Electrostatics-Dependent Shape of the Intrinsically-Disordered Protein Sic1

Abstract

For intrinsically disordered proteins (IDPs) containing a high density of charged residues, a Gaussian chain is an inadequate description of the protein's shape. One such IDP is Sic1, a highly positively charged IDP in the budding yeast Saccharomyces Cerevisiae which prevents the cell cycle from entering the S-phase from the G1-phase. Sic1's binding affinity for Cdc4 is highly phosphorylation state dependent, although the corresponding physical basis is not fully understood. NMR data supports the presence of a dynamic complex of Sic1:Cdc4 and a poly-electrostatic model has been proposed by Forman-Kay and coworkers.We studied the Sic1 N-terminal targeting region (1-90) to better understand the role of intrachain electrostatics, and to compare with the structural ensembles calculated from NMR and SAXS data. Sic1 is characterized using time-resolved fluorescence anisotropy (TRFA), which is sensitive to rotational and conformational dynamics. Sic1 exists in a dynamic ensemble of conformations and ensemble-averaged experiments have limitations in identifying and characterizing static and/or dynamic inhomogeneity in the motional dynamics. Therefore, we also performed burst spectroscopy and single-molecule TRFA on freely diffusing Sic1. Shape and flexibility were probed by varying the degree of intrachain repulsion screening through adjusting salt concentrations and described within a polymer physics framework, the polyelectrolyte model. To investigate the relative contributions of “global rotation” and “conformational flexibility”, Sic1 was labelled at three different sites. Sic1 is found to be well modelled as a prolate ellipsoid with internal flexibility.The single-molecule derived TRFA distribution data may be valuable as a constraint incorporated in future conformational ensemble calculations, complementary to SAXS and NMR data. Additionally, these measurements raise questions about the accuracy of highly charged IDP's radii of gyration when calculated from sm-FRET derived end-to-end distances, which often assume a Gaussian chain model for the end-to-end distance probability distribution.