The Mixed Lineage Leukemia gene (MLL1, previously termed HRX, ALL-1, MLL) was identified in humans as a gene at the site of common chromosomal translocations in infant leukemia.1 It is now clear that MLL1 represents a family of histone methyltransferases (HMTs) including its closest relative MLL2 and MLL3, MLL4, SETD1A, SETD1B and possibly MLL5.2 Most of these proteins possess histone H3, lysine 4 (H3K4) HMT activity encoded by the SET domain (for Suvar 3–9, Enhancer of Zeste, Trithorax, epigenetic regulators). Studies of the Drosophila homeotic regulator and MLL1 ortholog, Trithorax, established the paradigm for mechanisms by which these proteins function in vivo.3 For example, the fact that Trithorax regulates clustered homeodomain (Hox) genes prompted this line of investigation in the first loss-of-function mouse Mll1 mutants, which were demonstrated to mis-express multiple Hox genes.4, 5, 6 Since the discovery that the SET domain encodes HMT activity,7, 8 great attention has been paid to this particular enzymatic activity in regulating downstream target genes. To determine the function of MLL1 in hematopoietic stem/progenitor cells (HSPCs), we previously analyzed conditional knockout animals and determined that deletion of Mll1 rapidly reduced HSPCs through proliferation and cell fate defects.9 Using this model system, we identified genes that were deregulated upon Mll1 deletion in the hematopoietic stem cell (HSC)-enriched LSK/CD48neg population. Small-scale chromatin immunoprecipitation experiments identified a series of transcriptional regulators that were direct MLL1 target genes including Mecom, Pbx1 and Prdm16, which are known to control HSC self-renewal and quiescence.10 To determine whether maintenance of these target genes required the HMT activity of MLL1, a mouse strain in which homozygous germline deletion of the SET domain of MLL1 (ΔSET) was used for the analysis of the hematopoietic system.6 Surprisingly, we found that H3K4 methylation at target genes such as Hoxa9, Mecom and Prdm16 was absolutely normal in hematopoietic cells of ΔSET-mutant animals, as was the expression level of these genes. We also found that retroviral MLL-AF9 expression generated leukemia in ΔSET HSPCs that was indistinguishable in latency or phenotype from leukemia generated in wild-type HSPCs. Collectively, none of the known biological functions of MLL1 in the hematopoietic system required the HMT activity. To understand how MLL1 maintains the expression of its target genes, a survey of histone modifications that undergo changes upon deletion of Mll1 was performed. We identified only histone H4 lysine 16 acetylation (H4K16Ac) as immediately corresponding to the decrease in transcriptional initiation upon MLL1 removal from the target gene promoter regions. This modification is performed by the histone acetyl transferase MOF (also called Myst1 or KAT8), previously discovered as part of the MLL1 protein complex.11 Furthermore, inhibition of SIRT1, a deacetylase that removes H4K16Ac, was sufficient to prevent the loss of transcription at MLL1 targets. These observations together with our data from the MLL1▵SET knockout mouse provide a clear but unexpected mechanism for the maintenance of target gene expression by MLL1 in the hematopoietic cells.12 These studies highlight differences between endogenous MLL1 and MLL-fusion oncoprotein gene regulation mechanisms, suggesting that differential targeting in leukemia may be feasible.
Read full abstract