Throughout all domains of life, glycosylation represents an important biological process that regulates the properties of proteins. The end products of glycosylation are especially important for the existence of multicellular organisms, as glycoprotein are involved in the processes of signaling, immune response, cell adhesion, etc. Although the study of glycans is promising for new discoveries in medicine, biochemistry, pharmacology, and glycobiology, the complexity of the glycan forest as a whole poses major challenges for research in this field. The development of appropriate glycan detection methods is crucial in understanding the glycome. Although destructive approaches to determining the sequences and structures of oligosaccharides (mass spectrometry, isoelectric focusing, capillary electrophoresis, etc.) are most commonly used, the use of so called glycan-binding probes has also been well established in recent decades. This method mainly uses lectins and antibodies that bind oligosaccharides. Both groups of proteins have their advantages and disadvantages that determine their usefulness.
In this thesis, we inquired about the possibility of obtaining a third group of glycan-binding probes through the use of modified glycosyltransferases. We focused on N-acetylglucosaminyltransferase I, a key enzyme in the synthesis pathway of N-glycans. To determine the feasibility of preparing an in vitro system for the detection of homogeneously glycosylated proteins based on this glycosyltransferase, we used a number of in silico bioinformatics approaches. Throughout the course of our work, we developed a methodology for computer prediction of the binding of oligosaccharide ligands to glycan-binding proteins. Thus destilling information about our protein, which had not yet been obtained by crystallographic approaches. We have shown that the forming glycan moieties of glycoproteins are most likely to bind to our enzyme by directing their reducing ends towards the flexible loop of the catalytic site. We were able to show that we selected meaningful modifications to the initial sequence for human N-acetylglucosaminyltransferase I. By shortening the sequence, we were able to predict the catalytic domain with higher accuracy, and by mutating its catalytic residue D289 to alanine, we regained numerous interactions, which the protein forms only upon binding both substrates. In addition, we procured starting points (in the form of composite interaction networks) for further modifications that could serve to improve our synthetic glycan-binding probes.
|
