Traffic between cells occurs every single second inside our bodies. By so many simple but sophisticated ways, cells communicate with each other in an instance. Looking closer, lipid bilayers or cell membranes surround eukaryotic cells and serve as the barriers when anything tries to get in or out of the cells. Although these cell membranes are usually intrinsically flat, they are very flexible in shapes. At the surface of these cell membranes, we could find massive interactions between macro-biomolecules, such as protein-lipid and protein-protein interactions. Peripheral membrane proteins refer to proteins which associate with cell membranes but do not embed inside the cell membrane. I have spent the last four years studying both interactions between peripheral membrane protein themselves and interactions between the protein and different lipids using computer simulations, mainly molecular dynamics simulations.
One of the proteins that I have studied extensively is the BAR superfamily, especially the BAR-PH domain found in ACAP1 protein. BAR proteins have been reported to be recruited for inducing cell membrane curvatures during vesicle budding processes. Two molecular mechanisms have already been discovered for BAR proteins to induce local cell membrane curvatures, namely scaffolding and wedging. The BAR-PH domain caught people’s attention due to its unconventional behavior when interacting with the cell membrane. Given the symmetrical chemical component, the BAR-PH domain binds to the cell membrane asymmetrically.Such binding conformation has never been observed for other members of the BARsuperfamily. With the valuable experimental results on single protein crystal structures and multiple protein assembly structures from electron microscope provided by the collaborators, I have conducted studies on protein-protein interactions inside BAR-PH lattices and predicted important interacting residues for lattice formation. Hinted by the key interactions, I pictured the self-assembly process of the BAR-PH domains. I have also explained the unconventional binding conformation by studying the internal dynamics of the protein. Furthermore, I have carried out free energy calculations, also using molecular dynamics simulations, to quantify the interactions between the BAR-PH domain and different lipids to account for its lipid specificities.In this thesis, I will show all my studies on BAR proteins in details.
Computer simulations of biomolecules have been developed for more than 50 years. Specifically, molecular dynamics simulations of biomolecular structures have been termed “computational microscope” due to its capability of studying functionally important events in atomic details. Nevertheless, in this thesis, not only the power of this “microscope” but also the major precautions and obstacles when one deals with microseconds of molecular dynamics simulation trajectories are shown with examples.
Last but not least, state-of-art coarse-graining schemes of biomolecules has drawn much attention in recent decades and made multiscale first-principle-based classical simulation studies possible. In addition to elongating simulation temporal and spatial scales, coarse-graining of biomolecules also reveals essential, thus important, slow and global dynamics of proteins, which usually relates to their physiological functions.
This thesis covers a wide range of time, space, and resolution of the latest development of molecular dynamics simulations of biomolecules. It also summarizes my theoretical predictions of functional states of some peripheral proteins from their atomic details in molecular dynamics simulations, confirmed by experimental observations.