Fast, Three-Dimensional Fluorescence Imaging of Living Cells
by Hongje Jang
Date of Examination:2021-12-13
Date of issue:2022-01-21
Advisor:Prof. Dr. Jörg Enderlein
Referee:Prof. Dr. Jörg Enderlein
Referee:Prof. Dr. Alexander Egner
Referee:Prof. Dr. Andreas Janshoff
Referee:Prof. Dr. Timo Betz
Referee:Prof. Dr. Sarah Koester
Referee:Prof. Dr. Simone Techert
Files in this item
Name:PhD_Thesis_HongjeJang.pdf
Size:20.9Mb
Format:PDF
Abstract
English
This thesis focuses on multi-plane fluorescence microscopy for fast live-cell imaging. To improve the performance of multi-plane microscopy, I developed new image analysis methods. I used these methods to measure and analyze the movements of cardiomyocytesand Dictyostelium discoideum cells.The multi-plane setup is based on a conventional wide-field microscope using a custom multiple beam-splitter in the detection path. This prism creates separate images of eight distinct focal planes in the sample. Since 3D volume is imaged without scanning, three-dimensional imaging at a very high speed becomes possible. However, as in conventional wide-field microscopy, the "missing cone" of spatial frequencies along the optical axis in the optical transfer function (OTF) prevents optical sectioning in such a microscope. This is in stark contrast to other truly three-dimensional imaging modalities like confocal and light-sheet microscopy. In order to overcome the lack of optical sectioning, I developed a new deconvolution method. Deconvolution describes methods that restore or sharpen an image based on physical assumptions and knowledge of the imaging process. Deconvolution methods have been widely used to sharpen images of microscopes and telescopes. The recently developed SUPPOSe algorithm is a deconvolution algorithm that uses a set of numerous virtual point sources. It tries to reconstruct an image by distributing these point sources in space and optimizing their positions so that the resulting image reproduces as good as possible the measured data. SUPPOSe has never been used for 3D images. Compared to other algorithms, this method has superior performance when the number of pixels is increased by interpolation. In this work, I extended the method to work also with 3D image data. The 3D-SUPPOSe program is suitable for analyzing data of our multi-plane setup. The multi-plane setup has only eight vertically aligned image planes. Furthermore, for accurate reconstruction of 3D images, I studied a method of correcting each image plane's relative brightness constituting an image, and I also developed a method of measuring the movement of point emitters in 3D space. Using these methods, I measured and analyzed the beating motion of cardiomyocytes and the chemotaxis of Dicyosteilium discoidem. Cardiomyocytes are the cells of the heart muscle and consist of repetitive sarcomeres. These cells are characterized by fast and periodic movements, and so far the dynamics of these cells was studied only with two-dimensional imaging. In this thesis, the beating motion was analyzed by tracing the spatial distribution of the so-called z-discs, one of the constituent components of cardiomyocytes. I found that the vertical distribution of $\alpha$-actinine-2 in a single z-disc changed very rapidly, which may serve as a starting point for a better understanding the motion of cardiomyocytes. \textit{Dictyostelium discoideum} is a well established single cell model organism that migrates along the gradient of a chemoattractant. One has conducted much research to understand the mechanism of chemotaxis, and many efforts have been made to understand the role of actin in the chemotactic motion. By suppressing the motor protein, myosin, a cell line was created that prevented the formation of normal actin filaments. In these myosin null cells, F-actin moves in a flow-like behaviour and induces cell movement. In this study, I imaged the actin dynamics, and I analyzed the flow using the newly created deconvolution and flow estimation methods. As a result of the analysis, the spatio-temporal correlation between pseudo-pod formation and dynamics and actin flow was investigated.
Keywords: 3D Fluorescence Microscopy; Deconvolution; image processing; Live cell imaging; cardiomyocyte; Dictyostelium Discoideum