VISUALIZATION OF SUBCELLULAR DYNAMICS IN MULTICELLULAR ORGANISMS

摘要 当代荧光显微镜连接了我们对分子事件的理解 生物化学和结构生物学及其在活细胞中的活动。晶格光片显微镜 (LLSM)使我们能够跟踪内吞囊泡组装或脂质等现象 在整个细胞中以高分辨率、在空间和时间上以及以几乎 单分子敏感性。点阵光片显微镜近一年来的发展 自适应光学(AO-LLSM)已经克服了到目前为止限制了大多数 对培养中的单个细胞进行研究，使我们能够在 完整的、活的、多细胞有机体的复杂光学环境。它承诺将弥合这一差距 在细胞和生物体之间，通过高灵敏度的体积成像，具有衍射限制 活体组织和发育中的胚胎的分辨率。我们提出了一个分三个阶段的研究计划 重叠阶段：实施AO-LLSM(与其开发者合作)，开发 新的可视化和分析软件的规模和复杂性所需要的 数据集，并使用AO-LLSM来解决脊椎动物发育中的一个问题。为了满足 分析4D数据集的计算挑战(从低信噪比，通常是非 被研究对象的点状特征，时间变化的空间复杂性 数据和数据集的大小)，我们将使用深度学习和相关的 算法，并咨询专家。作为一个范例应用，我们将研究其后果 Notch信号和相关的细胞膜交通和蛋白质转位事件 斑马鱼早期神经发生的分化。AO-LLSM将第一次允许我们联系 在细胞界面上发生的以秒为时间尺度的分子信号事件最终将导致 几个小时后的子代细胞。因此，我们预计在解决一些长期的-- 在细胞命运测定方面存在的问题，我们将发展显微镜方法和计算机 可广泛应用于一系列模型系统和生物学问题的可视化工具。

Abstract Contemporary fluorescence microscopy connects our understanding of molecular events from biochemistry and structural biology with their activities in living cells. Lattice light sheet microscopy (LLSM) has made it possible for us to track phenomena such as endocytic vesicle assembly or lipid kinase recruitment across an entire cell with high resolution, in both space and time, and with nearly single-molecule sensitivity. Development during the past year of lattice light sheet microscopy with adaptive optics (AO-LLSM) has overcome the optical limitations that have so far restricted most studies to individual cells in culture, allowing us to achieve comparable resolution and sensitivity in the complex optical environment of an intact, living, multicellular organism. It promises to bridge the gap between cells and organisms, through high sensitivity, volumetric imaging, with diffraction-limited resolution, of living tissues and developing embryos. We propose a research program in three overlapping stages: implementation of AO-LLSM (in collaboration with its developer), development of the new kinds of visualization and analysis software required by the scale and complexity of the datasets, and use of AO-LLSM to solve a problem in vertebrate development. To meet the computational challenges of analyzing the 4D data sets (from low signal-to-noise, the often non- punctate characteristics of the objects being studied, the temporally varying spatial complexity of the data, and the size of the data sets), we will develop new approaches using deep learning and related algorithms, with consultation from experts. As a paradigm application, we will study the consequences of Notch signaling and the related membrane-traffic and protein translocation events for cell differentiation in zebrafish early neurogenesis. AO-LLSM will for the first time allow us to relate molecular signaling events occurring on a timescale of seconds at cell interfaces to the ultimate fate of daughter cells many hours later. We therefore expect that in the course of resolving some long- standing issues in cell fate determination, we will develop microscopy approaches and computer visualization tools that are widely applicable to a range of model systems and biological questions.