key: cord-262958-tmp6yxlv
authors: Pinto, Dora; Park, Young-Jun; Beltramello, Martina; Walls, Alexandra C.; Tortorici, M. Alejandra; Bianchi, Siro; Jaconi, Stefano; Culap, Katja; Zatta, Fabrizia; De Marco, Anna; Peter, Alessia; Guarino, Barbara; Spreafico, Roberto; Cameroni, Elisabetta; Case, James Brett; Chen, Rita E.; Havenar-Daughton, Colin; Snell, Gyorgy; Telenti, Amalio; Virgin, Herbert W.; Lanzavecchia, Antonio; Diamond, Michael S.; Fink, Katja; Veesler, David; Corti, Davide
title: Structural and functional analysis of a potent sarbecovirus neutralizing antibody
date: 2020-04-09
journal: bioRxiv
DOI: 10.1101/2020.04.07.023903
sha: 
doc_id: 262958
cord_uid: tmp6yxlv

SARS-CoV-2 is a newly emerged coronavirus responsible for the current COVID-19 pandemic that has resulted in more than one million infections and 73,000 deaths1,2. Vaccine and therapeutic discovery efforts are paramount to curb the pandemic spread of this zoonotic virus. The SARS-CoV-2 spike (S) glycoprotein promotes entry into host cells and is the main target of neutralizing antibodies. Here we describe multiple monoclonal antibodies targeting SARS-CoV-2 S identified from memory B cells of a SARS survivor infected in 2003. One antibody, named S309, potently neutralizes SARS-CoV-2 and SARS-CoV pseudoviruses as well as authentic SARS-CoV-2 by engaging the S receptor-binding domain. Using cryo-electron microscopy and binding assays, we show that S309 recognizes a glycan-containing epitope that is conserved within the sarbecovirus subgenus, without competing with receptor attachment. Antibody cocktails including S309 along with other antibodies identified here further enhanced SARS-CoV-2 neutralization and may limit the emergence of neutralization-escape mutants. These results pave the way for using S309 and S309-containing antibody cocktails for prophylaxis in individuals at high risk of exposure or as a post-exposure therapy to limit or treat severe disease.

SARS-CoV-2 is a newly emerged coronavirus responsible for the current COVID-23

19 pandemic that has resulted in more than one million infections and 73,000 24 deaths 1,2 . Vaccine and therapeutic discovery efforts are paramount to curb the 25 pandemic spread of this zoonotic virus. The SARS-CoV-2 spike (S) glycoprotein 26 promotes entry into host cells and is the main target of neutralizing antibodies. 27

Here we describe multiple monoclonal antibodies targeting SARS-CoV-2 S 28 identified from memory B cells of a SARS survivor infected in 2003. One 29 antibody, named S309, potently neutralizes SARS-CoV-2 and SARS-CoV 30 pseudoviruses as well as authentic SARS-CoV-2 by engaging the S receptor-31 binding domain. Using cryo-electron microscopy and binding assays, we show 32 that S309 recognizes a glycan-containing epitope that is conserved within the 33 sarbecovirus subgenus, without competing with receptor attachment. Antibody 34 cocktails including S309 along with other antibodies identified here further 35 enhanced SARS-CoV-2 neutralization and may limit the emergence of 36 neutralization-escape mutants. These results pave the way for using S309 and 37 S309-containing antibody cocktails for prophylaxis in individuals at high risk of 38 exposure or as a post-exposure therapy to limit or treat severe disease.

Coronavirus entry into host cells is mediated by the transmembrane spike (S) 41 glycoprotein that forms homotrimers protruding from the viral surface 3 . The S 42 glycoprotein comprises two functional subunits: S1 (divided into A, B, C and D domains) 43 that is responsible for binding to host cell receptors and S2 that promotes fusion of the 44 viral and cellular membranes 4,5 . Both SARS-CoV-2 and SARS-CoV belong to the 45 sarbecovirus subgenus and their S glycoproteins share 80% amino acid sequence 46 identity 6 . SARS-CoV-2 S is closely related to the bat SARS-related CoV (SARSr-CoV) 47

RaTG13 with which it shares 97.2% amino acid sequence identity 1 . We and others 48 recently demonstrated that human angiotensin converting enzyme 2 (hACE2) is a 49 functional receptor for SARS-CoV-2, as is the case for SARS-CoV 1,6-8 . The ranging between 1.4 and 6,100 ng/ml, and 0.8 and 254 ng/ml, respectively (Fig. 1a-94 b). MAbs were further evaluated for binding to the SARS-CoV-2 and SARS-CoV S B 95 domains as well as to the prefusion-stabilized OC43 S 28 , MERS-CoV S 29,30 , SARS-96 CoV S 30 and SARS-CoV-2 S 6 ectodomain trimers. None of the mAbs studied bound to 97 prefusion OC43 S or MERS-CoV S ectodomain trimers, indicating a lack of cross-98 reactivity outside the sarbecovirus subgenus (Extended Data Fig.1) . MAbs S303, 99 S304, S309 and S315 recognized the SARS-CoV-2 and SARS-CoV RBDs. In 100 particular, S309 bound with nanomolar affinity to both S B domains, as determined by 101 biolayer interferometry (Fig. 1c-d, Extended Data Fig. 2) . Unexpectedly, S306 and 102 S310 stained cells expressing SARS-CoV-2 S at higher levels than those expressing 103 SARS-CoV S, yet it did not interact with SARS-CoV-2 or SARS-CoV S ectodomain 104 trimers and RBD constructs by ELISA. These results suggest that they may recognize 105 post-fusion SARS-CoV-2 S, which was recently proposed to be abundant on the 106 surface of authentic SARS-CoV-2 viruses 31 (Fig. 1a-b and Extended Data Fig.3) . 107 the S 3-fold molecular axis. CDRH3 and CDRL2 sandwich the SARS-CoV-2 S glycan 142 at position N343 through contacts with the core fucose moiety (in agreement with the 143 detection of SARS-CoV-2 N343 core-fucosylated peptides by mass-spectrometry 34 ) 144 and to a lesser extent with the core N-acetyl-glucosamine (Fig. 2d) . These latter 145 interactions bury an average surface of ~170 Å 2 and stabilize the N343 oligosaccharide 146 which is resolved to a much larger extent than in the apo SARS-CoV-2 S structures 6,9 . 147

The structural data explain the S309 cross-reactivity between SARS-CoV-2 and 148 SARS-CoV as 19 out of 24 residues of the epitope are strictly conserved ( Fig. 2f and  149 Extended Data Fig. 6a To further investigate the mechanism of S309-mediated neutralization, we 175 compared side-by-side transduction of SARS-CoV-2-MLV in the presence of either 176 S309 Fab or S309 IgG. Both experiments yielded comparable IC50 values (3.8 and 3.5 177 nM, respectively), indicating similar potencies for IgG and Fab (Fig. 3d) . However, The 178 S309 IgG reached 100% neutralization, whereas the S309 Fab plateaued at ~80% 179 neutralization (Fig. 3d) . This result indicates that one or more IgG-specific bivalent 180 mechanisms, such as S trimer cross-linking, steric hindrance or aggregation of 181 virions 37 , may contribute to the ability to fully neutralize pseudovirions. 

To gain more insight into the epitopes recognized by our panel of mAbs, we 205 used structural information, escape mutants analysis 23,27,30 , and biolayer 206 inteferometry-based epitope binning to map the antigenic sites present on the SARS-207

CoV and SARS-CoV-2 S B domains ( Fig. 4a and Extended Data Fig.10 ). This analysis 208 identified at least four antigenic sites within the S B domain of SARS-CoV targeted by 209 our panel of mAbs. The receptor-binding motif, which is targeted by S230, S227 and 210 S110, is termed site I. Sites II and III are defined by S315 and S124, respectively, and 211 the two sites were bridged by mAb S304. Site IV is defined by S309, S109, and S303 212 mAbs. Given the lower number of mAbs cross-reacting with SARS-CoV-2, we were 213 able to identify sites IV targeted by S309 and S303, and site II-III targeted by S304 and 214 S315 (Fig. 4b) . Movie frame alignment, estimation of the microscope contrast-transfer function 439 parameters, particle picking and extraction were carried out using Warp 49 . Particle 440 images were extracted with a box size of 800 binned to 400 yielding a pixel size of 1.05 441 Å. For each data set two rounds of reference-free 2D classification were performed 442 using cryoSPARC 50 to select well-defined particle images. Subsequently, two rounds 443 of 3D classification with 50 iterations each (angular sampling 7.5˚ for 25 iterations and 444 1.8˚ with local search for 25 iterations), using our previously reported closed SARS-445

CoV-2 S structure 6 as initial model, were carried out using Relion 51 without imposing 446 symmetry to separate distinct SARS-CoV-2 S conformations. 3D refinements were 447 carried out using non-uniform refinement along with per-particle defocus refinement in 448 cryoSPARC 50 . Particle images were subjected to Bayesian polishing 52 before 449 performing another round of non-uniform refinement in cryoSPARC 50 followed by per-450 particle defocus refinement and again non-uniform refinement. supplemented with 30% glycerol. The dataset was collected at ALS beamline 5.0.2 472 and processed to 3.3 Å resolution in space group P41212 using mosflm 64 and 473

Aimless 65 . The structure of Fab S309 was solved by molecular replacement using 474

Phaser 66 and homology models as search models. The coordinates were improved 475 and completed using Coot 55 and refined with REFMAC5 67 . Crystallographic data 476 collection and refinement statistics are shown in Table 3 . S309 mAb. a-b , Ribbon diagrams of S309 and ACE2 bound to SARS-CoV-2 S B . This composite model was generated using the SARS-CoV-2 S/S309 cryoEM structure reported here and a crystal structure of SARS-CoV-2 S bound to ACE2 16 . c, Competition of S309 or S230 mAbs with ACE2 to bind to SARS-CoV S B (left panel) and SARS-CoV-2 S B (right panel). ACE2 was immobilized at the surface of biosensors before incubation with S B domain alone or S B precomplexed with mAbs. The vertical dashed line indicates the start of the association of mAb-complexed or free S B to solid-phase ACE2. d, Neutralization of SARS-CoV-MLV by S309 IgG1 or S309 Fab, plotted in nM (means ±SD is shown, one out of two experiments is shown). e, mAb-mediated ADCC using primary NK effector cells and SARS-CoV-2 S-expressing ExpiCHO as target cells. Bar graph shows the average area under the curve (AUC) for the responses of 3-4 donors genotyped for their FcgRIIIa (mean±SD, from two independent experiments). f, Activation of high affinity (V158) or low affinity (F158) FcgRIIIa was measured using Jurkat reporter cells and SARS-CoV-2 Sexpressing ExpiCHO as target cells (one experiment, one or two measurements per mAb). g, mAb-mediated ADCP using Cell Trace Violet-labelled PBMCs as phagocytic cells and PKF67labelled SARS-CoV-2 S-expressing ExpiCHO as target cells. Bar graph shows the average area under the curve (AUC) for the responses of four donor (mean±SD, from two independent experiments). h, Activation of FcgRIIa measured using Jurkat reporter cells and SARS-CoV-2 S-expressing ExpiCHO as target cells (one experiment, one or two measurements per mAb). Fig. 9) . b, Competition of mAb pairs for binding to the SARS-CoV-2 S B domain . c-d, Neutralization of SARS-CoV-2-MLV by S309 combined with an equimolar amount of S304 or S315 mAbs. For mAb cocktails the concentration on the x axis is that of the individual mAbs. Table 1 : Characteristics of the antibodies described in this study. VH and VL % identity refers to V gene identity compared to germline (IMGT). 

Isolation and characterization of a bat SARS-like coronavirus 599 that uses the ACE2 receptor

Site-specific 602 analysis of the SARS-CoV-2 glycan shield. bioRxiv

Genomic characterisation and epidemiology of 2019 novel 605 coronavirus: implications for virus origins and receptor binding

Structure of SARS coronavirus spike 608 receptor-binding domain complexed with receptor

Occupancy and mechanism in antibody-611 mediated neutralization of animal viruses

IgG Fc engineering to modulate antibody 614 effector functions

Alveolar macrophages are critical for broadly-reactive antibody-617 mediated protection against influenza A virus in mice

Differential Fc-Receptor Engagement Drives an 620

A neutralizing antibody selected from plasma cells that binds to 623 group 1 and group 2 influenza A hemagglutinins

Fc receptor but not complement binding is important in 626 antibody protection against HIV

Human monoclonal antibody combination against SARS 629 coronavirus: synergy and coverage of escape mutants

Specificity, cross-reactivity, and function of antibodies elicited 632 by Zika virus infection

Enhanced antibody half-life improves in vivo activity

Potent binding of 2019 novel coronavirus spike protein by a SARS 637 coronavirus-specific human monoclonal antibody

Automated molecular microscopy: the new Leginon system

Addressing preferred specimen orientation in single-particle 642 cryo-EM through tilting

Real-time cryo-electron microscopy data 645 preprocessing with Warp

cryoSPARC: 648 algorithms for rapid unsupervised cryo-EM structure determination

New tools for automated high-resolution cryo-EM structure 651 determination in RELION-3

A Bayesian approach to beam-654 induced motion correction in cryo-EM single-particle analysis

Prevention of overfitting in cryo-EM structure 657 determination

Visualizing density maps with UCSF 659 Chimera

Features and development 662 of Coot

Tools for macromolecular model building and refinement into 665 electron cryo-microscopy reconstructions

Automated structure refinement of macromolecular 668 assemblies from cryo-EM maps using Rosetta

Automatically fixing errors in glycoprotein structures with 672

MolProbity: all-atom structure validation for macromolecular 674 crystallography

EMRinger: side chain-directed model and map validation for 677 3D cryo-electron microscopy

Macromolecular structure determination using X-rays, 681 neutrons and electrons: recent developments in Phenix

Privateer: software for the conformational validation of 684 carbohydrate structures

UCSF ChimeraX: Meeting modern challenges in 687 visualization and analysis

689 iMOSFLM: a new graphical interface for diffraction-image processing with 690 MOSFLM

How good are my data and what is the 693 resolution?

Phaser crystallographic software

REFMAC5 for the refinement of macromolecular crystal 698 structures

Identifying SARS-CoV-2 related coronaviruses in Malayan 701 pangolins