Computational modeling of viral capsid and bacterial microcompartment assembly

在许多病毒家族中，复制需要数百至数千个蛋白质围绕病毒组装。 核酸形成一个蛋白质外壳，称为衣壳。了解衣壳形成的组装途径 了解抗病毒药物如何阻断或改变这些途径将为开发新的 抗病毒策略，并改善现有的。类似地，至少20%的细菌物种具有基于蛋白质的 细胞器称为细菌微区室，这是蛋白质外壳，围绕一组 内切酶由于微区室对于细菌生长和发病机理是必不可少的， 控制它们组装的机制将为开发新的抗生素提供信息， 抑制微区室形成。 仅从实验中推断的组装机制是不完整的，因为中间体是短暂的， 因此不容易观察到。因此，本项目开发并应用衣壳蛋白的计算模型， 核酸（NAs）、推定的抗病毒剂和揭示组装细节的微区室组分 无法进行实验。第一个目标是研究NA如何引导组装途径向特定的方向发展。 衣壳结构目标将包括理解实验中，衣壳蛋白形成不同的 二十面体形态以适应具有不同物理性质（例如，序列长度和 碱基配对相互作用），并对合作者的实验进行模拟路径测试。后者 研究工作将包括开发一种计算工具，通过模拟来预测小角度X射线散射剖面 轨迹以及开发模型，解释新颖的光散射实验，监测组装 单个的衣壳。第二个目标将研究小分子如何干扰蛋白质-蛋白质或蛋白质- NA相互作用重定向组装途径。目标是学习如何设计最佳的抗病毒药物。我们将 开发由原子模拟提供信息的粗粒度计算模型，并预测组装 途径和产物作为推定的抗病毒剂的量和类型的函数。预测将 与我们的实验合作者关于B型肝炎病毒（HBV）组装的大量数据进行比较 蛋白质在潜在的抗病毒剂的存在下。最后，第三个目标将研究细菌如何 微区室聚集在它们的酶货物周围。模拟将确定组装路径， 微区室组装的关键控制参数，同时了解蛋白质壳组装如何 促进和调节细胞内的液-液相分离。 为了能够模拟装配的长度和时间尺度，我们的模拟采用了先进的GPU计算 和一种方法，应用马尔可夫状态建模的组装反应，我们的小组开发。此外，委员会认为， 我们使用由实验和原子模拟告知的粗粒度模型。

In many virus families, replication requires that hundreds to thousands of proteins assemble around the viral nucleic acid to form a protein shell called a capsid. Understanding the assembly pathways for capsid formation and learning how antiviral drugs can block or alter these pathways would provide information to develop new antiviral strategies and improve existing ones. Similarly, at least 20% of bacterial species have protein-based organelles called bacterial microcompartments, which are protein shells that assemble around a group of enzymes. Since microcompartments are essential for bacterial growth and pathogenesis, understanding the mechanisms that control their assembly would provide information for developing novel antibiotics that work by inhibiting microcompartment formation. Assembly mechanisms inferred from experiments alone are incomplete because intermediates are transient and thus not readily observed. Therefore, this project develops and applies computational models for capsid proteins, nucleic acids (NAs), putative antiviral agents, and microcompartment components that reveal details of assembly not accessible to experiments. The first aim will study how NAs guide assembly pathways toward particular capsid structures. Goals will include understanding experiments in which capsid proteins form different icosahedral morphologies to accommodate NAs with different physical properties (e.g. sequence length and base-pairing interactions), and testing simulated pathways against experiments from collaborators. The latter effort will include developing a computational tool to predict small angle x-ray scattering profiles from simulation trajectories as well as developing models that interpret novel light scattering experiments that monitor assembly of individual capsids. The second aim will examine how small molecules that perturb protein-protein or protein- NA interactions redirect assembly pathways. The goal is to learn how to design optimal antiviral agents. We will develop coarse-grained computational models that are informed by atomistic simulations and predict assembly pathways and products as a function of the amount and type of putative antiviral agent. Predictions will be compared against extensive data from our experimental collaborator on assembly of hepatitis B virus (HBV) proteins in the presence of potential antiviral agents. Finally, the third aim will study how bacterial microcompartments assemble around their enzyme cargos. The simulations will identify assembly pathways and critical control parameters for microcompartment assembly, while learning how protein shell assembly can promote and regulate liquid-liquid phase separation within cells. To enable simulating the length and time scales of assembly, our simulations employ advanced GPU computing and an approach to apply Markov state modeling to assembly reactions developed by our group. Furthermore, we use coarse-grained models that are informed by experiments and atomistic simulations.