Molecular Simulation and Modeling Group, National Institutes for Quantum and Radiological Science and Technology, 8-1-7 Umemidai, Kizugawa, Kyoto, 619-0215, Japan.
Institute for Protein Research, Laboratory of Protein Synthesis and Expression, Osaka University, 3-2 Yamadaoka, Suita, Osaka, 565-0871, Japan.
Sci Rep. 2017 Mar 23;7(1):377. doi: 10.1038/s41598-017-00337-y.
In this study, we develop an approach termed "2D hybrid analysis" for building atomic models by image matching from electron microscopy (EM) images of biological molecules. The key advantage is that it is applicable to flexible molecules, which are difficult to analyze by 3DEM approach. In the proposed approach, first, a lot of atomic models with different conformations are built by computer simulation. Then, simulated EM images are built from each atomic model. Finally, they are compared with the experimental EM image. Two kinds of models are used as simulated EM images: the negative stain model and the simple projection model. Although the former is more realistic, the latter is adopted to perform faster computations. The use of the negative stain model enables decomposition of the averaged EM images into multiple projection images, each of which originated from a different conformation or orientation. We apply this approach to the EM images of integrin to obtain the distribution of the conformations, from which the pathway of the conformational change of the protein is deduced.
在这项研究中,我们开发了一种称为“2D 混合分析”的方法,通过生物分子的电子显微镜(EM)图像的图像匹配来构建原子模型。其主要优点是适用于柔性分子,这些分子难以通过 3DEM 方法进行分析。在所提出的方法中,首先通过计算机模拟构建具有不同构象的大量原子模型。然后,从每个原子模型构建模拟 EM 图像。最后,将它们与实验 EM 图像进行比较。两种模型都被用作模拟 EM 图像:负染色模型和简单投影模型。虽然前者更逼真,但后者被采用以进行更快的计算。使用负染色模型可以将平均 EM 图像分解为多个投影图像,每个图像都源自不同的构象或方向。我们将此方法应用于整合素的 EM 图像,以获得构象的分布,从中推导出蛋白质构象变化的途径。