Protein−carbohydrate interactions play a crucial role in many biological processes such as cell−cell recognition and receptor−ligand interactions and catalysis. This thesis explores the possibilities of engineering the protein–carbohydrate interactions between carbohydrate-binding proteins and their ligands.
Protein–carbohydrate interactions were characterised using different techniques, such as atomic force microscopy (AFM) and surface plasmon resonance (SPR).Two different chitinoligosaccharide-binding proteins were used in the presented experiments, e.g. the plant lectin wheat germ agglutinin (WGA) and a neolectin created by inactivation of the fungal chitin‑degrading enzyme Chit42.
The structures of WGA and Chit42, which were available as X-ray diffraction data and a homology model, respectively, revealed differences between their binding site architectures. The interaction of monomeric and polymeric N‑acetyl-d‑glucosamine (GlcNAc) with WGA, that normally binds chitinoligosaccharides on cell surfaces, was probed in nanomechanical force measurements using AFM.
These measurements aimed at determining the effect of ligand length and binding site clustering (multivalent binding) on binding strength between the protein and the carbohydrate ligand. Here, the GlcNAc ligand was presented at the surface of a planar self-assembled monolayer (SAM) of neoglycoconjugates and in a polymeric form as chitin beads. In adsorption measurements using the quartz crystal microbalance (QCM) it was shown that alkanethiols adsorb rapidly to a gold surface covering it completely within a few minutes as a SAM. In AFM force measurements, GlcNAc-specific binding events were detected with a WGA-modified probe on a GlcNAc-neoglycoconjugate SAM at bond rupture forces of 47 ± 15 pN. When this experiment was repeated on a polymeric substrate, multiple times higher forces were detectable. This indicated a high increase of affinity with additional binding subsites which were able to interact with the chitinous ligand. SPR measurements confirmed that WGA has higher affinity towards the immobilized GlcNAc-neoglycoconjugate SAM than towards the soluble free monosaccharide providing evidence of a significant affinity increase as a result of binding site cooperativity.
Furthermore, two different SAM formats have been tested for their suitability for studying protein–carbohydrate interactions. Here, non‑specific adhesion was identified as a critical factor that was mainly related to the hydrophobic parts of the neoglycoconjugates and could be attenuated by the introduction of a tetraethylene glycol spacer into the neoglycoconjugate. The binding measurements on neoglycoconjugate SAMs demonstrated that this type of carbohydrate presentation is a useful reference for interactions on naturally occurring two‑dimensional glycan arrays and demonstrated that the minimization of non‑specific adhesion of proteins is often required in order to obtain meaningful binding data.
Apart from WGA–carbohydrate interactions this thesis deals also with the carbohydrate binding cleft of the glycoside hydrolase Trichoderma harzianum chitinase Chit42, which solubilises as wild‑type form polymeric crystalline chitin (composed of â‑4 linked GlcNAc units). Nine different catalytically inactive Chit42 variants having amino acid alterations along the binding site cleft (at sugar-binding subsites -4 to +2) were created. These Chit42 variants were characterised with regard to their affinity towards chitinous and non‑chitinous oligosaccharides by SPR. As a result, hydrogen bonding between subsites -2/-3 and particularly stacking interactions by tryptophanes at subsites -3 and +2 were seen to be very important for the carbohydrate binding.
The exchange of the corresponding amino acids did not cause a change of binding specificity, however the selective binding of GlcNâ‑4(GlcNAc)4 could be improved by providing a counter charge through an amino acid substitution at subsite -3, replacing threonine with aspartic acid. In addition the introduction of glutamine and particularly an asparagine residue at this position seemed to broaden the substrate preference towards Galâ‑4(GlcNAc)4 and was thereby implicated as a binding groove hot spot for creating binding proteins, or hydrolytic enzymes with novel substrate specificity towards e.g. medically related oligosaccharides. The analysis of the Chit42 variants with modified active sites showed how the binding selectivity and affinity of neolectins can be engineered and may thereby function as a model for further neolectins and glycoside hydrolases with new ligand and substrate specificities, respectively.