Results Regarding the Resolving of Membrane-Protein Complexes using Neutron Reflection in Molecular Dynamics

Abstract

Detailed characterization of membrane-associated proteins on fluid lipid bilayers is a challenge despite the availability of many biophysical techniques. Neutron reflectivity (NR) is a method that provides sub-nanometer resolution of such structures in biommetic environments. Analysis of NR data provides density distribution of protein along the membrane normal and coupling this data with complementing structural information from crystallography or NMR can reveal a full atomistic description of the protein-membrane complex. For systems in which this information is not available, because of disordered regions or ensembles of conformational states, additional work is required. Molecular dynamics (MD) simulations can provide the necessary structural information, but long equilibration times and sampling is an issue. Here we show a method to incorporate NR data into MD simulations to steer simulations toward configurations which reproduce experimental results. Biasing potentials are calculated from the one-dimensional densities derived in NR and evolve as the MD trajectory progresses. The bias directs the molecule toward the configurations that NR suggests and grows weaker as these conformations are adopted. Instead of rigidly confining the structure, the system is permitted to fluctuate about these configurations. We demonstrate applications to peptides and larger proteins with disordered regions.