| Summary: | 碩士 === 國立臺灣大學 === 物理學研究所 === 107 === To understand brain function, detailed anatomical mapping of neurons and their fiber distributions should be the first step. However, it is not an easy task since neuronal network is composed of tiny fibers that may closely entangle with each other. For example, the diameter of a dendrite in Drosophila brain can be as small as 100 nm, and when dendritic fibers interweave with each other, at least 20-nm resolution would be necessary to resolve these fibers (assuming fluorescent protein expressed in cytoplasm, and the thickness of a cellular membrane is 10 nm). What makes it even more difficult is that these fibers may extend three dimensionally throughout a brain, so we need not only a high-resolution technique, but also a technique that is able to penetrate the whole brain to track neuronal fibers.
Because of diffraction limit, the resolution of an optical microscope is confined around λ/2, which is roughly equal to 250 nm for visible light. Nobel Prize in Chemistry 2014 was awarded to three scientists due to their contribution on “superresolution microscopy” that is able to break the diffraction barrier. Among all superresolution techniques, localization microscopy achieves one order resolution enhancement by detection of >100 photons from each fluorescent protein (FP). Thus, to reach 20-nm resolution, localization microscopy should be the best option.
Nevertheless, to apply localization microscopy across a whole brain, there are three major challenges. First, localization microscopy requires a “blinking” contrast agent. Generally speaking, genetically encoded FP is used when labeling a thick brain tissue. However, blinking FP is scarce, while suffered from photo-bleaching. Second, due to lack of optical sectioning ability, wide-field-based localization microscopy cannot distinguish signals from different layers and thus imaging depth is confined to less than 10 μm. Third, tissue-induced scattering and aberration also limit the imaging depth.
In this study, we combine four techniques to achieve 20-nm superresolution across a whole Drosophila brain, including 1. localization microscopy to enhance resolution; 2. adding ME to enable blinking of genetically encoded, photo-convertible FP. Furthermore, we can use photo-conversion to supply the bleached FP; 3. spinning disk confocal microscope to provide optical sectioning; 4. tissue clearing to minimize tissue scattering/aberration. We name this combination COOL, i.e. spinning disk COnfocal lOcalization with tissue cLearing. COOL allows us to distinguish 3D entangled dendrites even at the bottom of the brain, paving the way toward whole-brain neural network analysis.
|
|---|
