Characterisation of the interaction interface of the MOZ/BRPF1 acetyltransferase complex and its role in gene regulation

Monocytic leukaemia zinc-finger protein (MOZ) is a histone lysine acetyltransferase, which plays important roles in haematopoiesis and tissue development. Alterations of the Lysine (K) acetyl-transferase 6A (KAT6A) gene encoding (MOZ) are associated with cancers and in particular, the expression of MOZ-fusion proteins in haematopoietic progenitors has been shown to be causative of acute myeloid leukaemia (AML). More recently, mutations in KAT6A have been associated with developmental syndromes, the clinical features of which are global developmental delay and intellectual disability. MOZ has been shown to associate with a protein complex that includes BRPF1 (bromodomain and PHD finger containing protein 1) and direct association of MOZ and BRPF1 stabilises this complex. Mutations in the BRPF1 gene are also associated with intellectual disability syndromes. Importantly, association with BRPF1 modulates the substrate specificity of MOZ and enhances its acetyltransferase activity for Histone H3 in nucleosomes. However, both MOZ and BRPF1 contain chromatin reader domains that allow them to associate with chromatin by their ability to recognise specific histone modifications. Thus normal MOZ function appears to be regulated by BRPF1, although they may also function independently in different tissues. In animal models, the induction of AML by the MOZ-Transcriptional Intermediary Factors 2 (TIF2) was shown to be dependent on its association with BRPF1. Therefore, in this study, I set out to investigate the molecular interface between human MOZ and BRPF1 proteins, to gain a better insight into how the complex is formed and to understand the functional cooperativity of these proteins. Firstly, to map the minimal sequences required for the interaction of the MOZ/BRPF1 complex, I performed yeast-two-hybrid mapping of protein-protein interactions. This was achieved by generating a series of deletion constructs of both proteins in yeast two hybrid expression vectors. This identified that the sequence BRPF1 59-190 is sufficient to bind the MOZ sequence 676-763, which encodes helix-turn-helix motif located C-terminus of the MOZ MYST acetyltransferase domain. Site-directed mutagenesis identified key residues in the MOZ 676-763 sequence that when altered/disrupted interaction with BRPF1.These results strongly suggest that MOZ 676-763 is a BRPF1 recruitment domain, consistent with a recent structure of the related HBO1/BRPF2 complex. A dual expression vector has been constructed in which the minimal interaction domains were successfully co-expressed in E.coli, which will be taken forward for large scale purification and crystallisation trials. To extend our study to full-length proteins, we performed co-localisation and co-immunoprecipitation studies with epitope-tagged proteins transiently expressed in mammalian HEK293 cells. Interestingly, while FLAG-MOZ localised to the nucleus, expression of HA-BRPF1 alone showed it to be located to the cytoplasmic compartment. In contrast co-expression of both wild type FLAG-MOZ and HA-BRPF1 resulted in their strong co-localisation to the nucleus, and co-immunoprecipitation as a complex. In contrast, a FLAG-MOZ mutant lacking a key region of the BRPF1 binding site, was unable to induce nuclear localisation of HA-BRPF1, suggesting that the helix turn helix region (676-767) is the major interface between these proteins. Surprisingly, both the wild type and mutant FLAG-MOZ proteins continued to co-precipitate together from cell extracts, but we show using a BRPF1 bromodomain-selective inhibitor that this may be in part explained by indirect mechanism involving co-localisation to chromatin sites involving the histone reader domains. To investigate the functional consequences of disrupting the MOZ/BRPF1 complex, rescue experiments were performed using CRISPR-CAS9 edited HEK293 cells developed in our laboratory in which MOZ expression was inactivated. Due to the repressive function of BRPF1, we focussed on genes that we upregulated after MOZ deletion. The RTqPCR data showed YFP-MOZ successfully repressed several of these upregulated genes, whereas a YFP-MOZ mutant deficient in BRPF1 binding was slightly less efficient. These results are consistent with a role for MOZ/BRPF1 complexes in suppression of gene expression, although further studies will be required to investigate this fully. In summary, the minimal sequences involved in the molecular interface between MOZ and BRPF1 have been defined and shown to be contained within a helix-turn helix domain conserved in MYST acetyltransferase domains. This interaction is essential for co-localisation of MOZ and BRPF1 to the nucleus, and for MOZ-mediated repression of several of its target genes, although histone modification reader domains also play a role in assembling these complexes on chromatin. Future work will focus on how specific MYST/BRPF1 complexes are assembled