Understanding SARS-CoV-2 spike proteins in vaccines

Characterisation and quality control of spike is a relatively small but - in our view - essential contribution to [fighting the pandemic]

I’ve been working in mass spectrometry for the last 24 years and I’m currently the head of mass spectrometry at the Centre for Medicines Discovery at Oxford University where I’m in charge of protein characterisation. The proteins we make here are used in drug discovery.

At this stage, we rely on a multi-pronged approach to fighting the pandemic. This approach includes epidemiology, diagnostics and the development of many different potential drugs and vaccines. Characterisation and quality control of spike is a relatively small but - in our view - essential contribution to this work.

Q: How is your team investigating spike proteins in relation to SARS-CoV-2?

A: In March 2020, the Centre for Medicines Discovery (CMD) was tasked with producing the SARS-CoV-2 spike protein for laboratories working on the development of diagnostic tests, vaccines and therapeutic antibodies. We quickly realised that there was significant variation in the glycosylation pattern, that this might be functionally important, and that the molecular characterisation of spike glycans would be technically challenging. At the same time, our longstanding collaborative partner, Agilent Technologies, offered to assist us in developing suitable analytical methods.

Although in-depth analysis of spike glycosylation was achieved early-on in the pandemic, few labs have the equipment, expertise or time to duplicate it. We are a high-throughput lab and we needed a method for analysis of spike batches that would be simple, reliable and above all fast. We also wanted a method that would be accessible to non-specialists.

Q: Why is the SARS-C0V-2 viral spike protein important?

A: We believe the spike protein has two essential functions for the virus. The first one is it allows it to dock. The virus needs to be able to recognise a suitable host cell and dock onto it so it can be taken in for the viral replication cycle to start.

The second function is that the spike protein is covered with complex glycan sugars, which we think is how it hides it from the immune system. Like all viruses, SARS-CoV-2 hijacks the biosynthetic machinery in the host cell, including glycosylation, allowing it to mimic the cell’s own surface glycans. We believe this is what makes it so pathogenic; and makes it such a deadly disease.

These complex sugars are difficult for the immune system to recognise as foreign and if the spike protein can’t be identified as foreign then it can replicate within cells with impunity. We know that glycosylation is functionally important in viruses, and other pathogens, and we believe the unusually heavy glycosylation of spike enables SARS-CoV-2 to evade the immune system.

The cell lines used for expression of spike in our lab, and in others, are not the same as the lung epithelium cells that are the primary target for spike in humans. It’s therefore important to know exactly which glycans are expressed and how abundant they are at the location site on the spike protein. The degree of glycosylation not only affects the functional properties of spike but also its physical properties, such as molecular weight and solubility.

Q: What technique did you use to characterise the SARS-CoV-2 spike protein?

A: We decided to generate and then analyse short glycopeptides as a means of gaining positional as well as structural information. An in-depth glycan discovery approach was used to generate a mass-retention time database. Once this had been populated, the much faster and simpler mass-retention time fingerprinting method was employed. A single enzyme, elastase, is used for overnight digestion, followed by a 65-minute LC-TOF-MS run.

The extracted data from this run is used to search the mass-retention time database, allowing glycans to be identified and localised by both accurate mass and HPLC retention time. During an MS/MS experiment, glycopeptides may not be selected for fragmentation, or may not fragment. LC-TOF-MS has the advantage that all ionisable species will generate a signal. In addition, this allows the method to run on any TOF mass spectrometer.

Q: What were the key findings of the in-depth glycan discovery research?

A: Using our reversed-phase HPLC separation, up to 27 glycoforms for each elastase glycopeptide eluted within a predictable four-minute retention window. We identified 140 spike glycopeptides belonging to 13 sites, with a further six sites unassigned. Our data compares favourably with nano LC-MS/MS based methods and is in general agreement with our intact mass measurement of the spike receptor binding domain. In addition, we were able to infer the accurate mass and retention time for a further 306 spike glycopeptides.

The database allows any LC-MS-equipped laboratory to perform rapid glycan characterisation for quality control on different batches of spike protein, requiring no specialised glycobiology expertise or software. This is the first use of the mass-retention time fingerprinting approach, but given the importance of glycan characterisation in biopharmaceuticals, we believe it could become widely adopted in good laboratory practice (GLP) and good manufacturing practice (GMP) settings.

Q: Do you plan to further this research into rapid glycan characterisation?

A: We would like to know if the patterns of glycosylation observed at well-characterised sites can be used to generalise regarding other less well-characterised sites. We also observed that elastase appears to favour cleavage at the C-terminus of an N-linked glycan consensus NX(S/T). We are interested to know if this phenomenon is a general one and whether this reveals something of interest regarding the mechanism of elastase.

I believe the more we know about spike glycans, how they're made, where the glycans occur and their profile and how that profile varies, the better antibody diagnostics we can produce. It may possibly allow us to generate better therapeutic antibodies, perhaps even better vaccines and drugs.

Author: Dr Rod Chalk is head of mass spectrometry at the Centre for Medicines Discovery at Oxford University: cmd.ox.ac.uk