H2A.Z Is Dispensable for Both Basal and Activated Transcription in Post-Mitotic Mouse Muscles

bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

H2A.Z is dispensable for both basal and activated transcription in post-mitotic mouse muscles

Edwige Belotti1, Nicolas Lacoste1, Thomas Simonet1, Christophe Papin2, Kiran Padmanabhan3, Lorrie Ramos4, Defne Dalkara4, Isabella Scionti1, Ali Hamiche2*, Stefan Dimitrov4*, Laurent Schaeffer1,5*

1. Institut NeuroMyogène, Université Claude Bernard Lyon 1, INSERM U1217, CNRS UMR5310, 8 avenue Rockefeller 69008 LYON, France. 2. Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, Parc d’innovation, 1 rue Laurent Fries, 67404 Ilkirch Cedex, France. 3. Institut de Génomique Fonctionnelle de Lyon, CNRS UMR 5242, École Normale Supérieure de Lyon, Université Claude Bernard Lyon 1, Lyon, France. 4. Institute for Advanced Biosciences (IAB), Université Grenoble Alpes, CNRS UMR 5309, INSERM U1209, Site Santé - Allée des Alpes, 38700 La Tronche, France 5. Centre de Biotechnologie Cellulaire, Hospices Civils de Lyon, Lyon, France. * Equal last and corresponding authors

ABSTRACT

The histone variant H2A.Z is enriched in nucleosomes surrounding the transcription start site of active promoters, suggesting that it might be implicated in transcription. It is also required during mitosis. However, evidences obtained so far mainly rely on correlative evidences obtained in actively dividing cells. We have defined a paradigm in which cell cycle cannot interfere with H2A.Z transcriptional studies by developing an in vivo systems to invalidate H2A.Z in terminally differentiated post-mitotic muscle cells to dissociate its role during transcription from its role during mitosis. ChIP-seq, RNA-seq and ATAC-seq experiments performed on H2A.Z KO post-mitotic muscle cells show that this histone variant is neither required to maintain nor to activate transcription. Altogether, this study provides in vivo evidence that in the absence of mitosis H2A.Z is dispensable for transcription and that the enrichment of H2A.Z on active promoters is rather a marker than an actor of transcriptional activity.

INTRODUCTION comprises at least the histone variants macroH2A, H2A.X, H2A.Z and H2A.Bbd11. Histone variants are non-allelic isoforms of In mammals, the H2A.Z variant comprises conventional histones1,2. Each histone has two proteins: H2A.Z-1 and H2A.Z-212,13. They histone variants even H4 but only in are encoded by two distinct genes located on hominids1,3,4. Histone variants differ in their two different chromosomes (Chr3 and Chr11, primary sequence, transcriptional regulation respectively) and they differ by only three and timing of expression during the cell cycle amino-acids12. H2A.Z is an essential protein compared to conventional histones1–3. in mammals14. ANP32E and YL1 are part of Incorporation of histone variants can confer large protein complexes and were identified novel structural properties to nucleosomes as H2A.Z specific chaperones, responsible and change the functional chromatin for either its eviction or its deposition in landscape3,5–7. The current view is that chromatin15,16. In vertebrates, genome-wide histone variants are involved in all aspects of studies have shown that H2A.Z is enriched at DNA metabolism including transcription, active promoters17–19, in facultative replication and repair1. They are likely heterochromatin20 and at centromeres21,22. implicated in numerous diseases, particularly Interestingly, +1 and -1 nucleosomes in cancer (for review see8,9,10). surrounding the Transcription Start Site The H2A family of histone variants is one of (TSS) are enriched in H2A.Z in proportions the richest. In addition to conventional H2A, it correlated with the strength of

1 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

transcription18,23. H2A.Z is also preferentially (MURF1, MAFBX, MUSA1)28–30. Depriving bound to other regulatory elements such as muscles of their innervation thus provides a enhancers and CTCF-binding sites, which convenient mean to evaluate the involvement mark insulator sites in the genome15,17,24 and of H2A.Z in de novo regulation of gene more generally in chromatin regions where expression. histones have a high turnover. From all these Our RNA-seq and ATAC-seq analysis of observations, a consensus emerged innervated and denervated H2A.Z dKO stipulating that H2A.Z was a key player in the muscles revealed that the absence of H2A.Z control of transcription17,18,25. affected neither steady state nor activated However, the available data on the link gene expression. between H2A.Z-containing nucleosomes and transcription remain mainly correlative and based on studies carried out in cultured cells. RESULTS The interpretation of such results have to take into account perturbations of replication Invalidation of H2A.Z-1 and H2A.Z-2 in and cell division that can indirectly affect MEFs cells. H2A.Z-1flox/flox:H2A.Z-2flox/flox gene expression. For example, depletion of mouse line was obtained with the strategy H2A.Z in vertebrates results in strong mitotic described in figure 1A and B. Briefly, loxP and cytokinesis defects and can activate an sequences were inserted before exon 2 and apoptotic program subsequently leading to after exon 4 of h2afz (H2A.Z-1) and before cell death26. Thus, in this case the effect of exon 4 and after exon 5 of h2afv (H2A.Z-2) H2A.Z on transcription might be indirect and using standard recombination strategy. To would be linked with mitotic functions of evaluate the efficacy of our inactivation H2A.Z. It appears that unravelling the exact strategy and the stability of H2A.Z we have function of H2A.Z in the control of first isolated Mouse Embryonic Fibroblasts transcription requires a system in which (MEF) cell lines from the double floxed H2A.Z can be inactivated in cells that do not mouse line. Upon Cre induction using an replicate DNA anymore. Skeletal muscle and adenoviral system, h2afz and h2afv knock- the Cre-/loxP system provide such a down was was evaluated by RT-qPCR (Fig. paradigm. 1C), and H2A.Z protein depletion was Transgenic mice, expressing the Cre evaluated by western blotting (Fig. 1D) and recombinase under the control of the human immunofluorescence (Fig. 1E). Two days α-skeletal actin promoter (HSA-Cre), allow to after infection by the adeno-Cre H2A.Z specifically express the recombinase in post- protein was already undetectable in cycling mitotic skeletal muscle cells27. Thus, cells. H2A.Z dKO cells did not survive more breeding HSA-Cre mice with H2A.Z- than 9 days after H2A.Z withdrawal, 1flox/flox:H2A.Z-2flox/flox mice specifically recapitulating the essential function of inactivates H2A.Z-1 and H2A.Z-2 in post- H2A.Z26. We then used RT-qPCR to evaluate mitotic skeletal muscle cells (hereafter the expression of a set of genes known to be named H2A.Z dKO). These cells provide a affected by the absence of H2A.Z in mouse convenient tool to analyse in vivo the cells25. As expected, they were misregulated requirement of H2A.Z for transcription in the absence of H2A.Z (Fig. 1F). We then maintenance and regulation. initiated the study of the role of H2A.Z in post- Gene expression in skeletal muscle cells is mitotic muscle cells in vivo to determine if in tightly controlled by innervation. Two days absence of replication and cell division the after denervation muscle cells have adopted same effects would be observed. a new gene expression program involving the activation and the repression of hundreds of H2A.Z is enriched at TSS of active genes genes28. The best characterized genes in mouse skeletal muscle. Currently, no activated upon denervation encode a variety data are available on the distribution of of proteins: transcription factors (MYOD, H2A.Z in vivo in skeletal muscle and more Myogenin), neurotransmitter receptors generally in post-mitotic cells. To analyse (acetylcholine receptor subunits), kinases H2A.Z genomic distribution in adult skeletal (MuSK, PAK1), histone deacetylases muscle, ChIP-seq was performed on tibialis (HDAC6, HDAC4) and ubiquitin ligases anterior (TA) muscles from 7 weeks old

2 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 1. Generation and characterisation of the H2A.Z-1 and H2A.Z-2 models. (A) (i) Wild-type H2afz gene structure. Black boxes indicate the exons. (ii) The conditional allele of the H2afz flox/flox gene. LoxP sites were inserted at the indicated position. (iii) Organization of the H2A.Z-1 knock-out (KO). Exons 2, 3 and 4 are deleted upon expression of Cre recombinase to generate the KO allele H2afz (-/-). (B) Same as (A), but for the H2afv gene. LoxP site was inserted on both ends of exon 4-5. After Cre-recombinase expression, exon 4 and 5 are deleted to create the KO allele H2afv (-/-). (C) Bar graphs representing the expression level of H2afz and H2afv genes measured by RT-qPCR analysis in MEFs infected with either control adenovirus or with adenovirus expressing Cre recombinase. (D) H2A.Z is efficiently depleted upon infection of MEFs with adenovirus expressing Cre recombinase. Western blot of MEF cells using anti- H2A.Z antibody. H4 was used as loading control. (E) Immunofluorescence detection of H2A.Z on proliferating MEFs 2 days after infection. (E) RT-qPCR of RNA isolated from CTL and H2A.Z double flox MEFs cells in absence or presence of adeno-Cre virus particles. On the left panel, some up-regulated genes in absence of H2A.Z and on the right panel some down-regulated genes in absence of H2A.Z. mouse. Results indicated that H2A.Z was Using ATAC-seq in control muscles to enriched around transcriptional start sites assess chromatin accessibility at the genome (TSS) as in most other cell types31 (Fig. 2A). level, we could readily detect a correlation RNA-seq was then performed to compare between the presence of H2A.Z and gene expression data to the ChIP-seq. The accessible regions of the DNA (Fig. 2C). results showed that the abundance of H2A.Z Since ChIP-seq experiments showed a was correlated to gene expression (Fig. 2B). correlation between the strength of Altogether, these results are in agreement transcription and H2A.Z enrichment at the with the numerous studies showing that TSS, we can conclude that in muscle cells, H2A.Z is enriched around the TSS of active as in any cell type, a correlation exists genes15,17,18. Skeletal muscle thus provides between transcription strength and chromatin an adequate system to evaluate the accessibility. ChIP-seq and ATAC-seq requirement of H2A.Z for gene expression in alignment further indicated that in muscle post-mitotic cells. fibers, as in other mouse tissues32, chromatin

3 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 2. H2A.Z localisation and DNA accessibility at genome-wide level according to the transcription. (A) Distribution of H2A.Z relative to the transcription start sites (TSS) (the distinct colors indicate the different levels of transcription) and (B) Heat map of input and H2A.Z ChIP-seq enrichments around the transcription start sites (TSSs) ranked according the RNA-seq levels, in muscle skeletal myofibers. (C) Heat map of the ATAC-seq signal and H2A.Z Chip-seq enrichment around the transcription start sites (TSSs) ranked according the RNA-seq levels. (D) Distribution of H2A.Z and DNA accessibility relative to the TSS.

was the most accessible between the +1 and RT-qPCR (Fig. 3B) and western blotting (Fig. the -1 nucleosomes surrounding the TSS 3C). In both cases, the remaining signal (∼15- (Fig. 2D). 20 %) reflects the presence of satellite cells and non-muscle cells e.g. fibroblasts, Invalidation of H2A.Z-1 and H2A.Z-2 in endothelial cells, macrophages34. mouse skeletal muscle fibers. The H2A.Z- Immunofluorescence experiments were 1flox/flox:H2A.Z-2flox/flox mouse line was crossed carried out to specifically visualize the decline with mice expressing the Cre recombinase in H2A.Z content in the nuclei of muscle under the control of the HSA promoter that fibers (Fig. 3A and D). One week after birth specifically expresses the Cre recombinase H2A.Z levels were already significantly in post-mitotic skeletal muscle cells27. In decreased (Fig. 3A) and by post-natal week HSA-Cre mice, the expression of the Cre 4, H2A.Z could not be detected anymore in recombinase is initiated in the second half of muscle nuclei. Consistently, 7 weeks after embryonic development and reaches its birth no H2A.Z could be detected in muscle plateau of expression around birth33. The nuclei whereas its expression remained depletion of both H2A.Z in the TA muscles of unaffected in non-muscle cells (Fig. 3D). For 7-8 weeks old mice was demonstrated by further analysis, muscles were collected

4 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 3. Validation of the H2A.Z cKO mouse model. (A) Immunofluorescence detection of H2A.Z on tibialis anterior muscle fibers from 1 week, 2 weeks and 4 weeks old CTL and H2A.Z dKO mice. DAPI was used to stain the DNA (Scale bar 2 µm). (B) H2A.Z-1 and H2A.Z-2 RT-qPCR of RNA isolated from the double KO mouse strain on 7 weeks aged mice (n = 3). (C) Western blotting of H2A.Z from nuclear extracts of skeletal muscles from CTL and double cKO H2A.Z-1(-/-) x H2A.Z-2(-/-) (H2A.Z dKO) on 7 weeks old mice. H4 is used as loading control. (C) Immunofluorescence detection of H2A.Z on EDL muscle fibers from 7 weeks old CTL and H2A.Z dKO mice. α-bungarotoxine is used to stain the neuromuscular junction and DAPI to stain the DNA (Scale bar 10 µm).

seven weeks after birth, thus several weeks formation of new muscle fibers from resident after the depletion of H2A.Z (Fig. 3A). adult muscle stem cells. Regenerating Phenotypically, seven weeks old H2A.Z dKO muscle fibers are easily detected by the mice were indistinguishable from control central position of their nuclei since in other mice (CTL). Body weight and muscle fibers the nuclei are located at the periphery, histology of these mice were similar to the directly beneath the plasma membrane. In control ones. Fibrosis, fiber type composition, healthy muscles the proportion of muscle mitochondria distribution and function were fibers with central nuclei does not exceed identical in control and H2A.Z dKO muscles 1%. This value was not increased in H2A.Z (Fig. 4A, see below supplementary Fig. 3). dKO muscles, indicating an absence of active Importantly, the distribution of fiber size regenerative process. Altogether, these data (cross-sectional area) was comparable revealed that H2A.Z depletion in skeletal between CTL and H2A.Z dKO muscles (Fig. muscle fibers did not induce any detectable 4B). In adult muscles, the loss of muscle histological alteration. fibers is usually compensated by the 5 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 4. Characterization of the H2A.Z cKO mouse model. (A) Histological analysis of tibialis anterior muscles by Hematoxyline and Eosin (HE) and Wheat Germ Agglutinin-Lectin (WGA) staining from 7 weeks old CTL and H2A.Z dKO mice (Scale bar 20 µm, arrows indicate centronucleated fibers). (B) Fiber size distribution analysed from 200 myofibers of each samples. (C) Scatter plots comparing global gene expression levels between CTL and H2A.Z dKO cells in muscle from 7 weeks-old mice. (D) Heat map of the ATAC-seq signal enrichment around the transcription start sites (TSSs) in CTL and H2A.Z dKO, the TSSs being ranked according to the CTL level. (E) Scatter plot comparing ATAC-seq peaks enrichment in CTL and H2A.Z dKO muscles.

The absence of H2A.Z perturbs neither and no clustering of these genes was steady state gene expression nor the observed (Fig. 4C, Table 1). Therefore, chromatin landscape. To evaluate the removing H2A.Z from skeletal muscle fibers impact of H2A.Z inactivation on steady state does not significantly perturb steady state gene expression, RNA-seq experiment was gene expression. performed on 7 weeks old CTL and H2A.Z ATAC-seq results sorted by transcriptional dKO TA muscles. Analysis of results strength generated identical patterns in confirmed the inactivation of both H2A.Z control and H2A.Z dKO muscles (Fig. 4D). isoforms in H2A.Z dKO mice (Supplementary This was further confirmed by directly Fig. 1). The transcriptomes of CTL and comparing the peaks from each experiment H2A.Z dKO muscles were almost identical (Fig. 4E). Altogether, these results indicate (Fig. 4C). In the absence of H2A.Z, the that in muscle fibers nuclei the absence of expression of only 24 and 25 genes was H2A.Z did not significantly change gene slightly activated or repressed, respectively, expression or DNA accessibility.

6 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 5. Characterization of the H2A.Z cKO under acute transcriptional response. (A, B) Scatter plots comparing global gene expression levels between innervated and denervated cells (in 7 weeks-old mice) in the presence (A) or in the absence (B) of H2A.Z. (C) Scatter plots comparing global gene expression levels 48h after denervation between CTL and H2A.Z dKO cells in muscle cells of 7 weeks-old mice. (D) Functional annotation clustering of differentially expressed genes in response to denervation.

The absence of H2A.Z perturbs neither carried out on RNA purified from either acute gene activation nor repression in innervated or denervated TA muscles. As mouse skeletal muscle. To evaluate the expected, the expression of Myogenin, possibility that H2A.Z would be required to MuSK, Acetylcholine receptor α and δ and prime transcriptional changes but would be Hdac4 was strongly upregulated (Table 1), dispensable once promoters were already thus confirming that muscles responded activated or repressed, we analysed the correctly to denervation. In agreement with effect of H2A.Z inactivation upon acute available data35, the genome-wide changes in gene expression. For this transcriptome analysis revealed that upon purpose, seven weeks old CTL and H2A.Z denervation hundreds of genes were either dKO TA muscles were denervated. upregulated (894 genes) or downregulated Denervation is carried out by section of the (944 genes) (Fig. 5A, B). Denervation sciatic nerve of one hind limb. 48h after induced the downregulation of many genes denervation, TA muscles were collected, the involved in contractility and the upregulation contralateral TA muscles was used as of genes involved in muscle development innervated controls. RT-qPCR experiments and inflammation (Fig. 5D). In H2A.Z dKO showed that the strong post-denervation muscles, 889 genes were upregulated and upregulation of MyoD and Myogenin 876 downregulated. Up- and down-regulated expression took place normally in the genes were the same in CTL and dKO absence of H2A.Z (Supplementary Fig. 2). muscles, indicating that the vast majority of Note that no histological changes in genes is normally regulated by denervation in denervated H2A.Z dKO mice compared to the absence of H2A.Z (Fig. 5C and Table 1). CTL mice have been detected All together, the depletion of H2A.Z did not (Supplementary Fig. 3). significantly affect the genome–wide To have a global view of the transcriptional transcriptional response to denervation (Fig. response to denervation, RNA-seq was 5C). We conclude that in post-mitotic skeletal

7 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Figure 6. H2A.Z at DNA repetitive elements. (A) Scatter plot showing the average enrichment of H2A.Z in repeat families. (B) Scatter plots comparing global transcription of repetitive elements between CTL and H2A.Z dKO myofibers in innervated muscle of 7 weeks-old mice. (C) Scatter plots comparing global transcription of repetitive elements in innervated and denervated muscle cells (7 weeks-old mice) in the presence of H2A.Z. (D) Scatter plots comparing global transcription of repetitive elements between CTL and H2A.Z dKO myofibers in denervated muscle of 7 weeks-old mice. (E) Bar graphs representing the expression level of the MLTH1 and RLTR6_Mm retroelements, the two most overexpressed retroelements in response to denervation.

muscle cells, H2A.Z is not required to expression of these repetitive DNA elements. activate or to repress transcription to adapt to To test this hypothesis, we analysed how the a particular physiological context. absence of H2A.Z affected the expression of repetitive DNA elements in innervated and H2A.Z is enriched at DNA repetitive denervated muscles of seven weeks old mice elements. Since H2A.Z is also known to be (Fig. 6). The expression of repetitive enriched at pericentromeric regions in the elements was indistinguishable in CTL and mouse genome, we specifically analysed H2A.Z dKO innervated muscle (Fig. 6B). repetitive sequences that represent a very Interestingly, denervation altered the high proportion of these regions20,21. As expression of several types of repetitive expected, our ChIP-seq data indicated that elements in control muscles. 72 families were numerous families of DNA repetitive up-regulated and 3 were down-regulated 48 elements were enriched in H2A.Z (Fig. 6A). hours after denervation (Fig. 6C). These The families that exhibited the higher changes were similar in H2A.Z dKO muscles enrichment are listed in Table 2. Of note, the (Fig. 6D, E, Table 3). The MLTH1 and RLTR6 majority of repeats significantly enriched in families of repetitive elements were among H2A.Z (log2 enrichment > 0.5 and P value < the most up-regulated by denervation. H2A.Z 10-2) belonged to the simple repeats and to inactivation affected neither their expression the endogenous retrovirus classes of in innervated muscle nor their activation in repetitive elements. This suggested that denervated muscle (Fig. 6E). Taken as a H2A.Z might be involved in the control of the whole, these results indicate that H2A.Z is

8 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

neither required for normal expression nor for immunofluorescence was therefore the best the activation of repetitive DNA elements. way to evaluate the efficiency of H2A.Z knock-down in muscle fibers. In H2A.Z dKO muscles, H2A.Z staining had already DISCUSSION decreased below the detection level by 2 weeks of age and H2A.Z remained The goal of this study was to evaluate in vivo undetectable in muscle fibers since then. in post-mitotic tissues the function of H2A.Z When analysed seven weeks after birth, in transcription. As model system we have H2A.Z dKO mice and H2A.Z dKO muscles first isolated MEF cells derived from did not show any detectable alteration (Fig. H2AZ.1flox/flox:H2AZ.2flox/flox mice. In cycling 4A, B). Strikingly, RNA-seq analysis revealed MEFs, H2A.Z was depleted in less than two no significant effect of H2A.Z depletion on the days after adeno-Cre infection and many transcriptomic profile of H2A.Z dKO muscles. genes were up and down regulated (Fig. 1F Induction of a new transcriptional program by and data not shown) according to the denervation was not perturbed either by the literature25,36,37. We next sought to avoid a absence of H2A.Z. Moreover, the chromatin possible influence of the cell cycle on gene accessibility analysed by ATAC-seq did not expression in the absence of H2A.Z. An in reveal any differences between control and vivo approach in skeletal muscle was chosen H2A.Z dKO muscles. because it provides access to long lived post- H2A.Z is also preferentially enriched at mitotic cells in which Cre recombinase regulatory elements such as enhancers and expression can be efficiently targeted and CTCF-binding sites, which mark insulator where consequences can be analysed sites in the genome17. ATAC-seq results several weeks after gene inactivation. In would have shown changes in chromatin addition, denervation provides a convenient accessibility if H2A.Z depletion had perturbed means to rapidly change the transcriptional their function. We can therefore conclude that program of the targeted cells. as for proximal promoters regulation, H2A.Z ChIP-seq analysis confirmed that in skeletal is not required for enhancers and insulators muscle fibers, H2A.Z was enriched at function. transcriptional start sites and on regulatory The transcriptional function of H2A.Z is elements as previously described in other probably not as important as initially thought systems15,17,18,23. To specifically inactivate but it is still an essential protein in cycling H2A.Z in the skeletal muscle fibers of H2A.Z- cells. It is known that it is enriched at centric 1flox/flox:H2A.Z-2flox/flox mice, HSA-Cre mice and pericentromeric heterochromatin21,43,44 were used. This Cre driver mouse line is and it is also, somehow, involved in mitosis21. probably the best characterized muscle If H2A.Z has key roles in organizing the specific Cre-driver mice and in skeletal centrosome or other chromatin loci crucial for muscle they allow to selectively target post- mitosis, they are important enough to lead to mitotic cells27. cell death when perturbed. The fact that none Skeletal muscles contain a large variety of of these functions are required in post-mitotic cell types (fibroblasts, satellite cells cells would explain why H2A.Z is not an macrophages, endothelial cells, …)34. The essential protein in post-mitotic muscle cells. Human Skeletal Actin promoter drives the It would be interesting to investigate the expression of the Cre recombinase only in nucleosome composition at TSS in the differentiated muscle cells. When performing absence of H2A.Z. Three non-exclusive muscle extracts from the whole tissue non possibilities would fit with our data. The first, muscle cells such as fibroblasts are inevitably and most likely hypothesis, is that H2A.Z present. These cells represent less than 20% would be replaced by canonical H2A. H2A of the muscle tissue but they are sufficiently expression is stronger during S phase45, represented to generate detectable however, low level of H2A expression is expression of inactivated genes in muscle probably sufficient in post-mitotic cells. This extracts. This is true for H2A.Z dKO muscles is the case for histone H4 for which there is and for all muscle specific inactivation of no known variant in the mouse genome (to ubiquitously expressed genes38–42. By our knowledge H4 has a variant only in allowing to visualize individual nuclei, hominids4). If H4 is expressed at low levels

9 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

from one of the histone clusters, H2A might role. Altogether, these results suggest that be as well. The second possibility is that H2A.Z is rather a marker than an actor of H2A.Z would be replaced by another H2A transcriptional activity. Recently, several variant. There are many H2A variants which reports on the essential variant H3.3 similarly can be incorporated in nucleosomes. claimed that it is dispensable for However, the expression of none of them transcription47,48. This raises the question of was upregulated in H2A.Z dKO muscles. The the role of histone variants in transcriptional third possibility is that H2A.Z would not be regulation as a whole. replaced by another histone, thus creating an incomplete nucleosome e.g. a tetrasome at the TSS (see46) or even gaps in chromatin if MATERIALS AND METHODS the tetrasome was too unstable. However, ATAC-seq results did not reveal any increase Generation of H2A.Z-1/H2A.Z-2 HSA-Cre in chromatin accessibility, arguing against mice this hypothesis. The most likely explanation The H2A.Z mouse line was generated at the is therefore that canonical H2A replaces “Institut Clinique de la Souris” (ICS-MCI; missing H2A.Z molecules, but ruling out the Illkirch, France). The targeting vector other possibilities will require further in depth containing a Frt Neomycin resistance investigations. cassette and (i) H2afz: in which exons 2-4 It has long been thought that since H2A.Z was flanked by loxP sites or (ii) H2afv: in was enriched at TSS, it was playing an active which exons 4-5 was flanked by loxP sites role in transcriptional regulation. Our results was generated. Each constructs were showing normal gene expression and electroporated into mouse embryonic stem regulation in post-mitotic muscle cells (ES) cells. Targeted ES cells were injected challenge this view. Almost all data available into C57BL/6 blastocysts which were so far were obtained using cycling cells and implanted in pseudo-pregnant females. could not rule out the possibility that H2A.Z Removal of the Neomycin cassette in the functions in other processes than targeted allele was accomplished by crossing transcription. A role in replication or DNA the chimeric males giving germline repair could easily explain why H2A.Z total transmission with Flp transgenic females to KO in mouse was lethal during early generate mice with the conditional allele (Fig. embryonic development. Our results suggest 1A and B). Each strain was backcrossed for that the requirement of H2A.Z for cellular 10 generations on C57BL/6N mice. Finally, viability in vivo is not linked to transcription. H2A.Z-1flox/flox:H2A.Z-2flox/flox mice were Alternatively, a differential involvement of obtained by crossing the mice with HSA-Cre H2A.Z in transcription in cycling and post- transgenic strain27 to generate the muscle mitotic cells could be envisaged. Even specific conditional double KO. Mice were though, H2A.Z accumulation at active TSS genotyped by PCR amplification of genomic and regulatory elements with no apparent DNA extracted from newborn biopsies. A 40 function in post-mitotic cells would still be µl volume of extraction buffer (0.5 mg/ml puzzling. proteinase K, 0.2% SDS, 0.2 M NaCl, 100 If not required for a specific transcriptional mM Tris, 5 mM EDTA, pH 8.0) was added to function, the accumulation of H2A.Z at TSS biopsies and incubated at 54 °C overnight. and regulatory regions is intriguing. A simple After DNA precipitation and resuspension in explanation of this phenomenon could be 200 µl of water, 1 µl of DNA was taken as provided by the fact that overall, H2A.Z template in a 25 µl PCR reaction using the accumulates at sites where nucleosome Go Taq DNA polymerase (Promega) turnover is high. H2A.Z would be according to the provider’s recommendation. preferentially incorporated in new Genotyping was performed with primers nucleosomes at these sites because it is listed in supplementary Table 1. more available than its canonical counterpart mainly expressed during the S phase. H2A.Z Cell culture would therefore be used to locally Mouse embryonic fibroblasts (MEFs) were reconstitute chromatin and to protect the derived from E13.5 WT and dKO embryos. DNA from damage without playing an active Heads and internal organs were removed,

10 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

the torso was minced into chunks of tissue. culture plate that have been washed once Cells were immortalized with a retrovirus before with PBS. Cells are then scrapped and expressing the large SV40 T antigen and place in a 1.5 ml tube and processed cultured in high glucose Dulbecco’s modified following provider’s instructions. To extract Eagle medium (DMEM), with sodium total RNA from muscles, 500 µl of TRI pyruvate, Glutamax (Gibco), 10% fetal Reagent (Sigma) was added to individual bovine serum (FBS), and penicillin– frozen Tibialis Anterior in tubes containing streptomycin in a humidified incubator at ceramics beads (Lysing Matrix D – MP 37°C and a 5% CO2 atmosphere. biomedicals) for homogenization in a Immortalized H2A-Z.1flox/flox:H2A-Z.2flox/flox PreCellys (Bertin Technologies) (6500 RPM, MEFs were plated at 30% of confluency and 3 × 10 s). After centrifugation, TRI Reagent were infected with recombinant adenoviruses was removed and beads were washed once encoding either GFP or a Cre-GFP fusion with 500 µl of TRI reagent. Total RNA was (Ad-CMV-GFP or Ad-CMV-Cre-GFP, Vector extracted following provider’s instruction. To Biolabs, Philadelphia, Pennsylvania, USA) to generate cDNA, total RNA was treated with respectively generate control (CTL) and DNase (Ambion) and reverse transcribed H2A.Z double KO (H2A.Z dKO) cell lines. with RevertAid H Minus Reverse Viruses were diluted in the culture media for Transcriptase (ThermoFisher) primed with overnight infection. Cells were then washed random hexamers. RT–qPCR was and were diluted 4 times every 2 days to have performed using SYBR Green Mastermix them always below a confluence of 70%. (Qiagen) in the CFX-connect system (Bio- Cells were collected every 2 days until 8 days Rad). Relative expression levels were after infection. normalized to GusB and Rpl41 housekeeping genes expression using the ΔΔCt method. Mice care Primers are listed in Supplementary Table 2. Animals were provided with mouse chow and water ad libitum in a restricted-access, RNA-seq specific pathogen–free animal care facility at Libraries of template molecules suitable for the Ecole Normale Supérieure of Lyon strand-specific high-throughput DNA (Plateau de Biologie Expérimentale de la sequencing were created using a TruSeq Souris). All procedures were performed in Stranded Total RNA with Ribo-Zero Gold accordance with national and European Prep Kit (RS-122-2301; Illumina) as legislation on animal experimentation. previously described49. The libraries were sequenced on Illumina Hiseq 4000 Denervation of hindlimb muscles sequencer as single-end 50 bp reads Section of the left sciatic nerve was used to following Illumina's instructions. Image induce denervation. Briefly, after the mice analysis and base calling were performed were anesthetized with an intraperitoneal using RTA 2.7.3 and bcl2fastq 2.17.1.14. injection of ketamine (100 mg.kg−1) and Adapter dimer reads were removed using xylazine (10 mg.kg−1), the sciatic nerve was DimerRemover v0.9.2. Reads were mapped exposed in the thigh and doubly cut (5 mm to the mouse genome (mm9) using Tophat48 apart) just distal to the sciatic notch. v2.0.14 and Bowtie50 v2-2.1.0. Quantification Contralateral muscles that were not operated of gene expression was performed using and were used as controls. 48 hours post- HTSeq51 v0.6.1 and gene annotations from denervation, non-denervated and Ensembl release 67. denervated tibialis anterior (TA) muscles were collected and snap frozen in liquid Nuclei preparation nitrogen. All operative procedures were Muscles of individual hind limbs were performed using aseptic techniques and collected and put in 4 ml of cold buffer A (300 according to the ethical committee mM sucrose, 10 mM NaCl, 1.5 mM MgCl2, 15 recommendations (Ceccap-ENS-2014-019). mM Tris HCl pH 8.0 and protease inhibitor cocktail (Sigma)). Muscles were minced with Quantitative RT-PCR scissors and transfered in a dounce For RNA extraction from MEF cells, 1ml of homogenizer where 50 strokes of the loose TRI reagent (Sigma) was added to 10 cm cell pestle were slowly applied.

11 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

The extract was then passed through a 100 to the mouse genome (mm9) using Bowtie2 µm cell strainer to remove myofibrils debris v1.0.0 with the following arguments: -m 1 -- and placed back into a clean cold dounce strata --best -y -S –l 40 -p 2. homogenizer. Nonidet P-40 was then added to the extract to a final concentration of 0.5%, ATAC-seq which was then incubated for 15 min at 4°C. Purified nuclei are quantified using DAPI and Subsequently, 50 strokes with the tight pestle 50 000 of them are taken to performed the were applied. The nuclei suspension was ATAC-seq protocol as described in finally clarified through a 40 µm cell strainer, Buenrostro et al52. Nextera adapters were centrifuged at 2000 g for 10 min at 4°C and trimmed from reads with Trimmomatic53. washed in cold buffer A. Purified nuclei were Remaining paired reads longer than 20bp resuspended in a small volume of buffer A for were mapped to the mouse genome (mm9) further ChIP-seq, ATAC-seq or immunoblot using Bowtie22, using the local alignment analysis. mode (--local) and a maximum fragment length set to 2000bp (--maxins). ChIP-seq Purified nuclei were crosslinked with 0.2% Immunoblot formaldehyde for 1 min before quenching Chromatin preparations were separated on a with 125 mM glycine for 5 min at room 18% SDS-PAGE and transferred onto temperature. The buffer was then changed to polyvinylidene fluoride (PVDF) Immobilon-P 300 mM NaCl, 300 mM sucrose 0.1% NP-40, membranes (Millipore). Immunoblots were 2 mM CaCl2, 15 mM Tris/HCl pH 8.0 following performed with enhanced a centrifugation at 2000 g for 10 min. MNase chemiluminescence (ECL) PLUS reagent (Roche diagnostic) was then added to the (GE Healthcare) according to the nuclei and incubated at 37°C for a period of manufacturer’s instructions. Antibodies used time long enough to generate a majority of were as follows: rabbit monoclonal anti- mono- and di-nucleosomes. The reaction Histone H4 pan (04-858, Merck Millipore); was stopped by increasing EDTA rabbit polyclonal anti-H3 (#61277); mouse concentration to 10 mM. After a 5 min monoclonal anti-α-Tubulin (T6074, Sigma); centrifugation at 20 000 g, 20 µg of chromatin the rabbit polyclonal anti-H2A.Z15. from the supernatant were incubated overnight at 4°C with an immunopurified Histology highly specific polyclonal anti-H2A.Z rabbit Hind limb muscles surrounding the tibial bone antibody generated in-house15, followed by 2 (Gastrocnemius, Plantaris, Soleus (GPS), TA hours incubation with Protein A and Extensor Digitorum Longus (EDL) were dynabeads (Invitrogen). Immunoprecipitated collected from 7 weeks-old CTL and H2A.Z material was sequentially washed with buffer dKO mice, frozen in isopentane cooled on dry containing low salt (150 mM NaCl), high salt ice, and cross-sectioned at 10 µm thickness (500 mM NaCl) and finally a high stringency in a cryostat. Transverse sections were wash (500 mM NaCl, 0.25 M LiCl and 1% stained with Hematoxylin and Eosin for sodium deoxycholate) in 10 mM Tris/HCl pH immunohistochemistry and with Wheat Germ 8, 1 mM EDTA buffer. DNA was released Agglutinin (WGA) / DAPI for from immunoprecipitated complexes by immunofluorescence, analysed using an overnight incubation at 65°C and finally Axio Scan.Z1 slide scanner (Zeiss isolated over Qiagen PCR purification Microscopy). Morphometrics analysis were columns. performed with the ImageJ software. Libraries were prepared using the Diagenode MicroPlex Library Preparation kit v2, and Immunofluorescence sequenced on Illumina Hiseq 4000 Cross sections were rehydrated in PBS, fixed sequencer as single-end 50 bp reads with 4% PFA for 10 min, permeabilized in 1X following Illumina's instructions. Image PBS, 0.1% Triton for 30 min at room analysis and base calling were performed temperature then saturated in 1X PBS 1% using RTA 2.7.3 and bcl2fastq 2.17.1.14. BSA 1% normal goat serum for 1 hour. Adapter dimer reads were removed using Staining with green WGA (W11261, DimerRemover v0.9.2. Reads were mapped Molecular Probes) and DAPI were performed

12 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

overnight at 4°C. Coverslips were then of the tn5 transposon binding event, as mounted with Vectashield (Vector suggested58; the narrow model was the tn5 Laboratories) and sealed with nail polish. transposon occupancy only (ie --nomodel -- For myofibers staining, an EDL muscle was shift -5 --extsize 9). Mitochondrial peaks and collected and fixed in 4% paraformaldehyde peaks overlapping blacklisted regions were 10 min for tissue dissection and staining. filtered out59. In order to compare Muscles were slowly teased with fine forceps accessibility between both conditions, we to isolate individual fibers and small fiber focused on the union of their peaks, bundles. Fibers were permeabilized in 1X analysing those called with the wide or PBS, 0.1% Triton for 30 min at room narrow model separately. As an example, for temperature, saturated in 1X PBS 1% BSA the peaks produced with the wide calling for 1 hour before overnight incubation with model, we first defined true enriched regions the primary antibody (home-made anti- in each condition as MACS2 peaks shared by H2A.Z 1/100) diluted in blocking buffer. After the three replicates, taking the union of the washing, fibers were incubated with a intervals, if overlapped by at least 25% of secondary antibody conjugated with FITC their length; we then merged these true (Molecular probes). DNA and neuromuscular enriched regions from both conditions, taking junctions were respectively stained with DAPI the union of regions overlapping by more (2 µg/ml) and α-bungarotoxin Alexa-555 than 50% of their length, or keeping not (Molecular probes) for 2 hours at room overlapped regions with their boundaries temperature. After washing, fibers were unchanged60. We quantified the fragments mounted on glass slide with Vectashield ends tn5-shifted and 37bp extended falling (Vector Laboratories) and analysed using a within these 52142 union intervals. confocal laser scanning microscope (Leica Differential peak enrichment was then SP5). Images were processed using ImageJ computed using DESeq261, Heatmaps were software. performed using deeptools262, plotting the means of the fold enrichment files generated Computational analyses by MACS2 from the three replicates for each Repeat analyses of RNA-seq and ChIP-seq condition, using the narrow peak calling, the datasets were performed as previously TSSs being sorted either by the RNA-seq described1. Processed datasets were expression data from the CTL condition, or by restricted to repeat families with more than the CTL ATAC-seq TSS enrichment. 500 mapped reads per ChIP sample or more than 5 reads per million mapped reads per Data access RNA sample to avoid over- or under- The ChIP-seq and RNA-seq datasets have estimating fold enrichments due to low been deposited in the Gene Expression sequence representation. Omnibus (GEO; Heatmaps and quantitative analysis of the http://www.ncbi.nlm.nih.gov/geo/) under the ChIP-seq data were performed using accession number GSE111576. seqMINER (http://bips.u- strasbg.fr/seqminer/). As reference coordinates, we used the Ensembl 67 SUPPLEMENTARY DATA database of the mouse genome (mm9). Supplementary Data are available online. For ATAC-seq data not properly paired, supplementary alignments and putative PCR duplicates were removed54–56. Enriched ACKNOWLEDGEMENTS regions were identified with MACS257 at q Animal breeding and H2A.Z muscle-specific value 0,01, with either a wide or narrow inactivation were performed at the animal assumption of how accessibility is inferred facility (PBES) of the research federation from fragments ends : a wide model was a SFR Biosciences (UMS3444). Microscopy half nucleosome length (73bp) centered on was performed on the microscopy facilities of the fragment ends (ie --nomodel --shift -37 -- SFR Biosciences (PLATIM, UMS3444) and extsize 73), these ends having been shifted SFR Santé Lyon-Est (CIQLE, UMS 3453). 4bp forward for (+) reads and 5bp backward for (-) reads, in order to represent the center

13 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

FUNDING 6. Ausió, J. Histone variants--the structure behind This work was supported by a grant from the the function. Brief. Funct. Genomic. Proteomic. Association Française contre les Myopathies 5, 228–243 (2006). (AFM) through MyoNeurAlp alliance, the 7. Thambirajah, A. A., Li, A., Ishibashi, T. & Ausió, Agence Nationale pour la Recherche (ANR- J. New developments in post-translational 10-LABX-0030, ANR-12-BSV5-0017, ANR- modifications and functions of histone H2A variants. Biochem. Cell Biol. Biochim. Biol. Cell. 14-CE09-0019, ANR-16-CE12-0013, ANR- 87, 7–17 (2009). 17-CE11-0019 and ANR-18-CE12-0010), La 8. Zink, L.-M. & Hake, S. B. Histone variants: Ligue Nationale contre le Cancer [Equipe nuclear function and disease. Curr. Opin. Genet. labellisée (A.H.) USIAS (2015-42)], Dev. 37, 82–89 (2016). Fondation pour la Recherche Médicale 9. Boulard, M., Bouvet, P., Kundu, T. K. & Dimitrov, (FRM, ‘‘Epigénétique et Stabilité du S. Histone variant nucleosomes: structure, Genome’’ Program), Institut National du function and implication in disease. Subcell. Cancer, Association pour la Recherche sur le Biochem. 41, 71–89 (2007). Cancer, Inserm, CNRS, Université de 10. Corujo, D. & Buschbeck, M. Post-Translational Strasbourg and Université Grenoble Alpes. Modifications of H2A Histone Variants and E.B. benefited of an AFM fellowship. Their Role in Cancer. Cancers 10, (2018). 11. Sarma, K. & Reinberg, D. Histone variants meet their match. Nat. Rev. Mol. Cell Biol. 6, 139–149 (2005). AUTHOR CONTRIBUTIONS 12. Eirín-López, J. M., González-Romero, R., L.S., S.D., A.H., E.B., N.L. conceived the Dryhurst, D., Ishibashi, T. & Ausió, J. The research. L.R. and D.D. participated to the evolutionary differentiation of two histone generation of KO mice, E.B. performed the H2A.Z variants in chordates (H2A.Z-1 and conditional KO, MEF cells and mouse H2A.Z-2) is mediated by a stepwise mutation breeding. E.B., N.L. performed cell biology, process that affects three amino acid residues. RT-QPCR, immunofluorescences, histology, BMC Evol. Biol. 9, 31 (2009). nuclei isolation and biochemistry analysis. 13. Dryhurst, D. et al. Characterization of the K.P. performed ChIP experiments. I.S., C.P., histone H2A.Z-1 and H2A.Z-2 isoforms in T.S. performed bioinformatics analysis. E.B., vertebrates. BMC Biol. 7, 86 (2009). N.L., L.S., S.D., A.H. wrote the paper with 14. Faast, R. et al. Histone variant H2A.Z is input from all authors. required for early mammalian development. Curr. Biol. 11, 1183–1187 (2001).

15. Obri, A. et al. ANP32E is a histone chaperone that removes H2A.Z from chromatin. Nature CONFLICT OF INTEREST 505, 648–653 (2014). None declared. 16. Latrick, C. M. et al. Molecular basis and specificity of H2A.Z-H2B recognition and deposition by the histone chaperone YL1. Nat. REFERENCES Struct. Mol. Biol. 23, 309–316 (2016). 17. Barski, A. et al. High-Resolution Profiling of 1. Maze, I., Noh, K.-M., Soshnev, A. A. & Allis, C. D. Histone Methylations in the Human Genome. Every amino acid matters: essential Cell 129, 823–837 (2007). contributions of histone variants to 18. Nekrasov, M., Soboleva, T. A., Jack, C. & mammalian development and disease. Nat. Tremethick, D. J. Histone variant selectivity at Rev. Genet. 15, 259–271 (2014). the transcription start site: H2A.Z or H2A.Lap1. 2. K. E. Van Holde. Chromatin. Series in molecular Nucleus 4, (2013). biology. Springer-Verlag, New York. 1988. 530 19. Nekrasov, M. et al. Histone H2A.Z inheritance pp. $98.00. J. Mol. Recognit. 2, i–i (1989). during the cell cycle and its impact on 3. Talbert, P. B. & Henikoff, S. Histone variants on promoter organization and dynamics. Nat. the move: substrates for chromatin dynamics. Struct. Mol. Biol. 19, 1076–1083 (2012). Nat. Rev. Mol. Cell Biol. 18, 115–126 (2017). 20. Sarcinella, E., Zuzarte, P. C., Lau, P. N. I., Draker, 4. Long, M. et al. A novel histone H4 variant H4G R. & Cheung, P. Monoubiquitylation of H2A.Z regulates rDNA transcription in breast cancer. Distinguishes Its Association with Nucleic Acids Res. doi:10.1093/nar/gkz547. Euchromatin or Facultative Heterochromatin. 5. Cheema, M. S. & Ausió, J. The Structural Mol. Cell. Biol. 27, 6457–6468 (2007). Determinants behind the Epigenetic Role of 21. Greaves, I. K., Rangasamy, D., Ridgway, P. & Histone Variants. Genes 6, 685–713 (2015). Tremethick, D. J. H2A.Z contributes to the

14 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

unique 3D structure of the centromere. Proc. 36. Li, B. et al. Preferential occupancy of histone Natl. Acad. Sci. 104, 525–530 (2007). variant H2AZ at inactive promoters influences 22. Boyarchuk, E., Filipescu, D., Vassias, I., local histone modifications and chromatin Cantaloube, S. & Almouzni, G. The histone remodeling. Proc. Natl. Acad. Sci. U. S. A. 102, variant composition of centromeres is 18385–18390 (2005). controlled by the pericentric heterochromatin 37. Zhang, H., Roberts, D. N. & Cairns, B. R. state during the cell cycle. J. Cell Sci. 127, 3347– Genome-Wide Dynamics of Htz1, a Histone 3359 (2014). H2A Variant that Poises Repressed/Basal 23. Weber, C. M., Henikoff, J. G. & Henikoff, S. H2A.Z Promoters for Activation through Histone nucleosomes enriched over active genes are Loss. Cell 123, 219–231 (2005). homotypic. Nat. Struct. Mol. Biol. 17, 1500– 38. Risson, V. et al. Muscle inactivation of mTOR 1507 (2010). causes metabolic and dystrophin defects 24. Svensson, J. P. et al. A nucleosome turnover leading to severe myopathy. J. Cell Biol. 187, map reveals that the stability of histone H4 859–874 (2009). Lys20 methylation depends on histone 39. Zhang, Q. et al. Lack of muscle mTOR kinase recycling in transcribed chromatin. Genome activity causes early onset myopathy and Res. 25, 872–883 (2015). compromises whole-body homeostasis. J. 25. Giaimo, B. D. et al. Histone variant H2A.Z Cachexia Sarcopenia Muscle 10, 35–53 (2019). deposition and acetylation directs the 40. Bentzinger, C. F. et al. Skeletal Muscle-Specific canonical Notch signaling response. Nucleic Ablation of raptor, but Not of rictor, Causes Acids Res. 46, 8197–8215 (2018). Metabolic Changes and Results in Muscle 26. Rangasamy, D., Greaves, I. & Tremethick, D. J. Dystrophy. Cell Metab. 8, 411–424 (2008). RNA interference demonstrates a novel role 41. Castets, P. et al. mTORC1 and PKB/Akt control for H2A.Z in chromosome segregation. Nat. the muscle response to denervation by Struct. Mol. Biol. 11, 650–655 (2004). regulating autophagy and HDAC4. Nat. 27. Miniou, P. et al. Gene targeting restricted to Commun. 10, (2019). mouse striated muscle lineage. Nucleic Acids 42. Zhao, K. et al. Muscle Yap Is a Regulator of Res. 27, e27–e30 (1999). Neuromuscular Junction Formation and 28. Raffaello, A. et al. Denervation in murine fast- Regeneration. J. Neurosci. 37, 3465–3477 twitch muscle: short-term physiological (2017). changes and temporal expression profiling. 43. Rangasamy, D., Berven, L., Ridgway, P. & Physiol. Genomics 25, 60–74 (2006). Tremethick, D. J. Pericentric heterochromatin 29. Méjat, A. et al. Histone deacetylase 9 couples becomes enriched with H2A.Z during early neuronal activity to muscle chromatin mammalian development. EMBO J. 22, 1599– acetylation and gene expression. Nat. Neurosci. 1607 (2003). 8, 313–321 (2005). 44. Ryan, D. P. & Tremethick, D. J. The interplay 30. Thomas, J.-L. et al. PAK1 and CtBP1 Regulate between H2A.Z and H3K9 methylation in the Coupling of Neuronal Activity to Muscle regulating HP1α binding to linker histone- Chromatin and Gene Expression. Mol. Cell. Biol. containing chromatin. Nucleic Acids Res. 46, 35, 4110–4120 (2015). 9353–9366 (2018). 31. Nekrasov, M., Soboleva, T. A., Jack, C. & 45. Harris, M. E. et al. Regulation of histone mRNA Tremethick, D. J. Histone variant selectivity at in the unperturbed cell cycle: evidence the transcription start site: H2A.Z or H2A.Lap1. suggesting control at two posttranscriptional Nucleus 4, 431–437 (2013). steps. Mol. Cell. Biol. 11, 2416–2424 (1991). 32. Liu, C. et al. An ATAC-seq atlas of chromatin 46. Luger, K., Dechassa, M. L. & Tremethick, D. J. accessibility in mouse tissues. Sci. Data 6, New insights into nucleosome and chromatin (2019). structure: an ordered state or a disordered 33. Rao, P. & Monks, D. A. A tetracycline-inducible affair? Nat. Rev. Mol. Cell Biol. 13, 436–447 and skeletal muscle-specific Cre recombinase (2012). transgenic mouse. Dev. Neurobiol. 69, 401–406 47. Jang, C.-W., Shibata, Y., Starmer, J., Yee, D. & (2009). Magnuson, T. Histone H3.3 maintains genome 34. Giordani, L. et al. High-Dimensional Single-Cell integrity during mammalian development. Cartography Reveals Novel Skeletal Muscle- Genes Dev. 29, 1377–1392 (2015). Resident Cell Populations. Mol. Cell 74, 609- 48. Ors, A. et al. Histone H3.3 regulates mitotic 621.e6 (2019). progression in mouse embryonic fibroblasts. 35. Magnusson, C., Svensson, A., Christerson, U. & Biochem. Cell Biol. 95, 491–499 (2017). Tågerud, S. Denervation-induced alterations in 49. Papin, C. et al. Combinatorial DNA methylation gene expression in mouse skeletal muscle. Eur. codes at repetitive elements. Genome Res. 27, J. Neurosci. 21, 577–580 (2005). 934–946 (2017).

15 bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

50. Trapnell, C., Pachter, L. & Salzberg, S. L. TopHat: 57. Zhang, Y. et al. Model-based Analysis of ChIP- discovering splice junctions with RNA-Seq. Seq (MACS). Genome Biol. 9, R137 (2008). Bioinformatics 25, 1105–1111 (2009). 58. Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, 51. Anders, S. & Huber, W. Differential expression H. Y. & Greenleaf, W. J. Transposition of native analysis for sequence count data. Genome Biol. chromatin for fast and sensitive epigenomic 11, R106 (2010). profiling of open chromatin, DNA-binding 52. Buenrostro, J., Wu, B., Chang, H. & Greenleaf, W. proteins and nucleosome position. Nat. ATAC-seq: A Method for Assaying Chromatin Methods 10, 1213–1218 (2013). Accessibility Genome-Wide. Curr. Protoc. Mol. 59. Amemiya, H. M., Kundaje, A. & Boyle, A. P. The Biol. Ed. Frederick M Ausubel Al 109, 21.29.1- ENCODE Blacklist: Identification of 21.29.9 (2015). Problematic Regions of the Genome. Sci. Rep. 9, 53. Bolger, A. M., Lohse, M. & Usadel, B. 1–5 (2019). Trimmomatic: a flexible trimmer for Illumina 60. Quinlan, A. R. & Hall, I. M. BEDTools: a flexible sequence data. Bioinforma. Oxf. Engl. 30, 2114– suite of utilities for comparing genomic 2120 (2014). features. Bioinformatics 26, 841–842 (2010). 54. Li, H. et al. The Sequence Alignment/Map 61. Love, M. I., Huber, W. & Anders, S. Moderated format and SAMtools. Bioinforma. Oxf. Engl. 25, estimation of fold change and dispersion for 2078–2079 (2009). RNA-seq data with DESeq2. Genome Biol. 15, 55. Tarasov, A., Vilella, A. J., Cuppen, E., Nijman, I. J. 550 (2014). & Prins, P. Sambamba: fast processing of NGS 62. Ramírez, F. et al. deepTools2: a next generation alignment formats. Bioinforma. Oxf. Engl. 31, web server for deep-sequencing data analysis. 2032–2034 (2015). Nucleic Acids Res. 44, W160-165 (2016). 56. Picard Tools - By Broad Institute. https://broadinstitute.github.io/picard/.

was notcertifiedbypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmade bioRxiv preprint

Rrad 5,352 3,40E-58 Atp1b4 -4,134 2,88E-101 Myog 5,700 7,61E-243 Kbtbd13 -4,017 2,90E-75 Fbxo47 1,624 1,49E-07 Cntnap2 -1,884 7,15E-19 Srrm4 1,307 3,32E-06 Morn4 -1,618 2,32E-08 doi: Chrna9 5,336 1,08E-43 Ntf5 -3,911 3,97E-51 Krt18 5,647 8,74E-47 Atp1b4 -3,934 2,97E-98 Ighg2c 1,519 4,10E-04 Dach2 -1,630 1,23E-15 2410137F16Rik 1,300 2,91E-06 Hnmt -1,486 2,39E-13 Krt18 5,194 5,02E-40 Sec14l5 -3,833 3,17E-117 Rrad 5,476 6,09E-61 Asb18 -3,894 9,60E-69 U12 1,426 9,04E-06 Syt3 -1,610 1,16E-13 Csrp3 1,218 4,62E-05 Syt3 -1,397 1,10E-09

Myog 5,172 2,09E-202 Asb18 -3,808 2,71E-71 Prrt3 5,188 5,22E-56 Sec14l5 -3,745 3,13E-109 SNORA53 1,389 1,04E-12 Aldh1a7 -1,483 9,58E-05 Kcne1l 1,200 3,81E-05 Tph1 -1,366 6,29E-05 https://doi.org/10.1101/823526 Syt4 5,070 3,07E-63 Kbtbd13 -3,800 9,32E-72 Gadd45a 4,915 3,45E-191 D230002A01Rik -3,691 3,20E-30 Apitd1 1,370 1,50E-05 Xlr3b -1,475 1,65E-06 Casr 1,168 2,50E-07 Dach2 -1,340 1,04E-09 Gm15935 5,001 2,05E-54 Kcnn2 -3,618 2,33E-50 Cd101 4,847 4,59E-72 2310010M20Rik -3,663 7,43E-117 mt-Tt 1,303 5,22E-04 Nat1 -1,255 8,39E-04 BC048679 1,157 5,69E-07 Pld5 -1,279 9,92E-06 Gadd45a 4,892 6,37E-189 Lhb -3,607 1,87E-29 Chrna9 4,737 5,37E-35 Zfp750 -3,628 3,79E-20 Gins1 1,288 2,27E-08 Zfp286 -1,188 6,47E-07 Itgb1bp3 1,152 6,37E-07 Lrp2bp -1,239 6,90E-07 Prrt3 4,816 2,29E-48 D230002A01Rik -3,589 1,72E-30 1300002K09Rik 4,733 6,42E-68 Tmem25 -3,574 2,50E-67 SNORA66 1,179 1,35E-05 BC048355 -1,182 7,08E-10 Serpinb1a 1,142 9,01E-07 Ctnna3 -1,191 9,50E-20 Ckap2 4,771 4,59E-59 Zfp750 -3,568 8,10E-20 Ces2c 4,694 2,68E-34 Lrrc52 -3,533 2,32E-20 E130012A19Rik 1,174 1,92E-04 A930003A15Rik -1,175 2,33E-05 Card14 1,094 8,34E-05 Kcnc3 -1,172 4,96E-04 Runx1 4,505 7,95E-262 A230009B12Rik -3,525 6,88E-24 Gm15935 4,604 4,58E-58 Tecta -3,532 4,35E-32 Kif20a 1,112 3,39E-07 Tle6 -1,170 4,39E-12 9430073C21Rik 1,063 3,06E-05 Amy1 -1,169 4,36E-06 Cd101 4,499 2,32E-62 Golga7b -3,499 3,24E-47 Runx1 4,568 3,09E-268 A230009B12Rik -3,486 1,21E-26 Snora23 1,091 1,92E-05 Plekhb1 -1,156 1,22E-22 Gm13010 1,002 3,61E-04 Tle6 -1,169 1,36E-12 Plekhh1 4,368 7,00E-35 Tecta -3,498 2,40E-29 Syt4 4,415 7,80E-55 Gm9392 -3,440 1,28E-18 D630024D03Rik 1,086 6,10E-04 Ankrd34a -1,147 1,53E-04 Car11 -1,164 1,21E-07 Srxn1 4,364 5,12E-106 Hs3st5 -3,477 1,42E-106 Chrnd 4,321 7,02E-213 BY080835 -3,434 1,70E-51 mt-Tc 1,059 4,22E-04 Col9a1 -1,139 1,11E-04 Mtap9 -1,094 7,39E-08 Eda2r 4,290 1,58E-92 Kcna7 -3,360 6,70E-104 Gm16150 4,294 1,79E-50 2310016D23Rik -3,413 2,99E-25 mt-Tq 1,057 3,22E-05 Car11 -1,084 8,93E-07 Col9a1 -1,072 8,04E-04

Blnk 4,288 2,04E-177 2310016D23Rik -3,346 1,14E-26 Ckap2 4,184 5,22E-48 Golga7b -3,350 1,48E-40 Hephl1 1,045 7,37E-04 Lrp2bp -1,080 3,61E-07 Fam83h -1,063 2,24E-04 available undera Ampd3 4,199 2,79E-215 2310010M20Rik -3,309 2,90E-97 Blnk 4,178 9,60E-171 Hs3st5 -3,311 3,01E-100 Nudt10 1,043 4,02E-05 Ctnna3 -1,074 6,01E-17 BC048355 -1,051 1,59E-08 Stx11 4,181 4,94E-90 Dok5 -3,290 4,19E-41 Ampd3 4,167 1,89E-211 Kcnn2 -3,296 5,60E-39 Mthfd2 1,039 6,99E-10 Pak1 -1,072 3,60E-09 Zfp454 -1,006 1,94E-05 Lpin2 4,116 5,01E-154 BY080835 -3,284 4,57E-48 Eda2r 4,072 1,01E-83 Mir133a-1 -3,234 4,96E-16 Snora74a 1,014 5,57E-12 Fam189b -1,034 3,39E-07 Chrnd 4,111 2,24E-199 B230206H07Rik -3,240 4,07E-21 Fgf21 4,055 6,25E-27 Dusp15 -3,174 1,99E-16 Fam55c 1,005 2,83E-07 Hnmt -1,009 4,99E-07 Ccl7 4,106 1,14E-33 Sema6c -3,191 1,13E-103 Srxn1 4,055 5,88E-92 Pdzd7 -3,160 2,24E-32 BC048679 1,002 1,60E-05 Kcnf1 -1,003 6,30E-08 Tgif1 3,947 0,00E+00 B230312C02Rik -3,140 5,52E-18 Ttyh1 4,041 1,09E-88 Lhb -3,125 2,06E-21 Ush1c -1,001 2,06E-04 Gm16150 3,944 2,92E-43 Helt -3,126 2,62E-22 Dusp5 4,007 1,57E-82 Gm3235 -3,123 2,20E-29

Fgf21 3,883 5,54E-27 Dusp15 -3,111 6,63E-16 Lpin2 3,964 6,20E-143 Spink11 -3,118 7,09E-16 ;

Klhdc8a 3,818 9,89E-46 Lrrc52 -3,102 5,30E-16 Cd82 3,905 1,06E-176 Sema6c -3,103 1,21E-97 this versionpostedOctober30,2019. Ces2c 3,817 3,46E-23 Klhl34 -3,101 6,52E-49 Plekhh1 3,849 2,03E-27 Frat2 -3,101 9,90E-59

Ttyh1 3,809 4,95E-81 Pdzd7 -3,081 6,64E-32 Tbx10 3,849 7,22E-23 Lmod3 -3,100 1,41E-151 CC-BY-NC-ND 4.0Internationallicense 1110059M19Rik 3,732 6,44E-20 Mpz -3,080 2,38E-20 Musk 3,743 8,27E-151 Kcna7 -3,076 2,19E-87 Cd82 3,714 6,10E-161 Frat2 -3,052 5,53E-59 Elf3 3,725 5,23E-20 Gm889 -3,017 4,67E-66 Ccl2 3,702 2,59E-37 Spinkl -3,041 1,44E-16 Stx11 3,697 1,20E-72 Musk 3,691 2,22E-146 Itgb6 -3,030 6,91E-84 Tgif1 3,626 0,00E+00 Rhou 3,674 2,93E-105 Tmem25 -3,025 7,42E-54 Acox2 3,595 8,59E-38 Dusp5 3,631 7,99E-67 Lmod3 -3,006 1,79E-143 Tekt1 3,561 7,64E-69 Gdnf 3,597 1,62E-72 1110059M19Rik 3,535 1,92E-17 Fhl2 3,592 3,42E-52 Rhou 3,521 7,06E-97 Dusp9 3,540 2,21E-20 Gdnf 3,435 5,28E-76 Cdkn1a 3,516 4,29E-48 Atp13a4 3,430 5,87E-17 Tbx10 3,415 9,91E-19 E130218I03Rik 3,419 1,88E-16 Slc22a3 3,407 6,83E-30 Dusp4 3,395 2,60E-79 Elf3 3,381 9,39E-17 Dusp9 3,378 1,44E-18 Hdac4 3,379 4,56E-287 Chrna1 3,358 4,33E-92 Nrcam 3,352 6,47E-34 Hdac4 3,294 2,68E-274 Dusp4 3,346 1,13E-78 Ankrd1 3,273 2,37E-16 Tekt1 3,317 9,45E-60 Ggt6 3,208 2,87E-15 Acox2 3,264 1,65E-33 Il5 3,195 4,36E-16 Chrna1 3,220 1,03E-84 Fhl2 3,183 4,56E-42 Kcnn3 3,213 1,76E-113 Kcnn3 3,171 2,58E-111 Odc1 3,168 1,49E-60 Gm14029 3,170 1,02E-29

1300002K09Rik 3,118 3,78E-32 D730045A05Rik 3,135 1,64E-16 The copyrightholderforthispreprint(which Ggt6 3,091 4,67E-14 Atf3 3,106 9,95E-23 Ankrd1 3,048 2,21E-14 Shroom3 3,087 2,39E-41 A330102I10Rik 3,036 1,49E-32 Trim63 3,085 4,61E-47 . 4931413K12Rik 3,036 1,39E-24 Dio3 3,074 7,60E-20 Dio3 3,017 2,27E-19 A330102I10Rik 3,034 2,06E-33 Id2 3,027 1,37E-33

Table 1. Main down- and up-regulated genes in innervated and denervated CTL and dKO H2A.Z skeletal muscle from 7 week-old mice. bioRxiv preprint doi: https://doi.org/10.1101/823526; this version posted October 30, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

Family log2(H2A.Z/Input) P value Class (CG)n 1,771249278 2,51E-23 Simple repeats (microsatellites) (GGCTG)n 1,608493329 1,04E-19 Endogenous retrovirus (CCCTG)n 1,495038741 8,32E-21 rRNA (CTGGGG)n 1,470260675 4,28E-17 Other (CCCCAG)n 1,38147185 5,70E-09 (CAGGG)n 1,379605423 8,33E-16 (CAG)n 1,293622361 2,23E-12 C-rich 1,261165933 9,78E-08 (CTG)n 1,258188705 1,69E-10 G-rich 1,254904537 1,38E-07 (C)n 1,202679859 5,65E-17 (CACG)n 1,194718257 3,09E-46 (CAGCC)n 1,179626538 8,72E-19 (CTGTG)n 1,162942871 7,22E-39 (CGTG)n 1,154228446 3,34E-36 (CACCC)n 1,073961571 2,41E-18 (CGGGG)n 1,067749127 1,08E-05 (GGGTG)n 1,063288362 2,37E-17 GC_rich 1,046008066 1,02E-05 (CAGC)n 1,040762506 9,36E-14 (CCCCG)n 1,033504095 2,41E-05 (CCG)n 1,03257663 1,04E-05 (G)n 1,028187694 9,45E-13 (CGG)n 1,014904867 1,37E-05 (CCCCCA)n 1,008089232 2,91E-18 (CCTG)n 0,999185043 4,28E-28 (CAGG)n 0,97956483 1,79E-16 (GCTG)n 0,947769449 5,25E-08 (CGGG)n 0,941824291 4,84E-05 (TGGGGG)n 0,934067466 1,91E-14 (CACAG)n 0,911819856 2,67E-19 RLTR46 0,905195918 6,76E-13 (TCCCC)n 0,90101027 6,51E-16 (GGGGA)n 0,887502752 1,29E-10 (CAGAG)n 0,86138966 7,20E-09 (CTCTG)n 0,848075565 1,04E-06 (TCTCCC)n 0,841957779 2,49E-12 (GGGAGA)n 0,828462885 4,78E-15 (GGA)n 0,810793783 2,54E-14 (TCC)n 0,786075131 5,90E-07 (CCCCCT)n 0,785615531 3,89E-08 RLTR44E 0,783672874 3,94E-13 LSU-rRNA_Hsa 0,74486744 0,00086467 RLTR44D 0,744837852 7,93E-09 SSU-rRNA_Hsa 0,740274863 0,00283071 RLTR44C 0,735063197 2,47E-15 (TCCC)n 0,730496418 1,21E-13 (GGGA)n 0,724782145 5,29E-16 (AGGGGG)n 0,718581047 4,31E-06 RLTRETN_Mm 0,675861344 2,35E-28 RLTR10A 0,673919393 1,71E-29 RLTR44A 0,656089238 6,56E-14 (TCTCC)n 0,64864711 5,46E-08 (GGAGA)n 0,630983743 2,43E-05 (CAGTG)n 0,627964021 0,00011819 RLTR44B 0,62531813 3,53E-11 RLTR27 0,624233966 6,64E-29 IAPLTR1a_Mm 0,613729876 1,86E-45 (TGG)n 0,606200114 4,82E-37 RMER17B 0,602500281 1,04E-35 MMVL30-int 0,60030683 1,66E-09 RLTR4_MM-int 0,592998726 2,70E-36 (TTCCC)n 0,59246458 2,28E-06 (CCA)n 0,589743151 1,00E-34 ZP3AR 0,57874439 3,43E-26 (GTGTG)n 0,558201688 8,09E-06 RMER21A 0,556509167 2,91E-15 RMER17D2 0,553115718 4,79E-17 RMER13B 0,549217425 1,13E-43 (GGGAA)n 0,543996414 1,98E-05 RLTR4_Mm 0,530676013 1,69E-10 RLTR9F 0,529513642 1,56E-13 IAPLTR2_Mm 0,521689077 5,66E-15 RMER17A 0,518821297 3,99E-20 RMER17A2 0,513811041 1,66E-20 IAPLTR2a 0,509541256 8,47E-18 (CACAC)n 0,50797793 8,35E-05 IAPLTR1_Mm 0,504040512 1,36E-22

Table 2. Repetitive elements family enriched of H2A.Z. The majority of significantly H2A.Z-enriched repeats (log2 enrichment > 0.5 and P value < 10-2) belongs to the simple repeats and endogenous retrovirus classes of repetitive elements. was notcertifiedbypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmade bioRxiv preprint

Innervated vs Denervated (WT) Innervated vs Denervated (H2A.Z dKO) WT vs H2A.Z dKO (Innervated) WT vs H2A.Z dKO (Denervated) Most up-regulated in response to denervation Most down-regulated in response to denervation Most up-regulated in response to denervation Most down-regulated in response to denervation Most up-regulated in the absence of H2A.Z Most down-regulated in the absence of H2A.Z Most up-regulated in the absence of H2A.Z Most down-regulated in the absence of H2A.Z Name log2(Den./Inn.) P value Name log2(Den./Inn.) P value Name log2(Den./Inn.) P value Name log2(Den./Inn.) P value Name log2(dKO/WT) P value Name log2(dKO/WT) P value Name log2(dKO/WT) P value Name log2(dKO/WT) P value MLT1H 1,826960357 4,07E-30 L1Md_F3 -0,659305156 9,75E-03 MLT1H 1,709799489 2,252E-26 (TCG)n -0,692553204 4,138E-04 tRNA-Gln-CAA_ 0,742036902 1,15E-02 RLTR6_Mm 1,464874768 7,17E-18 MER45B -0,571375842 4,31E-02 RLTR6_Mm 1,291321451 7,524E-14 tRNA-Gln-CAA_ -0,582719435 7,331E-03 U3 1,441010749 7,50E-03 (TCG)n -0,542592501 3,63E-03 MIR3 0,889934668 1,809E-11 GSAT_MM 1,368589835 5,75E-03 RMER6D 0,884996763 9,247E-10 RMER6D 1,118608502 1,07E-15 MLT1H-int 0,826780023 1,115E-06 MLT1H-int 0,991095576 1,19E-09 YREP_Mm 0,756144618 4,295E-09 doi: RLTR6-int 0,915692564 1,90E-09 Charlie13a 0,743264167 3,713E-06 YREP_Mm 0,910180593 3,71E-13 (TCCCC)n 0,724618575 2,025E-06 https://doi.org/10.1101/823526 MIR3 0,895115705 1,06E-11 MURVY-int 0,713678946 5,217E-04 (TCTCC)n 0,87867813 2,97E-08 IAPEz-int 0,699908521 3,437E-07 RLTR13D4 0,848539316 7,17E-05 (TCTCC)n 0,699214542 2,975E-05 MURVY-int 0,845615076 1,22E-05 RLTR1D 0,687934738 8,024E-04 (GGAGAA)n 0,841644504 1,19E-09 MuLV-int 0,67597855 2,043E-04 RNERVK23-int 0,83655495 1,04E-05 (GGAGAA)n 0,675421024 2,880E-06 MERVK26-int 0,829113617 1,43E-04 RLTR6-int 0,606915108 2,832E-04 IAPLTR3-int 0,818179053 1,75E-04 L1MD3 0,573797994 1,514E-05 IAPEz-int 0,807746875 1,19E-09 BLACKJACK 0,566281512 2,321E-03 Charlie13a 0,799232449 3,31E-07 IAPLTR1_Mm 0,565959422 5,755E-04 RMER6C 0,785058923 1,74E-07 RMER6C 0,553852574 6,854E-04 MuLV-int 0,776690834 7,62E-06 MER58D 0,533192897 4,042E-03 available undera U2 0,774922812 2,57E-06 U2 0,533027595 3,045E-03 MamRep605 0,77180962 1,28E-05 LTR50 0,528508615 5,232E-04 ETnERV3-int 0,756188006 6,32E-05 RLTR13D4 0,522997159 3,335E-02 L1MD3 0,74815904 1,90E-09 (CAT)n 0,517345586 1,672E-04 Charlie6 0,724489122 1,27E-05 (CACAG)n 0,512335445 1,828E-03 BLACKJACK 0,720120103 2,68E-05 LTR33A_ 0,506063071 8,790E-04 RLTR44-int 0,703749697 1,60E-03 IAPLTR1_Mm 0,682065487 1,12E-05

MLT2C1 0,680450521 3,62E-05 ; (CAT)n 0,663952378 2,50E-07 this versionpostedOctober30,2019. IAPEY2_LTR 0,659444386 1,85E-02

LTR33A_ 0,65729853 3,87E-06 CC-BY-NC-ND 4.0Internationallicense L1M3e 0,653214603 5,05E-06 RMER17D 0,641630674 8,77E-06 IAP-d-int 0,636535861 1,51E-03 RLTR45-int 0,635843549 4,40E-05 RLTR31B_Mm 0,625034612 1,43E-04 ETnERV2-int 0,621384499 9,09E-04 Lx2A 0,612643764 6,56E-04 U17 0,598039484 2,07E-02 MER47B 0,592658377 2,68E-04 L1_Mur2 0,583284843 1,31E-03 ORR1A2-int 0,575305769 2,07E-04 RLTR13D2 0,568714456 3,18E-04 RMER20B 0,562755392 3,69E-04 RLTR26 0,560225931 1,50E-04 MLT2C2 0,557371916 1,39E-03 Looper 0,555918514 1,63E-03 MMERGLN-int 0,555194623 8,18E-04 LTR50 0,554622862 1,43E-04 7SLRNA 0,552893277 2,51E-02 RLTR31_Mur 0,550670836 2,97E-03 MER58D 0,540135699 1,65E-03 Zaphod 0,53995821 1,60E-03

MER68 0,53972248 7,32E-05 The copyrightholderforthispreprint(which L1_Mur1 0,537208317 2,11E-03 MurERV4_19-int 0,534785352 1,49E-03

RLTR12B 0,53462758 3,69E-04 . Cheshire 0,532077279 2,68E-05 RLTR9D 0,53089414 6,78E-04 ORR1D-int 0,530607188 4,24E-03 Charlie12 0,528481151 3,11E-04 RLTR21 0,525932491 1,31E-03 LTR37A 0,519033099 5,27E-03 MER44C 0,518986067 1,77E-03 RLTR31A_Mm 0,518301346 4,11E-04 MYSERV16_I-int 0,516425599 1,75E-03 (TGGGGG)n 0,513265511 2,68E-04 RLTR18-int 0,509868604 1,75E-03 MERVL-int 0,508553579 1,42E-04 ORR1A3-int 0,503634177 9,31E-04 U6 0,501718073 5,63E-04

Table 3. Main down- and up-regulated repetitive DNA element families in innervated and denervated CTL and dKO H2A.Z skeletal muscle from 7 week-old mice.