Integrated Super Resolution Fluorescence Microscopy and Transmission Electron Microscopy

Abstract

Science is developing at a very high speed. New discoveries are made while old approaches are speeded up, made more precise or at least easier to use. This is partly due to the broader use of computers and automated machines. The analysis strength of modern computers and ability of automated machines to repeat the same procedure with minimum error is beyond human power. Therefore, scientists are more and more in favor of leaving the boring and time-consuming tasks to machines and use their time for more interesting work. This is not always an easy target. Each task needs to go through several stages before becoming ready for machine use. Novel ideas need to be produced to make it fast, accurate and easy to be done by the user or the machine. Correlative Light and Electron Microscopy (CLEM) is a robust method for studying microscopic objects and structures. On the one hand, light microscopy (usually Fluorescence Microscopy (FM)) provides high sensitivity, great specificity and a large field of view. On the other hand, electron microscopy (EM) has a magnificent ability to study nanometer scale structures in great detail. Combining these two techniques benefits from their individual strengths and has been done for several decades. However, to bring correlative microscopy to the world of high speed, high accuracy and user friendliness, several challenges have to be overcome. First, even though fluorescence microcopy is a very powerful technique, its resolution is far worse than the resolution of electron microscopy. This resolution gap hinders the proper comparison between target structures in FM and EM. Second, accurately overlaying fluorescence and electron microscopy images, essential to correlative microscopy, can be difficult. Finally, the speed of correlative microscopy experiments is usually low due to use of separate setups. The first challenge has been tackled in recent years by the arrival of super resolution fluorescence microscopy (SR-FM) techniques. These new methods decrease the resolution difference between fluorescence and electron microscopy by about one order of magnitude. This is achieved by using methods based on manipulating the excitation pattern or on determining the positions of individual fluorophores. This fascinating development has now proven itself in studies on biological and non-biological samples. In this thesis, I focus on several challenges in correlative microscopy. I first introduce a novel method for accurately overlaying correlative microscopy images using fiducial markers and image processing algorithms. Next, I describe the successful integration of a super resolution fluorescence microscope inside a transmission electron microscope (TEM) and two proof of principle applications in biology and material science. Investigation of sample support properties and a promising dye for future applications using this integrated system form the last part of my study. The integrated system together with the high accuracy overlay method has great potential for applications that require high accuracy correlative microscopy. More research, however, is required to optimize sample preparation and labeling to get the most of the integrated approach.