Dissecting The Structural Contribution of The Cofilin N‐Terminus to Actin Filament Severing and Phosphorylation by LIMK

Abstract

Cofilin/ADF (actin depolymerizing factor) is a small actin‐binding protein important for cytoskeletal regulation within eukaryotic organisms. Cofilin promotes actin filament severing by disrupting interactions between adjacent actin subunits. Localized regulation of actin dynamics by cofilin phosphorylation is key to cell motility. Cofilin severing activity, and its regulation by phosphorylation, is mediated by an essential actin binding site within its N‐terminus (acetyl‐Ala2‐Ser3‐Gly4‐Val5 in humans). Phosphorylation of Ser3 by LIM domain kinases (LIMKs) is sufficient to inhibit actin binding and severing. Despite the importance of the cofilin N‐terminal peptide sequence, its functional contribution to the mechanism of actin severing has yet to be fully explored. We hypothesize that the cofilin N‐terminus contains evolutionarily conserved structural features that are necessary for actin severing activity and its regulation by LIMK. To identify key features of the cofilin N‐terminal sequence, we constructed a saturation mutagenesis library randomizing the first, third, and fourth residues of human cofilin, while maintaining a phosphorylatable Ser or Thr as the third residue. This combinatorial library of 16,000 (2 x 203) variants was screened as a pool for growth rescue in a strain of S. cerevisiaecontaining a silenced COF1gene. In parallel, we conducted the screen in yeast expressing exogenous human LIMK1 to identify cofilin variants that were inhibited by LIMK activity. In the absence of LIMK, the screen selected cofilin sequences with a Gly as the third residue and beta‐branched or hydrophobic residues for the fourth residue, including the native sequences from animal, plant, and yeast cofilins. In yeast expressing LIMK1, otherwise functional cofilin sequences containing a serine as the phosphoacceptor residue were unable to rescue growth. In contrast, Thr residues at the phosphoacceptor position and bulky hydrophobic residues at the third position imparted a protective effect that allowed cofilin variants to rescue growth despite the presence of LIMK activity. While LIMK1 was unable to phosphorylate a cofilin mutant with a Thr phosphoaccetor in vitro, large hydrophobic residues were tolerated at the third position, suggesting that they render cofilin functional even when phosphorylated. These results expand our understanding of actin regulation by providing structural insights about cofilin and by extension other proteins containing ADF‐homology domains.