Biophysical characterization of carbohydrate-lectin interactions

Improved knowledge of the biological role of lectins has raised the demand for carbohydrate-based therapeutics in recent years. The potential market is estimated to be greater than 20 billion dollars. However, lectins are challenging drug targets due to the unique binding properties of their extensively hydroxylated carbohydrate ligands. Hydroxyl groups provide directionality and therefore specificity, but are penalized with high desolvation costs. Consequently, monovalent carbohydrate-lectin interactions tend to be rather weak, often in the millimolar range. Moreover, the polar character of carbohydrates creates large obstacles for drug application regarding oral availability and long-lasting plasma levels. The key to the successful development of carbohydrate-based drugs is the simultaneous optimization of carbohydrate lead structures in terms of pharmacodynamics and pharmacokinetics. To further enhance the success rate of carbohydrate-based drug candidates, the understanding of carbohydrate-lectin interactions on a molecular basis has to be improved. For this purpose, we combined structural information (X-ray crystallography and nuclear magnetic resonance spectroscopy), binding data (isothermal titration calorimetry, microscale thermophoresis, and fluorescence polarization assay) and computational methods (quantum mechanical calculations and molecular dynamics simulations) to explore the lectins FimH and E-selectin and their interaction with carbohydrates and mimetics thereof.
FimH is a virulence factor of uropathogenic E. coli located at the tip of the bacterial type 1 pili. It interacts with the mannosylated glycoprotein uroplakin 1a in the urothelial mucosa and thereby mediates adhesion to the bladder wall as the initial step of urinary tract infections (UTI). In manuscript 1 we investigated the energy contribution to binding of the hydroxyl groups mediating the interaction between FimH and the carbohydrate moiety of its ligands. The rigidity of this bacterial lectin was demonstrated in manuscript 2, where could show that the affinity of a septanose as a mannose mimic is reduced by a factor of 10, mainly due to its flexibility in solution and the consequent conformational restrictions upon binding. In manuscript 3 we analyzed interactions between the tyrosine gate motif of FimH and the aglycones of different ligand classes. This motif (Tyr48, Tyr137) forms the entrance of the binding pocket and significantly contributes to binding affinity. In manuscript 4 we explored 2-C-branched mannosides as a novel family of FimH antagonists. In manuscript 5, a pharmacodynamically and pharmacokinetically optimized FimH antagonist was explored by oral application in a mouse model for UTI, resulting in a 1000-fold reduction of the bacterial load in the bladder. Finally, in manuscript 6 we reanalyzed ITC results from our previous studies with the novel analytical tool kinITC, allowing the determination of kinetics in addition to the thermodynamics of binding. The hydrophobic aglycone turned out to be mainly responsible for guiding the antagonist to its binding site whereas the hydrogen bond network between the mannose moiety and the protein had predominantly an impact on the off-rate.
E-selectin is a lectin expressed on the surface of vascular endothelial cells and is involved in the recruitment of leukocytes to the site of inflammation. By interacting with the tetrasaccharide epitope sialyl Lewis(x), E-selectin establishes the initial contact and enables leukocytes to roll along the endothelial surface. Whereas this process is a defense mechanism in case of infections and injuries, excessive extravasation of leukocytes can have deleterious consequences in case of numerous diseases with an inflammatory component, e.g. asthma, psoriasis or stroke. Thus, blocking the interaction of E-selectin with its physiological ligands is a promising strategy to suppress the inflammatory response at the beginning of the cascade. For reliable and materially efficient affinity measurements, we developed and evaluated a novel assay for E-selectin based on microscale thermophoresis technology in manuscript 7. In the subsequent manuscripts, we applied the microscale thermophoresis assay. In manuscript 8 we performed a competitive library screen, whereby four promising small-molecule fragments were identified for further development towards a non-carbohydrate E-selectin antagonist. Finally, in manuscript 9 we were able to improve the affinity of a sialyl Lewis(x) mimic to E-selectin by pre-organizing the acid in its bioactive conformation.