Histone variants in skeletal myogenesis

ABSTRACT

Histone variants regulate chromatin accessibility and gene transcription. Given their distinct properties and functions, histone varint substitutions allow for profound alteration of nucleosomal architecture and local chromatin landscape. Skeletal myogenesis driven by the key transcription factor MyoD is characterized by precise temporal regulation of myogenic genes. Timed substitution of variants within the nucleosomes provides a powerful means to ensure sequential expression of myogenic genes. Indeed, growing evidence has shown H3.3, H2A.Z, macroH2A, and H1b to be critical for skeletal myogenesis. However, the relative importance of various histone variants and their associated chaperones in myogenesis is not fully appreciated. In this review, we summarize the role that histone variants play in altering chromatin landscape to ensure proper muscle differentiation. The temporal regulation and cross talk between histones variants and their chaperones in conjunction with other forms of epigenetic regulation could be critical to understanding myogenesis and their involvement in myopathies.

Introduction

Chromatin in the nucleus is wrapped around basic, octameric protein structures known as nucleosomes. Around 147bp of genetic material is wrapped around one nucleosome, which allows for both compactions as well as regulation of gene transcription by controlling access to transcriptional machinery. Core nucleosomes consist of two copies of each of the histone subunits H2A, H2B, H3, and H4. There is also a linker arm, H1, which modulates interactions between the nucleosome and the associated DNA. Each of these subunits, with the sole exception of H4, have multiple variants in humans, many of which are also evolutionarily conserved in other species. Variants of a histone subunit are predominantly non-allelic, that is, produced by different genes, rather than through the production of splice isoforms. However, some variants of H2A are produced through alternative splicing. Histone variants are classified into one of two categories: replication-dependent or replication-independent. The synthesis of replication-dependent variants, as the name suggests, is linked to DNA replication. These variants form the bulk of nucleosomes and their synthesis is consequently tied to synthesis of DNA. In line with this role, these variants are often synthesized by several genes that are present in multiple copies within clusters, thus allowing for the production of a vast quantity in time for DNA replication and compaction. On the other hand, replication-independent variants can be synthesized at any time during the cell cycle, depending on context and requirement. These are often encoded by one or two genes, normally present outside the replication-dependent histone clusters. Some variants are limited to germ cells only. Replication-dependent histone genes lack introns and produce mRNA that contain a consensus stem-loop structure downstream of the stop codon, required for processing the 3’ ends. Conversely, replication-independent variants lack this structure. Some also possess introns that allows for alternative splicing and increased isoform diversity. The properties of histone variants, and the features of those that have been discovered thus far, have been summarized in several excellent reviews [2–8].

Histone variants can be exchanged within nucleosomes. This exchange can lead to changes in the structure, stability, and properties of the nucleosome. Therefore, the exchange of one variant for another, or the marking of a specific genetic locus with a particular variant or combination of variants, can alter the transcription of the target genes. This process is dependent on histone chaperone complexes, as each variant interacts preferentially with a few chaperones and chromatin remodellers. The complexes that have thus far been shown to associate with specific variants are shown in (Table 1). Both variants and their chaperones play important roles in modification of chromatin landscape and gene transcription. This process of modification of the chromatin landscape is one that is important to the regulation of skeletal myogenesis.

Table 1.

Variants of histone subunits

Variant	Associated chaperones & remodelling complexes	Known functions
H1.1	TAF1[163–165], NAP1[165,166], NPM1[167],ProTα[168–170], Nucleolin[171]	Regulator of gene expression[6,172]
H1.2	TAF1[163–165], NAP1[165,166], NPM1[167],ProTα[168–170], Nucleolin[171]	Transcriptional repression173, Transcriptional activation[174,175], Cell cycle progression[176]
H1.3	TAF1[163–165], NAP1[165,166], NPM1[167],ProTα[168–170], Nucleolin[171]	Repression of ncRNA expression[145]
H1.4	TAF1[163–165], NAP1[165,166], NPM1[167],ProTα[168–170], Nucleolin[171]	Heterochromatin formation[177], Transcriptional activation[174,177,178], Cell cycle progression[176]
H1.5	TAF1[163–165], NAP1[165,166], NPM1[167],ProTα[168–170], Nucleolin[171]	Transcriptional repression[123]
H1.0	TAF1[163,164]	Replacement of replication-dependent variants in differentiated tissue[179]
H1.X	TAF1[163]	Mitotic progression[180]
H1t	-	Testes specific form[181]
H1T2	-	Cell restructuring and polarity in spermiogenesis[182,183]
H1oo	-	Oocyte-specific form[184]
HILS	-	Spermatid-specific form[185,186]
H2A	FACT[187], NAP1[188–192], nucleolin[193]	Regulation of gene expression[194]
H2A.X	FACT[195]	Genome stability[196], proliferation[197], DNA damage response[198]
macroH2A.1.1	Nucleolin[193], ATRX[199]	Development[103,106,109,200], X-chromosome inactivation[105], myogenesis[116]
macroH2A.1.2	Nucleolin[193], ATRX[199]	Development[103,106,109,200], X-chromosome inactivation[105], myogenesis[116]
macroH2A2	-	Developmental genes[103]
H2A.Z.2.1	-	-
H2A.Z.2.2	NuA4/TIP60/P400[201], SRCAP[201]	-
H2A.Z.1	ANP32E[202,203], SRCAP[204,205], NuA4/TIP60/P400[75], INO80[203]	Transcriptional activation[205], Myogenesis[97], ES cell renewal and embryonic development [78,206]
H2A.Bbd	NAP1[207]	Stability of newly synthesized DNA[208], regulation of gene transcription[209]
H2B	FACT[187], NAP1[188–192], nucleolin[193]	Regulation of gene expression
TSH2B	-	Testis-specific form[210], reprogramming[211]
H2BFWT	SWI/SNF[212]	-
H3.1	FACT[187], NAP1[189,213-215], ASF1[216–218], NASP1[219,220], CAF1[221–223]	Repression of myogenesis[65], Regulation of gene expression[224]
H3.2	FACT[187], NAP1[189,213-215], ASF1[216–218], NASP1[219,220], CAF1[221–223]	Regulation of gene expression[224]
H3.3	FACT[187], ASF1[218,225], ATRX/DAXX [30–53–54–226–230], HIRA[30–42–50–230–234], CHD1[56,235,236], CHD2 [60,235], DEK[237], EP400[238], RPA[234]	Myogenesis [58,60], genome integrity[228], embryonic development[44], transcriptional activation[30,31], lineage determination and epigenetic memory [41,42]
H3.1 t	-	Testis-specific form[224]
H3.5	-	Transcriptional regulation in spermatogenesis[239,240]
H3.Y	-	Transcriptional response to stimuli[241,242]
H3.X	-	-
CENP-A	HJURP[243,244]	Centromere organization and chromosome segregation[244,245]
H4	FACT[187], NAP1[189,213-215], CAF1[221–223]	Regulation of gene expression

With the exception of H4, all core histone subunits have multiple variants in humans, which are deposited by specific chaperone complexes and associated chromatin remodellers. The known variants of all subunits, along with known physiological roles and associated chaperones/remodelling complexes are shown. For each subunit, dark colours indicate replication-dependent variants, whereas light colour indicate replication-independent variants. Variants within a single box are splice isoforms of each other.

Histone variants in skeletal myogenesis

Skeletal myogenesis is the process of formation of mature skeletal muscle tissue from precursor cells. While this occurs primarily during embryonic and foetal development, it is also be recapitulated in adult tissue, when resident muscle stem cells, known as satellite cells, are activated in response to injury or damage. Myogenic differentiation (Figure 1) is a tightly regulated process marked by the sequential expression of transcription factors and epigenetic complexes that regulate the expression of cell cycle and myogenic genes. Both foetal and adult stem cells express paired-box (Pax) transcription factors Pax7 and Pax3. Myoblast determination protein 1 (MyoD), the master regulator of myogenesis, and myogenic factor 5 (Myf5), are among the earliest expressed transcription factors involved in cell lineage specification [9–11]. Skeletal myoblasts irreversibly exit the cell cycle and fuse to form multinucleated myotubes. As differentiation proceeds, other myogenic regulatory factor (MRF) including myogenin (Myog) and Mrf4 as well as myocyte enhancing factor (MEF) drive the expression of subsequent genes, such as troponin, myosin heavy chain (MHC), muscle creatine kinase (MCK) and skeletal alpha (α)-actin. Terminal differentiation is marked by the presence of MHC (encoded by the Myh genes), skeletal muscle (α)- actin (encoded by Acta1) and MCK (encoded by the Ckm gene) [12–14].

Figure 1.

Histone variants and chaperones/remodelling complexes involved in myogenic differentiation. Skeletal myogenesis involves a series of well-defined steps the result in skeletal myofibres being produced from either foetal or adult stem cells. Each step is driven by the expression of transcription factors, and marked by the presence of certain components that can be used to determine the stage of differentiation. The histone variants and associated chaperones/remodelling complexes that are currently known to be involved at each stage are also shown. These are discussed later in this review

Gene expression is also controlled by epigenetic complexes, both activating and repressing. Many of these are histone-modifying enzymes. Histone modifications include methylation, demethylation, acetylation, deacetylation, phosphorylation, sumoylation, and ubiquitination among others [15]. Depending on the type and residue of modification, these marks can affect histone–chromatin interactions, or act as docking sites for enzymes and modifiers. Therefore, histone modifications are crucial to regulating gene expression, and many complexes play critical roles in regulating myogenesis [16–20].

ATP-dependent chromatin remodelling is also crucial to skeletal myogenesis [21]. SWItch/Sucrose Non-Fermentable (SWI/SNF)-linked chromatin remodelling is essential to MyoD-dependent activation of myogenic genes [22]. The activity of SWI/SNF complexes is regulated by external differentiation cues [23]. The SWI/SNF ATPase Brg1 interacts with both MyoD and Myog to induce expression of skeletal muscle genes [24]. During the onset of differentiation, there is extensive chromatin remodelling and altered expression of the HIST1 gene cluster on chromosome 13[1], which consists of a subset of 16 replication-independent histones. Downregulation of a majority of the genes in this locus (with the exceptions of HisIh2bc, Hist1h1a, and HistIh1c) corresponds with the idea that the expression of replication-dependent histones is progressively suppressed during myoblast differentiation [25].

Therefore, temporal regulation of chromatin landscape through a myriad of ways is critical to the process of myogenesis [26–28]. Our understanding of the role of histone variants in myogenesis is summarized below.

H3.3

Of all the replication-independent variants, the H3 variant H3.3 has been best-studied for its role in myogenesis. H3.3 varies from the two canonical subunits, H3.1 and H3.2, by only five and four amino acids, respectively, [29]. Nevertheless, in spite of the high degree of similarity between the canonical variants and H3.3, they diverge significantly in the roles they play when incorporated into a nucleosome. H3.3 is preferentially incorporated into promoters, although it has also been found in enhancers and in gene bodies of actively transcribed genes [30,31]. Unlike its canonical counterparts, it is enriched in activating histone marks and is thought to destabilize nucleosomes into which it is incorporated. Chromatin Immunoprecipitation sequencing (ChIP-seq) has shown that H3.3 enrichment correlates with markers of active transcription, such as active (Ser-5 phosphorylated) RNA polymerase II and H3K4me3, H3K36me3, and H3K4me1. In addition, H3.3 antagonizes binding of the repressive histone H1 [32,33]. Therefore, the role of H3.3 is predominantly activating in nature [30,34–40]. It is required for cellular ‘memory’ of lineage identity. This requires the accumulation of both activating (H3K4me3) and repressive (H3K27me3) marks [41]. Interestingly, H3.3 is required to maintain parental lineage early on during cellular reprogramming, but its role is later reversed to establish and maintain the newly established lineage [42]. Surprisingly, there is also some evidence to show that H3.3 accumulates at, and is required for, the maintenance of inactive regions of the chromatin, such as telomeres and pericentric heterochromatin [43].

Like all replication-independent variants, H3.3 is incorporated throughout the cell cycle, as opposed to the canonical form. The protein is encoded by two separate genes, H3F3A on chromosome 1 and H3F3B on chromosome 17. Double knockout of both genes leads to embryonic lethality, indicating the importance of H3.3 in embryonic development. While H3F3A-null mice are viable to adulthood, male fertility is impaired and there is some growth deficiency. Likewise, H3F3B-heterozygous mice show growth deficiency and male infertility. H3F3B-null mice show severe deficiencies in growth (both foetal and post-natal) and survival, resulting in embryonic or post-natal death [44]. Where mice do survive, knockout cells show karyotype defects, as well as defects in cell cycle. The defects are a result of ectopic localization of another variant of H3, CENP-A, in the absence of H3.3 [45]. H3.3 and its chaperone HIRA are also required for gastrulation in Xenopus [46,47]. GFP-tagged H3.3 knock-in mice are a useful tool to study genome-wide distribution of H3.3 and show correlation between H3.3 deposition and transcripts produced [48]. In addition, mouse models with SNAP-tagged H3.1/H3.3 have been used to study symmetric and asymmetric cell divisions in muscle stem cells. These have shown that muscle stem cell division patterns change when taken from in vivo to ex vivo, and that there is symmetric distribution of H3.1 and H3.3-containing nucleosomes among daughter cells [49].

Each variant has been shown to interact preferentially with different histone chaperones. H3.1 deposition is linked to the histone chaperone Asf1b and the CAF complex. H3.3 primarily interacts with the chaperone HIRA, responsible for its deposition at regions of active transcription, or those induced by stress or infection [30,50–52], as well as with the ATRX/DAXX complex, which is required for its incorporation into inactive regions of the genome [53–55]. H3.3 has also been previously shown to interact with the chromatin remodelling factor Chd1 [56].

Expression and function in myogenesis

Early studies in chicken myotubes showed that there are changes in the variants of H3 produced over the course of myogenic differentiation. H3.2 (the canonical subunit) is predominantly synthesized in proliferating myoblasts and a switch to H3.3 occurs differentiation proceeds. This is expected given that the switch to differentiation results in a permanent exit from the cell cycle, accompanied by a loss of synthesis of replication-dependent variants [57]. This pattern was also seen in mouse myotubes, where it was also seen that H2A.X, H2A.Z, macroH2A, and CENP-A were all downregulated upon cell cycle exit [58].

In C2C12 cells, similarly, the expression of H3.3 increased during the onset of differentiation, while the expression of the canonical variant of H3 decreased after cell cycle exit. In addition, the expression of HIRA and Asf1a were also maintained during myogenesis [58,59], as opposed to that of Chd1 and Asf1b.

An early study showed that H3.3 was deposited at the MyoD promoter, and promoted its expression, as well as that of Myf5 [41]. Later, the activation of MyoD was shown to be through the accumulation of H3.3 around the core enhancer region (CER), distal regulatory region (DRR) and proximal regulatory region (PRR) of the MyoD gene (Figure 2a). These have been shown to be important regulatory regions for MyoD expression. High levels of H3.3 were found around the CER and PRR prior to the onset of differentiation, during which process these levels increased even further. Combined loss of Chd1 and HIRA led to significant loss of H3.3 around these sites. Moreover, the role of H3.3 in regulating MyoD expression is dependent not on the changes in associated histone marks that usually accompany variant substitution, but through H3.3-dependent removal of repressive linker histone H1 and other structural changes to the chromatin around the PRR [58].

H3.3 deposition is also important for a later step in myogenic differentiation. In skeletal myoblasts, the SNF2-family chromatin remodelling enzyme chromodomain helicase DNA-binding domain 2 (Chd2) interacts with MyoD to deposit H3.3 at the loci of myogenic regulatory sequences. Binding of Chd2 to these myogenic loci is dependent on MyoD, and loss of either Chd2 or H3.3 reduced expression of the myogenic genes Myog, Ckm, Myh4 and Acta1, and impaired formation of myotubes. H3.3 was deposited around the transcription start site (TSS) of these myogenic loci prior to onset of differentiation, and levels of incorporation increased even further during differentiation. MyoD-dependent deposition of H3.3 at myogenic genes in myoblasts is a mechanism by which genes required for differentiation are marked prior to the induction of the process [60]. Thus, H3.3 deposition is important not only for MyoD expression (Figure 2(a)), but also for its subsequent activating function (Figure 3(a)).

Figure 2.

Regulation of MyoD expression through histone variant substitution. The effect of individual variants are shown separately (a) Replacement of H3.1 with H3.3 at the MyoD PRR and CER by a complex consisting of HIRA and Asf1a results in removal of the repressive linker histone H1. Removal of H1 also leads to removal of repressor Msx1.The resulting permissive chromatin structure causes activation of MyoD expression. (b) Around the transcriptional start site, canonical H2A is replaced with H2A.Z, which accumulates activating acetyl marks (Lys4, 7, 11, 13, and 15). This activates MyoD expression. It is important to note that the sequence of figures does not necessarily indicate a temporal sequence of events. The order in which these changes occurs is unclear as of now. The key for both parts (a) and (b) is shown on the right

Figure 3.

Regulation of Myog expression through histone variant substitution. Multiple histone variant substitutions are involved in the transcriptional activation of Myog. (a) Replacement of H3.1 with H3.3 by the HIRA-Chd2 complex, along with MyoD, results in the acquisition of activating H3K4me3 marks around the Myog promoter. Myog expression is activated. (b) Around the transcriptional start site, canonical H2A is replaced with H2A.Z, which results in accumulation of activating acetyl marks. This activates Myog expression. (c)

At the enhancer, replacement with macroH2A leads to an accumulation of H3K27Ac marks. This leads to recruitment of MyoD and co-activator Pbx1. Myog expression is activated. It is important to note that the sequence of figures does not necessarily indicate a temporal sequence of events. The order in which these changes occurs is unclear as of now. The key for parts (a), (b) and (c) is shown on the right

The forced expression of the canonical variant H3.1, which replaces H3.3 in the TSSs of myogenic genes when ectopically expressed, has a repressive effect on myogenesis. Overexpression of H3.1 in C2C12 myoblasts led to a reduction in the expression of myogenic genes Myog, Myh, Acta1 and Ckm, whereas overexpression of H3.3 led to increased expression of these genes. This was shown to be a result of ‘bivariant’ nucleosomes. There was an increase in H3K27me3 (repressive) and a decrease in H4K4me3 (activating) marks. The increase in the H3K27me3 mark was due to increased interaction between H3.1 and Ezh2, the catalytic subunit of the Polycomb Repressive Complex (PRC2). Conversely, knockdown of H3.3 led to loss of both H3K27me3 and H3K4me3, indicating that H3.3 is required for the maintenance of the bivalent state. Bivalent modifications have been shown to mark or prime the promoters of lineage-specific genes prior to induction [61–64]. In contrast, H3.1 leads to predominantly repressive marking. The recruitment of H3.3 to skeletal muscle genes is dependent on HIRA, the knockdown of which leads to reduction in the expression of genes such as Myog and Acta1 [65].

H3.3 is also involved in the reorganization of constitutive heterochromatin that takes place during myogenic differentiation. The ATRX/DAXX/H3.3 complex interacts with the muscle-specific long non-coding RNA ChRO1, which is induced during differentiation. This interaction and incorporation of H3.3 in satellite repeats of chromocenters is critical for chromocenter clustering and cell differentiation [66].

Satellite cells are capable of developing into mature muscle tissue in response to damage or injury [67–69]. One specific sub-variant, H3mm7, is expressed in satellite cells and is required for muscle regeneration. Although not much is known about the sub-variants of histone variants, due in part to the high degree of similarity between sub-variants, which makes detection difficult, one study has shown the presence of multiple sub-variants with tissue-specific distribution. The forms most highly expressed in mouse skeletal muscle are subvariants H3mm15, H3mm13, H3mm14, H3mm7, H3mm8, and H3F3A, whereas H3F3B is expressed at much lower levels [70]. H3mm7 is highly expressed in satellite cells, with reduced expression in muscle fibres. As a result of the alanine residue at position 57 (serine in other variants), H3mm7 forms weaker bonds with the arginine residue at position 40 on H4. The resulting nucleosome is less stable than the ones formed by H3.3-H4 subunits. Therefore, there is a more open chromatin structure around the promoter regions of many genes, including several myogenic genes. The difference between H3.3 and H3mm7 is the rate at which target genes are expressed upon incorporation, that is, how much nucleosomal instability and transcriptional accessibility each of these forms provides. Knockout of this sub-variant leads to defective adult skeletal muscle regeneration and differentiation, while there is no other dramatic phenotype. Thus, the incorporation of H3mm7 at myogenic genes is crucial for satellite cell differentiation [36].

H3.3 chaperones

In light of its role in depositing H3.3 in regions of active transcription at myogenic loci at the onset of differentiation, the HIRA chaperone complex has been recognized to be important for myogenesis. Loss of HIRA led to a reduction in the expression of MyoD, Myog, MCK, and MHC. In particular, loss of HIRA led to a loss of MyoD in differentiating cells, but did not significantly affect proliferating myoblasts [58,59]. Given that the HIRA complex localizes to myogenic genes before the onset of differentiation, after which there is not much change in the levels of protein, one possibility is that the activation of complex, rather than its levels, that are important for induction of differentiation. To this end, the regulation of HIRA has also been studied to a certain degree. Deposition of H3.3 during myogenesis is dependent on the activation of HIRA, which is phosphorylated at serine (S) 650 (in human) or 648 (in mouse) by Akt1. Phosphorylated HIRA is inactive and cannot undertake H3.3 deposition at myogenic genes and needs to be dephosphorylated before it can do so. Therefore, this provides a mechanism by which the deposition of H3.3, and the subsequent induction of myogenic genes can be regulated by differentiation cues such as the phosphorylation and activation of Akt [71].

Skeletal myocyte-specific HIRA depletion in mice resulted in myofibers hypertrophy, constant regeneration, sarcolemmal perforation and oxidative damage by 6 months of age. Fast-twitch muscle fibres of these knockout mice exhibited more profound effects. Hypertrophy and regeneration were accompanied by increased relative abundance of reactive oxygen species, which likely promoted muscle fibre degeneration and subsequent regeneration. Indeed, HIRA loss affects a large number of genes involved in cellular metabolism and DNA damage response [72]. HIRA also contributes to the activation of Mef2c-target genes during differentiation, and Asf1 was indispensable for transcriptional activation by this complex [73].

Interestingly, the H3.3 and macroH2A-linked chaperone ATRX also appears to be critical for muscle development; conditional skeletal muscle specific-knockout of ATRX in mice led to offspring that were viable at birth, but showed kyphosis and significant defects in muscle development within 3 weeks of birth due to profound genomic instability in myoblast cells, leading to poor expansion and regeneration [74].

H2A.Z

H2A.Z, one of the several replication-independent variants of the H2A subunit is, much like H3.3, linked to transcriptional activation [75–77]. H2A.Z constitutes about 10% of the total pool of H2A at any given time [78]. H3.3 and H2A.Z have been previously been shown to co-occupy nucleosomes that mark regions of transcriptional activity [37]. Like H3.3, H2A.Z is thought also to destabilize nucleosomes and confer a more open, accessible chromatin structure [78–81]. It is also thought to confer epigenetic memory of previous transcriptional states [82]. H2A.Z is required to maintain genomic stability and chromosome segregation [83]. It is enriched, and interacts with HP1α, at pericentric heterochromatic foci and is required to maintain heterochromatin through this association [83,84]. At the same time, it is present at boundary elements, where it plays a role in prevention of spread of silent heterochromatin [85]. Like H3.3, it can accumulate histone marks linked to ‘poised’ promoters, such as those involved in lineage specification and development [86].

H2A.Z deficiency results in embryonic lethality in mice, as in drosophila, indicating its importance in embryonic development [78,87]. Brain-specific knockouts of H2A.Z have been created, which exhibit altered proliferation and differentiation of neural progenitors [88]. ZNHIT1 knockouts in intestinal epithelial tissue have shown the importance of this SNF2‐related CBP activator protein (SRCAP) subunit in H2A.Z deposition and intestinal stem cell homoeostasis [89,90].

Recently, it has been shown that there are two isoforms of H2A.Z, H2A.Z.1 and H2A.Z.2, in vertebrates. Although these isoforms are deposited to the same extent by SRCAP, the two forms have some differences in the ways they affect nucleosomal stability and chromatin structure [91]. Nevertheless, most studies either take both to be one form, or focus on the form encoded by H2A.Z.1.

Expression and function in myogenesis

H2A.Z levels are low in proliferating myoblasts and accumulate rapidly in myotubes [57], as well as in dystrophic muscles [92]. However, downregulation of H3.1, H3.2, macroH2A, H2A.X, H2A.Z, and CENP-A has also been reported in C2C12 myotubes upon onset of differentiation [58].

H2A.Z is important for the expression of MyoD (Figure 2(b)). Ectopic expression of a non-acetylable form of H2A.Z in place of the wild-type form results in a less permissive chromatin structure around the MyoD transcriptional start site, reduced MyoD expression and a block in myogenic differentiation in C2C12 cells. There are five lysine residues that can be acetylated at the N-terminus of H2A.Z (Lysine-4, 7, 11, 13, and 15) and acetylation has been linked to its role in transcriptional activation, whereas ubiquitination (which usually occurs at Lysine-120, 121, and 125 at the C-terminus) results in transcriptional repression and polycomb signalling [93–95]. However, knockdown of H2A.Z did not result in the same phenotype, which could be due to the presence of isoforms that were not depleted by the chosen shRNA, resulting in residual H2A.Z [96].

The p38 MAPK substrate ZNHIT1 or p18^Hamlet, a subunit of the SRCAP chromatin remodelling complex, is required for the deposition of H2A.Z at the Myog promoter. This accumulation of H2A.Z during the early stages of differentiation is required for the activation of Myog expression (Figure 3(b)). This is similar to the HIRA/Chd2/Asf1a-dependent deposition of H3.3 at the same locus. Furthermore, similar to the Akt-dependent phosphorylation of HIRA, which acts as a signal for the induction of differentiation through H3.3 deposition, p38 MAPK‐mediated p18^Hamlet phosphorylation is also required for the deposition of H2A.Z [97]. Some work will be required to understand how the processes of H3.3 deposition and H2A.Z deposition interact at the Myog promoter and affect myogenic differentiation.

In post-mitotic cells, such as mature muscle, H2A.Z is not required either to activate or maintain transcription; it is a marker rather than a driver of transcription. RNA-sequencing and ATAC-sequencing showed that neither gene expression nor chromatin accessibility was significantly altered by loss of H2A.Z. Thus, while it is necessary for the transition from myoblasts to early differentiation, it is not required in differentiated tissue [98].

MacroH2A

MacroH2A is nearly thrice the size of canonical H2A due to the presence of an evolutionarily conserved non-histone macro domain of ∼25 kDa. This domain can interact with metabolites such as NAD, thus playing a role in regulation of energy and metabolism [99–102]. macroH2A is an important epigenetic regulator of developmentally regulated genes [103]. While histone variants tend to be non-allelic for the most part, macroH2A variants are produced by alternative splicing. In mammals, three macroH2A proteins-macroH2A1.1, macroH2A1.2 and macroH2A2- are produced through two genes and one mutually exclusive splicing event [101].

Knockout of both isoforms in mice leads to pre-natal and post-natal growth deficiencies and loss of reproductive capacity [104]. macroH2A plays an important role in X-chromosome inactivation in females [105]. Double knockout models show that these variants, along with H3K27me3, are found on important pluripotency genes, where they act as a barrier to induced pluripotency [106,107]. Cells harbouring the double knockout can differentiate, but retain the ability to return to a stem-cell like state [106]. While macroH2A1 has been mainly implicated in limiting chromatin accessibility [108] and promoting gene silencing, it has been suggested that it may have broader roles, including finetuning gene expression [109]. macroH2A is required to maintain nuclear organization and heterochromatin structure [110].

Expression and function in myogenesis

Both forms are found in proliferating myoblasts, but mRNA encoding macroH2A1.1 rapidly becomes predominant during differentiation. Indeed, while protein levels of macroH2A1.1 increase over differentiation, macroH2A1.2 levels rapidly decrease [111].

The macroH2A1.1 isoform binds to ADP-ribose, and in differentiating myotubes, where there is a switch towards increased carbon use for ATP production, macroH2A1.1 binds to and inhibits poly-ADP ribose polymerase (PARP-1) activity. Myotubes lacking macroH2A1.1 show defective mitochondrial respiratory capacity because the buffer of NAD+ precursors (NMN in particular) is disrupted [111]. Therefore, macroH2A plays an important role in protecting differentiated myotubes from oxidative stress [111–113].

Similar to H3.3 and H2A.Z, macroH2A1.2 is also recruited to the genes involved in the myogenic regulatory network and is required for the activation of the myogenic programme . Unlike H3.3 and H2A.Z, however, macroH2A is recruited to the prospective enhancers of genes such as Myog, where its presence is required for the activating H3K27 acetylation mark and the recruitment of transcription factor Pbx1 (Figure 3(c)), which cooperates with MyoD [114,115]. macroH2A1.2 ‘marks’ prospective enhancers, and although it is not critical for the maintenance of active enhancers, it is important for the activation of the myogenic gene network [116].

H1b

While the contribution of replication-independent histone variants in myogenic differentiation, for the most part, is activating in nature, there is also a repressive role played by the linker histone H1. H1 histones are ∼200 amino acid residues in length. H1 stabilizes the nucleosome through interactions with DNA and facilitates higher-order folding of nucleosomes [117]. There are 11 known variants histone H1 in humans. H1.1-H1.5 are replication-dependent, whereas H1.X and H1.0 are replication-independent. The remaining forms (H1oo, HILS, H1T, and H1T2) are specific to germ cells. Individual knockouts of the somatic variants are tolerated in mice [118], although combined knockout of three variants- H1c, H1d, and H1e- was embryonic lethal [119]. The loss of H10 in mice is also tolerated, and mice develop normally [120]. In Drosophila, however, H1 is important for embryonic development [121]. In Xenopus, H1 has been shown to be important for regulation of mesodermal differentiation [122].

Function in myogenesis

In muscle, H1b (H1.5), a variant of mouse histone H1, cooperates with the transcriptional repressor Msx1 to bind to the CER of MyoD, resulting in a closed chromatin conformation and preventing MyoD activation. There is consequently impaired myotube formation. The ability of Msx1 to repress myogenic differentiation depends on its interaction of H1b found at the CER of MyoD [123].

An extra-nuclear role for H1 in regulating skeletal muscle regeneration has also been shown. H1 is found at the plasma membrane and extracellular matrix in a pool of regenerating cells, where it interacts with the heparan sulphate chain of the proteoglycan perlecan. Perlecan is present on the plasma membrane of myoblasts, with reduced expression during differentiation [124]. H1-perlecan interaction results in a strong proliferative signal for myoblasts, indicating a role for H1 in skeletal muscle regeneration. The source of H1 in the extracellular matrix is hypothesized to be cell necrosis [125].

Conclusions

Hitherto, H3.3, H2A.Z, macroH2A, and H1b have been shown to be important for skeletal differentiation (Table 2). H3.3 is deposited at regulatory and enhancer regions of MyoD, whereas H2A.Z is deposited around the transcription start site. While H2A.Z is required for the accumulation of activating marks, H3.3 is required for the removal of H1. H1b has the opposite function; its presence serves to repress MyoD expression. Unlike MyoD where deposition of variants occurs at different loci, multiple variants appear to deposited around the promoter resulting in Myog expression. Loss of either of these variants shows a reduction in the expression of Myog and other myogenic genes. However, given the relationship between H3.3 and H2A.Z, one can predict the existence of nucleosomes containing both these variants [37,126]. macroH2A is also required at enhancer regions (Table 2).

Table 2.

Role of histone variants in expression of MRFs

MRF	RI Variant	Location	Deposition	Effect	References
MyoD	H3.3	CER, PRR	Prior to onset of differentiation; increases during differentiation	Removal of H1; gene activation	58
	H2A.Z	TSS	Prior to onset of differentiation	Accumulation of Ac marks; gene activation	96
	H1b	CER	-	Cooperates with MSX1; prevents activation	123
Myog	H3.3	Promoter	Prior to onset of differentiation; increases during differentiation	Accumulation of H3K4me3 marks; gene activation	60
	H2A.Z	Promoter	Onset of differentiation	Gene activation	97
	macroH2A	Enhancer	Onset of differentiation	Accumulation of H3K27Ac marks; recruitment of Pbx1 and MyoD; gene activation	116

The individual RI histone variants currently known to be involved in the in expression of two key myogenic regulatory factors, MyoD and Myog are shown here, along with location, time and effect of deposition.

While these studies have clearly started to shed light on their critical role in myogenesis, several questions remain: First, given their myriad roles in myogenesis, it is conceivable that the expression or function of histone variants may be deregulated in myopathies, but this has not been studied in any detail. Deregulation of the expression of certain variants, or their chaperone complexes, may have important pathological implications that are currently underappreciated. For instance, myogenesis is impaired in dystrophies, rhabdomyosarcoma, and age-dependent decline of muscle regenerative capacity. Congenital myopathy is characterized by impaired myotube formation as a result of loss of MyoD, Myog, dystrophin, and N-CAM expression. This defect is limited to skeletal muscle only [127]. Immunohistochemical staining of biopsies from patients with various forms of primary myopathies and neurogenic atrophy also showed varying degrees of loss of MyoD expression, mostly coinciding with the extent of atrophy [128]. It is unclear to what extent these defects are due to deregulation of histone variants that regulate MyoD expression and activity. The transcription factor Dux4, whose expression is deregulated in facioscapulohumeral dystrophy (FSHD), induces the expression of H3.X and H3.Y, which in turn are incorporated into the targets of Dux4, therefore enhancing a network of deregulated Dux4 target genes [129]. Other studies have shown the importance of deregulated histone modifications in certain myopathies [130]. Similarly, the age-related loss of muscle function is characterized by epigenetic changes that affect the ability of satellite cells to regenerate muscle tissue [131]. Small molecules targeting some epigenetic regulators that affect the differentiation process have been developed. For example, histone deacetylase (HDAC) inhibitors have been proposed as a therapeutic option for myopathies, but HDAC inhibitors often pose the problem of lack of specificity and off-target effects [132]. The requirement for certain histone variants in myogenic differentiation is somewhat overlooked. The Chaserr lncRNA inhibits Chd2 expression and thus targeting it could be a way to restore Chd2 levels [133] and Myog expression, as Myog is a specific target of the Chd2-HIRA-MyoD complex [60]. Similarly, it has been suggested that small-molecule inhibitors of the Asf1 histone chaperone may be used for therapeutic purposes in cancers, which harbour high levels of H3K56Ac marks and aberrant cell proliferation, or for treatment of fungal infections, as this histone mark is more abundant in fungal genomes [134,135]. The allosteric inhibitor, MK2206 can prevent the S650 phosphorylation of HIRA through inhibition of Akt phosphorylation. This is another potential, if somewhat complex, way to control HIRA and H3.3 deposition [71].

Second, while the expression of certain variants changes during myogenic differentiation, little is known about the upstream regulators of their expression. H3.3 is selectively incorporated into specific loci in skeletal muscle in response to exercise. It has been suggested that there is nucleosomal repackaging in response to exercise training. The kind and intensity of training are thought to determine target gene regulation [136–138]. However, the transcription factors and regulators involved are not yet clear. This is important if the expression of variants requires manipulation in the treatment of myopathies. Third, ChIP-sequencing and ATAC-sequencing have been done to study global changes upon perturbation of specific variants. Some targets genes such as MyoD and Myog have been identified. However, identification of additional downstream targets in physiological contexts as well as in myopathies may reveal novel genes that may be possible to target therapeutically in myopathies that are associated with a block of myogenic differentiation.

Several noncoding RNAs (ncRNAs), including microRNAs and lncRNAs have been shown to be critical to skeletal myogenesis (reviewed in [139–144]). However, the interaction between these ncRNAs and histone variants is unclear. It has been shown that Chro1 is important for H3.3 deposition [66]. H1.3 has also been previously shown to repress the expression of the h19 ncRNA in ovarian cancer [145], but the role of histone variants in either regulating expression of ncRNAs, or any interaction with ncRNAs, in the context of myogenesis, has not been shown. As these ncRNAs could also prove to be potential targets for regulating myogenesis, the interaction between ncRNAs and histone variants involved in skeletal myogenesis should be analysed. Similarly, DNA methylation is another important form of epigenetic regulation involved in myogenesis. Histone and DNA modifications work in conjunction to regulate gene expression [146–148]. DNA methylation has been shown to play an important role in skeletal myogenesis and response to exercise [149–152]. In addition, certain histone variants are known to regulate DNA methylation by regulating access to DNA methyltransferases and other modifiers (Table 3) [75,153–162]. However, the interaction between histone variants and DNA methylation has not been studied in the context of skeletal myogenesis. Such studies will shed light on mechanisms by which histone variants in combination with other epigenetic regulators, orchestrate chromatin landscaping during skeletal myogenesis.

Table 3.

Transcriptional properties of histone variants, PTMs, and associated epigenetic marks

Variant	Transcriptional function	Associated Epigenetic Marks/Modifications	Associated ncRNA
H1.1	Predominantly repressive	Multiple histone PTMs thought to be important for epigenetic regulation, including phosphorylation, methylation, acetylation, citrullination, ADP ribosylation, ubiquitination, carbonylation, formylation, denitration, crotonylation and lysine 2-hydroxyisobutyrylation[246,247] Enriched in regions of DNA methylation, recruit methyltransferases[161,162]
H1.2	Activating[173], Repressive[174,175]
H1.3	Repressive[145]		h19[145]
H1.4	Activating[174,177,178]
H1.5	Predominantly repressive
H1.0	Repressive
H1.X
H1t	-
H1T2	-
H1oo	-
HILS	-
H2A	-	Multiple histone modifications, including acetylation, phosphorylation, methylation, demethylation, deamination, β-N-acetylglucosamine, ADP-ribosylation, ubiquitination and sumoylation[15,248]
H2A.X	Predominantly repressive[249]	Specific histone residues include: K118Ub, K119Ub, K5Ac, K36Ac, K134me, K5Sumo, K9Sumo, K13Sumo, K15Sumo, K118Sumo, K119Sumo, K127Sumo, K133Sumo, K134Sumo[250]
macroH2A.1.1	Both activating and repressive[103,106,109,200,251]	Specific histone residues include: K17me, K122me2, K237me/me2, T128P, S137P, T101P, T136P, S139P, Y142P[250,252]
macroH2A.1.2
macroH2A2		-
H2A.Z.2.1	Activating[3]	Specific histone residues include: K4Ac, K7Ac, K11Ac, K13Ac, K15Ac, K120Ub, K121Ub, K125Ub, K4me, K7me, K101me2, K115Ub, K116Ub, K122Ub[250] Antagonizes DNA methylation[75]
H2A.Z.2.2	Activating[201]
H2A.Z.1	Predominantly activating[3,7596,205]
H2A.Bbd	Activating[207,253]	-
H2B	-	Multiple histone modifications, including acetylation, phosphorylation, methylation, demethylation, deamination, β-N-acetylglucosamine, ADP-ribosylation, ubiquitination and sumoylation[15,248] Deubiquitination/deacetylation required for DNA methylation[159,160]
TSH2B	-	-
H2BFWT	-	-
H3.1	Predominantly repressive[223,224]	Specific histone residues include: K27me3 [65,223,224], K9me2[254], K14Ac[246], K64me1[246]
H3.2	Predominantly repressive[224]	Specific residues include: K27me2/3, K36me1[246]
H3.3	Predominantly activating[34,35,37-40]	Specific histone residues include: K4me3, K36me3, K4Ac, K4me2, K9Ac, K27Ac, K79Ac, K14Ac, K18/K23Ac, K36me/me2 K79me/2[34,35,37-40,246,254] K27me3[41] Facilitates gene body and promoter DNA methylation[157,158]	Chro1[66]
H3.1 t	-	-
H3.5	-	-
H3.Y	-	-
H3.X	-	-
CENP-A	Unclear	Specific histone residues include: G1me3, S7P, S16P, S18P, S68P, K124Ub, K124Ac, K124me[255]
H4	-	Multiple histone modifications, including acetylation, phosphorylation, methylation, demethylation, deamination, β-N-acetylglucosamine, ADP-ribosylation, ubiquitination and sumoylation[15,248]

The key transcriptional role of histone variants, along with the PTMs and epigenetic marks associated with each, are shown here. DNA methylation is an important regulatory mechanism in skeletal myogenesis and histone variants involved in changes in DNA methylation are shown here (under Epigenetic changes). ncRNAs, which also play an important role in skeletal myogenesis, may well work in conjunction with histone variants and require further study.

Funding Statement

This work was supported by Ministry of Education grant (MOE2019-T2-1-024);MOE [MOE2019-T2-1-024].

Disclosure statement

No potential conflict of interest.