Mutational Analysis of the CBFβ-SMMHC Assembly Competence Domain Identifies a Surface Critical for Multimerization and Inhibition of RUNX1/AML1.

CBFβ complexes with RUNX1/AML1 to form Core Binding Factor. CBFβ-SMMHC is expressed from the inv(16) or t(16;16) chromosome in 8% of AML cases. This fusion protein contains the majority of CBFβ linked to the α-helical rod domain of smooth muscle myosin heavy chain. CBFβ-SMMHC is thought to contribute to leukemogensis by dominantly inhibiting RUNX1/AML1. Inhibition of AML1 depends upon the integrity of a 28 amino acid region near the C-terminus of the SMMHC segment termed the Assembly Competence Domain (ACD). A homologous region is present in multiple myosins and is required for optimal multimerization of their respective rod domains. The ACD is located within a 63 residue "extended" ACD, which includes 12 residues N-terminal and 23 residues C-terminal to the ACD. The extended ACD was noted to have a more neutral charge than other segments of myosin rods. We have now carried out a mutagenic analysis of individual α-helices within or near the extended ACD and have assessed the effect of these mutations on the ability of CBFβ-SMMHC to multimerize in vitro and to inhibit endogenous AML1 activities in the Ba/F3 cell line and in normal murine myeloid progenitors. The 7 amino acids constituting a single turn of the rod domain α-helix are designated abcdefg. The a and d residues form a hydrophobic surface that mediates coiled-coil dimerization, the e and g residues often form salt bridges that stabilize the dimer, and the b, c and f residues are on the outer surface of the helix and are the best candidates for mediating multimerization. We have therefore mutated the bcf residues as a group in ten helices, N3, N1, A, B, C, D, E, F, G, and H. A–D constitutes the core, 28 residue ACD. N3 and N1 are three or one helix N-terminal to helix A. Mutation of N3 or N1 did not affect multimerization in low ionic strength or the ability of CBFβ-SMMHC to inhibit AML1-mediated G1 to S cell cycle progression in Ba/F3 cells. In contrast, mutation of helices A, B, C, D, E, F, G, or H both impaired multimerization in vitro and prevented cell cycle slowing in Ba/F3 cells. Mutants A–E are each located predominantly in the cell nucleus. In transduced murine myeloid progenitors, mutant N3 again behaved similar to intact CBFβ-SMMHC, mutant A also markedly slowed proliferation, mutant B had an intermediate effect, and mutants C, D, or E did not slow proliferation, each in three independent experiments. The increased activities of mutants A or B in the latter setting may reflect the fact that Ba/F3 cells accumulate three times faster than myeloid progenitors and so perhaps are more sensitive to subtle effects. Sin3A, a co-repressor shown to interact with CBFβ-SMMHC, retained the ability to bind mutants A–E. Analysis of mutants N1 and F–H for mSin3A binding, nuclear localization, and their effects on normal progenitor proliferation is in progress. Together, these findings indicate that a surface near the C-terminus of the CBFβ-SMMHC rod domain, encompassing much of the "extended ACD", is required for multimerization and inhibition of AML1. Helices N1 and H demarcate the boundaries of this surface, with helix H been the very last helix of the rod domain. Further characterization of the molecular interactions which allow this surface to mediate SMMHC multimerization may enable the rationale design of drugs for the therapy of AML associated with inv(16).