key: cord-0839264-t0zpz3hv
authors: Supasa, Piyada; Zhou, Daming; Dejnirattisai, Wanwisa; Liu, Chang; Mentzer, Alexander J.; Ginn, Helen M.; Zhao, Yuguang; Duyvesteyn, Helen M.E.; Nutalai, Rungtiwa; Tuekprakhon, Aekkachai; Wang, Beibei; Paesen, Guido C.; Slon-Campos, Jose; López-Camacho, César; Hallis, Bassam; Coombes, Naomi; Bewley, Kevin; Charlton, Sue; Walter, Thomas S.; Barnes, Eleanor; Dunachie, Susanna J.; Skelly, Donal; Lumley, Sheila F.; Baker, Natalie; Shaik, Imam; Humphries, Holly; Godwin, Kerry; Gent, Nick; Sienkiewicz, Alex; Dold, Christina; Levin, Robert; Dong, Tao; Pollard, Andrew J.; Knight, Julian C.; Klenerman, Paul; Crook, Derrick; Lambe, Teresa; Clutterbuck, Elizabeth; Bibi, Sagida; Flaxman, Amy; Bittaye, Mustapha; Belij-Rammerstorfer, Sandra; Gilbert, Sarah; Hall, David R.; Williams, Mark A.; Paterson, Neil G.; James, William; Carroll, Miles W.; Fry, Elizabeth E.; Mongkolsapaya, Juthathip; Ren, Jingshan; Stuart, David I.; Screaton, Gavin R.
title: Reduced neutralization of SARS-CoV-2 B.1.1.7 variant by convalescent and vaccine sera
date: 2021-02-18
journal: Cell
DOI: 10.1016/j.cell.2021.02.033
sha: 911eb529be00b6cc6b7590f852beb5e24446fa8d
doc_id: 839264
cord_uid: t0zpz3hv

SARS-CoV-2 has caused over 2M deaths in little over a year. Vaccines are being deployed at scale, aiming to generate responses against the virus spike. The scale of the pandemic and error-prone virus replication is leading to the appearance of mutant viruses and potentially escape from antibody responses. Variant B.1.1.7, now dominant in the UK, with increased transmission, harbours 9 amino-acid changes in the spike, including N501Y in the ACE2 interacting-surface. We examine the ability of B.1.1.7 to evade antibody responses elicited by natural SARS-CoV-2 infection or vaccination. We map the impact of N501Y by structure/function analysis of a large panel of well-characterised monoclonal antibodies. B.1.1.7 is harder to neutralize than parental virus, compromising neutralization by some members of a major class of public antibodies through light chain contacts with residue 501. However, widespread escape from monoclonal antibodies or antibody responses generated by natural infection or vaccination was not observed.

Since its first appearance in Wuhan in December 2019 SARS-CoV-2 rapidly spread around the world leading the WHO to declare a Pandemic on 11 March 2020. Since then, drastic public health measures, including draconian lock-downs with severe economic cost, have been enacted to contain virus spread. Although initially successful at containing disease many countries are now experiencing further waves of infection, coinciding with winter in the Northern hemisphere, with infections in some countries outpacing those seen during the first wave (Kröger and Schlickeiser, 2021) . (https://blogs.sciencemag.org/pipeline/archives/2021/01/29/jj-and-novavax-data). In parallel, a number of potently neutralizing monoclonal antibodies have been developed which are in late stage trials to be used prophylactically or therapeutically (Yang et al., 2020) .

SARS-CoV-2 is a large positive-stranded RNA virus; the major virion surface glycoprotein is the trimeric spike which attaches the virus to host cells via the ACE2 receptor and, through a series of conformational changes, allows fusion of host and virion membranes releasing the virus RNA into the cell to start the infection cycle (Hoffmann et al., 2020; Ou et al., 2020) .

Spike is the target of RNA (Polack et al., 2020; Baden et al., 2020) , viral vectored (Voysey et al., 2020) and inactivated virus and recombinant protein based vaccines (Yadav et al., 2020) .

Because of the huge number of genome replications that occur in infected populations and error-prone replication, viral mutations do and will continue to occur (Robson et al., 2020) .

Although the vast majority will be inconsequential or detrimental to viral fitness, a few may give the virus a competitive advantage and be the subject of rapid natural selection relating to transmission advantage, including enhanced replication and immune evasion. This leads to the emergence of dominant new variant viruses. Coronaviruses, as we are seeing with COVID-19, have the potential to alter their proteins with dramatic effect (Denison et al., 2011) .

In recent months a number of mutations in the spike protein have been exemplified by viruses which have grown in alternative hosts such as mink and transmitted back to humans or in immunocompromised chronically infected individuals (Kemp et al., 2020; Oude Munnink et al., 2020; Hayashi et al., 2020) . Whilst most of these mutations currently show little evidence of a selective advantage in humans, variants have been identified with multiple mutations in spike which appear to have distinct selective advantages and have rapidly expanded in prevalence, notably that first identified in Kent in the UK (lineage B.1.1.7) and unrelated variants detected in South Africa (501Y.V2 also known as B. 1.351) and Manaus in Brazil (P.1). All of these contain mutations in the ACE2 receptor binding footprint of the RBD, one in B.1.1.7, three in 501Y.V2 and three in P.1, with the N501Y mutation being common to all. It is believed that these mutations in the ACE2 receptor binding domain increase the affinity for ACE2 (Zahradník et al., 2021) . These mutations also J o u r n a l P r e -p r o o f fall within the footprint of a number of potent neutralizing antibodies likely to afford vaccine induced protection and of several candidate therapeutic monoclonal antibodies (Cheng et al., 2021; Nelson et al., 2021) , thus potentially affording mutant viruses greater fitness to infect new hosts and also to escape from pre-existing antibody responses.

Such variants will continue to appear, indeed global surveillance by sequencing of viral isolates is wholly inadequate, and many may already be present but undetected. The B.1.1.7

variant was first identified in a sequence taken from a patient at the end of Sept 2020 (Rambaut et al., 2020) . The variant has rapidly become dominant in many areas of the UK which has coincided with a rapid increase of infections during the second wave of the pandemic, with cases and hospitalizations in excess of those seen during the first phase. The B.1.1.7 variant is estimated to be 30-60% more infectious than strains encountered in the first wave (Walker et al., 2021) and able to overcome public health efforts to contain infection. B.1.1.7 contains a total of 9 changes in the spike protein: N501Y, A570D, D614G, P681H, T716I, S982A, D1119H and deletions of residues 69-70 and 144. Mutation N501Y perhaps gives the greatest concern as it has the potential to increase RBD/ACE2 affinity whilst also disrupting the binding of potent neutralizing antibodies ( Figure. 1A) .

Here, we describe analysis of the cross reactivity of the antibody responses to earlier SARS- 

Analysis of 180,000 sequences from the COG-UK database (https://www.cogconsortium.uk)

showed one major and two minor subgroups ( Figure S1A ) of the ~13700 identifiable variants harbouring N501Y distinguishable by cluster analysis (Methods). The major strain was the Δ69-70 B.1.1.7 strain, whereas the two minor groups lack this deletion and either had wild-type or the S982A mutation ( Figure S1B) . A mixture of all three subgroups at the outset in late October resolved into the dominance of the B.1.1.7 variant over the course of about two months ( Figure S1A ), suggesting the Δ69-70 mutation may be driving the evolutionary advantage over these two other forms. On the week of Dec 24 th 2020, the N501Y variant represented 53% of the sequenced isolates in Great Britain.

The RBD may be likened to a classical human torso, in this analogy the shoulders and neck are involved in interactions with the ACE2 receptor ( Figure 1B ,C) (Dejnirattisai et al., 2020) .

In this context residue 501 lies within the footprint of the receptor on the right shoulder and J o u r n a l P r e -p r o o f is involved in hydrophobic interactions, especially with the side chains of residues Y41 and K353 of ACE2 with the 501 mutation from N to Y offering the opportunity for enhanced interactions ( Figure 1C,D) .

It has been reported that mutations at 501 can increase affinity for ACE2 (Starr et al., 2020; Gu et al., 2020) , although these data are not for the mutation to Y. In contrast Zahradník et al., (Zahradník et al., 2021) report direct selection of N501Y when evolving the RBD to enhance affinity. We therefore investigated the effect of this mutation on ACE2 binding by RBD using biolayer interferometry (BLI) ( Figure 1E) . The results indicate a marked (7-fold) increase in binding affinity due to a slower off-rate: WT RBD(501N)-ACE2: K D 75.1 nM (K on 3.88E4 /Ms, K off 2.92E-3 /s), RBD(501Y)-ACE2: K D 10.7 nM (K on 6.38E4 /Ms, K off 6.85E-4/s). This is in-line with enhanced interactions of the tyrosine side-chain with the side chains of residues Y41 and K353 of ACE2 ( Figure 1D ). In the context of a multivalent interaction at the cell surface this effect would be amplified. This alone might account for the selection of the N501Y mutation and an increase in transmission.

To investigate the effect of the N501Y mutation on antibody binding we took advantage of our set of 377 monoclonal antibodies (80 of which mapped to the RBD) generated from SARS-CoV-2 cases infected during the first wave of the pandemic in the UK using samples collected before June 2020 (Dejnirattisai et al., 2020) . In that study neutralization titres established that all 20 potent neutralizing antibodies (FRNT50 < 0.1µg/ml for the Victoria virus) bound the RBD; with the single exception of mAb 159, which bound the NTD. As (Yuan et al., 2020; Liu et al., 2020) .

Analysis of the position of the N501Y change with respect to the binding of all structurally characterised potent monoclonal antibodies suggests that the binding of over half of the antibodies would be unaffected by the change (Figure 2A ). However, one class of public antibodies have attracted particular attention, those using IGHV3-53 (Dejnirattisai et al., 2020; Yuan et al., 2020; Wu et al., 2020) . For these, and the IGHV3-66 antibodies, the mode of binding is dictated by the HC CDR1 and CDR2, which orientates the antibody such that the light chain CDR1 region lies atop residue 501. We would expect the majority of these antibodies to be affected by the mutation, since for them, unlike ACE2, the interaction with the asparagine is strongly favourable ( Figure 2B) .

To examine the effects on antibody binding we performed BLI experiments comparing the binding of potently neutralizing mAbs to RBDs containing 501Y and 501N (Methods, Figure   2C ). The results are mapped to the RBD in Figure 2D . As expected, there is little effect on J o u r n a l P r e -p r o o f many potent antibodies, for instance the IGVH1-58 antibodies: 55, 165, 253 and 318. There is a marked ~3-fold effect for mAb 40 (IGHV3-66) and for most of the important IGHV3-53

antibodies (150, 158 and 175) . However, there is a correlation between the LC for the IGHV3-53 antibodies and the magnitude of the effect, thus the common IGLV1-9 antibodies (mAbs 150 and 158) show a consistent reduction in affinity of roughly 3-fold ( Figure 3A) . In contrast mAb 222 which pairs IGHV3-53 with IGLV3-20 shows no reduction. When modelled using the most similar light chain from the PDB, IGLV3-20 does not contact residue 501 ( Figure 3B ). However, a survey of the various structures determined shows that IGHV3-53

frequently pairs with IGLV3-20 and often results in 501 contacts. mAb 269, which is also a VH3-53 mAb, paired with the IGLV1-9 light chain however, appears hyper-sensitive to the mutation (30-fold effect). The structure of Fab 269 in complex with WT RBD determined at 1.8 Å resolution (Methods, Table S1, Figure 3C ) shows similar interactions to those observed for mAbs 150 and 158. In order to understand this further, we determined the crystal structure of Fab 269 in complex with RBD harbouring 501Y at 2.2 Å resolution (Methods, Table S1 ). The result is shown in Figure 3C . The mutation introduces a rather small displacement of the L1 loop ( Figure 3D REGN10933 showed a slight reduction but still retained potent activity ( Figure 4B , Table S2 ).

The neutralisation of the AZ antibodies AZD1061 and AZD8895 was similarly little affected.

During the first wave of infection, before the emergence of B.1.1.7 strain, we collected a number of samples from cases at convalescence (4-9 weeks following infection) for the generation of monoclonal antibodies. Stored plasma from these cases was used in neutralization assays comparing Victoria and B.1.1.7 ( Figure 5A ). We analysed 34 convalescent samples including the WHONIBSC 20/130 reference serum and although a few sera showed near identical FRNT 50 values, the FRNT50 dilutions for the B.1.1.7 strain were 2.9-fold lower (geometric mean) than those for the Victoria strain (p<0.0001).

We also assayed neutralization of the B.1.1.7 and Victoria strains using serum obtained from recipients of the Oxford-AstraZeneca and Pfizer vaccines. For the AstraZeneca AZD1222 vaccine, serum was obtained at baseline and at 14 and 28 days following the second dose.

For the Pfizer vaccine, serum was obtained 7-17 days following the second dose of vaccine which was administered 3 weeks after the first dose (participants were seronegative at entry). Neutralization assays against B.1.1.7 and Victoria strains showed a 2.5-fold (geometric mean n=15 p=<0.0001) and 2.1-fold (geometric mean, n=10 p<0.002) reduction in the neutralization titres between B.1.1.7 and Victoria strains for the AstraZeneca vaccine after 14 and 28 days following the second dose respectively ( Figure 5B ). For the Pfizer-BioNTech vaccine BNT162b2, the reduction was 3.3-fold (geometric mean, n=25 p<0.0001) ( Figure 5C ).

Finally, we obtained plasma from 13 patients infected with B. Figure 6A ). At early time points neutralization titres were low or absent except in 1 case taken at day 1 of illness who showed identical neutralization of both viruses and was the highest titre of all the samples we have measured in this study, we speculate that this may represent a reinfection with B.1.1.7. For these samples as a whole there was a no significant difference between the neutralization titres for the two viruses ( Figure 6B ) meaning that infection with B.1.1.7 will afford protection from infection with earlier variants.

In conclusion, the neutralization assays on convalescent and vaccine serum revealed that (Dejnirattisai et al., 2020) the binding site has not been completely disrupted.

The level of expression of ACE2 has been shown to correlate with likelihood of infection by SARS-CoV-1 (Jia et al., 2005) and the higher affinity for ACE2 of SARS-CoV-2 has been imputed to underlie its greater transmission. It is reasonable to assume that a further increase in affinity will increase the likelihood of the stochastic events of virus attachment resulting in localisation for sufficient time to trigger, perhaps by the recruitment of J o u r n a l P r e -p r o o f additional receptors, internalisation of the virus. As noted by Zahradník, J. et al. (Zahradník et al., 2021) in a situation where public health measures reduce R0 to below 1 there will be selective pressure to increase receptor affinity.

Here we show that this increase in transmission is compounded by the reduction in neutralization potency of antibodies generated by prior infection. Modification of the ACE2 binding surface of the RBD would be predicted to directly disrupt the binding of antibodies that lose affinity to the mutated residues. However, antibodies that neutralise by ACE2 competition, even if not directly affected by the mutation will have to compete with ACE2

for binding to the RBD, and mutations of RBD that increase the affinity of ACE2 will tip the equilibrium away from mAb/RBD interaction toward RBD/ACE2 making the virus more difficult to neutralize.

Mutation at 484 of the spike likely has a similar dual effect and Zahradník, J. et al. (Zahradník et al., 2021) report that further affinity increase in ACE2 binding is possible.

Although most effort has been directed at generating antibodies that neutralise by blocking ACE2 binding, other mechanisms are possible (Huo et al., 2020; Zhou et al., 2020) and

indeed partial or non-neutralising antibodies may confer protection (Dunand et al., 2016) .

Such antibodies would likely be unaffected by mutations in the ACE2 binding site and they deserve more thorough investigation since they would form excellent components in therapeutic cocktails. In addition, natural exposure and vaccination may confer protective immunity against symptomatic and severe COVID-19 via memory T cell responses (Sariol and Perlman, 2020; Altmann and Boyton, 2020) .

The recent description of a number of virus variants which appear to have developed independently is a cause for concern as it may signal the emergence of strains able to evade vaccine induced antibody responses. There is now an imperative to closely survey the emergence of novel SARS-CoV2 strains on a global basis and to quickly understand the consequences for immune escape. There is a need to define correlates of protection from SARS-CoV-2 and also to understand how T cells contribute to protection in addition to the antibody response. It is also imperative to understand whether the newly emerging strains There is already work underway to modify vaccines directed toward viral variants, these modified vaccines can be built quickly to incorporate new strains but they will be given to individuals with pre-existing immunity to ancestral strains, whether the modified vaccines will be able to effectively redirect the antibody response to the areas of difference in the novel strains rather than simply boosting the pre-existing response will need intensive study.

In summary, we describe here a modest reduction in the neutralization titres against B. 2Fo-Fc maps are contoured at 1.2 σ and coloured in blue in both panels. The negative density (red) in (A) is contoured at -3 σ, and the positive density (green) in (B) at 3 σ.

Resources, reagents and further information requirement should be forwarded to and will be responded by the Lead Contact, David I Stuart (dave@strubi.ox.ac.uk).

Reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.

The coordinates and structure factors of the SARS-CoV-2 RBD-N501/Fab269 and SARS-CoV-2 J o u r n a l P r e -p r o o f

Pfizer vaccine serum was obtained 7-17 days following the second dose of vaccine which was administered 3 weeks after the first dose (participants were to the best of their knowledge seronegative at entry). Data from vaccinated volunteers who received two vaccinations are included in this paper.

Vaccine doses were either 5 × 10 10 viral particles (standard dose; SD/SD cohort n=21) or half dose as their first dose (low dose) and a standard dose as their second dose (LD/SD cohort n=4). The interval between first and second doses was in the range of 8-14 weeks. Blood samples were collected and serum separated on the day of vaccination and on pre-specified days after vaccination e.g. 14 and 28 days after boost.

Plasma and peripheral blood mononuclear cells were collected from individual with SARS- 

J o u r n a l P r e -p r o o f

All COG-UK sequences were downloaded on 24 th January 2020, and the translated protein sequences were roughly to the wild-type reference from start and stop codons between nucleotides 21000-25000, and filtered on the mutation 501Y. Sequence alignment was carried out, and identified mutations were plotted as red balls (single point mutations) or black balls (deletions) on the modelled C-alpha positions of the spike structure, size

proportional to the logarithm of the incidence. Residues which mutated at an incidence greater than 0.3% compared to the wild-type were labelled explicitly.

The constructs of native RBD and ACE2 are the same as in Zhou et al., (Zhou et al., 2020) . To clone RBD N501Y, a construct of native RBD was used as the template and two primers of RBD (Forward primer 5'-CTACGGCTTTCAGCCCACATACGGTGTGGGCTACCAGCCTT-3' and reverse primer 5'-AAGGCTGGTAGCCCACACCGTATGTGGGCTGAAAGCCGTAG-3') and two primers of pNEO vector (Forward primer 5'-CAGCTCCTGGGCAACGTGCT-3' and reverse primer 5'-CGTAAAAGGAGCAACATAG-3') were used to do PCR. Amplified DNA fragments were digested with restriction enzymes AgeI and KpnI and then ligated with digested pNEO vector. This construct encodes exactly the same protein as native RBD except the N501Y mutation, as confirmed by sequencing.

Protein expression and purification were performed as described in Zhou et al. (Zhou et al., 2020) and Dejnirattisai et al. (Dejnirattisai et al., 2021) . Briefly RBD and mAb were expressed in 293T cells, His-tagged RBD was purified on Ni-NTA and mAb on protein-A. The Regeneron and AstraZeneca antibodies were supplied by AstraZeneca.

J o u r n a l P r e -p r o o f

Fab fragments of 269 antibody were digested and purified using Pierce Fab Preparation Kit, following the manufacturer's protocol.

BLI experiments were run on an Octet Red 96e machine (Fortebio). To measure the binding affinities of monoclonal antibodies with native RBD and RBD N501Y, RBD and RBD N501Y

were immobilized onto AR2G biosensors (Fortebio) separately. Monoclonal antibodies were used as analytes. To measure the binding affinities of native RBD and RBD N501Y with ACE2, native RBD and RBD N501Y were immobilized onto AR2G biosensors separately. Serial dilutions of ACE2 were used as analytes. Data were recorded using software Data Acquisition 11.1 (Fortebio) and analysed using software Data Analysis HT 11.1 (Fortebio) with a 1:1 fitting model.

269 Fab was mixed with RBD or N501Y RBD in a 1:1 molar ratio with a final concentration of 9.9 mg ml −1 . After incubation at room temperature for 30 min, the sample was used for initial screening of crystals in Crystalquick 96-well X plates (Greiner Bio-One) with a Cartesian Robot using the nanoliter sitting-drop vapor-diffusion method as previously described (Walter et al., 2003) . Figure S2) . Mass spectrometry and biolayer interferometry data confirmed the presence of tyrosine at 501. Data collection and structure refinement statistics are given in Table S1 . Structural comparisons used SHP (Stuart et al., 1979) , residues forming the RBD/Fab interface were identified with PISA (Krissinel and Henrick, 2007) and figures were prepared with PyMOL (The PyMOL Molecular Graphics System, Version 1.2r3pre, Schrödinger, LLC).

The neutralization potential of an antibody was measured using a Focus Reduction Neutralization Test (FRNT), where the reduction in the number of the infected foci is compared to a no antibody negative control well. Briefly, serially diluted Ab or plasma was mixed with SARS-CoV-2 strain Victoria or B.1.1.7 and incubated for 1 hr at 37 °C. The mixtures were then transferred to 96-well, cell culture-treated, flat-bottom microplate containing confluent Vero cell monolayers in duplicate and incubated for further 2 hrs, followed by the addition of 1.5% semi-solid carboxymethyl cellulose (CMC) overlay medium to each well to limit virus diffusion. A focus forming assay was then performed by staining Vero cells with human anti-NP mAb (mAb206) followed by peroxidase-conjugated goat antihuman IgG (A0170; Sigma). Finally, the foci (infected cells) approximately 100 per well in the absence of antibodies, were visualized by adding TrueBlue Peroxidase Substrate. Virusinfected cell foci were counted on the classic AID EliSpot reader using AID ELISpot software.

The percentage of focus reduction was calculated and IC 50 was determined using the probit program from the SPSS package.

J o u r n a l P r e -p r o o f Statistical analyses are reported in the results and figure legends. Neutralization was measured by FRNT. The percentage of focus reduction was calculated and IC 50 was determined using the probit program from the SPSS package.The Wilcoxon matched-pairs signed rank test was used for the analysis and two-tailed P values were calculated and geometric mean values. BLI data were analysed using Data Analysis HT 11.1 (Fortebio) with a 1:1 fitting model.

Centre For Tropical Medicine and Global Health

PHASER (McCoy et al., 2007) using search models of SARS-CoV-2 RBD/scFv269 complex (PDB ID, 7BEM) and the ChCl domains of SARS-CoV-2 RBD/158 complex

Cyclic model rebuilding with COOT (Emsley and Cowtan, 2004) and refinement with PHENIX

SARS-CoV-2 T cell immunity: Specificity, function, durability, and role in protection

Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine

Structures of Human Antibodies Bound to SARS-CoV-2 Spike Reveal Common Epitopes and Recurrent Features of Antibodies

Antibody cocktail to SARS-CoV-2 spike protein prevents rapid mutational escape seen with individual antibodies

Isolation and rapid sharing of the 2019 novel coronavirus (SARS-CoV-2) from the first patient diagnosed with COVID-19 in Australia

Potent SARS-CoV-2 Neutralizing Antibodies Directed Against Spike N-Terminal Domain Target a Single Supersite Lead Contact

Impact of South African 501.V2 Variant on SARS-CoV-2 Spike Infectivity and Neutralization: A Structure-based Computational Assessment', bioRxiv

A neutralizing human antibody binds to the N-terminal domain of the Spike protein of SARS-CoV-2'

The Antigenic Anatomy of SARS-CoV-2 Receptor Binding Domain

Coot: Model-building tools for molecular graphics

Comprehensive mapping of mutations to the SARS-CoV-2 receptorbinding domain that affect recognition by polyclonal human serum antibodies', bioRxiv

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

Highly conserved binding region of ACE2 as a receptor for SARS-CoV-2 between humans and mammals

Both Neutralizing and Non-Neutralizing Human H7N9 Influenza Vaccine-Induced Monoclonal Antibodies Confer Protection

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Neutralization of SARS-CoV-2 by Destruction of the Prefusion Spike

ACE2 Receptor Expression and Severe Acute Respiratory Syndrome Coronavirus Infection Depend on Differentiation of Human Airway Epithelia

Neutralising antibodies in Spike mediated SARS-CoV-2 adaptation', medRxiv

Inference of Macromolecular Assemblies from Crystalline State

Forecast for the second Covid-19 wave based on the improved SIR model with a constant ratio of recovery to infection rate'

Molecular determinants and mechanism for antibody cocktail preventing SARS-CoV-2 escape'

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor'

Macromolecular structure determination using X-rays, neutrons and electrons: Recent developments in Phenix

Potent neutralizing antibodies against multiple epitopes on SARS-CoV-2 spike'

Phaser crystallographic software

Molecular dynamic simulation reveals E484K mutation enhances spike RBD-ACE2 affinity and the combination of E484K, K417N and N501Y mutations (501Y.V2 mutant', bioRxiv

Characterization of spike glycoprotein of SARS-CoV-2 on virus entry and its immune cross-reactivity with SARS-CoV'

Transmission of SARS-CoV-2 on mink farms between humans and mink and back to humans

Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

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

Coronavirus RNA Proofreading: Molecular Basis and Therapeutic Targeting

Lessons for COVID-19 Immunity from Other Coronavirus Infections

Deep Mutational Scanning of SARS-CoV-2 Receptor Binding Domain Reveals Constraints on Folding and ACE2 Binding

Crystal structure of cat muscle pyruvate kinase at a resolution of 2.6 Å'

Neutralizing and protective human monoclonal antibodies recognizing the N-terminal domain of the SARS-CoV-2 spike protein', bioRxiv

Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against

-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK', The Lancet

Increased infections, but not viral burden, with a new SARS-CoV-2 variant', medRxiv

A procedure for setting up high-throughput nanolitre crystallization experiments. I. Protocol design and validation'

SARS-CoV-2 501Y.V2 escapes neutralization by South African COVID-19 donor plasma'. bioRxiv

Xia2: An expert system for macromolecular crystallography data reduction

DIALS : implementation and evaluation of a new integration package

An Alternative Binding Mode of IGHV3-53 Antibodies to the SARS-CoV-2

Receptor Binding Domain

Recombinant vaccines for COVID-19

COVID-19 antibody therapeutics tracker: a global online database of antibody therapeutics for the prevention and treatment of COVID-19

Structural basis of a shared antibody response to SARS-CoV-2'

SARS-CoV-2 RBD in vitro evolution follows contagious mutation spread, yet generates an able infection inhibitor', bioRxiv

This work was supported by the Chinese Academy of Medical Sciences (CAMS) Innovation