Development of an optical system for three-dimensional multicolour super-resolution microscopy and its application to the study of subcellular compartmentalization

Over the last decades, super-resolution techniques have completely revolutionized the world of microscopy and have become increasingly relevant for the biological research by enabling the observation of unresolved details of cellular structures and by elucidating many biological processes at the molecular scale. However, there are still some challenges to be faced. Currently, performing multicolour super-resolution imaging is not trivial. In particular, fine correction of all potential aberrations and artifacts, which could significantly impact the outcome, is necessary to extract reliable information from multicolour single molecule images. This is even more crucial when imaging small samples, such as bacteria, where single molecule imaging is challenged by the limited volume. This thesis reports my work onto the development of an optical system to perform multicolour 2D and 3D super-resolution microscopy and its application to the study of bacterial subcellular organization. In the first part of the thesis, after addressing the technical challenges of single molecules multicolour imaging (mechanical and thermal drifts, chromatic aberrations and crosstalk signal), I show how I overcame them by i) developing custom algorithms and ii) a custom Single-Molecule Localization Microscopy (SMLM) setup to perform three-dimensional multicolour super-resolution imaging with nanometers accuracy. In the second part of the thesis, I report the application of multicolour super-resolution imaging to the study of subcellular organization in bacteria. In fact, despite bacteria apparently lack any subcellular organization, recent studies have suggested the presence of a molecular compartmentalization within enzymes of the same metabolic pathways. In order to study the spatial distribution of enzymes within bacteria and their co-localization, I first validated my method on positive and negative control samples of both simulated and real data. To study the mutual subcellular organization of biological samples I developed an analysis procedure that, thanks to a fine correction for crosstalk and chromatic aberrations, allows to localize photoactivable fluorescent proteins with 6 nm accuracy in 2D. Moreover, by combining a local density-based colocalization index and a nearest neighbor search, I was able to identify co-localizing pairs of molecules and measure the distance between them with an error of about 10 nm. The consistency of the results obtained from the analysis of simulations and acquisitions demonstrated that my method is solid as it can quantify significant differences between co-localizing and non-colocalizing distributions of molecules. Finally, in the last part of the thesis, I show the effort to further improve the resolution of the system by combining two super-resolution techniques: Expansion Microscopy (ExM) sample preparation protocol and Photo-Activated Localization Microscopy (PALM) imaging process. Thanks to the 4x physical isotropic expansion, the chromophores in the bacteria volume are put far apart, thus restoring the possibility to distinguish single astigmatic PSFs to perform 3D super-resolution. The application of ExM to bacteria is not trivial, even if few protocols are reported in literature, and I show how I set a custom protocol to work on our experimental conditions, and how I achieved two-colours 3D super-resolution imaging in expanded bacteria. This represents a valuable method to perform single molecule co-localization imaging on samples that are very limited in volume size, such as bacteria, where 3D imaging is hindered by the high density of proteins.