Computational modeling of viral assembly: encapsulation of nucleic acids and env

描述（由申请人提供）：在许多病毒家族中，复制需要数百到数千个蛋白质聚集在病毒核酸（NA）周围，形成一个称为衣壳的蛋白质外壳。此外，许多动物病毒利用蛋白质组装来驱动衣壳从细胞膜出芽。了解控制NAs周围和膜上组装的机制将确定抑制NA包装或出芽的新型抗病毒疗法的目标，并将指导利用病毒作为靶向运输工具的努力。仅从实验中推断出的组装机制是不完整的，因为中间体是短暂的。因此，本项目开发并应用了衣壳蛋白、NAs和脂质的计算模型，揭示了实验无法获得的组装和膜出芽的细节。为了了解病毒NAs的特性是如何促进组装的，我们开发了衣壳蛋白和NAs的模型，这些模型从线性聚电解质（没有碱基配对）开始，然后系统地添加碱基配对产生的NAs的几何和静电特征。对每个模型的预测装配动力学和热力学进行比较，确定碱基配对对装配的贡献。每个模型的预测都通过合作实验室对相应分子（例如合成聚电解质、异源NAs和病毒基因组NAs）周围衣壳组装的实验进行测试。研究了衣壳形成不同二十面体形态以容纳不同大小的NAs的机制。所采用的模拟技术包括布朗动力学和平衡计算。对于一些包膜病毒（如HIV），衣壳组装驱动从细胞膜出芽，而对于其他病毒（如甲病毒），膜蛋白组装驱动预组装衣壳出芽。模拟用于研究这两类组装驱动的出芽过程如何依赖于蛋白质相互作用和膜刚性等特性，以及为什么许多病毒优先从特定的膜微域发芽。预测将与甲病毒出芽的实验相比较。除了确定可以被操纵来阻止或利用病毒组装的因素外，提出的模拟将阐明生物学如何利用膜和丝状支架来组装多大分子复合物。该研究结合了原子模拟和实验的粗粒度模型，以及gpu和分布式计算的最新进展，以模拟相关的时间和长度尺度。提出了一种将马尔可夫状态模型应用于装配反应的新方法。

DESCRIPTION (provided by applicant): In many virus families, replication requires that hundreds to thousands of proteins assemble around the viral nucleic acid (NA) to form a protein shell called a capsid. Furthermore, many animal viruses use protein assembly to drive budding of the capsid from a cell membrane. Understanding the mechanisms that control assembly around NAs and on membranes would identify targets for novel antivirus therapies that inhibit NA packaging or budding, and would guide efforts to exploit viruses as targeted transport vehicles. Assembly mechanisms inferred from experiments alone are incomplete because intermediates are transient. Therefore, this project develops and applies computational models for capsid proteins, NAs, and lipids that reveal details of assembly and membrane budding not accessible to experiments. To understand how the properties of viral NAs facilitate assembly, models are developed for capsid proteins and NAs that begin with a linear polyelectrolyte (without base-pairing) and then systematically add the geometric and electrostatic features of NAs that arise due to base-pairing. Comparison of predicted assembly kinetics and thermodynamics for each model identifies the contributions of base-pairing to assembly. Predictions for each model are tested against experiments performed by collaborating labs on capsid assembly around corresponding molecules (e.g., synthetic polyelectrolytes, heterologous NAs, and viral genomic NAs). The mechanism by which capsids form different icosahedral morphologies to accommodate NAs with different sizes is also studied. Employed simulation techniques include Brownian dynamics and equilibrium calculations. For some enveloped viruses (e.g., HIV) capsid assembly drives budding from a cell membrane, while for others (e.g., alphaviruses) assembly of membrane proteins drives budding of a pre-assembled capsid. Simulations are used to investigate how these two classes of assembly-driven budding processes depend on properties such as protein interactions and membrane rigidity, and why many viruses preferentially bud from particular membrane microdomains. Predictions will be compared to experiments on alphavirus budding. In addition to identifying factors that can be manipulated to prevent or exploit viral assembly, the proposed simulations will elucidate how biology employs membranes and filamentous scaffolds to assemble multi- macromolecular complexes. The research combines coarse-grained models that are informed by atomistic simulations and experiments with recent advances in GPUs and distributed computing to simulate relevant time and length scales. A new method to apply Markov state models to assembly reactions is developed.