Abstract:
Polysialic acid (polySia) is a polymeric glycan made of α2,8-linked sialic acid moieties. It is found as a posttranslational modification on few proteins throughout the human body, but its prevalence is upregulated during embryogenesis and very low in adults. In addition, it is found on the surface of some tumours. Due to its low prevalence in healthy adults, polySia makes a promising target for gene therapy against these tumours. To date, the recognition of polySia is poorly understood and there is only limited structural data. One of the structures is of the human adenovirus 52 (HAdV-52) short fibre knob (SFK), which binds polySia, with a degree of polymerisation (DP) of three or more, at its non-reducing end. Adenoviruses have been previously used for gene therapy and vaccines, making HAdV-52 a promising vector. However, the affinity for polySia is only in the low millimolar range. Lenman et al. found a single point mutant, K349R, with increased affinity for polySia. In this thesis, multiple additional mutants were designed to further increase the affinity towards polySia. Most mutants, including K349R, were structurally characterised in a ligand-bound form. Out of all mutants, only the mutant K349R forms some additional contacts to polySia, thereby pulling it closer to the SFK surface, which explains the increased affinity. To evaluate the effects of the mutations on polySia binding, I also aimed at determining a dissociation constant (KD) with several different methods including surface plasmon resonance spectroscopy, biolayer interferometry and microscale thermophoresis (MST).
PolySia is recognised by the sialic acid binding immunoglobulin-like lectins (Siglecs) 11 (human) and E (mouse), which upon binding downregulate the immune system to prevent an autoimmune response. Siglecs are membrane proteins, and their extracellular part contains an N-terminal V-set domain and several C2-set domains. Usually, Siglecs recognise their ligand with their N-terminal V-set domain. Because polySia can form long polymers, it is unclear whether the C2-set domains also take part in binding to polySia. To explore this, different Siglec-E constructs, comprising different numbers of extracellular domains, were designed and an expression and purification protocol was established in this thesis. Even after multiple improvements to the expression and purification protocol, the yields from each purification were low. Siglec-E is a glycoprotein, but the flexible glycosylation of proteins can prevent crystallisation, thus multiple methods to deglycosylate Siglec-E were explored here. Unfortunately, crystallisation trials with several different Siglec-E proteins, and variants thereof, did not yield any protein crystals to date. I encountered the same problems with determining the affinity between polySia and Siglec-E, as for HAdVs-52 SFK, hence, no KD was determined. I could show with saturation transfer difference-nuclear magnetic resonance spectroscopy that Siglec-E binds to polySia DP5.Because Siglec-E has multiple possible ligands beside polySia, MST experiments with other glycans were conducted, and were successful in case of 3’-sialyl Lewis X (3’sLeX). Siglec-E binds to 3’sLeX with low millimolar affinity and the glycosylation on Siglec-E has a drastic positive effect on the affinity. In contrast, the dimerization of Siglec-E and the presence of the second C2-set affects the affinity only slightly.
The purification of Siglec-11 and an unliganded structure of the V-set domain was already established. In this thesis, new Siglec-11 V-set crystals were observed. The structure revealed that the putative polySia binding site is blocked by a crystal contact. In addition, the crystal packing was very tight, making Siglec-11 crystals unsuitable for soaking. Therefore, no structure of Siglec-11 with polySia could be obtained.
In summary, this thesis highlights the difficulties of characterising glycan-protein interactions, both in structural characterisation as well as affinity determination. Nevertheless, the insights of this thesis provide a starting point for future research.