Glycan Biosynthesis: Structure, Information, and Heterogeneity

The surfaces of all living cells are decorated with branched sugar polymers known as glycans. These information-rich structures confer cells with a recognizable molecular identity, and underlie many specific cell-cell interactions. Analytic methods - including NMR, mass-spectrometry, and glycan arrays - now permit the routine profiling of glycans associated with various cells or proteins. This has stimulated efforts to build comprehensive and searchable glycan databases. However, from an informatics perspective glycans present multiple challenges. First, whereas nucleotide and amino acid chains are efficiently represented as strings, sugars can polymerize into complex tree-like objects. The potential combinatorial space of glycans is therefore much larger than that of proteins. Second, many specific molecular interactions appear to be mediated by groups of closely-related glycan variants rather than by a single well-defined structure. This phenomenon of “micro-heterogeneity” makes it difficult to rigorously characterize the glycan repertoire of a cell. In this workshop, I will use ideas from algorithmic self-assembly to show that glycan structure and diversity are best understood through the lens of glycan biosynthesis. I will demonstrate that a specific glycan structure is the outcome of glycosyltransferase enzymes acting according to simple rules in a specific order, like workers on a factory floor. Errors in this process produce a well-defined spectrum of glycan by-products, precisely matching the observed micro-heterogeneity in real glycan profiles. This predictive theoretical framework allows us to use glycans as sensitive cell-biological probes. It provides a unifying perspective within which the rich and growing datasets of glycan structures can be organized and fully utilized.