Although the molecular basis of carbohydrate-protein interactions has been widely studied, the thermodynamics of these interactions remain complex and not well-understood. Consequently, the prediction of binding constants is still a difficult and unreliable process. The ability to design high affinity ligands for carbohydrate-binding enzymes is of critical importance for carbohydrate-based drug development. Several of our projects in the Boons Group are focused on the elucidation of the important factors for carbohydrate-protein interactions. These projects are highly multidisciplinary and often involve a design stage relying heavily on computer modeling and a synthesis stage, followed by biological or biophysical evaluation. The research often involves conformational analysis using high field NMR spectroscopy and a measuring of binding affinities using ELISA assays or microcalorimetry. Furthermore, some of our programs also use electron microscopy and Langmuir blodging, X-ray powder diffraction and protein crystallography. Currently, we are involved in the synthesis of a cyclic conformationally constrained

saccharide (Figure 1). These compounds are used to probe the importance of the loss of flexibility of a saccharide ligand when bound to a protein. In another project, we are modifying saccharides in a manner such that a conserved water molecule that is normally present in the

crystal structure of carbohydrate-protein complex is displaced. Elegant studies by Lemieux and others have demonstrated that water molecules in the binding site of lectins are of high energy. We anticipate that the displacements of these water molecules may result in a high affinity interaction.

Most eukaryotic cellular proteins, with the exception of certain hormones and enzymes, are reliant on their covalently bound sugar units (glycoproteins) to confer a broad range of important biological functions. These functions include immunogenicity, solubility, recognition, protection from proteolytic attack, induction and the maintenance of the protein conformation in a biologically active form. The biosynthesis of N-linked glycoproteins is complex and results in a micro-heterogeneity of the carbohydrate structure (formation of glycoforms). Several studies have highlighted that each glycoform should be interpreted as a unique population of

molecular species possessing particular biological properties. We are developing a semi-synthetic method to obtain glycoproteins with well-defined saccharide attachments. Our approach is based on an asymmetric disulfide conjugation (a S-S glycosidic linkage)

between a 5-nitro-2-pyridinesulfenyl activated thioglycoside (Scheme 1) and a protein or pre-assembled peptide sequence in which the natural asparagine glycosylation site is replaced by a cysteine residue(R'SH). The merit of such an asymmetric disulfide linkage is that it represents a sound structural mimic for natural N-linked glycoproteins as the disulfide linkage is flexible enough to adopt conformations imposed by natural amide linkages at glycosylation sites. Currently, we are involved in the construction of an aglycosylated IgG recombinant antibody by site-directed mutagenesis, which bears a cysteine residue in lieu of Asn-297. Covalent introduction of defined and activated oligosaccharide thiol derivatives provides neoglycoconjugates that will enable us to study in detail the relationship between carbohydrate microstructure and the biological properties of IgG molecules.