This overall project focuses on the study of membranes, proteins and carbohydrates by molecular dynamics computer simulation. Progress is reported under each Aim listed above Aim 1. Understand Model Membranes. This first project in this Aim concerned a simulation-based test of a near-universal assumption made in the experimental determination of spontaneous curvatures of lipids: that the spontaneous curvature of a lipid mixture can be obtained by simply adding the spontaneous curvature of each component in the pure state. Our simulations indicated that this assumption is not correct in some very important cases, including bilayers containing sphingomyelin (which can hydrogen bond to each other as concentration increases) and DPPC/cholesterol (where condensation changes the spontaneous curvature). The implication of this observation is critical for understanding rafts in cell membranes, because a shift in the lipid composition can change the sign of the spontaneous curvature. The simulations indicate that the liquid disordered phase of DPPC/DOPC/cholesterol has negative curvature (consistent with the additive assumption) while liquid ordered phase has positive curvature (inconsistent with the additive assumption). Paper 3. To change conformation, a protein must deform the surrounding bilayer. This deformation is modulated by curvature and compression stress. The second project in this Aim involved the development of a new (3-dimensional) continuum elastic model (CEM) to describe the deformation. The 3D-CEM is a substantial improvement over the traditional 2D-CEM (where only the surface is considered) because it does not require arbitrary assumptions regarding the slope of the bilayer surface in contact with the peptide or protein. The model was applied to estimate the enrichment of hydrophobically matched lipids near the gramicidin A (gA) channel in a lipid mixture. The results agree with single-channel experiments and extended molecular dynamics simulations, and pave the way for quantitative modeling of the interactions of membrane proteins with the surrounding membrane. Papers 5 and 6.
Aim 2. Develop Simulation Methodology. The first project of his Aim involved a systematic test of the periodic Saffman-Delbruck (PSD) model described in the 2016 Annual Report. Diffusion constants from coarse grained and all-atom simulations with different numbers of lipids were shown to agree well with the predictions of the PSD model, and a Bayesian-based method to extrapolate simulated diffusion constants to infinite system size was used to compare these diffusion constants with experiment. Such comparisons with experiment will be critically important when validating the next generation of force fields for membrane simulation. Paper 7. A second methodological project used Bayesian-based methodology for estimating diffusion tensors of permeants in membranes. The resulting diffusivity profiles characterize the membrane transport dynamics as a function of the position across the membrane, discriminating between diffusion normal and parallel to the membrane. The method was applied to simulations of O2 in a pure POPC bilayer and one with the lipid composition of the mitochondrial membrane. The results indicate that oxygen can diffuse considerable distances in the bilayer midplane before exiting, an observation that is relevant to the function of cytochrome C oxidase. Permeation of other solutes in different bilayers will provide a more quantitative test of the Overton Model. Paper 8.
Aim 3. Simulate Complex Membranes The hypothesis that the antimicrobial peptide (AMP) piscidin 1 forms pores to disrupt membranes was tested using multimicrosecond simulations on the Anton supercomputer. Starting from the well-characterized known pore-former alamethicin as a control, the simulations showed insertion from the membrane surface and formation of stable barrel-stave pores, in agreement with experiment. In contrast, different sized pores of piscidin 1 were not stable, and the peptide did not insert when placed on the surface. The simulations provide strong support for the hypothesis that the disruptive mechanism of piscidin 1, and likely many other AMPs does not involve formation of stable pores. Rather, it is likely short-lived defects allow leakage and facilitates peptide translocation. Papers 1 and 2. A second project in this Aim involved the conformation B38, an apolipoprotein B mimetic peptide, on a trilayer modeling a low density lipoprotein (LDL) particle. The insertion of the peptide was much shallower that the AMP discussed above, highlighting the variability of binding motifs of superficially similar amphipathic peptides. Paper 4.
Showing the most recent 10 out of 52 publications
