Computational Tools for Protein Complex Structure Prediction from MS Data

TR&D 5：项目总结。原生质谱指导结构生物学的建议资源 旨在开发先进的MS技术，用于生物大分子的结构表征，如 蛋白质：蛋白质、膜蛋白：脂质和RNA：蛋白质复合物。资源的实验开发 将专注于有效的分离方法，以纯化和提供天然蛋白质的MS，有效的表面 非共价界面裂解的诱导解离方法和共价断裂的UVPD方法 天然蛋白质复合物，以及完整复合物和解离产物（亚复合物）的测量 和共价片段）与离子迁移率MS（用于构象和构象变化，在配体上 结合）和/或高分辨率MS。有关大分子复合物的有价值的结构信息将 得到了然而，目前还没有从MS数据生成结构约束的自动化方法， 并且这些约束通常不足以单独从数据生成高精度复杂结构。 在TR&D 5中，我们提出，结合新的计算方法，SID和 IM，结合来自已建立的方法的限制，例如氢氘交换（HDX）和 共价标记（CL）足以改善大分子复合物结构预测。我们将开发 自动从实验MS数据中提取约束并将其纳入Rosetta的工具 结构预测工具，指导蛋白质复合物结构预测。拟议的研究结构分为 两个主要阶段。 目标1。我们将开发从溶液中预测大分子复合物结构的计算工具 通过MS（H/D交换和共价标记）监测测量。我们将实施量化 将共价标记和HDX暴露约束引入Rosetta对接算法，使得其由以下驱动： 与对接的子单元的暴露图案一致。这一目标使用复合体作为试验平台， 应用于从DBPs 1、2、3、7和8的复合物的HDX和CL数据预测结构 目标2.我们将开发用于从表面预测大分子复杂结构的计算工具- 从离子迁移率实验的诱导解离和碰撞截面。我们将开发新的罗塞塔 对接分数，用于测量复杂模型与SID和IM CCS数据的一致性。TR&D 5与 与其他TR& D集成，因为它旨在扩展开发的实验的适用性 方法通过定制计算方法，允许基于实验数据的结构建模。这 aim将使用SID起始能量、生成的低聚产物和CCS值来测试该程序， 通过使用DBPs 1、2、3、7和10的数据预测结构。

TR&D 5: Project Summary. The proposed Resource for Native Mass Spectrometry Guided Structural Biology aims to develop advanced MS techniques for the structural characterization of biomacromolecules such as protein:protein, membrane protein:lipid, and RNA:protein complexes. Experimental development in the resource will focus on effective separations methods to purify and deliver native proteins to the MS, effective surface induced dissociation methods for non-covalent interface cleavages and UVPD for covalent fragmentation of native protein complexes, and measurement of the intact complexes and dissociation products (subcomplexes and covalent fragments) with ion mobility MS (for conformations and conformational changes e.g., upon ligand binding) and/or high resolution MS. Valuable structural information about macromolecular complexes will be obtained. However, there is currently no automated way of generating structural restraints from the MS data, and those restraints are generally insufficient to generate high accuracy complex structures from the data alone. In TR&D 5, we are proposing that, in combination with novel computational methods, the restraints from SID and IM, combined with restraints from established methods such as hydrogen deuterium exchange (HDX) and covalent labeling (CL), are sufficient for improved macromolecular complex structure prediction. We will develop tools to automatically extract restraints from experimental MS data and incorporate them into the Rosetta structure prediction tools to guide protein complex structure prediction. The proposed research is structured into two main stages. Aim 1. We will develop computational tools for macromolecular complex structure prediction from solution measurements that are monitored by MS (H/D exchange and covalent labeling). We will implement quantitative covalent labeling and HDX exposure constraints into the Rosetta docking algorithm, such that it is driven by agreement with the exposure pattern of the docked subunits. This aim use complexes as testbeds or will be applied to predict structures from HDX and CL data for complexes from DBPs 1, 2, 3, 7 and 8 Aim 2. We will develop computational tools for macromolecular complex structure prediction from the surface- induced dissociation and collision cross sections from ion mobility experiments. We will develop new Rosetta docking scores that measure the agreement of complex models with the SID and IM CCS data. TR&D 5 is tightly integrated with the other TR&Ds because it aims to extend the applicability of the developed experimental methods by tailoring computational methods that allow structural modeling based on the experimental data. This aim will use SID onset energies, oligomeric products generated, and CCS values to test the procedure and to predict structures by using data from DBPs 1, 2, 3, 7 and 10.