Protein structure determination from low-resolution experimental data

摘要 确定蛋白质的三维结构在生物学中至关重要， 为药物设计提供对生物机制和重要靶点的见解。虽然很高- 分辨率X射线衍射数据为许多人提供了细胞成分的原子视图 有趣的和生物相关的络合物，可能只能获得低分辨率 结构信息。近年来，低温电子显微镜(Cryo-EM)作为一种 获得对这些大分子机器的结构洞察的强大方法。然而，对于 许多感兴趣的复合体，数据往往是：a)分辨率极其有限，缺乏原子 详细信息，以及b)数据质量方面的高度异质性。这两种影响都是由于 这些复合体中的许多具有固有的灵活性。从中提取详细的原子信息 对理解功能、突变的影响或设计药物至关重要的数据是 由于观察次数少和构象空间蛋白大而不可能 可能会被收养。我们提出了一套用于推断高精度原子的计算方法 来自低分辨率数据的模型，揭示了对这些复合体的详细结构洞察。 在之前的授权期内，我们：a)开发了一套工具，用于准确地构建和 将蛋白质模型提炼成低分辨率的实验性低温电子显微镜地图，b)发布了这些工具 作为免费提供的软件，并提供了教程，使低温电子显微镜工作者能够使用 这些方法，以及c)与数十名显微镜专家密切合作，开发出定制 针对他们感兴趣的特定系统的方法。在本提案中，我们进一步介绍了这些方法 在几个不同的方向。我们开发了准确识别配体和 准确模拟低分辨率数据集中的配基构象。我们建立在机器的基础上- 学习引导的蛋白质结构预测，并开发出更快速、更稳健的方法 解释低分辨率低温电磁数据集。我们觉得这些方法应该足够快 适用于在线数据处理流水线中使用。最后，我们开发了方法论来 通过构建一组最好的异质蛋白质模型来解决结构异质性 解释单粒子图像中的变化，提供对构象的结构洞察 单粒子数据的异质性。 拟议研究的总体目标是稳健和可访问的方法来解释-到 原子精确度水平-低分辨率和异质低温EM数据集。加在一起， 这项提议中的三个目标将导致我们推断原子的能力的显著提高 来自这样的数据集的交互。随着显微镜工作者继续研究更复杂的蛋白质 复合体，这些方法将需要揭示原子水平的洞察如何生物医学 重要的蛋白质复合体发挥它们的功能，以及在人类疾病中出了什么问题。

Abstract Determination of a protein’s three-dimensional structure is of critical importance in biology, providing insights to biological mechanisms and important targets for drug design. While high- resolution X-ray diffraction data provides an atomic view of cellular components, for many interesting and biologically relevant complexes, it may only be possible to obtain low-resolution structural information. In recent years, cryo-electron microscopy (cryo-EM) has emerged as a powerful method to gain structural insights into these large molecular machines. However, for many complexes of interest, data is often: a) extremely limited in resolution, lacking atomic details, and b) highly heterogeneous in terms of data quality. Both effects are due to the inherent flexibility of many of these complexes. Extracting detailed atomic information from this data, critical in understanding function, the effects of mutation, or in designing drugs is impossible due to the low number of observations and the large conformational space proteins may adopt. We propose a suite of computational methods for inferring high-accuracy atomic models from this low-resolution data, revealing detailed structural insights into these complexes. In the previous granting period, we: a) developed a set of tools for accurately building and refining protein models into low-resolution experimental cryoEM maps, b) released these tools as freely available software and gave tutorials to enable cryo-electron microscopists to use these methods, and c) closely collaborated with dozens of microscopists to develop custom methodology for their particular systems of interest. In this proposal, we further these methods in several distinct directions. We develop methodology for accurately identifying ligands and accurately modelling ligand conformations in low-resolution datasets. We build off of machine- learning-guided protein structure prediction, and develop methods to more rapidly and robustly interpret low-resolution cryo-EM datasets. We feel these methods should be fast enough to be suitable for use in an on-line data processing pipeline. Finally, we develop methodology to untangle structural heterogeneity by building a heterogeneous set of protein models that best explains variations in single-particle images, providing structural insights into the conformational heterogeneity of single-particle data. The overall goal of the proposed research is robust and accessible methods to interpret – to the level of atomic accuracy – low-resolution and heterogeneous cryo-EM datasets. Combined, the three aims in this proposal will lead to dramatic improvements in our ability to infer atomic interactions from such datasets. As microscopists continue tackling more complicated protein complexes, these methods will be needed to reveal atom-level insights into how biomedically important protein complexes perform their function and what goes wrong in human disease.