Statistical Convergence of Equilibrium Properties in Simulations of Molecular Solutes Embedded in Lipid Bilayers

详细信息    查看全文

• 作者：Chris Neale ; W. F. Drew Bennett ; D. Peter Tieleman ; R茅gis Pom猫s
• 刊名：Journal of Chemical Theory and Computation
• 出版年：2011
• 出版时间：December 13, 2011
• 年：2011
• 卷：7
• 期：12
• 页码：4175-4188
• 全文大小：1244K
• 年卷期：v.7,no.12(December 13, 2011)
• ISSN：1549-9626

文摘

In recent years, atomistic molecular simulations have become a method of choice for studying the interaction of small molecules, peptides, and proteins with biological membranes. Here, we critically examine the statistical convergence of equilibrium properties in molecular simulations of two amino acid side-chain analogs, leucine and arginine, in the presence of a hydrated phospholipid bilayer. To this end, the convergence of the standard binding free energy for the reversible insertion of the solutes in the bilayer is systematically assessed by evaluating dozens of separate sets of umbrella sampling calculations for a total simulation time exceeding 400 渭s. We identify rare and abrupt transitions in bilayer structure as a function of solute insertion depth. These transitions correspond to the slow reorganization of ionic interactions involving zwitterionic phospholipid headgroups when the solutes penetrate the lipid鈥搘ater interface and when arginine is forced through the bilayer center. These rare events are shown to constitute hidden sampling barriers that limit the rate of convergence of equilibrium properties and result in systematic sampling errors. Our analysis demonstrates that the difficulty of attaining convergence for lipid bilayer-embedded solutes has, in general, been drastically underestimated. This information will assist future studies in improving accuracy by selecting a more appropriate reaction coordinate or by focusing computational resources on those regions of the reaction coordinate that exhibit slow convergence of equilibrium properties.