Molecular dynamics at the atomistic level using the OPLS-AA force field MEMPEPT

Tampereen teknillinen yliopisto / Ilpo Vattulainen

Functions of many membrane proteins are known to be dependent on membrane composition. Interactions of proteins with lipids have therefore been investigated intensively by different computational and experimental methods, and the effects of some related phenomena such as hydrophobic mismatch are partly understood. However, in a more general context the implications of lipid-protein interactions on membrane organization and functions are only weakly known. For example, while cholesterol interacts favorably with a number of membrane proteins and peptides, the effects of cholesterol on nano-scale membrane domain formation around the proteins is not understood at all, and the hydrophobic mismatch is one of the phenomena related to this process. Meanwhile, some hypotheses have been proposed on the basis of recent unpublished experimental data (personal communication by the group of Prof. Kai Simons/MPI Dresden): it has been suggested that certain membrane proteins partition specifically into cholesterol-enriched membrane compartments, but the exact mechanisms are not known. Understanding of these phenomena is very important for better description of the formation of membrane rafts ? one of the main research areas in modern biophysics. Such membrane regions play a very significant role in many physiological processes like signal transduction, membrane fusion, neuronal maturation, lipid sorting and protein trafficking. The objective of this computational project is to focus on these lipid-peptide and cholesterol-peptide interactions and couple the simulations to the recent and on-going experiments in the Simons laboratory. The project will employ atomistic molecular dynamics simulations of many systems consisting of lipids, cholesterol and a model peptide with the sequence KKWWLLLLLLLLALLLLLLLLWWKK. The systems which have been simulated so far are: 1. C16:1 without cholesterol, 1 peptide 2. C24:1 without cholesterol, 1 peptide 3. C16:1 with 10 mol% cholesterol, 1 peptide 4. C24:1 with 10 mol% cholesterol, 1 peptide 5. C16:1 with 25 mol% cholesterol, 1 peptide 6. C24:1 with 25 mol% cholesterol, 1 peptide 7. C16:1 without cholesterol, 4 peptides 8. C24:1 without cholesterol, 4 peptides 9. C16:1 with 25 mol% cholesterol, 4 peptides 10. C24:1 with 25 mol% cholesterol, 4 peptides Here, C16:1 corresponds to a glycerophospholipid (PC) with acyl chains that have 16 carbons and one double bond; the other cases similarly. All systems were simulated for 200 ns. Our preliminary results show that cholesterol distributes differently in the two different lipid matrixes. Importantly, these results will be compared with experimental data (of similar systems that were simulated) provided by the group of Prof. Kai Simons at the Max Planck Institute (Germany) - our main partner in this project. However, as the time scale of lipid redistribution is long due to slow diffusion, we decided to extend the simulations to see full separation of the system components, facilitating comparison to experiments where the time scales are several orders of magnitude longer. In practice, we want to extend the simulations to at least 500 ns. Therefore, we need approximately 200 000 billing units to broaden our view of the research topic and to compare it more accurately with the experimental results. Key questions are: 1. How does the hydrophobic mismatch affect distribution of membrane components? 2. What are the mechanisms of the peptide-induced lipid redistribution with cholesterol? 3. How do the conformations of lipids in the vicinity of the transmembrane helix compare to bulk lipids? 4. How does the difference in the mismatch within the different PCs affect lipid conformations? 5. How do the conformations of the annular lipids compare when cholesterol is present? 6. Can one deduce a (relative) energetic parameter for the mismatch? 7. How rigid is the peptide backbone and the side chains? 8. Do the conformations of the side chains change in the presence of cholesterol?