The 'Sweet Tooth' of Corona Viruses

The researchers at the Department of Chemical Biology and Drug Discovery (CBDD) and Veterinary Science, Utrecht University, joined hands to discover that common cold coronaviruses employ a special sugar molecule linked to lipids, termed gangliosides, to initiate invasion. The team also managed to depict the interaction between viruses and sugars at atomic scale using biophysical and computational approaches. The scientific results were published in Nature Chemistry¹ and Journal of American Chemical Society.²

The groundwork for this advancement was the assembly of a type of highly complex and sensitive carbohydrate molecules by the team of Prof. Dr. Geert-Jan Boons at CBDD, termed O-acetylated sialosides, which were once considered ‘almost impossible to synthesize’.³ First author of the two publications, Zeshi Li (Jack) cracked the decade long challenge by combining the power of chemistry and biocatalysts derived from viruses (Figure 1). These synthetic sugars were printed on a glass slide with a special coating as microarrays, which allowed Jack and his teammate, Dr. Lin Liu (CCRC, University of Georgia, USA) to systematically investigate how proteins from coronaviruses interact with these molecules.

These proteins include the spike protein (Figure 2), which is responsible for coronaviruses to latch onto host cell surface to initiate invasion, and the hemagglutinin-esterase (HE), which not only helps the virus to traverse through the mucus layer before they find host cells, but also serves to release the virus progeny from the cell for further invasion. Jack and Dr. Lin Liu were surprised to find that the spikes proteins of common cold causing coronaviruses OC43 and HKU1, as well as influenza C virus – another human virus causing mild respiratory symptoms, all bind highly selectively to the same O-acetylated sialoside, 9-O-acetylated α2,8-linked disialic acid.

However, closely related viruses infecting other animals do not show such binding. This particular sugar structure is found at the terminus of gangliosides, a sugar-bearing lipid which resides on the surface of the cell (Figure 3). Jack and Dr. Yifei Lang, graduated PhD at Veterinary Science on the team of Dr. Raoul de Groot, continued to hypothesized that these gangliosides may be the entry receptor for these viruses, with which the viruses burglarize into host cells particularly as these sugar-bearing molecules are known to mediate cellular update of biomacromolecules such as bacterial toxins.

To test out this hypothesis, Yifei genetically engineered cells to express gangliosides containing 9-O acetylated α2,8-linked disialic acid, the particular sugar structure strongly bound by OC43 and HKU1, on the cell surface. As expected, only when this sugar structure was present on the cell surface were the viruses able to invade the cells, whereas when the receptor is absent from the cell surface, the infection dropped substantially. Moreover, Jack and Yifei also examined formalin preserved respiratory tissues from human donors, and consistently with their hypothesis, 9-O-acetylated α2,8-linked disialic acid structure was found to be rich in these virus-target tissues. This collaboration has answered a long-standing question in coronavirus infections and identified a functional receptor for OC43 and HKU1. This knowledge will facilitate predicting future viral cross-species transmissions and will be important for the implementation of prevention and intervention strategies.

O-acetylated sialosides are very labile molecules and decays over time due to the presence of unstable ester bonds. In a follow-up study, Jack aimed to stabilize these molecules, as well as further simplifying the preparation of acetylated sialosides by making their structural analogs (Figure 4). He switched one or more oxygen atoms in the sialic acid residue into nitrogen (referred to as acetamido-deoxy analogs), such that the ester bond became an amide, which is much more stable to various chemical conditions.

However, these structural analogs were not compatible to the parent molecule, as the oxygen-to-nitrogen change was not tolerated by the coronaviral proteins. Jack teamed up with Dr. Luca Unione, post-doc of Boons Lab, now working at CIC bioGUNE, Spain to rationalize why these acetamido-deoxy analogs were not recognized by spikes and HEs from coronaviruses, despite the analog sugar and the parent O-acetylated sialosides being structurally very similar.

They used the HE as a model protein and mixed it with the acetamido-deoxy analog in a specialized tube. In another tube, the same HE was mixed with the parent molecule. By using a technique called ‘saturation transfer difference nuclear magnetic resonance’ (STD-NMR), they found the oxygen-to-nitrogen swapped part of the sugar severely disrupted the recognition of the HE, whereas the parent O-acetylated molecule docked into the HE protein like a plug into a socket (Figure 5). Based on these observations, they generated a structural model using molecular dynamics to depict how HE interplays with O-acetylated sialosides. From the structural model, they could conclude that the positioning of O-acetyl groups have to match with the hydrophobic pockets present in the HE. They expected the plug-in-socket model can be generalized to other proteins that bind to O-acetylated sialosides.

These studies have provided fundamental understanding of virus-sugar interaction and lend important insights into virus’ ability to adapt to specific pools of complex carbohydrates expressed in the corresponding host species.

References

1          Li, Z. et al. Synthetic O-acetylated sialosides facilitate functional receptor identification for human respiratory viruses. Nat. Chem. 13, 496-503, doi:10.1038/s41557-021-00655-9 (2021).

2          Li, Z. et al. Synthetic O-Acetylated Sialosides and their Acetamido-deoxy Analogues as Probes for Coronaviral Hemagglutinin-esterase Recognition. J. Am. Chem. Soc., doi:10.1021/jacs.1c10329 (2021).

3          Chen, X. & Varki, A. Advances in the Biology and Chemistry of Sialic Acids. ACS Chem. biol. 5, 163-176, doi:10.1021/cb900266r (2010).