Non-Catalytic Role of SETD1A Regulates DNA Repair in Leukemia

Methylation of Histone Lysine 4 is a known histone modification associated with active gene transcription and it is modified by MLL/SET methyltransferases, including MLL1-5 and SETD1A/B. Here we show that a non-catalytic function of SETD1A is required for the survival and self-renewal of MLL-AF9 leukemic cells as well as normal HSC. Using a shRNA library targeting MLL/SET-COMPASS complex subunits, we found SETD1A is an indispensable factor for cell survival and self-renewal in MLL-AF9 leukemic cells in vitro as well as in vivo. In normal hematopoiesis, Setd1a is highly expressed in primitive hematopoietic stem and progenitor cells, and disruption of Setd1a in mice results in pancytopenia. Setd1a disruption severely decreases cell number and reconstitution ability of HSC in a dose-dependent manner. Surprisingly, Setd1a knockdown does not reduce the global H3K4me3 level in leukemia cells. We evaluated the functional domains of SETD1A using deletion mutants, and this study revealed a novel functional domain of SETD1A in leukemic cell survival that is independent of the SETD1A catalytic domain. Here, we name this novel domain FLOS (Functional Location on SETD1A). Domain-focused CRISPR sgRNAs against the FLOS domain show strong toxicity in leukemic cells, but less toxicity in fibroblasts or normal hematopoietic cells in vitro. In contrast, sgRNAs targeting the catalytic domain of SETD1A do not affect the growth of these cells. In addition, RNA-seq analysis demonstrated that Setd1a deficiency dramatically impairs the expression of DNA repair-associated genes, including Fancd2, whichis a critical factor for DNA interstrand cross-link repair. Here we show that Fancd2 is functionally required for MLL-AF9 leukemic cells. Exogenous expression of wild-type or catalytic-domain deletion mutant, but not FLOS domain mutant of SETD1A, restores Fancd2 expression in Setd1a-deficient cells. Consistent with the impaired expression of DNA repair associated genes, Setd1a-deficient leukemic cells show chromosomal instability and p53 accumulation, and p53 disruption rescues the apoptosis initiated by loss of the SETD1A FLOS domain. Through mass spectrometry analysis of SETD1A binding proteins, we found the FLOS domain acts as a binding site for the BuGZ/BUB3 complex. Immunoprecipitation assays indicated that these proteins form a tertiary complex via the SETD1A FLOS domain. The BuGZ/BUB3 complex is known as a regulator for chromosomal segregation and chromosomal stability, but we observed co-localization of these molecules at microtubules in mitotic phase as well as at the nucleus in interphase, suggesting a mitotic phase-independent role of the complex. Both BuGZ and BUB3 were essential for leukemia cell growth as well as for the expression of Fancd2. Taken together, this study indicates that SETD1A has a catalytic-domain-independent function to regulate DNA-repair in leukemia, and imparts this function through a protein complex that includes BuGZ/BUB3. These data suggest that targeting this FLOS domain in SETD1A and the complex more generally may represent a unique therapeutic opportunity for AML and other cancers.