Defining the impact of flavivirus envelope protein glycosylation on sensitivity to broadly neutralizing antibodies

Antibodies targeting the so-called envelope dimer epitope (EDE) cross-neutralize Zika virus (ZIKV) and all four dengue virus (DENV) serotypes and have thus inspired an epitope-focused vaccine design against these flaviviruses. There are two EDE antibody subclasses (EDE1, EDE2) distinguished by their dependence on viral envelope (E) protein N-linked glycosylation at position N153 (DENV) or N154 (ZIKV) for binding. Here, we determined how E glycosylation affects neutralization by EDE and other broadly neutralizing antibodies. Consistent with structural studies, mutations abolishing the N153/N154 glycosylation site increased DENV and ZIKV sensitivity to neutralization by EDE1 antibodies. Surprisingly, these mutations also increased sensitivity to EDE2 antibodies although they occurred at predicted contact sites. Despite preserving the glycosylation site motif (N-X-S/T), substituting the threonine at ZIKV E residue 156 with a serine resulted in loss of glycan occupancy accompanied with increased neutralization sensitivity to EDE antibodies. For DENV, the presence of a serine instead of a threonine at E residue 155 retained glycan occupancy, but nonetheless increased sensitivity to EDE antibodies, in some cases to a similar extent as mutation at N153, which abolishes glycosylation. E glycosylation site mutations also increased ZIKV and DENV sensitivity to other broadly neutralizing antibodies, but had limited effects on ZIKV-or DENV-specific antibodies. Thus, E protein glycosylation is context-dependent and modulates the potency of broadly neutralizing antibodies in a manner not predicted by existing structures. Manipulating E protein glycosylation could be a novel strategy for engineering vaccine antigens to elicit antibodies that broadly neutralize ZIKV and DENV. IMPORTANCEAntibodies that can potently cross-neutralize Zika (ZIKV) and dengue (DENV) viruses are attractive to induce via vaccination to protect against these co-circulating flaviviruses. Structural studies have shown that viral envelope protein glycosylation is important for binding by one class of these so-called broadly neutralizing antibodies, but less is known about the determinants of neutralization. Here, we investigated how envelope protein glycosylation impacts broadly neutralizing antibody potency. By characterizing a panel of ZIKV and DENV variants encoding envelope protein glycosylation site mutations, we found that glycan occupancy was not always predicted by an intact N-X-S/T sequence motif. Moreover, envelope protein glycosylation status alters the neutralization potency of broadly neutralizing antibodies in a manner unexpected from their predicted binding mechanism as determined by existing structures. We highlight the complex role and determinants of envelope protein glycosylation that should be considered in the design of vaccine antigens to elicit broadly neutralizing antibodies.