Toward the Next-Generation COVID-19 Vaccines That Circumvent Antigenic Drift while Defusing Viral Infection

Abstract

The emergence of vaccine-resistant variants suggests a complicated endemic scenario in the vaccination aftermath for COVID-19. The situation prompts us to enquire whether the antigen adopted by extant vaccines, the trimeric spike (S) protein, is the optimal in the sense of inducing an immunity that leaves the virus with no evolutionary route of evasion. The patterns of glycosylation camouflage suggest that the answer is negative while also suggesting an alternative antigen that appears to be better optimized, eliciting an additional immune attack as the virus gets primed for cell penetration. This type of vaccine is expected to induce antibodies capable of defusing the virus during the priming phase while also circumventing antigenic drift.

A Complicated Postvaccination Scenario

While there are reasons to remain cautiously optimistic, the postvaccination scenario for COVID-19 is anyone’s guess. Even at the current early stages of vaccination, when selection pressure has not yet intensified, the virus is already selecting variants seemingly capable of dodging immune surveillance. In particular, the E484K mutation in combination with N501Y is suspected to partially neutralize vaccine-elicited immune responses.¹ The emergence of these resistant phenotypes suggests a complicated endemic scenario in the vaccination aftermath. The situation prompts us to enquire whether the antigen adopted by extant vaccines, the trimeric spike (S) protein,² is optimal in the sense of inducing an immunity that leaves the virus with no evolutionary route for evasion. The answer is probably negative. Here we argue that had vaccine developers known then what they know now, they would have proceeded differently.

In addressing the issue of antigen optimality, we noticed that there is an elephant in the room: SARS-CoV-2 is coated with a layer of sugars, a chemical modification known as glycosylation carried out by host enzymes that modulates or camouflages most vaccine epitopes from immune surveillance.^3,4 The extent of glycosylation was probably not known at the time when the vaccines were being designed but has now been demonstrated to be extensive.

There is one vaccine epitope that does not get camouflaged in the virus for obvious reasons; it is found in the receptor binding domain (RBD), a region in the S1 subunit of the S protein that is involved in anchoring the virus to the host cell.⁴ So, essentially, glycosylation funnels immune surveillance toward the RBD region, while vaccine-induced antibodies for the other spike epitopes are essentially neutralized in the virion. This leads us to conclude that, in retrospect, adopting the trimeric S protein as antigen may have not been the optimal choice. This is so because a large portion of vaccine epitopes sidetracks the induced immune response toward regions on the virion surface that are not susceptible of being recognized.

On the other hand, recent evolutionary outcomes during early stages of vaccination are showing that the virus is capable of impairing antibody recognition at the RBD region without decreasing (or perhaps even enhancing) viral affinity for the human receptor (ACE2), that is, without paying a significant fitness cost.^1,5 This constitutes a disturbing trend that somewhat diminishes confidence in the postvaccination outcome.

COVID-19 Vaccines That Avert Antigenic Drift while Defusing the Virus

Extant vaccines have focused on eliciting the immune attack toward one phase of the virus, the virion, with its intact capsid.² Yet the preactivation state, where the virus gets primed to enter the human cell, may offer additional epitopes with antibody exposure.⁶ The virus enters this phase when the host enzymology cleaves the junction between the two subunits S1, S2 of the S-protein, enabling each subunit to play its distinctive role in cell penetration: S1 as anchor and S2 as harpoon.⁶ Consequently, a set of epitopes in S1 previously occluded at the S1/S2 interface becomes exposed in the preactivation phase. Since the RBD is fully contained in S1, which will present additional antibody-exposed epitopes during the preactivation phase, why not adopt the S1 subunit as antigen, instead of the full trimeric S protein?

We can confirm the presence of immunogenic regions on S1 at the S1/S2 interface. B-cell epitopes in that region are identified using linear epitope predictors from the public domain.⁷ Thus, the FASTA file for the amino acid sequence of the free monomeric S unit may be inputted into the predictor at http://tools.iedb.org/main/bcell/. The inferred epitopes shown below are subsumed in the vaccine S1 antigen, and their targeting with vaccine-induced antibodies is expected to promote S1 shedding during the virus priming phase,⁶ as well as promoting virus annihilation. The amino acid sequence of the S1-vaccine epitopes (bold-italic) exposed during the priming phase is as follows:

₆₂₄IHADQLTPTWRVYSTGSNVFQTRAGCLIGAEHVNNSYECDIPIGAGICASYQTQTNSPSGAGSVASQ₆₉₀

The reliability of the predictions for epitopes 624–644, 655–690 is supported by the strength of the signal (residue score above 0.6) and the length (n > 20). Targeting these epitopes with the S1-vaccine should in addition promote destabilization of the quaternary assemblage of the spike protein during the priming phase, thereby defusing viral infection.

Adopting the S1 subunit as antigen may improve vaccine efficacy, since S1 not only incorporates additional epitopes with antibody exposure but also promotes a two-stage immune attack at two key junctures during infection. Thus, should the virus evolve a vaccine-resistant virion phase, as we suspect it is capable of doing,¹ a contingency attack is in place so the immune system can respond during the virus preactivation (cell-penetration priming) phase.

We are well aware that the virus is capable of mutating to retain (or even enhance) affinity toward the hACE2 receptor while partially neutralizing vaccine-induced antibodies targeting the RBD (Figure 1).¹ We also know it is also capable of partially recovering the quaternary structure at the S1/S2 interface during the penetration-priming phase via the D614G mutation.⁶ What we are betting on is that it will not do both things better and simultaneously as it builds up its evasion route from the S1-antigen vaccine (cf. Figure 1).

Figure 1.

Antigen optimization. Two phases of SARS-CoV-2 with different levels of vaccine-induced antibody exposure. The virion (left) presents a glycosylation camouflage on the spike protein that modulates immune surveillance. The exposed ACE2-RBD region is the largest epitope with maximum antibody accessibility. As the virus gets primed for cell penetration (right), proteolytic activation frees an antigenic region in the S1 subunit accessible to antibodies. Identified mutations in the virus variants have been shown to partially neutralize RBD-targeting antibodies (N501Y, E484K) or partially occlude the epitopes at the S1/S2 protein–protein interface (D614G) by partially restoring the spike quaternary structure disrupted when the virus is primed for cell penetration.

The likelihood of dodging a vaccine-induced attack with multiple epitopes that become exposed at different stages of infection is probably insignificant, in contrast with current vaccines that focus solely on eliciting an immune response toward the extensively glycosylated virion. We are thus advocating that COVID-19 vaccines adopt the S1 subunit as an optimized antigen with a better chance of averting the selection of resistant variants while also being capable of eliciting an immune response that will defuse the virus during the penetration-priming phase. In this way, we advocate for the next-generation COVID-19 vaccines that circumvent antigenic drift while also defusing viral infection.

The author declares no competing financial interest.