Enhancement of inclined optical sheet illumination and its application to quantitative super-resolution microscopy

Over the last couple of decades, the field of optical microscopy has been revolutionized by the possibility of breaking the diffraction limit, which ushered in the new era of “super-resolution”. Optical techniques can now produce images of biological systems with a level of detail that was previously only accessible via specialized methods, such as electron or scanning probe microscopy. The 2014 Nobel Prize in Chemistry, won by super-resolution pioneers William Moerner, Eric Betzig and Stefan Hell, vastly popularized these techniques and sprouted a widespread interest in both finding useful applications for super-resolution microscopy and expanding its capabilities. Among the various existing super-resolution microscopy techniques, localization-based methods such as PALM and STORM are the ones that can achieve the highest image resolutions (a few tens of nanometers), and they also provide information at the single-molecule level. However, they suffer from two fundamental drawbacks: the first is the long time necessary to acquire these kinds of images (ranging from minutes to hours), which severely limits the applicability of PALM and STORM to non-fixed, living biological samples; the other is the high susceptibility of these techniques to factors of noise influencing the signal-to-background ratio, such as signal coming from out-of-focus planes in the acquired images: as the amount of photons produced by individual fluorescent molecules is very low, fluorescence background in the image can easily mask the on-focus signal, which in turn leads to inefficient sampling and image artifacts, especially for densely labelled, highly scattering or intrinsically fluorescent specimens. Localization-based super-resolution techniques also carry an intrinsic capability of providing not only qualitative information on the underlying structure of a fluorescent specimen, but also quantitative information on the number and position of the fluorescent molecules composing it. However, this capability is seldom exploited by researchers applying PALM and STORM to their research, as common experimental designs do not allow for it – e.g., because they employ labelling strategies that insert an arbitrary amount of fluorescent molecules in the sample, limiting the obtainable information to qualitative structural data. This doctoral dissertation reports our work in addressing the aforementioned issues of (i) improving single-molecule localization microscopy techniques when it comes to acquisition times and susceptibility to background fluorescence, and (ii) employing localization-based methods to obtain quantitative single-molecule data with an ad hoc experimental design. This thesis is therefore divided in two main parts. Part I deals with our efforts to characterize and improve the fluorescence excitation methods employed in PALM and STORM imaging. After introducing and explaining in detail the principles behind these techniques, we present our newly developed methods for measuring some crucial parameters of a fluorescence excitation system (e.g.: the thickness of the specimen volume undergoing excitation, the achievable signal-to-background ratio, and the sampling efficiency in single-molecule localization experiments). We also report on our improvements on the current state-of-the-art, showing that reducing the fluorescence excitation volume to subcellular sizes leads to a massive improvement in fluorophore sampling when applied to STORM imaging. Part II deals with our efforts towards obtaining quantitative single-molecule information from PALM imaging, and with our application of a quantitative PALM technique to the study of an open biological problem. We introduce the issue of bacterial efflux pumps, which are a class of proteins that mediate antimicrobial resistance in a growing number of pathogens. We report on our development of a genomically engineered Escherichia coli strain, encoding for a fluorescent version of the AcrB efflux pump, that we expressly designed to be used in quantitative PALM imaging. We describe our method for directly counting the number of copies of AcrB contained in individual bacteria via super-resolution microscopy, and we report our preliminary findings on the varying expression of this protein under planktonic and biofilm-associated growth. Moreover, we describe the application of our experimental strain and technique to elucidate the mechanism of action of a novel antimicrobial drug.