Atomic Resolution Cryo-EM using Model-Based Constraints

描述（由申请人提供）：大分子组装几乎参与细胞的每一个过程，从运动到信号传导。典型地，这种组合是由多种蛋白质组成的，并被组织成复杂的机器，在它们的组织和组成中经历动态和短暂的变化。因此，了解大分子组装对于增加我们对基本细胞事件的认识和发现疾病预防的目标至关重要。虽然电子冷冻显微镜（cryo-EM）能够对这些大型且通常难以研究的组件进行成像，但目前的限制阻碍了原子分辨率的能力，在原子分辨率上，这些结构可以揭示复杂化学的复杂相互作用和协调。最先进的低温电子显微镜成像最近达到了接近原子的分辨率（3.5-5¿），这样就可以构建整个大分子组装的全部或部分的基本模型。这些“第一方法”模型描述了蛋白质成分的组织和基本折叠，但没有描述大分子组装的必要的原子水平细节特征，而这些特征对于设计有效的预防疾病和改善人类健康的措施至关重要。由于需要新的工具来推动低温电镜中的原子分辨率结构，我们建议开发一个全面的大分子细化和建模协议。我们的最终目标是为社区提供一种新技术，将近原子分辨率的结构推进到原子分辨率，这样就可以直接从密度图中构建完整的原子模型。为了实现这一目标，我们将开发一种新的自动化方法来生成简单的结构模型，从而可以用于改进和提高低温电镜密度图的分辨率。由此，不仅可以构建填充原子模型，还可以对低温电镜成像捕获的结构动力学进行建模。我们相信这项工作将彻底改变冷冻电子显微镜领域，并导致我们对所有生物过程的理解发生重大飞跃。

DESCRIPTION (provided by applicant): Macromolecular assemblies are involved in nearly every process in the cell, from motility to signaling. Typically, such assemblies are composed of a variety of protein and organized into intricate machines that undergo dynamic and transient shifts in their organization and composition. As such, understanding macromolecular assemblies is paramount for increasing our knowledge about basic cellular events and discovering targets for disease prevention. While electron cryo-microscopy (cryo-EM) is capable of imaging these large and often difficult to study assemblies, limitations currently hinder the ability to atomic resolutions, at which point these structures can reveal the intricate sets of interactions and cordination of complex chemistry. State-of-the-art cryo-EM imaging has recently achieved near-atomic resolutions (3.5-5¿), such that basic models for all or parts of an entire macromolecular assembly could be constructed. These "first- approach" models describe the organization and basic fold of the protein constituents, but fall short of describing the necessary atomic-level detail features of macromolecular assemblies that is critical for designing effective measures for disease prevention and improving human health. Motivated by the need for new tools to push towards atomic resolution structures in cryo-EM, we propose developing a comprehensive macromolecular refinement and modeling protocol. Our ultimate goal is to provide the community with a new technique to push near-atomic resolution structures to atomic resolutions such that full atomic models can be built directly from a density map. To accomplish this goal, we will develop a novel automated method for generating simple structural models that, in turn can be used to refine and improve resolution in a cryo-EM density map. From this, it will not only be possible to construct fill atomic models, but also to model the structural dynamics captured by cryo-EM imaging. We believe this work will revolutionize the field of cryo-EM, as well as lead to significant leaps in our understanding of all biological processes.