Centre for Integrative Biology (CBI), Department of Integrated Structural Biology, IGBMC (Institute of Genetics and of Molecular and Cellular Biology), 1 rue Laurent Fries, Illkirch, France; Centre National de la Recherche Scientifique (CNRS) UMR 7104, Illkirch, France; Institut National de la Santé et de la Recherche Médicale (Inserm) U1258, Illkirch, France; Université de Strasbourg, Strasbourg, France.
Max Planck Institute of Biochemistry, Department of Molecular Structural Biology, Am Klopferspitz 18, 82152 Martinsried, Germany.
J Struct Biol. 2018 Jun;202(3):191-199. doi: 10.1016/j.jsb.2018.01.003. Epub 2018 Jan 11.
A current bottleneck in structure determination of macromolecular complexes by cryo electron microscopy (cryo-EM) is the large amount of data needed to obtain high-resolution 3D reconstructions, including through sorting into different conformations and compositions with advanced image processing. Additionally, it may be difficult to visualize small ligands that bind in sub-stoichiometric levels. Volta phase plates (VPP) introduce a phase shift in the contrast transfer and drastically increase the contrast of the recorded low-dose cryo-EM images while preserving high frequency information. Here we present a comparative study to address the behavior of different data sets during image processing and quantify important parameters during structure refinement. The automated data collection was done from the same human ribosome sample either as a conventional defocus range dataset or with a Volta phase plate close to focus (cfVPP) or with a small defocus (dfVPP). The analysis of image processing parameters shows that dfVPP data behave more robustly during cryo-EM structure refinement because particle alignments, Euler angle assignments and 2D & 3D classifications behave more stably and converge faster. In particular, less particle images are required to reach the same resolution in the 3D reconstructions. Finally, we find that defocus range data collection is also applicable to VPP. This study shows that data processing and cryo-EM map interpretation, including atomic model refinement, are facilitated significantly by performing VPP cryo-EM, which will have an important impact on structural biology.
目前，通过冷冻电子显微镜（cryo-EM）确定大分子复合物结构的一个瓶颈是获得高分辨率 3D 重建所需的大量数据，包括通过先进的图像处理将其分类为不同的构象和组成。此外，可能难以可视化以亚化学计量水平结合的小分子配体。Voltap 相板（VPP）在对比度传递中引入相移，并在保持高频信息的同时，大大增加记录的低剂量 cryo-EM 图像的对比度。在这里，我们进行了一项比较研究，以解决不同数据集在图像处理过程中的行为，并量化结构精修过程中的重要参数。自动数据采集是从相同的人核糖体样本中进行的，要么是常规的离焦范围数据集，要么是接近焦点的 Volta 相板（cfVPP），要么是小离焦（dfVPP）。图像处理参数的分析表明，dfVPP 数据在 cryo-EM 结构精修过程中表现出更强的稳健性，因为粒子对准、欧拉角分配和 2D 和 3D 分类表现更稳定，收敛速度更快。特别是，在 3D 重建中达到相同分辨率所需的粒子图像更少。最后，我们发现离焦范围数据采集也适用于 VPP。这项研究表明，通过执行 VPP cryo-EM，数据处理和 cryo-EM 图谱解释，包括原子模型精修，都得到了显著的促进，这将对结构生物学产生重要影响。
