SARS-CoV-2 immune evasion by the B.1.427/B.1.429 variant of concern

Description

C OVID-19 is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and is associated with acute respiratory distress syndrome, as well as extrapulmonary complications such as vascular thrombosis, coagulopathy, and a hyperinflammatory syndrome contributing to disease severity and mortality. SARS-CoV-2 infects target cells using the spike glycoprotein (S), which is organized as a homotrimer with each monomer comprising an S 1 and an S 2 subunit (1, 2). The S 1 subunit harbors the receptor-binding domain (RBD) and the N-terminal domain (NTD), as well as two other domains designated here as C and D (3,4). The RBD interacts with the angiotensin-converting enzyme 2 (ACE2) entry receptor on host cells through a subset of amino acids forming the receptorbinding motif (1,2,(5)(6)(7). The NTD was suggested to bind DC-SIGN, L-SIGN, and AXL, which may act as attachment receptors (8,9). Both the RBD and the NTD are targeted by neutralizing antibodies (Abs) in infected or vaccinated individuals, and some RBD-specific monoclonal Abs (mAbs) are currently being evaluated in clinical trials or are authorized for use in COVID-19 patients (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24). The S 2 subunit is the fusion machinery that merges viral and host membranes to initiate infection and is the target of Abs cross-reacting with multiple coronavirus subgenera because of its higher sequence conservation compared with the S 1 subunit (25)(26)(27)(28).
The ongoing global spread of SARS-CoV-2 led to the fixation of the D614G substitution (29,30), as well as to the emergence of a large number of viral lineages worldwide, including several variants of concern (VOC). Specifically, the B.1.1.7, B.1.351, and P.1 lineages that originated in the United Kingdom, South Africa, and Brazil, respectively, are characterized by the accumulation of mutations in the S gene as well as in other genes (31)(32)(33). Some of these mutations lead to marked reductions in the neutralization potency of several mAbs, convalescent sera, and Pfizer/ BioNTech BNT162b2-or Moderna mRNA-1273-elicited Abs (19,(34)(35)(36)(37)(38)(39)(40). The B.1.1.7 variant has become dominant worldwide because of its higher transmissibility (33), underscoring the importance of studying and understanding the consequences of SARS-CoV-2 antigenic drift.

Results
The incidence of the B.  We also analyzed plasma from nine convalescent donors who experienced symptomatic COVID-19 in early 2020 (and consequently were likely exposed to the Wuhan-1 or a closely related SARS-CoV-2 isolate) collected 15 to 28 days after symptom onset (table S2) figs. S1 and S2; and table S3). In several cases, the level of neutralizing activity against the VOC was found to be below the limit of detection.
These findings show that the three mutations present in the B1.427/B.1.429 S glycoprotein decrease the neutralizing activity of vaccine-elicited and infection-elicited Abs, suggesting that these lineage-defining residue substitutions are associated with immune evasion. However, these data also underscore the higher quality of Ab responses induced by vaccination compared with infection and their enhanced resilience to mutations found in VOC.

B.1.427/B.1.429 S mutations reduce sensitivity to RBD-and NTD-specific Abs
To evaluate the contribution of RBD and NTD substitutions to the reduced neutralization potency of sera from vaccinees and convalescent plasma, we compared the neutralizing activity of 34 RBD and 10 NTD mAbs against the D614 S or B.1.427/B.1.429 S variant using a VSV pseudotyping system (1,43).
We subsequently used local refinement to account for the conformational dynamics of the NTD and S2L20 relative to the rest of S and obtained a cryoEM reconstruction of the NTD bound to S2L20 at 3.0-Å resolution (Fig. 4C, fig. S5, and table S4). The structure revealed that the B.1.427/B.1.429 NTD antigenic supersite is severely altered. The N terminus is disordered up to residue 27, as is the supersite b-hairpin (disordered between residues 137 and 158) and the supersite loop (disordered between residues 243 and 264) (Fig. 4F). These structural changes explain the abrogation of binding and neutralization of the panel of NTD-specific mAbs evaluated.
Overlaying an ACE2-bound SARS-CoV-2 RBD structure with the B.1.427/B.1.429 variant S structure showed that the R452 residue points away from and does not contact ACE2, suggesting that this substitution would not affect receptor engagement (Fig. 4G). We next evaluated binding of the monomeric human ACE2 ectodomain to immobilized B.1.427/ B.1.429 and WT RBDs using surface plasmon resonance ( fig. S6, A    NTD variants using ELISA. The S13I signal peptide mutation dampened binding of five mAbs and abrogated binding of five additional mAbs of the 11 neutralizing mAbs evaluated ( Fig. 5A and fig. S7). Furthermore, the W152C mutation reduced recognition of six NTD-neutralizing mAbs, including a complete loss of binding for two of them, with a pattern complementary to that observed for S13I ( Fig. 5A and fig. S7). The B.1.427/B.1.429 S13I/W152C NTD did not bind to any NTD-directed neutralizing mAbs, which are known to target a single antigenic site (antigenic site I) (12), whereas binding of the non-neutralizing S2L20 mAb to the NTD antigenic site IV was not affected by any mutants, confirming proper retention of folding, as supported by the structural data ( Fig.   5A and fig. S7). Binding of vaccine-elicited plasma to NTD mutants confirmed and extended these observations with polyclonal Abs by showing an increasingly marked reduction in binding titers caused by the W152C, S13I, and S13I/W152C residue substitutions ( Fig. 5B and fig. S8). We previously showed that disruption of the C15/C136 disulfide bond that connects the N terminus to the rest of the NTD by mutation of either residue or alteration of the signal peptide cleavage site abrogates the neutralizing activity of mAbs targeting the NTD antigenic supersite (site I) (12). Because the S13I substitution resides in the signal peptide and is predicted to shift the signal peptide cleavage site from S13-Q14 to C15-V16, we hypothesized that this substitution indirectly affects the in-tegrity of NTD antigenic site I, which comprises the N terminus. Mass spectrometry analysis of the S13I and S13I/W152C NTD variants confirmed that signal peptide cleavage occurs immediately after residue C15 (Fig. 5, C to E). As a result, C136, which would otherwise be disulfide linked to C15, is cysteinylated in the S13I NTD because of the presence of free cysteine in the expression medium ( Fig. 5D  and fig. S9). Likewise, the W152C mutation, which introduces a free cysteine, was also found to be cysteinylated in the W152C NTD (Fig. 5F). It is not clear whether cysteinylation would occur during natural infection with S13I or W152C mutants alone or what contribution cysteinylation plays in immune evasion of S13I or W152C mutants alone. Dampening of NTDspecific neutralizing mAb binding is stronger for the S13I mutant than for the S12P mutant, which we previously showed also shifts the signal peptide cleavage site to C15-V16 (Fig. 5A). Conversely, we did not observe any effect on mAb binding of the S12F substitution, which has also been detected in clinical isolates, in agreement with the fact that this mutation did not affect the native signal peptide cleavage site (i.e., it occurs at the S13-Q14 position), as observed by mass spectrometry (Fig. 5G). In the absence of the C15-C136 disulfide bond, the N terminus is no longer stapled to the NTD, consistent with the structural data showing that the N terminus of the B.1.427/B.1.429 variant becomes disordered relative to the rest of the NTD (Fig. 4C).
Although the S13I and W152C NTD variants were respectively cysteinylated at positions C136 and W152C, the double mutant S13I/ W152C was not cysteinylated, suggesting that C136 and W152C had formed a new disulfide bond (Fig. 5, D to F). Tandem mass spectrometry analysis of nonreduced, digested peptides identified linked discontinuous peptides con-taining C136 and W152C ( fig. S9), confirming that a disulfide bond forms between C136 and W152C in the S13I/W152C NTD of the B.1.427/ B.1.429 variant. W152C is in the b-hairpin of the antigenic supersite, and the formation of a new disulfide bond with C136 would move residues in the b-hairpin >20 Å. The local structure of the b-hairpin was disordered in the B.1.427/ B.1.429 variant (Fig. 4C).
Collectively, these findings demonstrate that the S13I and W152C mutations found in the B.1.427/B.1.429 S variant are jointly responsible for escape from NTD-specific mAbs because of deletion of the two SARS-CoV-2 S N-terminal residues and overall rearrangement of the NTD antigenic supersite. Our data support that the SARS-CoV-2 NTD evolved a compensatory mechanism to form an alternative disulfide bond and that mutations of the S signal peptide occur in vivo in a clinical setting to promote immune evasion. The SARS-CoV-2 B.1.427/B.1.429 S variant therefore relies on an indirect and unusual neutralization escape strategy.

Discussion
Serum-or plasma-neutralizing activity is a correlate of protection against SARS-CoV-2 challenge in nonhuman primates (47,48), and treatment with several neutralizing mAbs has reduced viral burden and decreased hospitalization and mortality in clinical trials (10,14,15,22,23,49 (54), is suggestive of positive selection, which might result from the selective pressure of RBD-specific neutralizing Abs (55).
The SARS-CoV-2 NTD undergoes rapid antigenic drift and accumulates a larger number of mutations and deletions relative to other regions of the S glycoprotein (12,56). For instance, the L18F substitution and the deletion of residue Y144 are found in 8% and 26% of viral genomes sequenced and are present in the B.1.351/P.1 lineages and the B.1.1.7 lineage, respectively. Both of these mutations are associated with reduction or abrogation of mAb binding and neutralization (12,34). The finding that multiple circulating SARS-CoV-2 variants map to the NTD, several of them in the antigenic supersite (site I), suggests that the NTD is subject to a strong selective pressure from the host humoral immune response. This is further supported by the identification of deletions within the NTD antigenic supersite in immunocompromised hosts with prolonged infections (57)(58)(59) and in the in vitro selection of SARS-CoV-2 S escape variants with NTD mutations that decrease the binding and neutralization potency of COVID-19 convalescent patient sera or mAbs (12,34,60,61). The data herein showing immune evasion of all tested NTDspecific mAbs by the B.1.427/B.1.429 variant also support that the NTD antigenic supersite is under host immune pressure.
Similar to how the S13I/W152C mutations facilitate evasion of all tested NTD-specific mAbs, E484K causes broad resistance to many RBD-specific mAbs. The independent acquisition of the E484K mutation in the B.1.351, P.1, and B.1.526 variants and, more recently, in the B.1.1.7 variant (34), suggests this could also occur in the B.1.427/B.1.429 lineages. Indeed, four genome sequences with the E484K RBD mutation in the B.1.427 variant have recently been deposited in GISAID. Alternatively, the S13I/W152C mutations could emerge in any of these variants. The S13I mutation was recently detected in the SARS-CoV-2 B.1.526 lineage, which was originally described in (antigenic site i) and one non-neutralizing (antigenic site iv) NTD-specific mAbs to recombinant SARS-CoV-2 NTD variants analyzed by ELISA displayed as a heatmap.
(B) Binding of plasma Abs from vaccinated individuals to recombinant SARS-CoV-2 NTD variants analyzed by ELISA. The mean dilution factor for each mutant was compared by the one-way ANOVA test against WT (*P < 0.05, **P < 0.001). (C to G) Deconvoluted mass spectra of purified NTD constructs, including the WT NTD with the native signal peptide (C), the S13I NTD (D), the S13I and W152C NTD (E), the W152C NTD (F), and the S12F NTD (G). The empirical mass (black) and theoretical mass (red) are shown beside the corresponding peak. An additional 119 Da were observed for the S13I and W152C NTDs, corresponding to cysteinylation of the free cysteine residue in these constructs (as L-cysteine was present in the expression media). The cleaved signal peptide (blue text) and subsequent residue sequence (black text) are also shown based on the MS results. Mutated residues are shown in bold. Cysteines are highlighted in light orange (unless in the cleaved signal peptide), and disulfide bonds are shown as dotted light orange lines between cysteines. Residues are numbered for reference.
New York (62,63). Understanding the newly found mechanism of immune evasion in emerging SARS-CoV-2 variants, such as the signal peptide modification described herein, is as important as sequence surveillance itself to successfully counter the ongoing pandemic.