Main

The evolution of RNA viruses can result in immune escape and modulation of binding to host receptors through the accumulation of mutations5. Previously emerged SARS-CoV-2 variants of concern (VOCs) have developed resistance to neutralizing antibodies, including some clinical antibodies that are used as therapeutics6,7,8. The B.1.351 (Beta) VOC showed the greatest magnitude of immune evasion from serum neutralizing antibodies6,7, whereas B.1.617.2 (Delta) quickly outcompeted all other circulating isolates through the acquisition of mutations that enhanced transmission and pathogenicity9,10,11 and eroded the neutralizing activity of antibody responses9.

The Omicron (B.1.1.529) variant was first detected in November 2021, was immediately declared to be a VOC by the World Health Organization (WHO) and quickly rose in frequency worldwide. The Omicron variant is substantially mutated compared to any previously described SARS-CoV-2 isolates, including 37 substitutions of residues in the spike protein in the predominant haplotype (Fig. 1a, Extended Data Figs. 14). Fifteen of the Omicron mutations are clustered in the RBD, which is the main target of neutralizing antibodies after infection or vaccination12,13, suggesting that Omicron might escape infection- and vaccine-elicited antibodies and therapeutic monoclonal antibodies. Nine of these mutations map to the receptor-binding motif (RBM), which is the RBD subdomain that directly interacts with the host receptor, ACE214.

Fig. 1: The Omicron RBD shows increased binding to human ACE2 and gains binding to mouse ACE2.
figure 1

a, Omicron mutations are shown in a primary structure of the SARS-CoV-2 spike protein, with domains and cleavage sites highlighted. BH, beta hairpin; C, C domain; CD, connector domain;  CH, central helix; D, domain D; FP, fusion peptide;HR1/2, heptad repeat 1/2; SH, stem helix; SP, signal peptide; TM, transmembrane domain;UH, upstream helix. b, Single-cycle kinetics SPR analysis of ACE2 binding to six RBD variants. ACE2 is injected successively at 11, 33, 100 and 300 nM (human) or 33, 100, 300 and 900 nM (mouse). Dashed black curves show fits to a 1:1 binding model. White and grey stripes indicate association and dissociation phases, respectively. KD, dissociation constant; RU, response units. c, Quantification of human ACE2 binding data (mean ± s.d. of three replicates). Asterisks indicate that Delta was measured in a separate experiment with a different chip surface and capture tag. Delta fold change (FC) is calculated relative to the affinity of Wuhan-Hu-1 measured in parallel (91 ± 1.6 nM). d, Entry of Wuhan-Hu-1, Alpha, Beta, Delta, Gamma, Kappa and Omicron spike VSV pseudoviruses into mouse ACE2-expressing HEK293T cells. RLU, luciferase relative light units. Shown are two biological replicates (technical triplicates). Lines, geometric mean.

Preliminary reports indicate that the neutralizing activity of plasma from individuals who had received the Pfizer–BioNTech BNT162b2 vaccine is reduced against the Omicron variant15,16, documenting a substantial—albeit not complete—escape from mRNA-vaccine-elicited neutralizing antibodies. Another report also shows that vaccine effectiveness against symptomatic disease induced by the Omicron variant is significantly lower than for the Delta variant17. The potential for booster doses to ameliorate this decline in neutralization is being investigated. In addition, the neutralizing activity of several therapeutic monoclonal antibodies appears to be decreased or abolished against Omicron16,18.

To understand the consequences of the high number of mutations found in the Omicron spike protein, we used a pseudovirus assay to study receptor use and neutralization mediated by monoclonal and polyclonal antibodies, as well as surface plasmon resonance (SPR) to measure binding of the RBD to human and mouse ACE2 receptors.

Omicron RBD binds ACE2 with increased affinity

At present, 23 out of the 37 amino acid mutations in the Omicron spike protein have been individually observed previously in SARS-CoV-2 variants of interest, VOCs or other sarbecoviruses, whereas the remaining 14 substitutions have not to our knowledge been described before (Extended Data Fig. 5a). Analysis of the GISAID database indicates that there are rarely more than 10–15 Omicron spike mutations present in a given non-Omicron haplotype or Pango lineage (Extended Data Fig. 5b–d). Although we have not formally assessed the possibility of recombination events, persistent replication in immunocompromised individuals or inter-species ping-pong transmission5 are possible scenarios for the rapid accumulation of mutations that could have been selected on the basis of viral fitness and immune evasion.

Several of the Omicron RBD mutations are found at positions that are key contact sites with human ACE2, such as K417N, Q493R and G496S19. Except for N501Y, which increases ACE2-binding affinity by sixfold20,21, all other substitutions were shown by deep mutational scanning either to reduce binding or to have no effect on human ACE2 affinity when present individually22, resulting in an overall predicted decrease of binding affinity (Supplementary Table 1). However, we found that the Omicron RBD has a 2.4-fold increased binding affinity to human ACE2 (Fig. 1b, c, Extended Data Fig. 6a), suggesting epistasis of the full constellation of RBD mutations. It remains to be determined whether and how the spike mutations in Omicron may influence the dynamics of RBD opening, which may also affect the engagement of the RBD with ACE2.

The presence of the N501Y mutation has previously been reported to enable some SARS-CoV-2 VOCs to infect mice23. As Omicron contains the N501Y mutation, along with 14 other RBD mutations, we investigated whether the Omicron RBD binds mouse ACE2 using SPR (Fig. 1b, Extended Data Fig. 6). The Omicron RBD binds mouse ACE2 with a 1:1 binding affinity of 470 nM (Fig. 1b), whereas weak binding of the Beta RBD and very weak binding of the Alpha RBD to mouse ACE2 was observed (Fig. 1b, Extended Data Fig. 6b), consistent with previous reports23,24. Conversely, our assay did not detect any binding of the Wuhan-Hu-1, Delta or K417N RBDs to mouse ACE2. The enhanced binding of the Omicron RBD to mouse ACE2 is likely to be explained by theQ493R substitution, which is similar to the Q493K mutation isolated in mouse-adapted SARS-CoV-219. Our binding data correlate with ourobservation of Omicron spike protein-mediated but not Wuhan-Hu-1/G614 spike protein-mediated entry of VSV pseudoviruses into mouseACE2-expressing cells (Fig. 1d), as recently reported25. Collectively, these findings highlight the plasticity of the SARS-CoV-2 RBM, which in the case of the Omicron VOC acquired enhanced binding to human and mouse ACE2 orthologues, relative to other SARS-CoV-2 isolates. The influence of these findings on viral load and replication kinetics in humans and animal models remains to be evaluated, owing to the interplay of additional factors besides receptor binding. Preliminary data suggest that Omicron appears to be attenuated in some laboratory mouse strains (M.S.D., personal communication) and that it replicates less efficiently in human lung tissue as compared to Delta26.

Extent of Omicron escape from plasma antibodies

To investigate the magnitude of immune evasion that is mediated by the 37 mutations present in the Omicron spike protein, we used Wuhan-Hu-1 and Omicron spike VSV pseudoviruses and compared plasma neutralizing activity in different cohorts of convalescent individuals (that is, individuals who had recovered from COVID-19) or individuals who had been vaccinated with six major COVID-19 vaccines (mRNA-1273, BNT162b2, AZD1222, Ad26.COV2.S, Sputnik V and BBIBP-CorV) (Fig. 2, Supplementary Figs. 13, Extended Data Table 1).

Fig. 2: Neutralization of Omicron SARS-CoV-2 VSV pseudovirus by plasma from convalescent and vaccinated individuals.
figure 2

Plasma neutralizing activity in convalescent or vaccinated individuals (mRNA-1273, BNT162b2, AZD1222, Ad26.COV2.S (single dose), Sputnik V and BBIBP-CorV). a, Pairwise neutralizing antibody titres (half-maximum inhibitory dose; ID50) against Wuhan-Hu-1 (D614G), Beta and Omicron VOCs, and SARS-CoV. Vero E6-TMPRSS2 cells were used as target cells. Data are the geometric mean of n = 3 biologically independent experiments except for SARS-CoV for which = 1. b, Pairwise neutralizing antibody titres of plasma (ID50) against Wuhan-Hu-1 and Omicron. Data are the geometric mean of n = 2 biologically independent experiments. c, Plasma neutralizing activity in individuals undergoing dialysis who received three doses of either the BNT162b2 or mRNA-1273. Pairwise neutralizing antibody titres of plasma (ID50) against Wuhan-Hu-1 and Omicron. One representative experiment out of two is shown. Vero E6 cells were used as target cells in b, c, Line, geometric mean of 1/ID50 titres. Shown is the percentage of samples that lost detectable neutralization against Omicron or SARS-CoV, excluding  samples with 1/ID50 below the limit of detection. The demographics of enrolled donors are provided in Extended Data Table 1. Statistical significance is set as P < 0.05 and P values are indicated with asterisks (*P = 0.033; **P = 0.002; ***P < 0.001), using a paired two-sided t-test (Wilcoxon rank test).

Convalescent individuals and individuals who had been vaccinated with Ad26.COV2.S (single dose), Sputnik V or BBIBP-CorV had no detectable neutralizing activity against Omicron, with only one convalescent subject and three BBIBP-CorV vaccinees exhibiting ID50 values above 10 (Fig. 2a). Individuals who were immunized with mRNA-1273, BNT162b2 and AZD1222 had more potent neutralizing activity against Wuhan-Hu-1 and retained detectable neutralization against Omicron, with decreases of 39-, 37- and 21-fold, respectively (Fig. 2a). The dampening of neutralizing activity against Omicron was comparable to that observed against SARS-CoV, a virus that differs from Wuhan-Hu-1 by 52 residues in the RBD. Reductions of neutralization potency were less pronounced in vaccinated individuals who had been previously infected (fivefold; Fig. 2b) and in individuals undergoing dialysis (fourfold; Fig. 2c) who were boosted with a third mRNA vaccine dose. In the same cohort of individuals undergoing dialysis, the levels of antibodies that neutralize the vaccine-matched Wuhan-Hu-1 strain were found to be low (less than 1/100) or undetectable in 44% of individuals after the second dose of mRNA vaccine27.

Collectively, these findings provide evidence of a substantial reduction in plasma neutralizing activity against Omicron as compared to the ancestral virus, with neutralizing activity probably falling below the protective threshold in several cases28. Our data further indicate that multiple exposures to the ancestral virus through infection or vaccination result in the production of antibodies that can neutralize divergent viruses, such as Omicron or even SARS-CoV, as a consequence of affinity maturation or epitope masking by immune-dominant RBM antibodies28,29,30.

Broadly neutralizing antibodies inhibit Omicron

Neutralizing monoclonal antibodies with demonstrated in vivo efficacy in the prevention or treatment of COVID-1931,32,33,34,35,36,37 can be divided into two groups on the basis of whether they do or do not block the binding of the spike protein to ACE2. Of the eight currently authorized or approved monoclonal antibodies, seven (LY-CoV555, LY-CoV016, REGN10987, REGN10933, COV2-2130, COV2-2196 and CT-P59; all synthesized on the basis of publicly available sequences) block the binding of spike protein to ACE2 and are often used as two-antibody cocktails8. They bind to epitopes that overlap with the RBM (Fig. 3a), which is structurally and evolutionary plastic38, as shown by the accumulation of mutations throughout the pandemic and the genetic diversity of this subdomain among ACE2-using sarbecoviruses39. Combining two such ACE2-blocking monoclonal antibodies can provide greater resistance to variant viruses that carry RBM mutations31. The second class of monoclonal antibodies, represented by sotrovimab, do not block ACE2 binding but neutralize SARS-CoV-2 by targeting non-RBM epitopes that are shared across many sarbecoviruses, including SARS-CoV4,40.

Fig. 3: Neutralization of Omicron SARS-CoV-2 VSV pseudovirus by clinical-stage monoclonal antibodies.
figure 3

a, RBD sequence of SARS-CoV-2 Wuhan-Hu-1 with highlighted footprints of ACE2 (light blue) and monoclonal antibodies (mAbs; coloured according to the RBD antigenic site recognized). The Omicron RBD is also shown, and amino acid substitutions are boxed. b, Neutralization of SARS-CoV-2 VSV pseudoviruses displaying Wuhan-Hu-1 (white) or Omicron (coloured as in Fig. 4b) spike proteins by clinical-stage monoclonal antibodies. Data are representative of at least two independent experiments. Shown is the mean of two technical replicates. c, Geometric mean half-maximum inhibitory concentration (IC50) values for Omicron (coloured as in Fig. 4b) and Wuhan-Hu-1 (white) (top), and geometric mean fold change (bottom). Vero E6 cells were used as target cells. Shown in blue (right) is neutralization of authentic virus by sotrovimab (WA1/2020 versus hCoV-19/USA/WI-WSLH-221686/2021). Non-neutralizing IC50 titres and fold change were set to 104 and 103, respectively. Orange dots for sotrovimab indicate neutralization of Omicron VSV pseudovirus carrying R346K (Omicron-R346K). Data are representative of n = 2 biologically independent experiments for most monoclonal antibodies; for sotrovimab against Omicron VSV n = 6 and for Omicron authentic virus n = 3.

We compared the in vitro neutralizing activity of these therapeutic monoclonal antibodies side-by-side against Wuhan-Hu-1 and Omicron spike proteins using VSV pseudoviruses (Fig. 3). Although sotrovimab had a threefold-reduced potency against Omicron and Omicron-R346K variant VSV pseudoviruses, all of the other (RBM-specific) monoclonal antibodies completely lost their neutralizing activity—with the exception of the combination of COV2-2130 and COV2-2196, for which we determined an approximately 100-fold-reduced potency (Fig. 3b, c). Moreover, sotrovimab exhibited a reduction of less than twofold in neutralizing activity against authentic Omicron SARS-CoV-2 as compared to the WA1/2020 D614G virus (Fig. 3c, Extended Data Fig. 7), consistent with recent reports on S309, the parent of sotrovimab41,42. The threefold and less-than-twofold decrease in the neutralizing activity of sotrovimab against pseudoviruses and authentic virus, respectively, is within the currently defined threshold of ‘no change’ as defined by the US Food and Drug Administration (FDA; the FDA fact sheet for sotrovimab denotes no change as a reduction of less than fivefold in susceptibility43). Overall, our findings agree with two preliminary reports16,18 and, together with serological data, support the conclusion that the Omicron VOC has undergone antigenic shift.

We next tested a larger panel of 36 neutralizing NTD- or RBD-specific monoclonal antibodies for which the epitopes have been characterized structurally or assigned to a given antigenic site through competition studies3,4,9,12,44,45 (Fig. 4a, Extended Data Table 2, Extended Data Fig. 8). The four NTD-specific antibodies completely lost activity against Omicron, consistent with the presence of mutations and deletions in the NTD antigenic supersite21,45,46. Three out of the twenty-two monoclonal antibodies that target the RBD antigenic site I (RBM) retained potent neutralizing activity against Omicron, including S2K146, which binds the RBD of SARS-CoV-2, SARS-CoV and other sarbecoviruses through ACE2 molecular mimicry1. Of the nine monoclonal antibodies that are specific for the conserved RBD site II (refs. 4,12), only S2X2593 retained activity against Omicron, whereas neutralization was decreased by more than tenfold or abolished for the remaining antibodies. Finally, the S2H97 monoclonal antibody retained neutralizing activity againstOmicron through recognition of the highly conserved cryptic site V (ref. 4). The panel of 44 monoclonal antibodies tested in this study includes members of each of the four classes of neutralizing monoclonal antibodies, defined by their cognate RBD-binding sites (sites I, II, IV and V)12. Our findings show that a member, or members, of each of the four classes can retain Omicron neutralization: S2K146, S2X324 and S2N28 targeting site I, S2X259 targeting site II, sotrovimab targeting site IV and S2H97 targeting site V (Fig. 4b). Several of these monoclonal antibodies cross-react with and neutralize sarbecoviruses beyond the SARS-CoV-2 clade 1b1,3,4, indicating that targeting of conserved epitopes can lead to neutralization breadth and resilience to antigenic shift associated with viral evolution.

Fig. 4: Neutralization of Omicron SARS-CoV-2 VSV pseudovirus by monoclonal antibodies.
figure 4

a, Top, mean IC50 values for Omicron (coloured as in b) and Wuhan-Hu-1 (white). Bottom, mean fold change for 4 NTD monoclonal antibodies and 32 RBD monoclonal antibodies. Non-neutralizing IC50 titres and fold change were set to 104 and 103, respectively. Triangles for S2K146 indicate neutralization of Omicron carrying R346K. Vero E6 cells were used as target cells. Data are representative of n = 2 biologically independent experiments (except for S2K146, for which n = 6). b, The RBD sites targeted by four monoclonal antibodies that cross-neutralize Omicron are annotated and representative antibodies (the Fv region) bound to spike proteins are shown as a composite. Coloured surfaces on the RBD depict the epitopes and the RBM is shown as a black outline.

Discussion

The high number of substitutions present in the Omicron spike protein marks a pronounced shift in antigenicity and is associated with immune evasion of considerable magnitude for SARS-CoV-2. Antigenic shift of the influenza virus is defined as genetic reassortment of the RNA genome segments, but the mechanism for the abrupt appearance of a large number of mutations in SARS-CoV-2 Omicron spike protein remains to be determined. Although recombination events are a hallmark of coronaviruses47, we and others48 propose that the Omicron shift may result from extensive viral replication in immunodeficient hosts47,49, although we cannot rule out the possibility of a contribution of inter-species ping-pong transmission5 between humans and rodents, as previously described for minks50.

Consistent with the variable decrease in plasma neutralizing antibody titres, we found that only 6 out of a panel of 44 neutralizing monoclonal antibodies retained potent neutralizing activity against Omicron. The monoclonal antibodies that retain neutralization recognize RBD antigenic sites that are conserved in Omicron and other sarbecoviruses. Notably, three of these antibodies bind to the RBM, including one that is a molecular mimic of the ACE2 receptor (S2K146)1. Collectively, these data may guide future efforts to develop SARS-CoV-2 vaccines and therapies to counteract antigenic shift and future sarbecovirus zoonotic spillovers.

Methods

Cell lines

Cell lines used in this study were obtained from ATCC (HEK293T and Vero E6), Thermo Fisher Scientific (Expi-CHO-S cells, FreeStyle 293-F cells and Expi293F cells), Takara (Lenti-X 293T cells) or generated in-house (Vero E6-TMPRSS2)40. Vero-TMPRSS251 cells were cultured at 37 °C in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 10 mM HEPES pH 7.3 and 100 U ml−1 of penicillin–streptomycin and supplemented with 5 µg ml−1 of blasticidin. None of the cell lines used was authenticated. Cell lines were routinely tested for mycoplasma contamination.

Omicron prevalence analysis

The viral sequences and the corresponding metadata were obtained from the GISAID EpiCoV project (https://www.gisaid.org/). Analysis was performed on sequences submitted to GISAID up to 20 December 2021. Spike protein sequences were obtained from the genomic sequences with the exonerate52 2 2.4.0–haf93ef1_3 (https://quay.io/repository/biocontainers/exonerate?tab=tags) using protein to DNA alignment with parameters -m protein2dna –refine full –minintron 999999 –percent 20 and using accession YP_009724390.1 as a reference. Multiple sequence alignment of all human spike proteins was performed with mafft53 7.475–h516909a_0 (https://quay.io/repository/biocontainers/mafft?tab=tags) with parameters –auto –reorder –mapout –keeplength –addfragments using the same reference as above. The –mapout parameter was used to retrieve insertions. Spike sequences that contained more than 10% ambiguous amino acids or that were less than 80% of the canonical protein length were discarded. Figures were generated with R 4.0.2 (https://cran.r-project.org/) using the ggplot2 3.3.2 and sf 0.9-7 packages. To identify each mutation prevalence, missingness (or ambiguous amino acids) was taken into account in both nominator and denominator.

Monoclonal antibodies

Sotrovimab and VIR-7832 (VIR-783254 is derived from sotrovimab, Fc further engineered to carry GAALIE) were produced at WuXi Biologics. Antibody VH and VL sequences for the monoclonal antibodies COV2-2130 (Protein Data Bank (PDB) ID 7L7E), COV2-2196 (PDB ID 7L7E, 7L7D), REGN10933 (PDB ID 6XDG), REGN10987 (PDB ID 6XDG) and ADI-58125 (PCT application WO2021207597, seq. IDs 22301 and 22311) were subcloned into heavy chain (human IgG1) and the corresponding light chain (human Ig κ-chain, Ig λ-chain) expression vectors respectively and produced in transiently transfected ExpiCHO-S cells (Thermo Fisher Scientific, A29133) at 37 °C and 8% CO2. Cells were transfected using ExpiFectamine. Transfected cells were supplemented 1 day after transfection with ExpiCHO Feed and ExpiFectamine CHO Enhancer. Cell culture supernatant was collected eight days after transfection and filtered through a 0.2-µm filter. Recombinant antibodies were affinity purified on an ÄKTA Xpress FPLC device using 5 ml HiTrap MabSelect PrismA columns followed by buffer exchange to histidine buffer (20 mM histidine, 8% sucrose, pH 6) using HiPrep 26/10 desalting columns. Antibody VH and VL sequences for LY-CoV555, LY-CoV016 and CT-P59 were obtained from PDB IDs 7KMG, 7C01 and 7CM4, respectively, and monoclonal antibodies were produced as recombinant IgG1 by ATUM. The remaining monoclonal antibodies were discovered at VIR and have been produced as recombinant IgG1 in ExpiCHO-S cells as described above. The identity of the produced monoclonal antibodies was confirmed by liquid chromatography–mass spectrometry (LC–MS) analysis.

IgG mass quantification by LC–MS intact protein mass analysis

Fc N-linked glycan from monoclonal antibodies was removed by PNGase F after overnight non-denaturing reaction at room temperature. Deglycosylated protein (4 µg) was injected to the LC–MS system to acquire intact MS signal. Thermo MS (Q Exactive Plus Orbitrap) was used to acquire intact protein mass under denaturing condition with m/z window from 1,000 to 6,000. BioPharma Finder 3.2 software was used to deconvolute the raw m/z data to protein average mass. The theoretical mass for each monoclonal antibody was calculated with GPMAW 10.10 software. Post-translational modifications such as N-terminal pyroglutamate cyclization, C-terminal lysine cleavage and formation of 16–18 disulfide bonds were added into the calculation.

Sample donors

Samples were obtained from SARS-CoV-2 convalescent and vaccinated individuals under study protocols approved by the local institutional review boards (Canton Ticino Ethics Committee, Switzerland, Comitato Etico Milano Area 1). All donors provided written informed consent for the use of blood and blood derivatives (such as peripheral blood mononuclear cells, sera or plasma) for research. Plasma samples from convalescent individuals and individuals who had been vaccinated with Ad26.COV2.S, mRNA-1273 or BNT162b2 were obtained from the HAARVI study, approved by the University of Washington Human Subjects Division Institutional Review Board (STUDY00000959). Samples from individuals who had been vaccinated with AZD1222 were obtained from INGM, Ospedale Maggio Policlinico of Milan and approved by the local review board Study Polimmune. Samples from individuals who had been vaccinated with Sputnik V were obtained from healthcare workers at the hospital de Clínicas ‘José de San Martín’, Buenos Aires, Argentina. Samples from individuals who had been vaccinated with Sinopharm were enrolled from Aga Khan University under the institutional review board of the UWARN (United World Antivirus Research Network) study.

Pseudovirus neutralization assays

Generation of VSV pseudovirus used in Vero E6 cells

The plasmid encoding the Omicron SARS-CoV-2 spike variant was generated by overlap PCR mutagenesis of the wild-type plasmid, pcDNA3.1(+)-spike-D1955. Replication defective VSV pseudovirus expressing SARS-CoV-2 spike proteins corresponding to the ancestral Wuhan-Hu-1 virus and the Omicron VOC were generated as previously described46 with some modifications. Lenti-X 293T cells (Takara) were seeded in 15-cm2 dishes at a density of 10 × 106 cells per dish and the following day were transfected with 25 µg of spike expression plasmid with TransIT-Lenti (Mirus, 6600) according to the manufacturer’s instructions. One day after transfection, cells were infected with VSV-luc (VSV-G) with a multiplicity of infection (MOI) of 3 for 1 h, rinsed three times with PBS containing Ca2+ and Mg2+, then incubated for an additional 24 h in complete medium at 37 °C. The cell supernatant was clarified by centrifugation, aliquoted, and frozen at −80 °C.

Generation of VSV pseudovirus used in Vero E6-TMPRSS2 cells

Comparison of Omicron SARS-CoV-2 spike VSV to SARS-CoV-2 G614 spike (YP 009724390.1) VSV and Beta spike VSV used pseudotyped particles prepared as described previously9,56. In brief, HEK293T cells in DMEM supplemented with 10% FBS, 1% penicillin–streptomycin seeded in 10-cm dishes were transfected with the plasmid encoding the corresponding spike glycoprotein using Lipofectamine 2000 (Life Technologies) following the manufacturer’s instructions. One day after transfection, cells were infected with VSV(G*ΔG-luciferase)57 and after 2 h were washed five times with DMEM before adding medium supplemented with anti-VSV-G antibody (I1-mouse hybridoma supernatant, CRL-2700, ATCC). Virus pseudotypes were collected 18–24 h after inoculation, clarified by centrifugation at 2,500g for 5 min, filtered through a 0.45-μm cut-off membrane, concentrated 10 times with a 30-kDa cut-off membrane, aliquoted and stored at −80 °C.

VSV pseudovirus neutralization

Assay performed using Vero E6 cells

Vero E6 cells were grown in DMEM supplemented with 10% FBS and seeded into clear bottom white 96 well plates (PerkinElmer, 6005688) at a density of 20,000 cells per well. The next day, monoclonal antibodies or plasma were serially diluted in pre-warmed complete medium, mixed with pseudoviruses and incubated for 1 h at 37 °C in round bottom polypropylene plates. Medium from cells was aspirated and 50 µl of virus–monoclonal antibody–plasma complexes was added to cells, which were then incubated for 1 h at 37 °C. An additional 100 µl of prewarmed complete medium was then added on top of complexes and cells were incubated for an additional 16–24 h. Conditions were tested in duplicate wells on each plate and eight wells per plate contained untreated infected cells (defining the 0% of neutralization, ‘MAX RLU’ value) and infected cells in the presence of S309 and S2X259 at 20 µg ml−1 each (defining the 100% of neutralization, ‘MIN RLU’ value). Virus–monoclonal antibody–plasma-containing medium was then aspirated from cells and 100 µl of a 1:2 dilution of SteadyLite Plus (PerkinElmer, 6066759) in PBS with Ca2+ and Mg2+ was added to cells. Plates were incubated for 15 min at room temperature and then were analysed on the Synergy-H1 (Biotek). The average relative light units (RLUs) of untreated infected wells (MAX RLUave) was subtracted by the average of MIN RLU (MIN RLUave) and used to normalize percentage of neutralization of individual RLU values of experimental data according to the following formula: (1 − (RLUx – MIN RLUave)/(MAX RLUave – MIN RLUave)) × 100. Data were analysed and visualized with Prism (v.9.1.0). IC50 (monoclonal antibodies) and ID50 (plasma) values were calculated from the interpolated value from the log(inhibitor) versus response, using variable slope (four parameters) nonlinear regression with an upper constraint of ≤100, and a lower constrain equal to 0. Each neutralization experiment was conducted on two independent experiments—that is, biological replicates—in which each biological replicate contains a technical duplicate. IC50 values across biological replicates are presented as arithmetic mean ± s.d. The loss or gain of neutralization potency across spike variants was calculated by dividing the variant IC50/ID50 by the parental IC50/ID50 within each biological replicate, and then visualized as arithmetic mean ± s.d.

Assay performed using Vero E6-TMPRSS2 cells

VeroE6-TMPRSS2 cells were cultured in DMEM with 10% FBS (Hyclone), 1% penicillin–streptomycin and 8 µg ml−1 puromycin (to ensure retention of TMPRSS2) with 5% CO2 in a 37 °C incubator (Thermo Fisher Scientific). Cells were trypsinized using 0.05% trypsin and plated to be at 90% confluence the following day. In an empty half-area 96-well plate, a 1:3 serial dilution of serum was made in DMEM and diluted pseudovirus was then added and incubated at room temperature for 30–60 min before addition of the serum–virus mixture to the cells at 37 °C. Two hours later, 40 μl of a DMEM solution containing 20% FBS and 2% penicillin–streptomycin (Thermo Fisher Scientific, 10,000 units per ml of penicillin and 10,000 µg ml−1 of streptomycin when undiluted) was added to each well. After 17–20 h, 40 μl per well of One-Glo-EX substrate (Promega) was added to the cells and they were incubated in the dark for 5–10 min before reading on a BioTek plate reader. Measurements were done at least in duplicate using distinct batches of pseudoviruses and one representative experiment is shown. RLUs were plotted and normalized in Prism (GraphPad). Nonlinear regression of log(inhibitor) versus normalized response was used to determine IC50 values from curve fits. Normality was tested using the D’Agostino-Pearson test and in the absence of a normal distribution, Kruskal–Wallis tests were used to compare two groups to determine whether differences reached statistical significance. Fold changes were determined by comparing individual IC50 and then averaging the individual fold changes for reporting.

Focus reduction neutralization test

The WA1/2020 strain with a D614G substitution was described previously58. The B.1.1.529 isolate (hCoV-19/USA/WI-WSLH-221686/2021) was obtained from a nasal swab and passaged on Vero-TMPRSS2 cells as described59. The B.1.1.529 isolate was sequenced (GISAID: EPI_ISL_7263803) to confirm the stability of substitutions. All virus experiments were performed in an approved biosafety level 3 (BSL-3) facility.

Serial dilutions of sotrovimab were incubated with 102 focus-forming units of SARS-CoV-2 (WA1/2020 D614G or B.1.1.529) for 1 h at 37 °C. Antibody–virus complexes were added to Vero-TMPRSS2 cell monolayers in 96-well plates and incubated at 37 °C for 1 h. Subsequently, cells were overlaid with 1% (w/v) methylcellulose in MEM. Plates were collected 30 h later (WA1/2020 D614G on Vero-TMPRSS2 cells) or 70 h later (B.1.1.529 on Vero-TMPRSS2 cells) by removal of overlays and fixation with 4% PFA in PBS for 20 min at room temperature. Plates with WA1/2020 D614G were washed and sequentially incubated with an oligoclonal pool of SARS2-2, SARS2-11, SARS2-16, SARS2-31, SARS2-38, SARS2-57 and SARS2-7160 anti-S antibodies. Plates with B.1.1.529 were additionally incubated with a pool of monoclonal antibodies that cross-react with SARS-CoV-1 and bind a CR3022-competing epitope on the RBD61. All plates were subsequently stained with HRP-conjugated goat anti-mouse IgG (Sigma, A8924) in PBS supplemented with 0.1% saponin and 0.1% bovine serum albumin. SARS-CoV-2-infected cell foci were visualized using TrueBlue peroxidase substrate (KPL) and quantitated on an ImmunoSpot microanalyser (Cellular Technologies). Antibody dose response curves were analysed using nonlinear regression analysis with a variable slope (GraphPad Software), and the IC50 was calculated.

VSV pseudovirus entry assays using mouse ACE2

HEK293T (293T) cells (ATCC CRL-11268) were cultured in 10% FBS, 1% penicillin–streptomycin in DMEM at 37 °C in a humidified 8% CO2 incubator. Transient transfection of mouse ACE2 in 293T cells was done 18–24 h before infection using Lipofectamine 2000 (Life Technologies) and an HDM plasmid containing full-length mouse ACE2 (GenBank: Q8R010, synthesized by GenScript) in Opti-MEM. After a 5-h incubation at 37 °C in a humidified 8% CO2 incubator, DMEM with 10% FBS was added and cells were incubated at 37 °C in a humidified 8% CO2 incubator for 18–24 h. Immediately before infection, 293T cells with transient expression of mouse ACE2 were washed with DMEM 1×, then plated with pseudovirus at a 1:75 dilution in DMEM. Infection in DMEM was done with cells between 60% and 80% confluence for 2.5 h before adding FBS and penicillin–streptomycin to final concentrations of 10% and 1%, respectively. After 18–24 h of infection, One-Glo-EX (Promega) was added to the cells and they were incubated in the dark for 5 min before reading on a Synergy H1 Hybrid Multi-Mode plate reader (Biotek). Cell entry levels of pseudovirus generated on different days (biological replicates) were plotted in GraphPad Prism as individual points, and average cell entry across biological replicates was calculated as the geometric mean.

Production of recombinant RBD proteins

SARS-CoV-2 RBD proteins for SPR binding assays (residues 328–531 of the spike protein from GenBank NC_045512.2 with N-terminal signal peptide and C-terminal thrombin cleavage site-TwinStrep-8×His-tag) were expressed in Expi293F (Thermo Fisher Scientific) cells at 37 °C and 8% CO2. Transfections were performed using the ExpiFectamine 293 Transfection Kit (Thermo Fisher Scientific). Cell culture supernatants were collected two to four days after transfection and supplemented with 10× PBS to a final concentration of 2.5× PBS (342.5 mM NaCl, 6.75 mM KCl and 29.75 mM phosphates). SARS-CoV-2 RBDs were purified using cobalt-based immobilized metal affinity chromatography followed by buffer exchange into PBS using a HiPrep 26/10 desalting column (Cytiva) or, for the second batch of Omicron RBD used for SPR, a Superdex 200 Increase 10/300 GL column (Cytiva).

The SARS-CoV-2 Wuhan-Hu-1 and Delta (B.1.617.2) RBD-Avi constructs were synthesized by GenScript into pcDNA3.1- with an N-terminal mu-phosphatase signal peptide and a C-terminal octa-histidine tag, flexible linker and avi tag (GHHHHHHHHGGSSGLNDIFEAQKIEWHE).The boundaries of the construct are N-328RFPN331 and 528KKST531-C (refs. 9,14). Proteins were produced in Expi293F cells (Thermo Fisher Scientific) grown in suspension using Expi293 Expression Medium (Thermo Fisher Scientific) at 37 °C in a humidified 8% CO2 incubator rotating at 130 rpm. Cells grown to a density of 3 million cells per ml were transfected using the the ExpiFectamine 293 Transfection Kit (Thermo Fisher Scientific) and cultivated for 3–5 days. Proteins were purified from clarified supernatants using a nickel HisTrap HP affinity column (Cytiva) and washed with 10 column volumes of 20 mM imidazole, 25 mM sodium phosphate pH 8.0 and 300 mM NaCl before elution on a gradient to 500 mM imidazole. Proteins were biotinylated overnight using the BirA Biotin-Protein Ligase Kit (Avidity) and purified again using theHisTrapHP affinity column. After a wash and elution as before, proteins were buffer-exchanged into 20 mM sodium phosphate pH 8 and 100 mM NaCl, and concentrated using centrifugal filters (Amicon Ultra) before being flash-frozen.

Recombinant production of ACE2 orthologues

Recombinant human ACE2 (residues 19–615 from Uniprot Q9BYF1 with a C-terminal AviTag-10×His-GGG-tag, and N-terminal signal peptide) was produced by ATUM. Protein was purified via Ni Sepharose resin followed by isolation of the monomeric hACE2 by size-exclusion chromatography using a Superdex 200 Increase 10/300 GL column (Cytiva) pre-equilibrated with PBS. The mouse (Mus musculus) ACE2 ectodomain construct (GenBank: Q8R0I0) was synthesized by GenScript and placed into a pCMV plasmid. The domain boundaries for the ectodomain are residues 19–615. The native signal tag was identified using SignalP-5.0 (residues 1–18) and replaced with an N-terminal mu-phosphatase signal peptide. This construct was then fused to a sequence encoding a thrombin cleavage site and a human Fc fragment or an 8×His tag at the C terminus. ACE2-Fc and ACE2-His constructs were produced in Expi293 cells (Thermo Fisher Scientific, A14527) in Gibco Expi293 Expression Medium at 37 °C in a humidified 8% CO2 incubator rotating at 130 rpm. The cultures were transfected using PEI-25K (Polyscience) with cells grown to a density of 3 million cells per ml and cultivated for 4–5 days. Proteins were purified from clarified supernatants using a 1-ml HiTrap Protein A HP affinity column (Cytiva) or a 1-ml HisTrap HP affinity column (Cytiva), concentrated and flash-frozen in 1× PBS, pH 7.4 (10 mM Na2HPO4, 1.8 mM KH2PO4, 2.7 mM KCl, 137 mM NaCl).

ACE2-binding measurements using SPR

Measurements were performed using a Biacore T200 instrument, in triplicate for monomeric human and mouse ACE2 and duplicate for dimeric mouse ACE2. A CM5 chip covalently immobilized with StrepTactin XT (IBA LifeSciences) was used for surface capture of TwinStrepTag-containing RBDs (Wuhan-Hu-1, Alpha, Beta, Omicron, K417N) and a Cytiva Biotin CAPture Kit was used for surface capture of biotinylated RBDs (Delta and Wuhan-Hu-1 used for fold-change comparison to Delta). Two different batches of Omicron RBD were used for the experiments. Running buffer was HBS-EP+ pH 7.4 (Cytiva) and measurements were performed at 25 °C. Experiments were performed with a threefold dilution series of human ACE2 (300, 100, 33, 11 nM) or mouse ACE2 (900, 300, 100, 33 nM) and were run as single-cycle kinetics. Monomeric ACE2-binding data were double-reference-subtracted and fit to a 1:1 binding model using Biacore Evaluation software. High concentrations of dimeric mouse ACE2 exhibited significant binding to the CAP sensor chip reference flow cell.

Statistical analysis

Neutralization measurements were performed in duplicate and RLUs were converted to per cent neutralization and plotted with a nonlinear regression model to determine IC50/ID50 values using GraphPad Prism software (v.9.0.0). Comparisons between two groups of paired two-sided data were made with Wilcoxon rank test.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this paper.