Main

The emergence of SARS-CoV-2 variants of concern (VOC) and interest (VOI) has been a hallmark of the COVID-19 pandemic during 20211,2. Classification of VOC and VOI by the World Health Organization has been highly dynamic with emergent SARS-CoV-2 variants such as Lambda (lineage C.37), Delta (lineage B.1.617.2), Mu (lineage B.1.621) and Omicron (B.1.1.529) being the most recently recognized, whereas others such as Epsilon (lineage B.1.427/29), Zeta (lineage P.2) and Theta (lineage P.3) are no longer considered as VOI3. VOC and VOI are characterized by the presence of mutations located in key positions of the spike protein, including those in the receptor-binding domain (RBD) such as N501Y (shared by VOC Alpha, Beta, Gamma and Omicron) associated with increased Angiotensin-Converting Enzyme 2 (ACE2) binding and infectivity, or K417N/T and E484K/A (shared by VOC Beta, Gamma and Omicron) associated with escape from neutralizing antibodies1. The spike protein of VOI Lambda, responsible for major outbreaks in South American countries, including Chile, has a pattern of seven mutations (G75V, T76I, Δ246–252, L452Q, F490S, D614G, T859N) of which L452Q is similar to the L452R mutation reported in the Delta and Epsilon variants4.

The presence of mutations in antigenic sites characteristic of SARS-CoV-2 VOC and VOI has raised concerns regarding their impact on the neutralizing capacity of antibodies elicited by currently available vaccines, which were designed using the ancestral spike as a reference. Indeed, several studies have shown that VOC and VOI escape neutralization from antibodies elicited by vaccines, although to different extents1. Interestingly, recent studies suggested that neutralizing antibody titres elicited by vaccines should be considered as a correlate of protection5,6,7,8,9. As such, the emergence of SARS-CoV-2 variants carrying mutations in antigenic sites within the spike protein and their impact on neutralization by antibodies elicited by vaccines requires continuous monitoring.

The virus-inactivated vaccine CoronaVac and the messenger RNA vaccine BNT162b2 are the most widely inoculated vaccines worldwide and were identified among those with lower and higher correlates of protection, respectively5,6,10. Therefore, we sought to determine and compare the neutralization capacity of plasma samples obtained from 109 healthcare workers from Clínica Santa María and the Faculty of Medicine at Universidad de Chile, both in Santiago, Chile, receiving the complete scheme of these vaccines during the Chilean Ministry of Health’s vaccination campaign (Supplementary Table 1).

We generated pseudotyped viruses carrying either the ancestral SARS-CoV-2 spike protein (Wuhan-Hu-1; lineage A)11, the spike protein carrying the D614G mutation (lineage B.1) or the spike from variants Alpha (lineage B.1.1.7), Gamma (lineage P.1), Lambda (lineage C.37) and Delta (lineage B.1.617.2). These variants have dominated the COVID-19 pandemic in Chile during 2021 (Extended Data Fig. 1a,b). Neutralization assays using the pseudotyped virus carrying the ancestral spike revealed that the geometric mean of the pseudotyped virus neutralization titre 50 (pVNT50) was 97.78 (95% confidence intervals (CI) 71.51 to 133.7) for individuals receiving the inactivated virus vaccine, and 720 (95% CI 567.2 to 913.9) for those inoculated with the mRNA vaccine (Fig. 1a and Supplementary Tables 2 and 3). These data indicate that the BNT162b2 vaccine elicits 7.4-fold higher neutralizing antibody titres against the ancestral spike protein compared with CoronaVac (Fig. 1a). Neutralization assays using pseudotypes carrying the D614G mutant or the Alpha, Gamma, Lambda and Delta variants resulted in geometric mean pVNT50 values of 104.9 (95% CI 78.9 to 139.5), 60.31 (95% CI 42.51 to 85.56), 58.27 (95% CI 40.05 to 84.79), 47.43 (95% CI 34.41 to 65.38) and 23.15 (95% CI 16.98 to 31.56), respectively, for the CoronaVac samples, and 692 (95% CI 560.6 to 854.6), 466 (95% CI 368.2 to 589.8), 389.7 (95% CI 301.6 to 503.7), 309.5 (95% CI 311.6 to 489.4) and 301.7 (95% CI 230.5 to 394.9) for the BNT162b2 samples (Fig. 1b and Supplementary Tables 2 and 3). Neutralizing antibody titres against spike variants were 6.6- to 13-fold higher for BNT162b2 vaccinees (Fig. 1b). Compared with the ancestral spike, neutralizing antibody titres elicited by CoronaVac decreased by a factor of 0.9 for the D614G mutant, 1.6 for the Alpha spike, 1.7 for the Gamma spike, 2.1 for the Lambda spike and 4.2 for the Delta spike (Fig. 1c,d and Supplementary Tables 2 and 3). Neutralizing antibody titres elicited by BNT162b2 decreased by a factor of 1.04, 1.54, 1.84, 1.84 and 2.38 for the D614G mutant or the Alpha, Gamma, Lambda and Delta variants, respectively (Fig. 1e,f and Supplementary Tables 2 and 3).

Fig. 1: Neutralization titres of plasma from individuals receiving CoronaVac and BNT162b2 vaccines.
figure 1

a,b, pVNT50 of recipients of the CoronaVac vaccine (n = 53) and of the BNT162b2 vaccine (n = 56) against (a) pseudotype Wuhan-Hu-1 reference strain (wild-type) or (b) D614G (B.1), Alpha (B.1.1.7), Gamma (P.1), Lambda (C.37) and Delta (B.1.617.2) variants. The statistical significance of the difference between the neutralization was calculated by Wilcoxon matched-pairs signed-rank test. Two-tailed P values are reported. Geometric mean titres and 95% CIs are indicated. c,d, Box plots indicate the median and interquartile range of pVNT50 for each SARS-CoV-2 pseudotype for the CoronaVac (c) and BNT162b2 (d) groups. Factor changes are shown in parentheses as the difference of the geometric mean titre in pVNT50 compared with that of the wild-type pseudotyped virus. Statistical analyses were performed with the two-tailed Kruskal–Wallis test after adjustment for the false discovery rate. Geometric mean titres are indicated. e,f, Results are shown as the difference in neutralization titres of matched samples for each SARS-CoV-2 pseudotype in the CoronaVac (e) and BNT162b2 (f) groups. P values for comparison of the pVNT50 were calculated with the Wilcoxon signed-rank test.

Source data

Full size image

Demographic analyses of our vaccinated cohort regarding sex, age, body mass index (BMI) or smoking status and neutralizing antibody titres showed positive correlations between age-CoronaVac-NAbs and Delta (P = 0.0221), age-BNT162b2-NAbs and D614G (P = 0.0285) and BMI-BNT162b2-NAbs and Delta (P = 0.025) (Extended Data Fig. 2a,d).

We also performed neutralization analyses using 30 samples from convalescent individuals infected during the first wave in Chile. This wave was dominated by the B.1 lineage and occurred before the circulation of SARS-CoV-2 variants in the country and the use of vaccines (Extended Data Fig. 3). Here geometric means of pVNT50 were 454.8 (95% CI 351.3 to 588.8), 361.4 (95% CI 251.6 to 519.1), 270.5 (95% CI 180.6 to 405.1), 208.4 (95% CI 145.7 to 298.1), 195.2 (95% CI 132.9 to 286.8) and 169.8 (95% CI 108 to 267.1) for ancestral lineage, D614G, Alpha, Gamma, Lambda and Delta, respectively. As such, neutralizing antibody titres were decreased by a factor of 1.26, 1.68, 2.18, 2.33 and 2.67 for D614G, Alpha, Gamma, Lambda and Delta, respectively (Extended Data Fig. 3 and Supplementary Table 4).

Comparative analyses of immunogenic regions showed that the four tested SARS-CoV-2 variants display mutations in immunodominant epitopes of the receptor-binding motif targeted by class I and class II antibodies (Extended Data Fig. 4a,b). Whereas Alpha and Gamma carry a polar to aromatic change (N501Y), and Gamma changes the charge of hydrophilic amino acids (E484K, K417T) in the receptor-binding motif within the RBD, the Delta variant carries the L452R mutation shown to confer both increased infectivity and escape from neutralizing antibodies12,13. The Lambda RBD variant carries the L452Q mutation that is similar to the L452R mutation described in Delta, and the F490S mutation, which has also been associated with escape from neutralizing antibodies12,14. On the other hand, all tested variants show substitutions in the N-terminal domain (NTD) antigenic supersite. However, whereas the Alpha variant carries a single amino acid deletion (ΔY144) in the β-hairpin, and Gamma displays two substitutions in the N terminus (L18F, T20N), the Lambda variant harbours a seven amino acid deletion of the supersite loop (Δ246–252) described as an antigenic supersite15. Indeed, both L452Q/F490S and Δ246–252 reside at the interface between the spike protein and monoclonal antibodies targeting the RBD and the NTD, respectively (Extended Data Fig. 5a,b). By contrast, the Delta variant shows changes at the N terminus (T19R) and in the β-hairpin (Δ156–157, R158G). This distinctive substitution pattern within antigenic sites may, at least in part, explain the different neutralizing activity of CoronaVac and BNT162b2-induced polyvalent sera against the tested variants.

Analyses using equivalent amounts of pseudotyped viruses in previously described HEK293T-ACE2 cells11, showed that the Lambda spike provides increased viral entry when compared with Alpha, Gamma and Delta spikes (Fig. 2a). However, viral replication assays using equivalent viral titres (multiplicity of infection (MOI) = 1) to infect Vero-E6 cells showed that Lambda replicates at lower rates compared with the Alpha, Gamma and Delta variants, as indicated by a significant higher cycle threshold value (which is inversely related to viral RNA levels) in the released genomic RNA recovered both at 24 and 48 hours post infection (Fig. 2b). Although some RBD mutations such as N501Y and L452R have been linked to increased infectivity1,12,13, mutations within intermediate folding domains may also change infectivity by affecting the spike molecular dynamics16,17. In this sense, the spike protein of the Lambda variant displays the substitution T859N that lies in the neighbourhood of G614 in the apposed protomer (Extended Data Fig. 6a). In the ancestral spike protein, the D614–T859 interaction was proposed to modulate the stability of the inter-protomer interface, and earlier molecular dynamics simulations suggest that fluctuations in the native contacts may facilitate transition between up and down states in the spike’s RBD18. In fact, the G614 form increases the occupancy of the one-up state, which accounts for its enhanced infectivity19. We also noticed that T859 forms hydrophobic contacts with F592 in the apposed protomer (Extended Data Fig. 6a). Interestingly, conservation analysis among betacoronavirus spikes revealed that when a phenylalanine is present at an equivalent 592 position, it is coupled to threonine at 859 exclusively in the sarbecovirus group, which includes SARS-CoV, PCoV (pangolin), RatG13 (bat) and SARS-CoV-2 (Extended Data Fig. 6b). Interestingly, we were not able to observe differences between pseudotyped viruses carrying the Lambda spike and the original residue (T859) or the T859N mutation (Fig. 2c). Considering the presence of the T859N mutation in several SARS-CoV-2 lineages, its effects on neutralization by antibodies warrant further investigation.

Fig. 2: Infectivity and replication of SARS-CoV-2 variant Lambda.
figure 2

a, Levels of infection of SARS-CoV-2 pseudotypes bearing the Alpha (B.1.1.7), Gamma (P.1), Lambda (C.37) and Delta (B.1.617.2) spike protein in HEK293T cells stably expressing ACE2. Cells were seeded in 96-well plates and infected with 0.5 RT units of the indicated pseudotyped viruses for 48 h. Two biological replicates were assessed in three independent experiments. Data are presented as the mean ± s.d. of three independent experiments. P values were determined by ANOVA. b, Genomic RNA levels Alpha, Gamma, Lambda and Delta viruses produced on infection of Vero-E6 cells. Cells were infected at a MOI = 1. Genomic RNA levels are expressed as cycle threshold (Ct) values, which are inversely related to viral load. Data are presented as the mean ± s.d. of three experiments. P values were determined by ANOVA from a sample size of n = 3. c, Levels of infection of SARS-CoV-2 pseudotypes bearing the Lambda (C.37) and Lambda mutant T859 spike protein in HEK293T-ACE2 cells. Cells were seeded in 96-well plates and infected with 0.5 RT units of the indicated pseudotyped viruses for 48 h. Two biological replicates were assessed in three independent experiments. Data are presented as the mean ± s.d. of three independent experiments. P values were determined by Mann–Whitney signed-ranked test.

Source data

Full size image

The data presented here confirm that neutralizing antibody titres elicited by the mRNA vaccine BNT162b2 are higher than those elicited by the inactivated virus vaccine CoronaVac in a real-life vaccination setting. Previous studies have shown that most of the neutralizing antibodies are directed to antigenic sites present in the RBD and, to a lesser extent, antigenic sites in the NTD1,2. In this sense, BNT162b2 delivers an mRNA sequence encoding a spike protein stabilized in the prefusion conformation that keeps intact the RBD and NTD-containing S1 domain20. Conversely, studies performed on a β-propiolactone inactivated virus like that used in the CoronaVac vaccine showed that 74.4% of the spike proteins are in the post-fusion conformation. Thus, a minor fraction (25.6%) of the spike proteins has the major antigenic sites present in the S1 domain21. Moreover, treatment of SARS-CoV-2 particles with β-propiolactone may induce virus aggregation, diminishing its antigenic capacity22. These differences in the availability of antigenic sites may explain, at least in part, the differences in the neutralizing antibody titres observed between both vaccine platforms.

Interestingly, a nationwide study for BNT162b2 effectiveness in Israel revealed that this vaccine was 97% efficient in preventing symptomatic COVID-19, 97.5% efficient against severe or critical disease and 96.7% efficient in preventing COVID-19-related deaths23. A similar study for CoronaVac performed in Chile showed that this vaccine was 65.9% efficient in preventing COVID-19, 90.3% efficient in preventing intensive care unit admission and 86.3% efficient in preventing COVID-19-related deaths24. The differences observed in the neutralizing antibody titres elicited by these two vaccines and the ability of these antibodies to neutralize SARS-CoV-2 variants may help to explain the differences observed, especially those relating to their effectiveness to prevent symptomatic disease (97% for BNT162b2 versus 65.9% for CoronaVac) in a real-life vaccination setting involving millions of individuals. Therefore, our data provide further evidence favouring the use of neutralizing antibody titres as a correlate of protection, as recently suggested5,6,7,8,9. The implementation of a correlate of protection is critical when analysing the impact of emerging SARS-CoV-2 variants such as Mu, Delta and now, Omicron.

The data presented here also show that the unique pattern of mutations present in the spike protein of VOI Lambda confers escape from neutralizing antibodies elicited by CoronaVac, BNT162b2 and convalescent plasma, which is in agreement with recent observations25. The escape by Lambda from neutralizing antibodies was higher than that observed for Alpha and Gamma variants, but lower than that observed for Delta, indicating that escape from neutralizing antibodies is a key feature of this globally dominant variant. Because the 246–252 deletion in the NTD and the L452Q/F490S mutations in the RBD of the Lambda spike are located in antigenic sites, these mutations may reduce the affinity of neutralizing antibodies. Our data also suggest that the unique set of spike mutations present in the Lambda variant confer increased viral entry, but that the replication fitness is lower compared with other variants. Together, these observations may help explain the rapid rise and high circulation rates of the Lambda variant in South American countries before emergence of the Delta variant.

Methods

Study cohort

Healthcare workers without previously diagnosed SARS-CoV-2 infection from Clínica Santa María and the Faculty of Medicine at Universidad de Chile, both in Santiago, Chile, were invited to participate. Volunteers received the two-dose scheme of CoronaVac or BNT162b2, each dose being administered 28 or 21 d apart, respectively, according to the Chilean vaccination programme. Plasma samples were collected between May and June 2021. Convalescent individuals were participating in the study COVID0920 carried out at Hospital Clínico Red de Salud UC Christus, Santiago, Chile. All participants signed an informed consent before any study procedure was undertaken.

Pseudotype infectivity assay

Pseudotyped viruses carrying different SARS-CoV-2 spike proteins were prepared as previously described11. Briefly, HIV-1-based SARS-CoV-2 pseudotypes were produced in HEK293T cells by transfecting the pNL4.3-ΔEnv-Luc together with the corresponding pCDNA-SARS-CoV-2 Spike coding vector in a 1:1 molar ratio. Plasmids coding a codon-optimized spike lacking the last 19 amino acids of the C-terminal end (SΔ19) known to avoid retention at the endoplasmic reticulum were obtained by gene synthesis or customized site-directed mutagenesis (GeneScript). They contained the following mutations: lineage A (reference sequence), lineage B (D614G), lineage B.1.1.7 (Δ69–70, Δ144, N501Y, A570D, D614G, P681H, T716I, S982A, D1118H), lineage P.1 (L18F, T20N, P26S, D138Y, R190S, K417T, E484K, N501Y, D614G, H655Y, T1027I), lineage C.37 (G75V, T76I, Δ246–252, L452Q, F490S, D614G, T859N) and lineage B.1.617.2 (T19R, del157/158, L452R, T478K, D614G, P681R, D950N). Each pseudotype preparation was cleared by centrifugation at 850g at room temperature, quantified using the HIV-1 Gag p24 Quantikine ELISA Kit (R&D Systems) or the EnzChek Reverse Transcriptase Assay (ThermoFisher Scientific), aliquoted in 50% fetal bovine serum (FBS; Sigma-Aldrich) and stored at −80 °C until use.

Different amounts of pseudotyped viruses (as determined by the levels of the HIV-1 p24 protein) were used to infect HEK-ACE2 cells, and 48 h later, firefly luciferase activity was measured using the Luciferase Assay Reagent (Promega) in a Glomax 96 Microplate luminometer (Promega).

SARS-CoV-2 variants isolation and infectivity assay

Vero-E6 cells (African green monkey) were maintained in DMEM (low glucose) containing 10% FBS, 1% antibiotics (penicillin and streptomycin) in a 5% CO2 atmosphere. Five isolates belonging to B.1.1 (D614G), P.1 (Gamma), B 1.1.7 (Alpha), C.37 (Lambda) and B.1.617.2 (Delta) were isolated from SARS-CoV-2-positive subjects and inoculated into Vero-E6 in a biosafety level 3 laboratory. Virus supernatants were harvested and sequenced using the ARTIC SARS-CoV-2 ONT (Oxford Nanopore Technologies) sequencing protocol; vials were stored at −80 °C and the viral titre was measured by plaque assay.

For infectivity assay, Vero-E6 cells were seeded in two 24-well plates at 80% confluence and, after 24 h, an MOI of 1 for each variant was inoculated in triplicate into cells. After, 1 h at 37 °C and 5% CO2, the inoculum was removed and monolayers were replenished with DMEM supplemented with 2% FBS and 1% antibiotics (penicillin and streptomycin). Supernatants (200 μl) were collected at 24 and 48 h post inoculation and mixed with 200 μl of lysis buffer before PCR with reverse transcription assay. The PCR with reverse transcription assay was carried out with LightMix from TibMolBiol (Roche) against RdRP and the crossing threshold of each variant was normalized with the crossing threshold of each initial inoculum.

Neutralization assay

Plasma samples inactivated for 30 min at 56 °C were threefold serially diluted (from 1:2 to 1:4,374 for the CoronaVac group, from 1:5 to 1:10,935 for the BNT162b2 group and from 1:20 to 1:43,740 for the convalescent plasma) in supplemented DMEM. In parallel, pseudoviruses were thawed on ice and diluted to equal the p24 concentration in DMEM 50% FBS. A subsequent dilution of pseudovirus was prepared in supplemented DMEM to obtain 3 ng of p24 in 50 μl. Then, 50 μl of sample dilutions were mixed with 50 μl of each variant pseudovirus in duplicate and the mixture was incubated for 1 h at 37 °C. Later, 1 × 104 HEK-ACE2 cells in 100 μl of supplemented DMEM were added to each well and HEK293T cells transduced with the wild-type (lineage A) pseudovirus were used as a control. Cells were lysed 48 h later, and firefly luciferase activity was measured using the Luciferase Assay Reagent (Promega) in a Glomax 96 Microplate luminometer (Promega).

Relative luminescence units (RLUs) of HEK293T cells transduced with wild-type pseudovirus were averaged and considered as 100% neutralization, whereas RLUs measured at the highest dilution of each sample were established as 0% neutralization. Thus, the percentage of neutralization of each one of the eight dilutions of a sample was calculated as the complement of the division between the corresponding RLUs and the RLUs obtained at the higher dilution after subtracting the background (HEK293T cells plus wild-type pseudotyped virus). This calculation was done independently for each technical replicate and each spike variant.

Pseudotyped virus-neutralizing antibody titre (ID50) was calculated in GraphPad Prism v.9.2.0 by modelling a four-parameter nonlinear regression with variable slope constraining top values to 100 and bottom values to 0. Samples showing an ID50 lower than the first dilution (1:4 for CoronaVac, 1:10 for BNT162b2 or 1:40 for convalescent plasma) were considered as 4, 10 or 40, respectively.

Spike protein and sequence analyses

Mutations and antigenic sites were mapped on previously determined crystal structures (PDB: 7BNN, PDB: 7BWJ, PDB: 7LY0). Structural alignments were performed using Swiss-PdbViewer, displayed by PyMOL v.2.4.1 (http://www.pymol.org/pymol), and rendered by Blender v.2.92.0 (http://www.blender.org). The reference sequence of spike protein included in the analysis was YP_009724390.1 (SARS-CoV-2). Multiple sequence alignment was performed with ClustalO v.1.2.4 and visualized with MEGA X v.10.2.6 (https://www.megasoftware.net/). NCBI sequences of spike proteins from betacoronaviruses included in the analysis were as follows: SARS-CoV-2 (YP_009724390.1), RatG13 (bat) (QHR63300.2), PCoV (pangolin) (QIQ54048.1), SARS-CoV (YP_009825051.1), WIV16 (bat) (ALK02457.1), BtRl (bat) (QDF43815.1), Khosta-2 (bat) (QVN46569.1), MERS-CoV (YP_007188579.1), HKU5 (bat) (YP_001039962.1), MCoV (mouse) (AGT17716.1), HKU23 (camel) (QEY10673.1), BCoV (bovine) (AMQ23500.1), OC43 (YP_009555241.1), HKU24 (rat) (YP_009113025.1) and HKU1 (AXT92528.1).

Statistical analyses

Statistical analyses were performed using GraphPad Prism software v.9.2.0. Multiple group comparisons for plasma neutralization titres against a panel of SARS-CoV-2 pseudoviruses were applied using the Kruskal–Wallis test with false discovery rate method, and multiple testing correction was performed for each comparison using the Benjamini–Hochberg procedure at a 5% false discovery rate threshold. Factor change was calculated as the difference of geometric mean titre in the ID50 compared with that of the wild-type pseudotyped virus. Comparisons of neutralizing antibody responses by sex and smoke status were carried out using a paired Wilcoxon signed-ranked test. Correlation analysis between neutralizing antibody titres and age or BMI was carried out using Spearman’s test. One-way analysis of variance (ANOVA) and Tukey’s multiple comparison test were performed for statistical analysis of infectivity. A P value ≤0.05 was considered as statistically significant.

Ethical approval

The study protocol was approved by the Ethics Committee of the Faculty of Medicine at Universidad de Chile (project numbers 0361-2021 and 096-2020), Clínica Santa María (project number 132604-21) and CEC-Salud-UC (project number 200513023). All donors signed an informed consent, and their samples were anonymized.

Reporting Summary

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