Fluorescence microscopy techniques, restricted by the diffraction limit of light, have seen a remarkable advancement in recent years. An approach called dSTORM (direct stochastic optical reconstruction microscopy) utilises the photoswitching capabilities of organic fluorophores when in the presence of special mounting media, the solution within which the sample is placed, to detect single molecule fluorescing events over time. The image that can be reconstructed from these events is not diffraction limited, but instead is limited by how well each event can be precisely localised. In Chapter 3 the importance of using a suitable mounting buffer in order to achieve super-resolution dSTORM is discussed in detail. A quantitative method for determining the reactivity of thiol dSTORM switching mountants was developed for use within the lab. Every fluorescent probe has different photophysical properties which can be manipulated by varying the composition of the switching buffer to enhance desirable qualities, such as; increased photon counts, faster switching rates, and longer survivability. In addition to investigating the effects of buffer composition the use of a near UV light-source was also explored as a means of manipulating the same properties to improve overall resolution and quality of the resulting images. A range of photoswitchable fluorescent dyes were tested including Alexa Fluor 660 which is a dye that to my knowledge has not been greatly tested for use in single molecule localisation microscopy by others to date. This dye performed strongly alongside the traditional Alexa Fluor 647 used for dSTORM imaging in optimal conditions. A relatively new approach to single molecule imaging which does not require the fluorophore to photoswitch, called DNA-PAINT (point accumulation for imaging in nanoscale topography), has been investigated throughout this thesis. This approach relies on the transient binding of small oligonucleotide sequences, called “Imagers”, to target docking strands anchored in positions of interest. These imagers have a photostable and bright fluorophore conjugated to the oligonucleotide. It is the transient immobilisation of the imager strand, as it binds to a fixed docking strand, which appears as stochastic blinks. The duration of these events, which can be extended by increasing the number of overlapping base pairs, is primarily responsible for improved localisation precision and therefore potentially overall resolution. At the end of Chapter 3 I compare this new pointillism microscopy approach, DNA-PAINT, with dSTORM using a set of custom-designed oligonucleotide sequences that allow both formats to be employed on the same target. The transient binding of small strands of oligonucleotides offers a far more controllable system for stochastic imaging. In Chapter 4 I use this superior approach to achieve greater resolution than other fluorescence techniques in biological samples, sufficient to visualise single ryanodine receptors (RyR). The RyR are extremely important in the contraction of muscle cells as they are capable of detecting transient changes to calcium concentration and are responsible for releasing large stores of calcium from the sarcoplasmic reticulum. With DNA-PAINT I observed that RyRs cluster into irregular arrays which contain significant gaps that are occupied by other proteins, including junctophilin (JPH). The stoichiometry of JPH with RyR varied cluster to cluster, exposing a new complexity in the regulation of RyRs. In Chapter 5, quantitative super-resolution is reliably achieved through the implementation of quantitative DNA-PAINT (qPAINT) within the Python Microscopy Environment (PYME) software. Quantitative measurements are possible because of the statistical predictability of DNA hybridisation and the near constant influx of fresh imager strands by diffusion. This results in limited photobleaching, a permanent dark state. The frequency with which a region of interest blinks is proportional to the number of binding sites available, and therefore the mean dark time between detected events is also inversely proportional. I validate my approach to qPAINT, which maintains the spatial information of individual structures, by using a DNA-origami test slide. Two distinguishable structures were present and an estimate for the ratio of available docking sites between them was satisfactorily established. I conclude that with this tool, molecule densities can be inferred and information about biological samples can be probed to new levels. The results of the full methodological approach to accomplish dual-colour super-resolution imaging of optically thick cardiac tissue, using both dSTORM and DNA PAINT techniques, is discussed in detail in Chapter 6. The current range of photoswitchable fluorophores limits the possible combination of molecular dyes for use with dSTORM and some compromise is made in their selection. For DNA-PAINT, the prospect of chromatic aberration is removed by imaging the same dye in subsequent rounds of imaging. The process, called Exchange-PAINT, allows the user to remove previously imaged imager strands, through a series of washes, and replace them with a complementary sequence for another target. I introduce the concept of using quencher strands to eliminate signal from unwanted imager sequences, accelerating their removal in samples of reduced diffusion and decreasing the risk of sample disturbance, in a process we termed Quencher Exchange-PAINT. Using this technique, I achieve superior super resolution results in optically thick samples. The results presented in this thesis are expected to (1) lead to a better understanding of the variables associated with single molecule localisation microscopy, (2) further reveal the complexity in cardiac protein distribution, (3) quantify relationships between co-localising proteins and other targets, and (4) apply DNA-PAINT to imaging in optically thick biological samples. This study shows promise for the future applications of the DNA-PAINT pointillism super-resolution method and its ability to investigate a multitude of biological questions.

No description available.