Prediction of 3D structures of oligosaccharides. Methodological development and studies on saccharide antigens of interest for vaccine design

This thesis presents 3D structure predictions of five immunologically active surfacepolysaccharides of pathogens, which are of interest as targets for development of glycoconjugatevaccines. Furthermore it describes two methods which have been used for theconformational predictionsThe Shigella dysenteriae type 2 Oantigenwas predicted to form a compact helicalstructure with three repeating pentasaccharide units per turn. Favoured conformations werealso predicted for the Shigella dysenteriae type 4 and Escherichia coli O159 Oantigens,which have been reported to be crossreactivewith antibodies. We propose a commonepitope at the branch region of the repeating units. The conformation of the Burkholderiacepacia exopolysaccharidewas also predicted and was shown to form an extended fibre,with the three branches of the repeating unit packing tightly against the main chain.Furthermore, the 3D structure of a Schistosoma mansoni cercarial surface polysaccharidethat has been reported to be immunologically reactive, was predicted. The fucosyl brancheswrap tightly around the main chain, reducing the molecule's flexibility.The polysaccharides studied are of interest for development of glycoconjugatevaccines,and the results suggest that oligosaccharide 3D structure prediction can be an importanttool for rational vaccine design. These rapid predictions provide detailed information aboutpotential epitopes and the basis of immunological crossreactivityas well as conformationsof the polymer.Methods used for the conformation predictions were filtered systematic search and geneticalgorithm search. The filtered systematic search method thoroughly explores theconformational space below a set filter cutoffenergy. It can be used for structures up totetrasaccharide size with reasonable computational requirements. It is guaranteed to findthe global minimum energy conformation and the results can be used to generate adiabaticenergy maps, clearly describing the conformational restrictions and flexibility of theoligosaccharide. The genetic algorithm search is a very fast conformational search method.It scales well and is suitable for large oligosaccharides approximately up to the size ofdecasaccharides. It finds most local energy minima in the conformational space quickly, butis not guaranteed to find the global minimum conformation