Journal Name: Journal of Proteome Research,
Published: (September 2019)
IF：3.78 （2018）
Unit: University of Basel, Switzerland
Species: human cell lines
Technology: cryo-electron microscopy (Cryo-EM), single-particle electron microscopy
I. Overview:
This paper describes a "visual differences proteomics" (DVP) method, structural changes in proteins and protein complexes of different proteomic samples level. After cell lysis, non-destructive manner to prepare a sample machine, scanning electron microscope image of a single particle. Single protein particle image is extracted, the arrangement, classification, to find the female quantitative analysis, the difference between particle structure of proteins a positive control sample. After the single-cell interference, the method can be used in proteomics structural change was found.
Second, the background:
Functional three-dimensional structure of proteins and protein complexes with a protein closely related. Gold standard proteomics, mass spectrometry, and can not fully reflect the structure of the protein. Frozen conventional electron microscopy (EM) or immunogold analysis of three-dimensional structures of proteins may be reacted, but there are some limitations, such as the flux is not high, the operation, the analysis of complex, much protein can be detected, and the like. This paper presents a method based on the comparison between the structural proteomics single cell single-particle electron microscopy, showing the sample processing and analysis algorithms, including cell lysis, microfluidics machine preparatory steps before an electron microscope, and the processing is not processing visualization cell line comparison process. By comparing different three-dimensional structures of proteins difference sample set you can find new proteins, protein complexes, protein complexes, or rearranged.
Third, the experimental design:
Fourth, the research results:
1. Analysis of frozen microscopy with electron microscopy data DVP simulation algorithm. Author has created three data sets, each data set has 100 microscope pictures, it contains eight kinds of proteins. FIG 2 a set of data shows a black frame partially enlarged three orders of magnitude. DVP algorithms to extract from the pictures of the microscope 300 58300 protein particles, the particles are arranged two-dimensionally to generate a two-dimensional analysis of protein-based particle classification family 44, FIG. 24 shows the class b.
2. simulation analysis algorithm DVP frozen microscopic electron microscopy data, and data on the same FIG. Figure a, from each data set to identify a protein particle classification, the number of statistics, the same protein category (excluding direction) of the particles are grouped into a category. FIG. B, proteins classified particles, and the particles in the same direction of the same protein is classified as a class, but the same protein in different directions particles are classified as a category.
3. Compared to large protein particles (Ursease, GroEL, etc.), a small protein particles (e.g., Hsc70, HscB of HspA and) difficult to be detected correctly, not classified. FIG analog data a set of statistics of the protein particles, b is the computer to retrieve FIG simulated data sets to identify a protein particle classification statistics. And a comparing Figures B, the number of particles of protein concentration distribution data and the retrieved algorithm DVP linear relationship between the number of protein present particle distribution, but there are small differences in protein particles (of HspA and HSCB) value and the detected actual values.
4. 用DVP算法分析批量的HEK293细胞。6份HEK293细胞的蛋白液（来自同一份母液）中加入了不同浓度的尿素酶，负染，用电镜检测。DVP算法从6份样品的120张电镜图片中提取到了96,000个蛋白质粒子（图a是尿素酶含量最高的样品图片示例），其中71,000个蛋白质粒子被准确检测后分成48类（图b展示了其中27类）。随后在PCA图中归纳统计了尿素酶蛋白质粒子的数量（图c中红框圈出了尿素酶粒子）。C1到C6样品中，DVP算法检测到的尿素酶蛋白粒子含量普遍与样品中实际添加的尿素酶浓度呈线性关系，粒子的数量随着浓度的降低而减少。只有C4样品例外，在电镜拍照时就发现尿素酶粒子的数量出乎意料的少。
5. 用DVP算法分析热休克处理的SH-SY5Y单细胞及其阴性对照。DVP算法自动从单细胞样本的200张显微图片中提取到120,000个蛋白质粒子。图a是热休克处理后的单细胞显微图片。处理细胞及其阴性对照样本中有5500蛋白粒子可被准确检测后分为28类，图b展示了其中14类。图c展示了两种样本间粒子丰度差别最大的几类蛋白质。红框圈出了一种环状的热休克蛋白，粒子数目明显在热休克处理后的单细胞（黄色）中更多。
五、文章亮点：
“视觉差异蛋白质组学”（DVP）可以视作常规蛋白质组学研究方法（比如纸质谱）的补充。DVP可以检测单细胞层面或更大量细胞的结构蛋白质组学差异。DVP算法与单粒子电子显微镜分析方法相结合，有效利用了电镜的高分辨率优势。作者承认DVP方法目前在小分子蛋白（＜100kDa）中应用得不太好，无法分辨小分子蛋白粒子的结构特征从而进行有效识别和归类。但随着电镜技术的飞速发展，相信这个难点会很快克服。
阅读人：李思奇
原文链接：https://pubs.acs.org/doi/abs/10.1021/acs.jproteome.9b00447