Structure and function of histone acetyltransferase MOF

2.1. Structure characteristics

MOF belongs to the MYST family of histone acetyltransferase. The name MYST stands for the first four founding members: MOZ, Ybf2 (Sas3), Sas2, and Tip60. Other important family members include MOF, Esa1, MORF, and HBO1 [17,18]. The common characteristic of this family is the catalytic MYST domain that transfers an acetyl group from acetyl-CoA to the lysine side chain of histone tail. Within the MYST family, MOF is closely related to Tip60 as they both contain a chromo-like domain adjacent to the N-terminal of the MYST domain.

The structure of human and Drosophila MOF are quite similar, sharing 50% identity and 69% similarity (Figure 1). The histone acetyltransferase (HAT) domain of MOF consists of a central core that mediates acetyl-CoA binding and an atypical C₂HC-type zinc finger within the flanking N-terminal region that mediates substrate recognition. Within the central core, a glutamate (E350) residue is crucial for the HAT activity, as the E350Q mutation essentially abolishes the MOF HAT activity [19]. A unique C₂HC-type zinc finger (CxxCx₁₂HxxxC) directly interacts with the globular part of the nucleosome and the histone H4 tail, and is crucial for HAT activity and substrate specificity [20]. Point mutations in the highly conserved cysteine or histidine residues completely abolish HAT activity as well as its interaction with the nucleosome [20].

Another common feature of the MOF/Tip60/Esa1 subfamily is the chromobarrel domain, named for its beta-barrel structure. Unlike canonical chromodomains that exhibit the structure of alpha + beta folds and bind to methylated lysine residues, the chromobarrel domains in MOF/Tip60/Esa1 lack the aromatic cage and consequently mediate the interaction between protein and nucleic acids [9,21]. Disruption of the chromobarrel domain compromises MOF HAT activity, resulting in a global reduction of H4K16 acetylation [22].

2.2. MOF and histone H4K16 acetylation

Multiple lines of evidence indicate that MOF is a H4K16 specific acetyltransferase. First, MOF is the only subunit in Drosophila MSL complex that has acetyltransferase activity [8,9]. Binding of MSL complex to the male X chromosome is correlated with a significant increase of H4K16 acetylation in that chromosome [8,23]. In addition, similar to its Drosophila ortholog, hMOF is a part of the human MSL complex, which is responsible for the majority of H4K16 acetylation in mammalian cells [24]. Moreover, knockdown of hMOF in HeLa and HepG2 cells leads to a strong decrease in H4K16 acetylation but other histone lysine acetylation remains unchanged [24,25]. Furthermore, the level of H4K16 acetylation is nearly undetectable in MOF knockout cells [15,16], while other histone acetylation marks including H4K5ac, H4K8ac, H4K12ac, H3K9ac and H3K14ac are not affected [16]. However hMOF is able to acetylate other histone lysines in in vitro assays, indicating the importance of protein complex, such as MSL or NSL, in regulating histone acetylation in cells. Finally, expressing MOF in cells where this gene was knocked out resulted in the recovery of H4K16 acetylation [26]. Taking together, these results indicate that MOF specifically and exclusively targets nucleosomal H4K16.

2.3. Non-histone targets of hMOF

Besides acting as a histone acetyltransferase, hMOF has been shown to acetylate the lysine residue in several non-histone proteins, including the subunits of multiprotein complexes (Msl3, TIP5) and crucial transcription factors (p53, Nrf2).

Drosophila MOF can acetylate MSL3 (a major component of the MSL complex) at lysine residue (K116) that is adjacent to one of its chromodomains [27]. MSL3 acetylation leads to a temporary loss of interaction between MSL3 and RNA, therefore contributing to the fine-tuning of dosage compensation levels in Drosophila [27]. However, it is not clear whether this function is conserved in the mammalian MSL complex. Interestingly, mammalian MOF is able to acetylate TIP5, the largest subunit of NoRC chromatin-remodeling complex, at lysine 633 that is adjacent to the TIP5 RNA-binding domain [28]. Acetylation of TIP5 also weakens its interaction with promoter-associated RNA and is required for NoRC-mediated rDNA silencing [28].

Transcription factors are also the target of MOF. As a master regulator of cellular response to genotoxic stress, p53 controls cells undergoing either cell cycle arrest or apoptosis by triggering different sets of downstream target genes [29,30]. It is reported that hMOF as well as its closely related MYST member TIP60, are capable of acetylating p53 at K120, a lysine residue located in the DNA-binding domain of p53. Acetylation of p53 K120 modulates its DNA-binding preference, leading to increased transcription of pro-apoptotic genes: PUMA and BAX [31]. Further studies show that p53K120 acetylation regulates cell apoptosis by both transcription-dependent and -independent mechanisms [31,32].

P53 is not the only transcription factor that can be targeted by MOF. A recent study on human non-small cell lung cancer reported the acetylation of NF-E2-related factor 1 (Nrf2) by hMOF [33]. Nrf2 is a crucial factor protecting cells against oxidative stress, for example, cancer cells require activated NrF2 in order to survive high levels of oxidative stress that exist in this disease [34]. hMOF is able to bind and acetylate Nrf2 at lysine 588, which facilitates its nuclear translocation and increases transcription of Nrf2 downstream genes. Nrf2 protein levels remain the same in the presence or absence of hMOF, suggesting that hMOF affects the Nrf2 transcription activity rather than its expression level [33].

3.1. MOF-containing protein complex

MOF exists in two distinct but highly conserved multiprotein complexes: MSL and NSL. MOF is best known as a major HAT in Drosophila MSL complex. Drosophila MSL complex contains five protein subunits and two non-coding roX RNAs [13,36]. Four other components include male-specific lethal 1-3 (MSL1, MSL2, MSL3) and the ATP-dependent RNA/DNA helicase maleless (MLE). While MSL1 and MSL3 are crucial for MOF activity and substrate specificity, MSL2, MLE and RNA components recruit MOF to specific chromatin locations [13,36]. Mammalian orthologs for these protein subunits have been identified in human, but not the RNA components (roX1/2). Interestingly, human MSL complex consisting only of MSL1-3 and MOF seem sufficient for acetylation of nucleosomal H4K16 [24]. Recently, MOF was found as a key component of another chromatin-modifying complex, nonspecific lethal (NSL) complex [6,24,37,38]. All the components of the NSL complex are highly conserved, consisting of at least 7 subunits: MOF/hMOF, NSL1/KANSL1, NSL/KANSL2, NSL3/KANSL3, WDS/WDR5, MCRS2/MCRS1, dHCF/HCFC1, OTG/OTG1 and MBD-R2/PHF20, in Drosphila and mammals, respectively [37,38].

MOF is the only common subunit shared by the MSL and NSL complexes. Both complexes are able to acetylate H4K16 in varied chromatin contexts, suggesting that MOF activity and substrate specificity are tightly controlled by the associated proteins in the complex. This is supported by the fact that Drosophila MSL complex is only associated with male X-chromosome as MSL2 is exclusively expressed in male cells [13,39]. On the other hand, the NSL complex exists in both male and female cells, and is responsible for global H4K16 acetylation [38]. In addition, while mammalian MSL complex only acetylates nucleosomal histone H4K16, the NSL complex can acetylate nucleosomal H4K5 and H4K8 as well as non-histone targets (p53 K120) [37]. Moreover, recent genome-wide studies have revealed different distribution patterns of the two complexes. In Drosophila, the MSL complex is localized in the coding region or 3′ end of X-linked genes, and the level correlates to gene transcription. The NSL complex is primarily located in the promoter region of more than 4000 genes and controls H4K16 acetylation and gene expression [38,40,41]. In mammals, the NSL complex is globally associated with the promoter region of housekeeping genes [40,42]. Interestingly, mammalian MSL complex binds to both promoter and gene bodies and regulates many embryonic stem cells-specific and bivalent genes [42,43].

3.2. Autoacetylation and phosphorylation

In 2011, three groups reported that MOF is capable of autoacetylation at lysine 274, a highly conserved lysine residue in the C2HC zinc-finger domain [19,44,45]. Two groups studied the crystal structure of MOF and found that a strong electron density at the tip of the K274 side chain could be modeled as an acetyl group [19,44]. The acetylation of K274 is critical for the proper position of the hairpin structure and affects MOF substrate specificity and HAT activity, which is supported by a lack of enzyme activity in various MOF K274 mutants (K274A, K274R, and K274Q) [19,44]. The third group [45] studied MOF autoacetylation based on its similarity to Tip60, where autoacetylation enhances its interaction with substrates [46]. Lu et al reported that both in vitro and in vivo MOF autoacetylation reduced its ability to bind to nucleosomes. Sirt1, a class III histone deacetylase, is able to remove the acetyl group from hMOF and thereby restore its interaction with nucleosomal substrates [45]. On the other hand, Yuan and his colleagues reported that K274 autoacetylation is required for MOF mediated acetylation of p53 and histone H4K16 [47].

In addition to acetylation, phosphorylation may also play a role in regulating MOF activity. A recent study found that the DNA double-strand break (DSB) induced ATM-dependent MOF phosphorylation at threonine 392 [48]. Phosphorylated MOF co-localized with γ-H2Ax, ATM, and 53BP1 foci, and is required for the recruitment of homologous repair proteins to DNA damaged sites during S/G2 phase.

4.1. hMOF and ATM

ATM (Ataxia Telangiectasia Mutated) protein is a serine protein kinase that has sequence homology to PI3K (Phosphoinositide 3-kinase) [49]. ATM gene was originally identified by position cloning of an altered protein in patients with Ataxia Telangiectasia, a neuron degeneration disorder characterized by cerebellar degeneration and neuromoter dysfunction. Later, numerous studies have established its critical role in DNA damage response and cell cycle checkpoint [49]. In the absence of DNA damage stress, ATM is presented as an inactive dimer or oligomer [50]. Following DNA damage, ATM autophosphorylates at serine 1981 and converts from an inactive dimer to an active monomer, which in turn phosphorylates and activates downstream effectors, such as p53 and checkpoint kinase 2 (Chk2).

In 2005, Gupta and colleague reported a direct physical interaction between hMOF and ATM [51]. This interaction was mediated by hMOF chromobarrel domain and the leucine zipper domain in ATM. How does this interaction affect hMOF or ATM function? The experimental data suggests that MOF is able to influence ATM function. First, ionizing radiation (IR) induced ATM-independent but MOF-dependent increase of H4K16 acetylation. In addition, inactivation of MOF using either MOF mutant or MOF knockdown resulted in a significant decrease of IR-induced ATM autophosphorylation. ATM kinase activity and phosphorylation of downstream effector Chk2 were also reduced. Interestingly, an increase in H4K16 acetylation using histone deacetylase inhibitors (trichostatin A or sodium butyrate) had no effect on ATM activity, suggesting that H4K16 acetylation alone is not sufficient to influence ATM function. These data suggest that hMOF is able to affect the early stages of DNA damage signaling through modulating ATM activation [51]. However, a recent study from the same group reported that ATM is able to phosphorylate hMOF at threonine 392, which is critical for cell survival and damage repair in the S and G2 phase [48]. Therefore, hMOF may act both upstream and downstream of ATM. It is worth noting that mice deficient in Purkinje-specific MOF displayed an ataxia-telangiectasia-like neurological phenotype [52], suggesting a functional interaction between MOF and ATM in vivo.

4.2. DNA damage response

Mammalian cells are constantly exposed to genotoxic stress from both internal and external sources. To protect genomic integrity, cells have developed a highly sophisticated defense response (DNA damage response, DDR) that consists of multiple key components: damage sensors, signal transducers, mediators, and effectors [53]. Successful DNA repair requires functional DDR to recruit specific DNA repair complexes as well as chromatin remodeling machineries to control DNA accessibility [53].

Given its fundamental role in modulating higher-order chromatin structure, several studies have investigated the involvement of hMOF in each step of DDR. MOF induced acetylation of H4K16 after DNA damage is subsequently modulated by the internucleosomal interaction between the acidic pocket of H2A-H2B surface and the basic patch on histone H4 tail [26]. In addition, it was reported that MOF is associated with ATM and the DNA-dependent protein kinase subunits (DNA-PKcs). Depletion of MOF reduced ATM autophosphorylation and abrogated ATM-dependent phosphorylation of Chk2 and DNA-PKcs [51,54]. Moreover, IR-induced recruitment of repair mediator Mdc1, 53BP1 and BRCA1 to DNA damage foci was completely abolished in MOF conditional knockout mouse embryonic fibroblasts (MEFs) [26]. These results indicate that hMOF is involved in nearly every step of DDR, and actively modulates DDR signaling as well as recruitment of DDR metiators and effectors (Figure 2). It is worth noting that early DNA damage sensing and ATM-mediated signaling were not affected in MOF conditional knockout MEFs [26]. This discrepency may result from the different ways that MOF was depleted or the type of cells used.

In addition to DDR signaling, hMOF is also crucial for DNA damage repair. DNA double-strand breaks can be repaired by either homologous recombination (HR) or nonhomologous end-joining (NHEJ) pathways, depending on the cell cycle phase as well as the opposing activity of 53BP1 and BRCA1 [55,56]. NHEJ is more active in G1 as 53BP1 inhibits DNA end resection of DSBs. On the other hand, HR is mostly active in the S/G2 phase due to inhibition of 53BP1 by BRCA1 [56]. Results from both MOF knockout and knockdown cells indicate that MOF is required for both HR and NHEJ [26,51,54]. Interestingly, phosphorylation of MOF by ATM facilitates the recruitment of BRCA1 as well as HR-related repair proteins to DNA damage sites in the S/G2 phase [48].

4.3. Cell survival and death

Since its identification, MOF is believed to play an important role in the life or death decision of cells and organisms. In Drosophila, mutation of MOF results in male-specific lethality [8]. In mammals, depletion of MOF leads to peri-implantation death in both males and females [15,16]. Additionally, it is impossible to derive MOF-deficient cell lines from MOF null embryo, as all MOF−/− cells died by apoptosis [15,16]. Moreover, Purkinje cell-specific MOF deficient mice exhibits A-T like phenotype as well as a significant loss of Purkinje cells [52]. Similarly, T-cell-specific deletion of MOF leads to significantly reduced T-cell number as well as defective T-cell maturation and differentiation [57]. These in vivo data suggest that, under normal condition, MOF is indispensible for cell viability.

However, MOF can be pro-apoptotic when cells are subject to ionizing radiation induced stress. MOF acetylates p53 at K120, promotes the expression of BAX and PUMA genes, and thereby induces apoptosis following radiation [31].

A recent study reported that MOF and H4K16 acetylation regulates the outcome of autophagy, a cellular process devoted to stress adaptation [58]. As a catabolic process in which the lysosome degrades damaged organelles, proteins, and other cell contents [59,60], autophagy normally protects cells from apoptosis. However in some cases, autophagy can also contribute to cell death [61,62]. MOF and H4K16 acetylation are down-regulated in cell autophagy induced by various stimuli, which subsequently lead to reduced expression of autophagy-related genes [58]. In addition, overexpression of MOF or inhibition of H4K16 deacetylation increased autophagy flux and can lead to cell death. These results support the existence of a feedback regulatory loop in cells to control autophagy flux. MOF, as a major player in this feedback loop, prevents excessive autophagy flux and fine-tunes survival versus death decision upon induction of autophagy.

4.4. Gene Expression

Acetylation of histones in nucleosomes can lead to the opening of the chromatin structure and increase its accessibility for other proteins to alter transcription. MOF has been shown to be involved in regulating gene expression by increasing H4K16 acetylation. On the Drosophila male X chromosome, MOF is found on both the 5′ and 3′ end of genes, suggesting that the MSL complex and H4K16 acetylation may increase gene expression through the transcriptional elongation process [23]. The NSL complex in Drosophila has also been reported to regulate gene expression through H4K16 acetylation on the non- X chromosome. The complex binds to the promoter region of over 4000 genes, of which 70% of these genes are actively transcribed [38]. Additionally, the loss of NSL as well as MOF binding led to a striking reduction of gene expression. Regulation of gene expression by MOF is also found in humans. Similar to its Drosophila othorlog, hMOF accumulates at both gene promoters and coding regions [63,64]. The NSL complex has recently been found to be a co-activator of c-Jun, a proto-oncogene. Knockdown of NSL complex component negatively affected c-Jun downstream gene expression while knockdown of MSL complex had little to no effect [65]. hMOF and H4K16 were also found at active enhancer sites [64,66]. Studies on ESCs further demonstrate that hMOF and H4K16 define a new set of enhancer elements that are independent of H3K27ac and EP300 [66]. Two recent studies investigated two different MOF complexes and gene expression in ESCs [42,43]. The MSL complex regulates expression of genes required for ESCs self-renewal and pluripotency, while the NSL complex regulates housekeeping genes as well as genes associate with growth and proliferation. All these studies suggest that MOF, whether in the MSL or NSL complex, is able to influence gene expression.