AlphaFold2-Assisted Domain Determination of the SARS-CoV-2 Protein NSP3 & Exploration of the Solvent Region of Macromolecular Crystals with Experimental Phasing

Proteins fulfil various essential functions in all known life forms and modern biotechnology enabled their use in experimental, medical, and industrial applications. Knowledge about the structure and function can be used to optimize proteins for these applications and finds use in drug discovery, if a the protein originates from a pathogen. In this work, the multidomain protein NSP3 from coronaviruses was examined with bioinformatical methods and with structure prediction. The results include the identification and experimental validation of a previously unknown domain and a new hypothesis about functions of additional domains. Furthermore, the manual process of a new domain determination technique was automated to enable use in any multidomain protein. In the second part of this dissertation, the electron density of solvent regions from macromolecular crystals was examined. Specifically, reflection data was experimentally phased in a newly developed automated pipeline. The resulting density maps were compared to published maps and the new insights enabled the construction of new water models, which improved one structure and lowered its R-values.