Nogal Bartek, Bowman Charles A, Ward Andrew B
From the Department of Integrative Structural and Computational Biology, The Scripps Research Institute, La Jolla, California 92037.
From the Department of Integrative Structural and Computational Biology, The Scripps Research Institute, La Jolla, California 92037
J Biol Chem. 2017 Nov 24;292(47):19400-19410. doi: 10.1074/jbc.M117.808352. Epub 2017 Sep 29.
Several biophysical approaches are available to study protein-protein interactions. Most approaches are conducted in bulk solution, and are therefore limited to an average measurement of the ensemble of molecular interactions. Here, we show how single-particle EM can enrich our understanding of protein-protein interactions at the single-molecule level and potentially capture states that are unobservable with ensemble methods because they are below the limit of detection or not conducted on an appropriate time scale. Using the HIV-1 envelope glycoprotein (Env) and its interaction with receptor CD4-binding site neutralizing antibodies as a model system, we both corroborate ensemble kinetics-derived parameters and demonstrate how time-course EM can further dissect stoichiometric states of complexes that are not readily observable with other methods. Visualization of the kinetics and stoichiometry of Env-antibody complexes demonstrated the applicability of our approach to qualitatively and semi-quantitatively differentiate two highly similar neutralizing antibodies. Furthermore, implementation of machine-learning techniques for sorting class averages of these complexes into discrete subclasses of particles helped reduce human bias. Our data provide proof of concept that single-particle EM can be used to generate a "visual" kinetic profile that should be amenable to studying many other protein-protein interactions, is relatively simple and complementary to well-established biophysical approaches. Moreover, our method provides critical insights into broadly neutralizing antibody recognition of Env, which may inform vaccine immunogen design and immunotherapeutic development.
有几种生物物理方法可用于研究蛋白质-蛋白质相互作用。大多数方法是在体溶液中进行的，因此仅限于对分子相互作用整体的平均测量。在这里，我们展示了单颗粒电子显微镜如何能增进我们在单分子水平上对蛋白质-蛋白质相互作用的理解，并有可能捕捉到用整体方法无法观察到的状态，因为这些状态低于检测限或未在适当的时间尺度上进行观察。以HIV-1包膜糖蛋白（Env）及其与受体CD4结合位点中和抗体的相互作用作为模型系统，我们既证实了源自整体动力学的参数，又展示了时间进程电子显微镜如何能进一步剖析用其他方法不易观察到的复合物的化学计量状态。对Env-抗体复合物的动力学和化学计量的可视化证明了我们的方法在定性和半定量区分两种高度相似的中和抗体方面的适用性。此外，实施机器学习技术将这些复合物的类平均值分类为离散的颗粒亚类有助于减少人为偏差。我们的数据提供了概念证明，即单颗粒电子显微镜可用于生成一个“可视化”的动力学概况，该概况应该适用于研究许多其他蛋白质-蛋白质相互作用，相对简单且是对成熟生物物理方法的补充。此外，我们的方法为广泛中和抗体对Env的识别提供了关键见解，这可能为疫苗免疫原设计和免疫治疗开发提供信息。
