Development of a standard methodology for cryogenic fixation, DNA-PAINT super-resolution microscopy, and Bayesian analysis of the internal structures of healthy and infected plant cells

Abstract

Microtubules contribute to plant cellular defence against pathogens. Identifying their location and structure within the cell is impacted by the diffraction limit of light imposed upon confocal microscopy. This places an upper limit of resolution at ~200 nm, whilst the width of microtubules is ~25 nm. DNA-PAINT is a new super-resolution microscopy technique that has not previously been successfully applied to the whole mount imaging of plant cells. This dissertation demonstrates the successful preservation of Arabidopsis thaliana root cell microtubule arrays through high-pressure freeze and quick freeze substitution fixation, followed by imaging using DNA-PAINT. The resolution of this image was 44.8 nm, successfully surpassing the diffraction barrier, but not achieving resolution increases reported in the literature. The second half of this project proposes a new sampling methodology, known as “propagating lune”, for use in Kalman filter dependent tracking of plant microtubules. The parameters of the Kalman filter were optimised through the tracking of synthetic data generated by functions developed over this course of work, displaying mean tracking accuracies of > 80% across three levels of data noise. However, the optimised Kalman filter showed mixed results when this was applied to real DNA-PAINT coordinate data. Further optimisation of the fixation and imaging methodology of whole mount plant cell DNA-PAINT imaging, as well as refinements to the Kalman filter architecture and training methodology, may show future promise in the automated analysis of plant microtubules as they pertain to cytoskeletal response to attacking pathogens.