Crystal structure of Formin Homology 2 domain of human DAAM1

Center for Biological Resources and Informatics, Tokyo Institute of Technology* Department of Biological Information, Tokyo Institute of Technology, Japan** Deparment of Cardiovascular Medicine, Kyoto University, Japan***
○Shuya Fukai* Masami Yamashita** Tomohito Higashi*** Yusuke Sato** Ryutaro Shirakawa*** Toru Kita*** Hisanori Horiuchi*** Osamu Nureki**

Formin proteins nucleate and elongate the unbranched actin filament under the regulation by the GTP-dependent interaction with the Rho/Rac GTPases. They are conserved beyond species and characterized by the formin homology domains, FH1 and FH2. FH1 binds with profilin, and enhances the elongation rate. FH2 forms a ring-shaped dimer in a head-to-tail manner, and processively moves towards the barbed end of the actin filamet to catalyze the actin polymerization. Dishevelled-associated activator of morphogenesis (DAAM) is a Rho-regulated formin implicated in the actin organization during neuronal development. In this study, we determined the crystal structure of the human DAAM1 FH2 dimer at 2.8 Å resolution, and performed a structure-based mutagenic analysis. This is the first FH2 dimer structure that is represented without an aid of the crystallographic symmetry. Two DAAM1 FH2 molecules were associated in a head-to-tail manner, and formed a rectangular ring in contrast to the parallelogram-shaped ring of yeast Bni1p FH2. Furthermore, the orientation of the DAAM1 FH2 ring was basically different from that of the Bni1p FH2 ring. Docking analysis using the structure of the yeast Bni1p FH2 in the complex with the filament-like actin showed that the present DAAM1 FH2 dimeric ring adopts a contracted conformation, which could cap the barbed end of the actin filament. These structural analyses suggested the functional importance of the length of the linker region, which connects the N-terminal “ lasso“ region and the C-terminal “core” region. This was shown by the polymerization assays using a series of FH2 mutants with various linker lengths. We propose the “expanded and contracted“ stair-stepping model for the actin polymerization by FH2.