primary antibodies (Table S2) were diluted in blocking solution and incubation took place overnight at 4°C. Secondary antibody incubation (Invitrogen, Alexa-Fluor) was performed 1h at RT followed by DAPI (Sigma, D9542) staining (1:5000) 10min at RT. EdU staining reaction was performed according to manufacturer’s guidelines (Invitrogen, EdU Click-iT PLUS Kit, C10640).

Western blot

Cells were lysed in RIPA buffer (50mM Tris HCl (Sigma, T5941), 150mM NaCl (Sigma, 9625), 1% Igepal (Sigma, I8896), 0.5% Sodium deoxycholate (Sigma, D6750), 0.1% SDS (Sigma, L3771), 1mM EDTA (Sigma, E5134), 1X Protease Inhibitor Cocktail (Roche, 04-693-116-001); pH8) for protein extraction. 10μg of protein were used per sample. The blocking of the membrane was performed in 0.5% of gelatin from cold water fish (Sigma, G7765) and 5% Tween 20 (Sigma, P9416) in PBS (Invitrogen, 003002) 1h at RT. Primary antibodies were diluted in blocking solution and incubated 2h at room temperature (Table S2). Secondary antibodies were diluted in blocking solution and incubated 1h at room temperature (Table S2). Revelation reaction was performed from 30s to 5min at room temperature using SuperSignal West Pico (ThermoScientific, 35060). Full gel images are present in Figure S6.

RNA extraction, cDNA synthesis and RT-qPCR

A minimum of 2 × 10⁵ C2C12 cells, cultured in high serum-containing medium and at low density, were collected per sample for RNA extraction (Macherey Nagel, 740955) following the manufacturer protocol. Reverse transcription was performed using the SuperScript III Reverse Transcriptase (Invitrogen, 18080-093) following the manufacturer’s guidelines. RT-qPCR was performed using the PowerUp SYBR Green Master Mix (Applied Biosystems, A25742). The relative mRNA levels were calculated using the 2^ˆ−ΔΔCt method.⁷¹ The ΔCt were obtained from Ct normalized to the housekeeping gene Tbp levels in each sample. The RT-qPCR primers used are listed in the Table S1.

RNA-sequencing

RNA was prepared as described for RNA extraction and sent to Integragen (Daxx KO sample, which was prepared and sequenced at the same time as the samples in GEO GSE161056) or to Institut Mondor de Recherche Biomédicale genomic platform (dKO and corresponding C2C12 control), deposited in GEO GSE277347. Libraries were prepared with TruSeq Stranded Total RNA Sample preparation kit according to supplier recommendations. Briefly, ribosomal RNA fraction was removed from 1μg of total RNA using the Ribo-Zero Gold Kit; fragmentation was performed using divalent cations under elevated temperature to obtain approximately 300bp pieces; followed by double strand cDNA synthesis Illumina adapters ligation and cDNA library amplification by PCR for sequencing. Sequencing for Daxx KO (with Hira KO and C2C12 control being in GEO GSE161056) was carried out on paired-end 75bp of Illumina HiSeq4000. For dKO and corresponding C2C12 control, samples sequencing was carried out on single-end 75 bp of Illumina NextSeq500 (GSE277347).

RNA-sequencing analysis

The RNA-seq analysis was performed using the Galaxy web platform,⁶¹ public server https://usegalaxy.org. The FASTQ files were uploaded in Galaxy and formatted as Sanger using the FASTQ groomer tool.⁶³ The quality of the data was analyzed with FastQC tool v0.72. FASTQ Sanger files were aligned to the mm10 mouse genome using the built-in index of BOWTIE2 v2.3.4.2.⁶² Genes were counted using featureCounts v1.6.3+galaxy2⁷² and differently expressed genes were determined by DESeq2 v2.11.40.6+galaxy1,⁶⁴ obtaining also the MA-Plot and the sample-to-sample distances plot. Transposable elements (TE) differential analysis was performed using TEtranscripts v2.2.3+galaxy0. Gene ontology analyses were performed on http://geneontology.org/ using the biological process option. The gene expression heatmap was created with displayR software with normalized reads for each triplicate.

Chromatin immunoprecipitation (ChIP)

Daxx KO samples GEO GSE277346 were prepared and sequenced at the same time as the samples in GEO GSE161056 and ChIP was performed as follows.⁷³ C2C12 cells, cultured in high serum-containing medium and at low density were trypsinized, washed and subjected to nuclei isolation, chromatin fragmentation with Micrococcal Nuclease (MNase) (Sigma, N5386) and nucleosome purification by Hydroxyapatite (BioRad, 158–2000) chromatography. Immunoprecipitation was performed overnight at 4°C with 6μg of chromatin and 5μg of antibody (Table S1), previously incubated with Protein-A- (for H3K27ac, H3K4me3, H3K9me3 and H3K27me3 antibodies) or Protein-G-coated (H3.3 antibodies) magnetic beads (Diagenode, C03010020 and C03010021, respectively) and analyzed by RT-qPCR as percentage of the input or by sequencing. For sequencing, samples were sent to the IRCM (Institut de recherches cliniques de Montréal) molecular biology platform. Library was prepared using KAPA Hyper Prep Kits with PCR Library Amplification/Illumina series (Roche, 07962363001) with IDT for Illumina TruSeq DNA-RNA UD 96 Indexes (UDI) (Illumina, 2023784) and quantified by RT-qPCR using NEBNext Library Quant Kit for Illumina (NEB, E7640AA). Sequencing was carried out on paired-end 50bp of Illumina HiSeq4000.

ChIP-sequencing analysis

The ChIP-seq analysis was performed using the Galaxy web platform⁶¹ public server https://usegalaxy.org. The FASTQ files were uploaded in Galaxy and formatted as Sanger using the FASTQ groomer tool.⁶³ Quality of ChIP-seq data was analyzed with FastQC tool v0.72. FASTQ Sanger files were aligned to the mm10 mouse genome using the built-in index of BOWTIE2 v2.3.4.2.⁶² The peaks were called with MACS2 v2.1.1.2⁶⁵ using paired-end BAM files with the cutoff q-value 5-e2 and the broad region calling off, the input was used as control. MACS2 resulting bedgraph file was converted to bigwig format using the wig/bedgraph-to-bigwig converter v1.1.1 for visualization in the Integrated Genome Browser v2.17.2.⁶⁶ The called peaks were annotated using the ChIPseeker v1.18.0 tool.⁶⁷ Average signal graphs were performed using computeMatrix Galaxy version 3.3.2.0.0⁶⁸ to prepare data for plotting using plotProfile Galaxy version 3.3.2.0.0.⁶⁸

ATAC-sequencing

Daxx KO cells GEO GSE277345 were prepared and sequenced at the same time as the samples in GEO GSE161056. C2C12 cells, cultured in high serum-containing medium and at low density, were collected and frozen in culture media containing 10% FCS (Eurobio, CVFSVF00-01) and 10% DMSO (ThermoScientific, 20688). Cryopreserved cells were sent to Active Motif. The cells were thawed in a 37°C water bath, pelleted, washed with cold PBS and tagmented as previously described,⁷⁴ with some modifications based on previous publication.⁷⁵ Cell pellets were resuspended in lysis buffer, pelleted and tagmented using the enzyme and buffer provided in the Nextera Library Prep Kit (Illumina). Tagmented DNA was purified using the MinElute PCR purification kit (Qiagen), amplified with 10 cycles of PCR and purified using Agencourt AMPure SPRI beads (Beckman Coulter). Resulting material was quantified using the KAPA Library Quantification Kit for Illumina platforms (KAPA Biosystems) and sequenced with PE42 sequencing on the NextSeq 500 sequencer (Illumina).

ATAC-sequencing analysis

Reads were aligned using the BWA algorithm (mem mode; default settings). Duplicate reads were removed, only reads mapping as matched pairs and only uniquely mapped reads (mapping quality ≥ 1) were used for further analysis. Alignments were extended in silico at their 3′-ends to a length of 200bp and assigned to 32-nt bins along the genome. The resulting histograms (genomic “signal maps”) were stored in bigWig files. Peaks were identified using the MACS 2.1.0 algorithm at a cutoff of p-value 1e-7, without control file and with the –nomodel option. Signal maps and peak locations were used as input data to Active Motif’s proprietary analysis program, which creates Excel tables containing detailed information on sample comparison, peak metrics, peak locations and gene annotations.

Quantification and statistical analysis

Image capturing and quantification

Image capturing was performed using a Zeiss LSM 800 confocal microscope with the associated Zeiss Zen Lite v2.3 software or a Zeiss Imager D.1 fluorescent microscope associated with the Zen 2 blue edition software. Quantifications were performed in at least 5 pictures taken randomly within each sample, per experiment, using ImageJ.⁶⁹

Statistical analysis

The statistical test performed in each analysis, the meaning and exact value of n and the definition of the error bar, are described in the associated figure legends. Datasets with 3 samples were analyzed with two-tailed unpaired t-test, datasets with more than 3 samples were analyzed with Mann-Whitney non-parametric test. Statistics was performed using GraphPad Prism version 8. p-values below 0.05 were considered significant.

Published: July 16, 2025