(466g) Prediction Of Configuration Of Membrane-Associated Peptides Through Simulated Self Assembly
Membrane associated peptides are important model systems to understand the larger and more complex membrane proteins. Antimicrobial peptides such as magainin, and designed model peptides such as WALP and KALP are good examples of membrane associated peptides. The nature of association of these peptides with lipid bilayers depends on a range of factors such as the sequence of the peptide itself, peptide/lipid ratio and length of the peptide relative to the bilayer width. At low concentrations, the two dominant states of these peptides are the surface orientation and the transmembrane orientation. At higher peptide/lipid ratios, more complex structures such as peptide oligomers and peptide-lined pores can be expected. Due to the relative difficulty in experimental characterization, little is known about the structure and lipid association of membrane associated peptides. Solid state NMR can typically determine the average orientation of these peptides relative to the bilayer normal, but more detailed conformational information such as oligomerization state are difficult to estimate. Molecular dynamics simulations of peptides with lipid bilayers can in principle predict the partitioning behavior of membrane associated peptides by adequately sampling all relevant states. However, due to extremely large energy barriers, a single simulation is typically limited to sampling either the transmembrane or the surface state. Self-assembly is an elegant method to estimate ensembles of equilibrium conformations. We have used atomic-level and coarse-grained self-assembly simulations of amphipathic and hydrophobic peptides in random lipid/water mixtures to study the partitioning behavior of these peptides in lipid bilayers. We find that it is possible to estimate the partitioning of hydrophobic peptides using atomic level simulations. However, amphipathic peptides, with charged residues typically lead to conformations that are kinetically trapped, showing the limitations of atomic-level simulations. A recently developed coarse-grained model for lipids and proteins is able to overcome the limitations of the atomic level simulations. Thousands of self-assembly simulations of amphipathic and hydrophobic peptides with lipid/water mixtures at different peptide/lipid ratios yield a clear picture of the partitioning behavior of the peptides in lipid bilayers. At low concentrations, hydrophobic peptides typically partition in a transmembrane orientation and at higher concentrations, they form peptide bundles. Amphipathic peptides typically prefer a surface orientation at low peptide/lipid ratios, but form complex structures such as transmembrane pores at higher concentrations of peptides. We have shown that self-assembly can be used as an accurate tool to predict the partitioning behavior of peptides in lipid bilayer.