UCSF

In this work, I will discuss the insight we can gain into biophysical processes by looking at atomistic details using computational methods. The focus will be on describing protein-membrane interactions using molecular dynamics simulations and continuum models. I will begin with the ongoing improvements being made to molecular dynamics simulations; specifically, one contribution made to improve a popular pressure-maintenance algorithm known as a barostat. I will then discuss two cases that employ the molecular dynamics combined with continuum modeling approach to fruitful ends. The first shows how a physics-based scale of protein insertion free energy applied to a model system comprising a dual-pass transmembrane helix reporter fused to a misfolded domain yielded membrane insertion predictions that were more accurate than existing bioinformatics approaches. Follow-up molecular dynamics simulations suggested that cases with unexpectedly high insertion resulted from stabilizing hydrogen bonds, emphasizing the need for further investigation of the properties that influence transmembrane helix insertion. The second case attempts to reconcile existing mechanosensation models with emerging data about the mechanosensitive channel of small conductance by comparing the open and closed states with molecular dynamics simulations, then constructing a continuum elasticity model to reproduce the dramatic membrane behavior. A quantitative analysis of this model highlights the importance of tension in membrane free energy and offers suggestions for bridging the apparent gap between the ‘force-from-lipids’ paradigm and the newer ‘membrane deformation’ model.