|dc.description.abstract||Galectins are an ancient class of lectins found in all forms of living organisms, responsible for modulating fundamental biological processes. They specifically recognise β-galactose containing glycans through their carbohydrate recognition domain (CRD) and mediate most galectin functions. Many galectins have emerged as novel therapeutic targets due to their association with the progression of various metabolic and disease conditions. Galectin-8, a tandem-repeat type member of the galectin family that contains two CRDs joined by an amino acid linker, is the focus of this research project. Galectin-8 plays a critical role in various biological processes such as cell adhesion and growth, immunomodulation, autoimmunity, inflammation, cancer and bone remodelling. Of interest to this project is the ability of galectin- 8 to increase bone resorption factors like RANKL that leads to a decrease in bone mass density. Inhibition of this role of galectin-8 can potentially be of therapeutic importance and therefore could be a novel approach for tackling diseases associated with bone-loss. The individual CRDs of the tandem-repeat lectin exhibit similar functions to that of the full-length galectin-8, however, with lesser potency. The N-terminal domain of galectin-8 (galectin-8N) preferentially recognises 3´-anionic saccharides, a feature not observed in the C-terminal domain of galectin-8 (galectin-8C) or any other galectin CRD. This research details the initiation of inhibitor design campaign by targeting the galectin-8N CRD of the tandem-repeat.
The main body of research uses structure-based ligand design approaches through a combination of computational and experimental techniques. Molecular dynamics (MD) simulations were employed to analyse galectin-8N-glycan interactions and to construct the ligand design hypothesis. The designed ligands were synthesised and subjected to evaluation for binding to galectin-8N using various experimental techniques. Subcloning (ligationdependant cloning) followed by protein expression (in E. coli) and affinity-based purification (using Lactosyl-Sepharose column) steps were performed to obtain the purified protein. Crystallographic investigations were conducted on various galectin-8N-glycan or galectin-8Nligand complexes to study their binding modes and interactions. Different techniques either in solution or solid state were used to assess the binding strength of the designed ligands towards galectin-8N.
Our crystal structures of galectin-8N bound to human milk glycans (LNnT and LNT; Chapter 2) not only provided a rationale for the difference in their affinity but also formed the basis for ligand design carried out in the project. The crystal structure of galectin-8N-LNT complex revealed a unique binding mode, wherein, for the first time a non-reducing end disaccharide of the tetrasaccharide was occupying the primary binding site. The presence of unique residue (Tyr141) in the extended galectin-8N binding site was mainly responsible for the observed binding mode. MD simulations including the in silico single residue mutations further supported the experimental findings and indicated the flipping of Tyr141 side chain governs the recognition of larger oligosaccharides. The galectin-8N-glycerol complex highlighted the minimum atomic features required for ligand recognition. Taken together, these crystal structures formed the basis to design interaction-based filters that guided the structurebased virtual screening.
A ligand design campaign was initiated by virtual screening a library of noncarbohydrate-small molecules using rigorous interaction-based criteria (Chapter 3). Available structural information including that generated in Chapter 2 and the preference of galectin-8N towards anionic saccharide was at the center of the screening. A library of compounds through iterative docking and molecular dynamics simulations was narrowed down to a small subset of molecules. The top fraction of the in silico analysis was purchased and evaluated for binding through saturation transfer difference (STD) nuclear magnetic resonance (NMR) and X-ray crystallography. These compounds did not bind to galectin-8N in our STD and X-ray experiments. However, the simulation outcome of one of the purchased compound provided the basis for exploiting unique amino acid residues in the galectin-8N binding site for ligand design.
Continuing the quest for identifying the ligands against galectin-8N, galactose as the core scaffold was taken forward for ligand design, primarily due to its inherent ability to interact with the conserved amino acid residues in the binding site. Based on our simulation results in Chapter 3 and available literature data, MB46A (compound 6) was designed. It contains a propionic acid side chain that is ether-linked to the 3′-position of galactose, the idea being to mimic the interactions of the carboxylic acid portion of a high-affinity 3′-O-sialylated lactose (3′-SiaLac) (Chapter 4). The designed molecule was synthesised and shown to bind (139 μM) galectin-8N through isothermal titration calorimetry (ITC) and X-ray crystallography. Our crystal structure confirmed the hypothesis and prompted for furthering the ligand design using compound 6 as a template molecule.
The encouraging outcome from the ligand design exercise so far in combination with molecular dynamics-based examination was employed to develop a library of compounds (Chapter 5). After considering the synthetic feasibility and ligand physicochemical properties, two closely related molecules (MB61B and MB63N) to the library of compounds were synthesised and experimentally evaluated. MB61B and MB63N contain benzoyl and napthoyl group respectively that are ester-linked to the 3′-position of the galactose (Chapter 5). The ligands were confirmed to be binding to galectin-8N by SPR (MB61B - 123.6 μM and MB63N - 124.4 μM). The obtained structure-activity insights will guide towards progressing the ligand design through ligand optimisation.
Overall the research presented in this thesis, demonstrate the successful rational medicinal chemistry application towards exploring the structure-activity landscape of galectin-8N. The information generated provides insight into the involvement of key binding site residues in recognising natural and synthetic ligands. Of encouragement is that the designed molecules (outlined in Chapter 4 and Chapter 5) inhibited galectin-8 mediated cell migration (preliminary results from our collaborator Prof. Y. Zick, Weizmann Institute, Israel) and are undergoing extensive dose-response analysis and possibly followed by in vivo investigations. These exciting preliminary in vitro results partially highlights the overall success of our ligand design campaign and further encourages the development of these leads into potent, efficient and selective ligands targeting galectin-8.||