Role of noncovalent interactions in protein peripheral membrane binding. Computational perspectives

Abstract

Noncovalent forces are important driving forces in nature particularly in biology, and they dictate many biological processes including the binding of peripheral protein to the cell membrane. The widely acknowledged models describe this process as electrostatics driven membrane adsorption followed by short-range protein-lipid interactions i.e. hydrogen bonds, hydrophobic interactions. Some of the key elements in such models are: clusters of basic residues are essential for electrostatic adsorption, and basic residues contribute equally to the membrane binding. Nevertheless, none of these models account for the role of cation-π interactions in membrane binding. With selected protein candidates, we further explore these models and work towards a generalized description of protein peripheral binding to membranes in terms of noncovalent forces. Our investigation highlights the limitations of these existing descriptions. We demonstrate that the requirement of having a cluster of basic residues is not essential. Further, we show that the contributions of basic residues are distance dependent. In other words, their localization in the membrane-water interface determines their strength and hence is not equal. We also establish the role of tyrosine-choline cation- π interactions in membrane binding of peripheral proteins. We explore in detail the nature of tyrosine-choline mediated cation-π interactions using high-level quantum mechanical calculations. Later, this information is used to improve the description of cation-π interactions in molecular simulation models. These improvements of force field parameters are further tested using molecular dynamics simulations. Finally, we used this information to build an interaction diagram that can be used to better describe the binding of peripheral proteins to the cell membrane. Future testing and the generalization of this diagram will further establish this as a common model.