BIOMOLECULAR KINETICS VIA DYNAMIC IMPORTANCE SAMPLING

我们建议将一种新的计算技术“动态重要性抽样”(DIMS)扩展到多肽和蛋白质的构象转变研究中。我们的长期目标是研究VI型转折的动力学和热力学，最终研究脂肪酸结合蛋白(FABP)的动力学和热力学。这些体系是DIMS技术的理想对象，因为已知了VI型转角中不同态的构象态和FABP中的配体结合/非结合态的三维结构。根据这种已知的起始和终止结构，DIMS能够计算(A)反应速率、(B)转变路径和平衡状态布居。该方法在小型系统中被证明是有效的，因为它收集了发生跃迁事件(即跨越障碍)的非常短的时间尺度上的所有必要信息，而不是关于跃迁之间的长得多的时间尺度--这限制了许多传统技术，如分子动力学。我们的具体目标是计算下列全原子体系的(A)-(C)，用“分子建模工具包”中的琥珀势模拟：(I)隐式溶剂化丁烷丙氨酸二肽中的异构化；(Ii)丁烷和丙氨酸二肽在显性水中的异构化；(Iii)VI型转折的构象转变；以及(Iv)脂肪酸在FABP中的进出。

We propose to extend a new computational technique, "Dynamic Importance Sampling" (DIMS) to the study conformational transitions in peptides and proteins. Our long term goals are to study the kinetics and thermodynamics of a type VI turn and, ultimately, of the fatty acid binding protein (FABP). These systems ideal subjects for the DIMS technique because the three-dimensional structures of the different states conformational states in the type VI turn and ligand bound/unbound states in FABP) are known. From such known start and end structures, DIMS is capable of computing (a) reaction rates, (b) transition paths, and equilibrium state populations. The method has proved efficient in small systems because it gathers all necessary information on the very short timescale in which a transition event (i.e., barrier crossing) occurs, rather than on the much longer timescale between transitions - which limits many traditional techniques like molecular dynamics. Our specific aims are to compute (a) - (c) for the following all-atom systems, modeled by the AMBER potential in the "Molecular Modelling Toolkit": (I) isomerizations in implicitly-solvated butane alanine dipeptide; (II) isomerizations in butane and alanine dipeptide in explicit water; (III) conformational transitions for a type VI turn; and, (IV) fatty acid entrance and exit in FABP.