Multiscale Modeling and Enhanced Sampling of Protein-Protein Recognition

该项目由化学系生命过程化学项目、分子和细胞生物科学系分子生物物理项目以及计算和数据驱动科学与工程项目共同资助。密歇根大学的布鲁克斯III开发了探索蛋白质相互作用的计算方法。蛋白质-蛋白质相互作用（PPI）是大多数生物过程的基础。蛋白质-蛋白质相互作用的关键特征的原子描述将在建立对支持所有生命系统的过程的基本理解以及获得操纵和控制这些过程的能力方面产生巨大的好处。实验方法，如核磁共振光谱（NMR）和X射线晶体学可以揭示相互作用的关键蛋白质-蛋白质相互作用的复合物和分离的单体的特征，但难以捕捉动态，灵活的系统。分子动力学（MD）是一个有用的工具，去卷积这些实验测量，并提供了一个详细的，分子相互作用的动态视图。虽然蛮力MD方法在研究单分子系统的动力学方面取得了重大进展，但由于这些系统的尺寸大、时间尺度长和灵活性高，它们在生物学相关PPI方面存在根本性困难。该项目将开发，测试和应用计算方法，使MD研究能够探测病毒衣壳糖蛋白血凝素与假定的人类抗体模拟物的相互作用，以及细菌酸休克分子伴侣HdeA的分子伴侣活性。布鲁克斯博士将为博士后、研究生和一些本科生提供培训，并指导他们发展物理理论和模拟方法，以研究蛋白质-蛋白质相互作用。此外，在这个项目中开发的发现，方法和技术将通过教程和案例研究分发，以集成到结构生物学多尺度建模工具（MMTSB）教程以及密歇根大学教授的课程中。该项目的具体成果将是计算工具和方法，i）从粗粒度（CG）模拟预测环境相关的NMR观测值; ii）开发专门为重现观察到的蛋白质-蛋白质界面而设计的可转移CG相互作用势; iii）开发增强的采样方法，以有效确定分子间相互作用的热力学和动力学性质。这项工作将集中在两个要求苛刻的基准情况下：病毒衣壳糖蛋白血凝素与假定的人类抗体模拟物的相互作用，和细菌酸休克分子伴侣HdeA的分子伴侣活性。前一种系统由于系统尺寸大（总共600个氨基酸）而困难，而后一种系统中的相互作用是高度动态的，对系统pH敏感，并且涉及HdeA和底物蛋白的多个拷贝。

This project is jointly funded by the Chemistry of Life Processes Program in the Division of Chemistry, the Molecular Biophysics Program in the Division of Molecular and Cellular Biosciences and the Computational and Data-Enabled Science and Engineering program.The Chemistry of Life Processes Program in the Chemistry Division is funding Dr. Charles L. Brooks III from the University of Michigan to develop computational methods to explore protein-protein interactions. Protein-protein interactions (PPIs) form the basis for most biological processes. An atomic description of the key features of protein-protein interactions will yield tremendous benefits in both establishing a fundamental understanding of the processes that support all living systems, and gaining the ability to manipulate and control those processes. Experimental approaches such as nuclear magnetic resonance spectroscopy (NMR) and X-ray crystallography can reveal features of interacting complexes and isolated monomers of key protein-protein interactions, but have difficulty capturing dynamic, flexible systems. Molecular dynamics (MD) is a useful tool to deconvolute these experimental measurements, and offers a detailed, dynamic view of molecular interactions. Although brute force MD approaches have made significant progress studying the dynamics of single molecule systems, they have fundamental difficulties with biologically relevant PPIs due to the large sizes, long time scales and high flexibility of these systems. This project will develop, test and apply computational approaches that enable MD studies to probe the interactions of the viral capsid glycoprotein Hemagglutinin with putative human antibody mimics, and the chaperone activity of the bacterial acid-shock chaperone HdeA. Dr. Brooks will provide training for a postdoc, graduate student and a number of undergraduates as well as mentoring them in the development of physical theories and simulation methods to study protein-protein interactions. Moreover, the findings, methods and techniques developed in this project will be distributed through tutorials and case studies to be integrated into the Multiscale Modeling Tools for Structural Biology (MMTSB) tutorials as well as into courses taught at the University of Michigan The specific outcome of the project will be computational tools and methods to i) predict environment-dependent NMR observables from coarse-grained (CG) simulations; ii) undertake the development of transferable CG interaction potentials that are specially designed for reproducing observed protein-protein interfaces; iii) undertake the development of enhanced sampling methods for efficient determination of thermodynamic and kinetic properties of intermolecular interactions. The work will focus on two demanding benchmark cases: the interactions of the viral capsid glycoprotein Hemagglutinin with putative human antibody mimics, and the chaperone activity of the bacterial acid-shock chaperone HdeA. The former system is difficult due to the large system size (600 amino acids total), while the interactions in the latter system are highly dynamic, sensitive to the system pH, and involve multiple copies of HdeA and substrate protein.