Updated restraint dictionaries and automated model building of pyranose carbohydrates

Carbohydrates are essential biomolecules, which facilitate biological processes involving other biomolecules, notably proteins. But, as opposed to proteins, carbohydrates can have complex stereochemistry, multiple native forms, polymeric branching and tight conformational preferences. The Protein Data Bank (PDB) has been shown to contain numerous errors in carbohydrate structures determined with X-ray crystallography or electron cryo-microscopy. This is partly because the software aimed at solving carbohydrate structures have yet to become as featureful as their protein counterparts. The aim of this thesis is to better understand the problems affecting carbohydrate structures in the PDB and develop software methods to address those. Following the analysis in the first chapter (also in Atanasova et al., 2020), two areas were targeted for improvement - model refinement and automated model building.

Refinement software uses dictionaries with chemical geometry data about molecules, such as bond lengths and angles, that can be used when the X-ray crystallography or electron cryo-microscopy data are unclear. However, carbohydrate dictionaries have been reported to contain errors that have led to incorrect structures. The aim of the work reported on the second chapter was to introduce a completely new set of restraint dictionaries that correct these errors and add new unimodal torsion restraints. These allowed for carbohydrate conformation to be fixed automatically, as evidenced by the results presented here (also in Atanasova et al., 2022).

Finally, a new method for building N-glycans into electron density/potential maps was explored. This new software, called Sails, uses fingerprint-assisted detection: it relies on a database of monosaccharide fingerprints, which it uses to scan a map and locate sugars. The method was found to be moderately successful at detecting monosaccharides at medium to high resolution, showing most hits at the beginning of N-glycans – this opens the door to the extension of the glycan chain by other approaches.