Search

Many signaling and trafficking events in cells are regulated by the binding of peripheral membrane proteins to specific lipids (e.g. phosphatidylinositol phosphates; PIPs) in cell membranes. This binding frequently occurs via lipid-binding modules such as the Pleckstrin Homology (PH) domain present in many peripheral proteins. It remains challenging to investigate the molecular mechanisms and the energetics by which these lipid-binding modules associate with lipid membranes. Here we will discuss the application of a high-throughput molecular dynamics simulation approach to investigate the association of peripheral membrane proteins with lipid bilayers of differing phospholipid composition. This computational approach was designed to perform large number of coarse-grained molecular dynamics simulation. It therefore enables detailed investigation of the molecular mechanisms by which peripheral membranes proteins associate with model membranes and construction of the energy landscapes for their binding to model membrane surfaces. In particular, we have determined the mechanism of localization of different PH domains to a PIP-containing bilayer and we have examined the extent of conservation/variation of the interactions of different members of the PH domain family with the lipids. The PH/PIP complexes obtained from our simulations are in good agreement with available structural data. By applying a potential of mean force approach we can also calculate the energy of binding of a PH domain to a PIP-containing bilayer. Additionally, we have extended our studies to other peripheral membrane proteins that contain FERM, PH, and C2 lipid-binding domains. Our results suggest novel models for the association and interactions of these lipid-binding domains with membranes.

Interactions of peripheral membrane proteins with membranes, often mediated by lipid-binding modules such as Pleckstrin Homology (PH) domains, are vital for many signal transduction and related processes in cells. PH domains have been shown to associate with phosphatidylinositol phosphate (PIP) lipids, binding to different PIP species with differing affinities and selectivity. However, the exact molecular and energetic details of the PH-membrane interactions remain in many cases elusive. Potential of mean force calculations, using an umbrella sampling approach, allow quantification of binding affinities. The umbrella sampling methodology has previously been largely limited to smaller systems, e.g. the binding of short peptides to proteins. In this study we have used coarse-grained molecular dynamics simulations and an umbrella sampling approach to quantify the binding of PH domains to model membranes containing PIP lipids, considering particularly the PH domain of GRP1. Out results show that the binding affinity of the GRP1-PH domain for PIP3 molecules in a lipid bilayer is greater than for PIP2, in line with previous studies, while other phospholipids did not exhibit favourable binding. Mutation of four key basic residues in the binding site reduced the affinity of GRP1-PH for PIP3. Analysis of the simulations shows that in addition to the canonical orientation, GRP1-PH can bind to the membrane in a secondary orientation via a separate PIP interaction site. This secondary site coincides with an ‘atypical’ site identified in several other PH domains. These results demonstrate how an umbrella sampling approach, combined with coarse-grained simulations, can be applied to peripheral protein-membrane systems in order to increase understanding of the energetics and molecular details of protein-lipid interactions.