New perspectives for the regulation of acetyltransferase MOF

Abstract

In higher eukaryotes, histone acetyltransferase MOF (male absent on the first) is the major enzyme that acetylates histone H4 lysine 16, a prevalent mark associated with chromatin decondensation. Recent studies show that MOF resides in two different but evolutionarily conserved complexes, MSL and MOF-MSL1v1. Although these two MOF complexes have indistinguishable activity on histone H4 K16, they differ dramatically in acetylating non-histone substrate p53. The regulation of MOF activity in these complexes remains elusive. Given the evolution conservation of MOF and the importance of H4 K16 acetylation in maintaining higher order chromatin structures, understanding the function and regulation of MOF bears great significance. Here, we discussed the key differences in two MOF complexes that may shed light on the regulation of their distinct acetyltransferase activities. We also discussed coordinated functions of two MOF complexes with different histone methyltransferase complexes in transcription regulation.

MOF, also called MYST1 or KAT8, is a MYST family (MOZ, YBF2, SAS2 and TIP60) histone acetyltransferase (HAT), featuring a highly conserved catalytic MYST domain.^1-3 Unlike most HATs that are often promiscuous for lysine acetylation on histone tails, MOF has strict substrate specificity for nucleosomal H4 K16 when in complex with several evolutionarily conserved proteins.^3-5 Deletion of MOF in Drosophila and mammals led to global reduction of H4 K16 acetylation, suggesting that MOF is a major HAT for this site.^4-7 It was shown that acetylation of H4 K16 by MOF perturbs inter-nucleosomal contacts between the “basic patch” of H4 N-terminal tail (K14 to K20)^8,9 and the “acid patch” on H2A/H2B tails.^10,11 This leads to reduced compaction of 30 nm chromatin fibers in vitro and chromatin decondensation in vivo.^3,11-15 Thus, MOF may regulate many chromatin-based activity such as transcription and DNA damage repair through H4 K16 acetylation.

Lessons from Drosophila Dosage Compensation by the MOF-MSL Complex

The MOF ortholog in Drosophila was originally described as an essential component of the X chromosome dosage compensation complex (DCC),⁷ causing a two-fold increase in the expression of X-linked genes in male flies.¹⁶ In DCC, MOF associates with several highly conserved proteins referred to as Male Specific Lethal (i.e., MSL1, 2 and 3), and other components including JIL1, MLE and two roX RNAs. Among these components, expression of MSL2 and two roX RNAs, which are critical for X-chromosome-wide targeting of DCC, is limited to male flies.^17-19

The function of DCC in Drosophila can be conceptually divided into two steps: First, the recruitment and spreading of DCC along the male X chromosomes. This step is initiated by binding of MSL1 and MSL2 to high affinity sites (∼150 entry sites) on the male X-chromosomes.²⁰ Binding of MSL1 and MSL2 activates expression of roX RNAs,^21,22 which are then incorporated into DCC. roX RNAs in coordination with RNA helicase MLE²³ and other components facilitate spreading of DCC to whole X-chromosome.²⁴ This function of DCC is not conserved in mammals, evident by the absence of roX RNAs and the presence of a different dosage compensation mechanism in mammals. The second step of DCC function involves transcription upregulation of X-linked genes mediated by MOF-dependent H4 K16 acetylation. Since DCC and H4 K16 acetylation are highly enriched in the coding regions of transcriptionally active X-linked genes,^20,25-28 it is likely that DCC regulates transcription by promoting transcription elongation. This was further supported by recent studies that the chromo-domain of MSL3 binds directly to methylated H3 K36, a histone mark that are important for transcription elongation.^29-31 This function of DCC is likely to be conserved, supported by the enrichment of H4 K16 acetylation at the coding regions of transcriptionally active genes in mammals.

Interestingly, it was recently shown that although DCC binding is specific for male X-chromosome, MOF was able to bind autosomes in both male and female Drosophila cells.^25,26 This binding, which is likely attributes to MSL independent mechanism, suggests that MOF may function beyond dosage compensation in Drosophila.

The Regulation of MOF Acetyltransferase Activity in Mammals

MOF and most MOF interacting proteins are highly conserved in higher eukaryotes. Biochemical purifications showed that MOF resided in two distinct complexes, which are separated into distinct size fractions on a gel filtration column with apparent mass of ∼1.2 mDa and ∼600 kDa respectively.⁵ These two MOF complexes are equally abundant in mammalian cells.^4,32 One MOF complex, referred to as the MSL complex, was purified through TAP-tagging of the mammalian MSL3 ortholog. Like Drosophila DCC, this complex contains core components MSL1, MSL2 and MSL3. Furthermore, specific interactions between MOF and the coiled-coil protein MSL1 as well as the chromodomain-containing protein MSL3 are important for nucleosomal H4 K16 acetylation (unpublished observation). However, no mammalian homologues for MLE or non-coding RNAs were identified in the purification. The second MOF complex was initially co-purified with WDR5, a key component of MLL family H3 K4 methyltransferase complexes.³³ This MOF complex, referred to as MOF-MSL1v1, does not include MSL1, MSL2 or MSL3 proteins. Instead, it contains MOF, MSL1v1 (previously labeled as LOC284058 or KIAA1267), PHF20 (also called C20orf104), MCRS1/2 and components of the MLL complex (e.g., WDR5, HCF1).^4,33 Several proteins in this complex (e.g., WDR5, PHF20) are capable of binding to nucleosomes, a feature analogous to MSL3. Consistent with distinct complex composition, the MSL and the MOF-MSL1v1 complexes have very different substrate specificity. MOF-MSL1v1 extends the substrate spectrum of MOF to non-histone proteins such as tumor suppressor p53 (K120)⁵ and also demonstrates weak activities on H4 K5 and K8 as well.³⁴

A detailed comparison of two mammalian MOF complexes provides mechanistic insights for the regulation of MOF acetyltransferase activity. By biochemical reconstitution, we found that MSL1v1 alone is necessary and sufficient to support MOF in acetylating nucleosomal H4 K16, a notable difference from the MSL complex.⁶ MSL1v1 is a paralog of MSL1. Genes encoding MSL1 and MSL1v1 were generated by duplication of a certain ancestor gene that emerged together with coelomate animals.³⁵ The primary sequence identity of canonical MSL1v1 with MSL1 is about 27%.³⁵ Both proteins contain a coiled-coil domain on the N-terminus and a highly conserved PEHE domain on the C-terminus (Fig. 1). Of note, MSL3-interacting domain in MSL1 is completely absent in MSL1v1. Instead, MSL1v1 contains a much larger N-terminal domain preceding PEHE. Comparison of MSL1 and MSL1v1 as well as their respective complexes pointed out several key differences: (1) Although both MSL1 and MSL1v1 contain PEHE domain, only MSL1v1 PEHE is able to directly interact with MOF. In contrast, mammalian MSL1 is not able to interact with MOF in the absence of MSL3 (unpublished observation). (2) Although both MSL1 and MSL1v1 contain the coiled-coil domain at the N-terminus of the protein, only MSL1 is able to interact with MSL2. No MSL2 or MSL2 homologues were identified in the MOF-MSL1v1 complex. (3) It was shown that non-histone acetylation is a unique function for the MOF-MSL1v1 complex. It requires a region preceding PEHE domain in MSL1v1,⁵ which is not conserved in MSL1 (Fig. 1). Detailed structure-function analyses and thorough comparative studies for MSL1v1 and MSL1 are likely to provide insights for the regulation of two MOF complexes in future.

Figure 1.

Schematics of mammalian MOF, MSL1 and MSL1v1. Secondary structures for these three proteins were highlighted. the known protein interaction partners for each domain were labeled on top. See text for references.

In the MSL complex, the specific and efficient acetylation of nucleosomal H4 by MOF depends on its interaction with both MSL1 and MSL3. It was proposed that MSL1 functions to mediate the interaction between MOF and MSL3, which enhances the activity of MOF by interacting with nucleosomes.⁴ In the MOF-MSL1v1 complex, there are two proteins, WDR5 and PHF20, which were reported to have high affinity for nucleosomes. However, neither protein was required for the nucleosomal activity by MOF-MSL1v1.⁵ The ability of MOF-MSL1v1 to acetylate nucleosomal H4 in the absence of MSL3 or MSL3 homologues suggests that MSL3 may not function to target MOF activity to nucleosomes. Instead, the relatively relaxed substrate specificity of MOF-MSL1v1,^5,34 suggests that MSL3 may function to restrict MOF activity to nucleosomal H4 K16.

Coordinated Functions of MOF and Histone Methylation

As previously stated, MOF-MSL1v1 was co-purified with WDR5, a key component of the MLL family H3 K4 methyltransferases,³² and co-fractionated with MLL on a gel filtration column.⁵ Importantly, some components of the MOF-MSL1v1 complex, but not those unique for the MSL complex, show characteristic homeotic expression patterns during mouse embryonic development. For instance, MOF,³⁶ MCRS1/2 and PHF20,³⁷ are expressed in limb bud, embryonic nervous system and somites. The distribution of their transcripts bears striking similarities to those of MLL and polycomb group proteins (PcGs) Bmi1 in mouse E10 embryos.^38-40 These results imply important functional interactions between MLL and MOF in vivo. In addition to co-localization of two enzymatic machineries, the functional coordination of H4 K16 acetylation and H3 K4 methylation was recently demonstrated at the promoters of transcriptionally active genes by genome-wide ChIP-sequencing studies.⁴¹ It was reported that H3 K4 methylation and H4 K16 acetylation are closely correlated with each other. Importantly, 73% of the promoters associated with H3 K4me were quickly acetylated at H4 K16 upon HDAC inhibitor treatment in contrast to only 0.5% of the promoters not associated with H3 K4 methylation.⁴¹ The rapid acetylation of H4 K16 at H3 K4 methylated promoters is likely mediated by WDR5, which was shown to bind di-methylated H3 K4 in vitro.⁴² Dual modified nucleosomes at gene promoters, in turn, may serve as multi-valent docking sites for the recruitment of downstream effector protein or protein complexes.⁴³ The function of MOF-MSL1v1 at gene promoters is further supported by the specific interaction of the MOF-MSL1v1 complex with transcription factors (e.g., p53).⁶

In addition to promoter enrichment of MOF and H4 K16 acetylation, MOF was also found in the coding region of actively transcribed genes. This binding, however, is not co-localized with peaks of H3 K4 methylation. Instead, it associates with 3′ end of the genes and probably a result of the binding of the MSL complex.²⁹ Although genome-wide mapping of the MSL binding sites have not been reported in mammals, preferred binding of MSL1 and MSL3 at sequences downstream of transcription start site were demonstrated at several specific gene loci.⁶ These results suggest the involvement of the MSL complex in transcription elongation, similar to DCC in Drosophila. Distinct functions of the MOF-MSL1v1 and the MSL complexes in transcription regulation are likely a general phenomenon, given the ubiquitous expression patterns of these two MOF complexes. Beside H3 K4 and K36 methylation, the MOF-MSL1v1 complex may also be functionally linked to H4 K20 methylation through PHF20, one of the complex components. PHF20 is a scaffold protein of 1,012 amino acids and contains three domains: a PHD finger, a tudor domain and a MBT domain.⁴⁴ The tudor domain of PHF20 is highly homologous to that of 53BP1 and interacts strongly with H4K20me2,^45,46 a mark in the vicinity of H4 K16. Both H4 K20me2 and 53BP1 were implicated in DNA damage repair.⁴⁵ It would be interesting to know whether PHF20 links MOF to H4 K20me2 mediated DNA repair process in future.

Conclusions and Perspectives

Although MOF was cloned and characterized in the dosage compensation process in male Drosophila more than a decade ago, recent purification of a distinct MOF-MSL1v1 complex showed yet how little we knew about this important histone acetyltransferase. Given the novel MOF-MSL1v1 complex and its expanded list of substrates, it is likely that MOF functions beyond regulating H4 K16 acetylation and chromatin structures in mammals. Thus, it is important to develop methodologies for systematic identification of novel substrates for MOF in future. It is also important to establish conditional knockout mouse models for Mof, which will overcome embryonic lethality caused by conventional Mof knockout.^47,48 Given that MOF is essential for normal development and is important for coordinated functions with Mll, a proto-oncogene for acute leukemia, understanding the in vivo function of MOF bears great significance.