All living cells are covered with a dense coat of complex carbohydrates also called glycans. The structural complexity of surface glycans is especially diverse in prokaryotes given the multiple involvements of glycans in regulating the interactions of prokaryotes with their environment. On prokaryotes such as bacteria, glycans are essential components of bacterial cell walls, they also act as receptors for bacteriophages and enable the evasion of bacteria from immune recognition.
In contrast to glycans covering animal cells, which consist of an alphabet of only 10 monosaccharides, bacterial glycans are built from an alphabet of nearly 200 monosaccharides. This tremendous diversity hinders the analysis and visualization of bacterial glycans. The development of mapping tools for bacterial glycans is essential for the identification of pathogenic strains (e.g. enterotoxic Escherichia coli O157:H7), for vaccine design (e.g. Neisseria meningitidis) and understanding the contribution of bacterial antigen mimicry in triggering autoimmune disease (e.g. Guillain-Barré syndrome).
The goal of the proposed project is to develop a visualization system for bacterial glycans, which enables a clear, yet thorough representation of complex structures. The visualization system will be used as a query language to search for related molecules through glycan databases (e.g., CSDB). Moreover, the visualization system will be linked to bacterial glycan arrays applied for determining the glycan-binding specificity of antibodies and other immune proteins. The combined expertise of the applicants in glycobiology and bioinformatics is critical to the success of the project. The areas of competence covered include the isolation and structural characterization of glycans, analysis of carbohydrate-binding proteins on glycan arrays, development of glycan databases and analysis software.
The first stage of the project will consist in establishing a catalog of bacterial glycans of interest, including O-antigens, capsular oligosaccharides, and cell wall glycoconjugates such as lipoarabinomannan. These structures need to be rationally encoded and translated into the visualization system to be used interactively.
An important aspect will be the preservation of information relative to the biosynthesis and degradation of carbohydrates. Glycan databases and catalogues are too often composed of independent items while common substructures and therefore common biosynthetic enzymes emphasise the relatedness of glycan molecules. The visualisation system will bring out the similarity of glycans on that basis. The CAZy database of carbohydrate-active enzymes is a rich source that will be integrated, while CSDB will provide useful structural information to explore the diversity of glycans.
In a second phase, the structures present on a bacterial glycan array will also be integrated into the visualization system. The application of the tool will allow the rapid delineation of glycan epitopes recognized by antibodies by parsing the global data obtained from bacterial glycan arrays. So far, such analyses can only be achieved through manual examination, which is slow and prone to mistakes. The application of automated visualization analysis will accelerate the discovery of bacterial epitopes of interest for vaccine development. The visualization tools developed will be accessible to the research community through the ExPASy portal, which is hosted at UniGE and maintained for the glycomics part by the Lisacek group.