key: cord-0973230-y03xz4ue
authors: Tada, Takuya; Dcosta, Belinda M.; Samanovic-Golden, Marie; Herati, Ramin S.; Cornelius, Amber; Mulligan, Mark J.; Landau, Nathaniel R.
title: Neutralization of viruses with European, South African, and United States SARS-CoV-2 variant spike proteins by convalescent sera and BNT162b2 mRNA vaccine-elicited antibodies
date: 2021-02-07
journal: bioRxiv
DOI: 10.1101/2021.02.05.430003
sha: e648a547bedba4a23667b7f3ce99b104c9ea75a4
doc_id: 973230
cord_uid: y03xz4ue

The increasing prevalence of SARS-CoV-2 variants with mutations in the spike protein has raised concerns that recovered individuals may not be protected from reinfection and that current vaccines will become less effective. The B.1.1.7 isolate identified in the United Kingdom and B.1.351 isolate identified in the Republic of South Africa encode spike proteins with multiple mutations in the S1 and S2 subunits. In addition, variants have been identified in Columbus, Ohio (COH.20G/677H), Europe (20A.EU2) and in domesticated minks. Analysis by antibody neutralization of pseudotyped viruses showed that convalescent sera from patients infected prior to the emergence of the variant viruses neutralized viruses with the B.1.1.7, B.1.351, COH.20G/677H Columbus Ohio, 20A.EU2 Europe and mink cluster 5 spike proteins with only a minor decrease in titer compared to that of the earlier D614G spike protein. Serum specimens from individuals vaccinated with the BNT162b2 mRNA vaccine neutralized D614G virus with titers that were on average 7-fold greater than convalescent sera. Vaccine elicited antibodies neutralized virus with the B.1.1.7 spike protein with titers similar to D614G virus and neutralized virus with the B.1.351 spike with, on average, a 3-fold reduction in titer (1:500), a titer that was still higher than the average titer with which convalescent sera neutralized D614G (1:139). The reduction in titer was attributable to the E484K mutation in the RBD. The B.1.1.7 and B.1.351 viruses were not more infectious than D614G on ACE2.293T cells in vitro but N501Y, an ACE2 contacting residue present in the B.1.1.7, B.1.351 and COH.20G/677H spike proteins caused higher affinity binding to ACE2, likely contributing to their increased transmissibility. These findings suggest that antibodies elicited by primary infection and by the BNT162b2 mRNA vaccine are likely to maintain protective efficacy against B.1.1.7 and most other variants but that the partial resistance of virus with the B.1.351 spike protein could render some individuals less well protected, supporting a rationale for the development of modified vaccines containing E484K.

Since the zoonotic transfer of SARS-CoV-2 to humans at the end of 2019, the virus has rapidly mutated to adapt to its new host. Such adaptations are a feature of viral zoonosis in which selective pressure drives viral proteins to be optimized for interaction with the host cell proteins of the new species. In addition, viral amino acid sequences are selected to escape the humoral and cellular adaptive immune responses which recognize a different set of epitopes. While all viral genes are subjected to evolutionary pressure, the viral envelope glycoprotein is selected both for optimal interaction with its cell surface receptor and for escape from neutralizing antibodies.

In January, 2020, a variant SARS-CoV-2 was identified in Germany and China with a D614G mutation 1 as compared to the presumptive zoonotic Wuhan isolate. By May, the variant had risen to a prevalence of >97% world-wide. Amino acid residue G614 lies in subdomain 2 near the S1:S2 processing site of the spike protein. The mutation was found to reduce S1 subunit shedding from virions leading to increased infectivity [2] [3] [4] and the viral isolate replicated to higher titers in the upper respiratory tract although was not associated with increased mortality. Additional variants with increasing prevalence were identified.

The B.1.1.7 lineage (VOC-202012/01) isolate which was identified in South East England, London and east of England [5] [6] [7] , was found to replicate with increased virus loads and to be increasing in prevalence with mathematical modeling suggested a 56% increase in transmissibility 6 . The B.1.1.7 lineage is defined by 23 mutations of which 8 are located in S (Δ69-70, Y144Del, N501Y, A570D, P681H, T716I, S982A and D1118H). N501Y is one of six contact residues with ACE2 8 and has been shown to increase affinity for ACE2 9 .

The Δ69-70, which was found in multiple independent lineages, increases viral infectivity and leads to immunoevasion in immunocompromised patients 10 . The P681H mutation lies adjacent to the furin cleavage site suggesting a possible role in spike protein processing.

In October, 2020, the B.1.351 lineage variant was identified in South Africa where it rapidly became the predominant circulating genotype 11 . The variant is more heavily mutated than B.1.1.7 with 9 mutations (L18F, D80A, D215G, L242-244del, R246I, K417N, E484K, N501Y and A701V) three of which (K417N, E484K and N501Y) are in the receptor binding domain (RBD). E484 and N501Y lie in the amino acid motif that directly contacts specific ACE2 residues. N501Y has been shown to enhance affinity of the spike protein for ACE2 by hydrogen bonding with ACE2 Y41 and is selected in a mouse model 12 .

K417N, while not contributing to ACE2 binding, is a key epitope for the binding of neutralizing antibodies, as is E484K, and thus these mutations may have been selected for evasion of the humoral response [13] [14] [15] [16] [17] . Based on phylogenetic tree branch-length, it has been suggested that the variant arose through the prolonged virus replication in an immunocompromised individual 10 .

Additional variants found to be circulating in the human population include 20A.EU2 which was identified in Spain 18 and later elsewhere in Europe. COH.20G/677H which was identified in Columbus, Ohio contains D614G, N501Y, Q677H but lacks the mutations present in the UK and South Africa variants suggesting an independent origin 19 .

In addition, isolates with variant spike proteins were found in domesticated minks in Denmark with the potential for transfer into humans 20 .

The rapid emergence of new viral variants is of concern both because of their increased transmissibility and the possibility that they may escape neutralizing antibodies. We report here on antibody neutralization of the European, UK, South Africa, Europe, Columbus, Ohio and mink spike variants by the sera of convalescent individuals and those vaccinated with BNT162b2. Neutralization was determined using lentiviral virions pseudotyped by the variant spike proteins. In addition, the infectivity, thermostability and affinity of the virions for ACE2 binding was determined. The results showed that the variants bound ACE2 with increased affinity and thermostability. The 

The trimeric SARS-CoV-2 spike protein is synthesized as a full-length precursor polypeptide that is processed by cellular proteases into S1 and S2 subunits (Fig 1A) . S1, which mediates cell attachment, consists of an amino-terminal domain, the RBD, a receptor binding motif (RBM) within the RBD that directly contacts ACE2, and subdomains 1 and 2. S2, which mediates membrane fusion, consists of a hydrophobic amino-terminal fusion peptide, heptad repeats 1 and 2, a hydrophobic transmembrane domain and a cytoplasmic tail. The location of point mutations and small deletions in the major SARS-CoV-2 variant spike proteins are shown (Fig 1A) . Mutations of concern are those lying in the RBD which is the binding site for most neutralizing antibodies and those within the RBM (453, 477, 484, 501) which directly contacts ACE2. Two mutations lie near the processing site and others are in the S1 NTD and S2. Whether the mutations act independently or in a coordinated fashion is not known and whether they were selected or are simply markers is not clear.

pseudotypes provide a rapid and accurate means to assess spike protein function.

Neutralizing antibody titers determined by lentiviral pseudotype assay closely mirror those measured by live SARS-CoV-2 assay 23 . To analyze the functional properties of the spike protein variants, we constructed cytomegalovirus (CMV) promoter-driven expression plasmids encoding the B.1.1.7, B.1.351, COH.20G/677H, 20A.EU2 and mink cluster 5 spike proteins or the component mutations, singly and in combination. Vector coding sequences were based on the Wuhan Hu-1 S gene with a deletion of the carboxy-terminal 10 19 amino acids that increases virion (Fig. 1B) . In this study, the D614G spike protein is considered "wild-type" and the variants tested contain G614. The expression vectors were used to generate lentiviral pseudotypes with a packaged genomic RNA encoding both GFP and nanoluciferase. Immunoblot analysis of the spike proteins showed that each was expressed in cells and incorporated into virions at levels comparable to wild-type D614G spike protein, with the exception of T716I and the fully mutated B.1.1.7 spike proteins which were expressed at lower levels ( Supplementary Fig. 1A and C) . Because Fig. 1B) as was also the case for the 20A.EU2 and COH.20G/677H spike proteins ( Supplementary Fig. 1D ).

The infectivity of virus with each of the variant spike proteins was tested by infection of ACE2.293T cells, a cell-line that expresses high levels of ACE2 with normalized amounts of pseudotyped viruses. Analysis of the B.1.1.7 variant and its component mutations showed that the single point mutations had little effect on infectivity (Δ69-70, Y144Del, N501Y, A570D, P681H and D1118H) except for T716I which was low and S982A which significantly increased infectivity (Fig. 1B) . The double mutant Δ69-11 70/N501Y and triple mutant Δ69-70/N501Y/P681H also had increased infectivity, suggesting that the two mutations coordinate to increase infectivity. Analysis of the B.1.351 proteins showed that the individual point mutations had similar infectivity while the full complement was slightly reduced (Fig. 1C) . Spike proteins with the minkassociated mutations were fully infectious except for the protein containing all 4 mutations (Δ69-70/Y453F/I692V/M1229F) which was 2-fold reduced in infectivity. The COH.20G/677H and 20A.EU2 variants were fully infectious (Fig. 1D) .

To determine whether differences in infectivity could be caused by effects of the mutations on the stability of the spike proteins or their incorporation into virions, we analyzed the spike proteins produced in transfected cell lysates and incorporated into virions. The transfected cell lysates were analyzed on an immunoblot probed for the spike protein S2, which allows for detection of the full-length spike protein and processed S1 protein, and for the HIV-1 capsid protein P24 as a means of normalizing for particle number. Analysis of B.1.1.7 showed that each of singly mutated proteins was expressed in cells at similar levels and processed to a similar extent ( Supplementary Fig. 1A ). In contrast, some of the point mutations appeared to affect amount of spike on the virion and the extent of spike protein processing. T716I was expressed at significantly lower level, accounting for the decreased infectivity of this spike protein. P681H was present at high copy number but processed more efficiently than wild-type as demonstrated by a lower ratio of full-length to S2 protein level. The N501Y and A570D mutations resulted in a small decrease in copy number on virions. When combined with Δ69-70, processing returned to wild-type level. Addition of the P681H mutation in the triple mutant increased (Fig. 2B) . This was also the case for COH.20G/677H, 20A.EU2 and mink cluster 5 spike proteins which were more easily neutralized than D614G virus. A detailed analysis of one donor serum chosen at random showed that the 13 sera neutralized B.1.1.7 and its constitutive point mutations similarly, with the exception of T716I that was more easily neutralized than D614G (Supplementary Fig. 2) .

Analysis of B.1.351 and its constituent E484K point mutation showed that both viruses were neutralized by convalescent sera with titers similar to that of D614G ( Fig.   3A and B) . For some donors, the neutralization curves were virtually identical (donors 1, 

The D614G mutation, which was first identified in January, 2020, caused a significant increase in viral infectivity and binding to ACE2. We previously reported on increased binding using an in vitro assay in which spike protein-pseudotyped virus was incubated with beads coated with soluble ACE2 (sACE2) 25 . Here, we used a more sensitive assay in which the pseudotyped viruses were incubated with free sACE2 and then allowed to infect ACE2.293T cells. In this assay, an increased sensitivity to sACE2 neutralization indicated higher affinity of the spike protein for ACE2. An analysis of viruses by this approach showed that the B.1.1.7 spike protein itself did not cause an increase in ACE2 binding; however, spike proteins containing the single N501Y mutation, as well as those that included N501Y (Δ69-70/N501Y and Δ69-70/N501Y/P681H), showed a significant increase in ACE2 binding compared to D614G (Fig. 5A) . Analysis of B.1.351 showed that 15 it had increased ACE2 binding (Fig. 5A) . The increase was due to N501Y as none of the other point mutations had an effect. The COH.20G/677H spike protein, which also contains N501Y, displayed increased ACE2 binding (Fig. 5C) . Analysis of the proteins in the ACE2 binding assay in which virions were incubated with matrix-bound sACE2 confirmed that those spike proteins that contained N501Y had increased affinity for ACE2 (N501Y, Δ69-70/N501Y, COH.20G/677H and B.1.351). (Fig. 5D) Transmissibility of the viruses is likely to be affected by their stability in aerosols and resistance to high temperature. A test of heat resistance over a short time period of viruses with spike proteins containing the critical mutations showed a gradual decrease in infectivity over 1 hour with increasing temperature (Supplementary Fig. 4) . Incubation of the virus for 1 hour at 50°C caused a 40-fold decrease in infectivity of the D614G virus.

Viruses with N501Y, S982A, Δ69-70/N501Y/P681H spikes and B.1.351 decreased their infectivity less than 20-fold, suggesting that the mutations increase spike protein stability.

The emergence of SARS-CoV-2 variants with mutations in the spike protein has raised concerns as to whether antibodies elicited by primary infection and by vaccination will remain protective. We show that convalescent sera from individuals who had been As SARS-CoV-2 continues to rapidly evolve following its zoonotic transfer to humans, it is likely that novel variants will continue to emerge. These may have mutations in the spike protein that are selected both for escape from humoral responses and for increased transmissibility resulting from further increases in ACE2 affinity or increased stability. Our study, and others recently reported, demonstrate the importance of worldwide surveillance of circulating viruses by nucleotide sequencing and for monitoring novel anti-GAPDH mAb (Life Technologies) followed by goat anti-mouse HRP-conjugated second antibody (Sigma). The membrane was treated with luminescent substrate (Millipore) and the band intensities were quantified on an iBright CL1000 Imager.

As previously described 25 , serially diluted recombinant soluble ACE2 protein containing carboxy-terminal His-tag were mixed with 20 µl Ni-NTA beads for 1 hour at 4˚C. Unbound protein was removed by washing with PBS. The coated beads were then mixed with 40 µl pseudotyped lentiviral virions and incubated at 4˚C for 1 hour. The beads were then washed with PBS, resuspended in reducing Laemmle loading buffer, heated to 90°C and separated by SDS-PAGE and analyzed on an immunoblot probed with anti-p24 antibody (AG3.0) followed by goat anti-mouse HRP-conjugated secondly antibody. The membrane was analyzed as described above.

As previously described 25 , serially diluted recombinant soluble ACE2 protein were mixed with pseudotyped virus for 1 hour at room temperature. The virus was added on the target cells and incubated for 2 days. After 2 days, the medium was removed and Nano-Glo luciferase substrate (Nanolight) was added to wells. The luminescence was read in an 23 Envision 2103 microplate luminometer (PerkinElmer).

All experiments were performed in technical duplicates or triplicates and data were analyzed using GraphPad Prism (Version 8 501Y.V2

Evolutionary and structural analyses of SARS-CoV-2 D614G spike protein mutation now documented worldwide

The D614G mutation in SARS-CoV-2 Spike increases transduction of multiple human cell types

Naturally mutated spike proteins of SARS-CoV-2 variants show differential levels of cell entry. bioRxiv

The D614G mutation in the SARS-CoV-2 spike protein reduces S1 shedding and increases infectivity

Transmission of SARS-CoV-2 Lineage B.1.1.7 in England: Insights from linking epidemiological and genetic data. medRxiv

Estimated transmissibility and severity of novel SARS-CoV-2 Variant of Concern

Preliminary genomic characterisation of an emergent SARS-CoV-2 lineage in the UK defined by a novel set of spike mutations

Deep mutational scanning of SARS-CoV-2 receptor binding domain reveals constraints on folding and ACE2 binding

The Fc-mediated effector functions of a potent SARS-CoV-2 neutralizing antibody, SC31, isolated from an early convalescent COVID-19 patient, are essential for the optimal therapeutic efficacy of the antibody. bioRxiv

Recurrent emergence and transmission of a SARS-CoV-2 Spike deletion ΔH69/ΔV70. bioRxiv

Emergence and rapid spread of a new severe acute respiratory syndrome-related coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in South Africa. medRxiv

Adaptation of SARS-CoV-2 in BALB/c mice for testing vaccine efficacy

mRNA vaccine-elicited antibodies to SARS-CoV-2 and circulating variants. bioRxiv

Comprehensive mapping of mutations to the SARS-CoV-2 receptor-binding domain that affect recognition by polyclonal human serum antibodies. bioRxiv

Escape from neutralizing antibodies by SARS-CoV-2 spike protein variants

ER stress-induced upregulation of NNMT contributes to alcoholrelated fatty liver development

We thank NYU Langone Vaccine Center staff Sara Hyman, Mahnoor Ali, Lisa Zhao, Heekoung Youn, Jimmy Wilson, Trishala Karmacharya, Joseph Allen, Sophie Gray-Galliard for their contributions. The work was funded by grants from the NIH to N.R.L. 

The authors declare no competing interests.