Multivalent protein-carbohydrate interactions are crucial in many aspects of biology as they initiate the first contact between pathogens and host cells which ultimately leads to infection. Developing an understanding of the structures and binding modes used for these interactions will help develop specific, potent multivalent glycan inhibitors that can block such interactions, thereby preventing infection. This work focuses on two closely related tetrameric C-type lectins, DC-SIGN and DC-SIGNR, collectively abbreviated as DC-SIGN/R. These proteins are known to bind multivalently to multiple glycans, specifically mannose, found within the surface glycoproteins of many viruses, such as human immunodeficiency virus (HIV), Ebola Virus and West Nile Virus, and facilitate their infections. Despite significant research over the past 20 years, knowledge on the binding characteristics and structural mechanisms of these proteins remains limited because their tetrameric structures remain unknown. This project aims to develop novel glyconanoparticle probes to establish a valid structural model of the binding domains with these proteins and to reveal the mechanisms behind the multivalent glycan – DC-SIGN/R interactions. This objective has been achieved through the development of two sensitive fluorescent based assays, Fӧrster resonance energy transfer (FRET) and fluorescence quenching, using polyvalent glycan coated quantum dots (QDs) and gold nanoparticles (AuNPs) by exploiting their unique strong fluorescence and fluorescence quenching properties respectively. In particular, the biocompatible, low cytotoxic AuNP-saccharide conjugates make them potentially suitable for in vivo studies in the future.
This thesis has specifically designed a series of multifunctional glycan ligands for nanoparticle surface coating. Each ligand consists of three unique functional domains, a dihydrolipoic acid (DHLA) for strong nanoparticle capping, an oligo(ethylene glycol, EGn) linker for promoting hydrophilicity and reducing non-specific adsorptions, and a terminal mannose sugar moiety for specific DC-SIGN/R binding. Both the monosaccharide and disaccharide forms of a mannose sugar were investigated. Our group has previously found that the mono-mannose capped QDs have much weaker affinity for DC-SIGNR than that of DC-SIGN, suggesting that the binding sites in DC-SIGN/R orient differently, although their exact binding modes remain unclear as DC-SIGNR binding is too weak to be distinguished from monovalent CRD. In this thesis, by displaying the dimannose ligands on the QD surface to increase monovalent binding affinity and developing a novel multimodal readout strategy consisting of FRET and TEM imaging, the exact binding modes of DC-SIGN/R have been revealed. The four carbohydrate recognition domains in DC-SIGN face in the same direction and bind tetravalently to a single glycan-QD; while those in DC-SIGNR are split into a pair of back-to-back dimers and bind bis-divalently with two different saccharides-QDs. Furthermore, QD-saccharides capped with two different ligand series have been prepared and their binding affinities, Kds, with DC-SIGN/R have been investigated. Initial studies using QD-EGn- Saccharide (where n = 3 or 11, Saccharide = -Man and Man--1,2-Man) to study their DC-SIGN/R interactions revealed that dilution of the surface saccharide ligand using an inert zwitterionic ligand resulted in a decrease in binding. So a second dendritic glycan ligand series, DHLA-(EGn-Glycan)m, (where n = 1 or 2; m =1, 2, or 3 and Glycan = -Man and Man--1,2-Man) was synthesised to further increase the nanoparticle surface glycan density to investigate how this affects DC-SIGN/R binding.
Significant results were observed using these ligand series and the different scaffold materials. In general, reducing the ethylene glycol linker length and using DiMan ligand capped nanoparticles led to enhanced binding affinity with DC-SIGN. For example, DC-SIGN binds more strongly to QD-EG3-DiMan than QD-EG11-DiMan (Kd: 0.61 ± 0.07 v.s. 2.1 ± 0.5 nM) and its monosaccharide equivalent, QD-EG3-Man (Kd = 35 ± 7 nM). Interestingly, further increasing the glycan density on the nanoparticle surface using dendritic sugar ligands weakened their binding with DC-SIGN but enhanced the binding with DC-SIGNR. The Kd values for DC-SIGN binding to QD-EG2-DiMan and QD-(EG-DiMan)3 were 1.7 ± 0.1 and 1.59 ± 0.2 nM, while those for DC-SIGNR were 1.59 ± 0.2 and 0.49 ± 0.6 nM respectively. The same binding affinity trend of was also observed for AuNPs capped with such glycan ligands, which gave one of the smallest reported Kd values for DC-SIGN/R interactions. Finally the potency of these materials in inhibiting DC-SIGN mediated Ebola virus glycoprotein driven viral infection of target cells were investigated, giving impressively low IC50 values, e.g. 0.7 ± 0.2 nM for QD-EG3-DiMan and 0.2 ± 0.04 nM for AuNP-EG2-DiMan. These IC50 values make these materials to be the most potent glyco-nanoparticle inhibitors.
