A Kirkwood-Buff Approach to Improved Cosolvent Force Fields

助溶剂对溶液中多肽和蛋白质的性质有显著影响。它们通常用于增强蛋白质的稳定性或使其变性以研究蛋白质折叠。助溶剂究竟如何影响多肽和蛋白质尚不清楚。对这些效应的理解将大大改善我们目前对蛋白质变性状态本质的认识。反过来，这将有助于通过提供一个良好定义的初始状态来澄清蛋白质折叠过程，从这个初始状态可以使温度或变性剂的影响合理化。由于蛋白质的许多构象，以及无法以原子水平分辨率确定蛋白质与共溶剂分子之间的相互作用，因此对蛋白质变性状态的实验研究受到限制。原则上，计算机模拟可以提供变性蛋白质及其与共溶剂相互作用的原子水平分辨率。然而，计算机模拟实验只能与描述不同物种之间相互作用的力场一样可靠。本文指出，为了定量再现共溶剂与水混合物的实验性质，需要改进现有的共溶剂力场。将分子动力学模拟与混合物的Kirkwood-Buff理论相结合，为与实验比较提供充分的数据，并保证力场的质量。将研究氯化钠、氯化胍、尿素和2,2,2三氟乙醇的有效对电位和极化模型；它们代表了对肽和蛋白质具有一系列作用的常见助溶剂。改进后的力场将用于全面表征这些溶液的性质，包括原子细节。这是确定它们对蛋白质的不同影响的第一步，如果要对这些系统的计算机模拟结果有任何信心，这是绝对必要的。

Smith, Paul E.MCB-0090483Cosolvents can have dramatic effects on the properties of peptides and proteins in solution. They are often used to enhance the stability of proteins or to denature them in order to study protein folding. Exactly how cosolvents affect peptides and proteins is unknown. An understanding of these effects will greatly improve our current ideas concerning the nature of the denatured state of proteins. In turn, this will help to clarify the protein folding process by providing a well defined initial state from which one can rationalize the effects of temperature or denaturants. Experimental studies of the denatured states of proteins are limited due to the many conformations available to the protein, and an inability of techniques to determine the interactions between proteins and cosolvent molecules with atomic level resolution. In principle, computer simulations can provide atomic level resolution of denatured proteins and their interactions with cosolvents. However, computer simulation experimentsare only as reliable as the force fields which describe the interactions between the different species. In this proposal, it is shown that current cosolvent force fields need to be improvedin order to quantitatively reproduce the experimental properties of cosolvent mixtures with water. A combination of molecular dynamics simulation and the Kirkwood-Buff theory of mixtures will be used to provide sufficient data for comparison with experiment, and to ensure the quality of the force field. Both effective pair potentials and polarizable models will be investigated for sodium chloride, guanidinium chloride, urea, and 2,2,2 trifluoroethanol; which represent common cosolvents displaying a range of effects on peptides and proteins. The improved force fields will then be used to fully characterize, with atomic detail, the properties of these solutions. This is the first step in determining their different effects on proteins, and is absolutely essential if one is to have any confidence in computer simulation results for these systems.