Structural modeling of regulation in α-actinin/F-actin interactions

The α-actinins (ACTNs) are a highly conserved family of actin-crosslinking proteins that are critical to various fundamental biological processes in eukaryotes, ranging from cell motility and surface remodeling to muscle contraction. Binding of ACTNs to actin filaments is regulated by several mechanisms: epidermal growth factor (EGF)-induced tyrosine phosphorylation, binding of calcium, limited proteolysis by calpain enzymes, and binding of phosphoinositide moieties. The molecular mechanisms by which these external cues drive the regulation of ACTN function are still not understood, however, largely because there is currently no high-resolution experimental structure that brings together the multiple domains that comprise ACTNs. An understanding of these molecular mechanisms should provide us with insights into how the cell is able to modulate actin cytoskeletal remodeling and give rise to complex cellular phenomena. In this thesis, we investigate how these external cues regulate the actin-binding function of human ACTN4, a non-muscle isoform that is essential to cell motility and has been implicated in cancer invasion and metastasis. First, we develop and validate an atomic model of the multi-domain assembly that makes up the full ACTN4 homodimer, with a novel ternary complex between CH2, neck, and CaM2 comprising the core of this assembly. Next, we show that a novel tandem phosphorylation mechanism in the disordered N-terminal region of ACTN4, where phosphorylation of the functional Y31 requires prior phosphorylation at Y4, is responsible for the regulation of ACTN4 function in the presence of EGF. This tandem mechanism can work in conjunction with m-calpain cleavage of the N-terminal to generate varied actin-binding responses at the front and rear ends of the cell during motility. Using our full structural model, we also show that: (i) Y265 phosphorylation eases ABD opening: (ii) binding of calcium may break the CaM2/neck complex; (iii) CaM2 protects the neck region from m-calpain cleavage; and (iv) binding of phosphoinositides to CH2 allows ACTN4 to crosslink actin filaments at the inner membrane. Finally, we bring together these structural insights to develop a preliminary network-level model that can serve as a computational tool for predicting the actin-binding response of ACTN4 in the presence of multiple external cues