key: cord-0985156-415qt2s2 authors: Banach, Bailey B.; Cerutti, Gabriele; Fahad, Ahmed S.; Shen, Chen-Hsiang; de Souza, Matheus Oliveira; Katsamba, Phinikoula S.; Tsybovsky, Yaroslav; Wang, Pengfei; Nair, Manoj S.; Huang, Yaoxing; Urdániz, Irene M. Francino; Steiner, Paul J.; Gutiérrez-González, Matias; Liu, Lihong; López Acevedo, Sheila N.; Nazzari, Alexandra; Wolfe, Jacy R.; Luo, Yang; Olia, Adam S.; Teng, I-Ting; Yu, Jian; Zhou, Tongqing; Reddem, Eswar R.; Bimela, Jude; Pan, Xiaoli; Madan, Bharat; Laflin, Amy D.; Nimrania, Rajani; Yuen, Kwon-Tung; Whitehead, Timothy A.; Ho, David D.; Kwong, Peter D.; Shapiro, Lawrence; DeKosky, Brandon J. title: Paired heavy and light chain signatures contribute to potent SARS-CoV-2 neutralization in public antibody responses date: 2021-01-03 journal: bioRxiv DOI: 10.1101/2020.12.31.424987 sha: 437267f17de3a840e6b7a61eb79c1cabc5dbaa2a doc_id: 985156 cord_uid: 415qt2s2 Understanding protective mechanisms of antibody recognition can inform vaccine and therapeutic strategies against SARS-CoV-2. We discovered a new antibody, 910-30, that targets the SARS-CoV-2 ACE2 receptor binding site as a member of a public antibody response encoded by IGHV3-53/IGHV3-66 genes. We performed sequence and structural analyses to explore how antibody features correlate with SARS-CoV-2 neutralization. Cryo-EM structures of 910-30 bound to the SARS-CoV-2 spike trimer revealed its binding interactions and ability to disassemble spike. Despite heavy chain sequence similarity, biophysical analyses of IGHV3-53/3-66 antibodies highlighted the importance of native heavy:light pairings for ACE2 binding competition and for SARS-CoV-2 neutralization. We defined paired heavy:light sequence signatures and determined antibody precursor prevalence to be ~1 in 44,000 human B cells, consistent with public antibody identification in several convalescent COVID-19 patients. These data reveal key structural and functional neutralization features in the IGHV3-53/3-66 public antibody class to accelerate antibody-based medical interventions against SARS-CoV-2. Highlights A molecular study of IGHV3-53/3-66 public antibody responses reveals critical heavy and light chain features for potent neutralization Cryo-EM analyses detail the structure of a novel public antibody class member, antibody 910-30, in complex with SARS-CoV-2 spike trimer Cryo-EM data reveal that 910-30 can both bind assembled trimer and can disassemble the SARS-CoV-2 spike Sequence-structure-function signatures defined for IGHV3-53/3-66 class antibodies including both heavy and light chains IGHV3-53/3-66 class precursors have a prevalence of 1:44,000 B cells in healthy human antibody repertoires 2020c). dhACE2 competition ELISA assays at pH 5.5 and 4.5 showed that IGHV3-53/3-66 class 288 members compete in a concentration-dependent manner with dimeric human ACE2 for binding 289 to SARS-CoV-2 S2P spike, and to the D614G S2P spike (Fig. 4B, Supplemental Fig. 3) . Using 290 single-cycle surface plasmon resonance, we found that the extremely potent mAb 1-20 291 recognized S protein and RBD with no loss in affinity at endosomal pH, whereas the less potent 292 antibodies 910-30 and B38 showed reduced affinity in the endosomal pH range (Fig. 4C, 293 Supplemental Fig. 4) . We compared authentic virus neutralization IC50 potencies (from Fig. 2D ) 294 to the ratio of mAb-Spike affinity (Supplemental Fig. 4 ) divided by reported dhACE2-Spike affinity 295 (Zhou et al., 2020b) , which suggested that potent mAb neutralization was correlated with mAb 296 affinity across all pH values tested (Fig. 4D) . Finally, a qualitative Octet pH series analysis using 297 D614 S2P spike showed that as the pH reduces (and RBDs preferentially rotate down), the potent 298 neutralizer mAb 1-20 exhibited strong recognition of D614 S2P spike for pH³6.0, whereas 910-299 30 showed reduced binding below pH=6.5, and the least potent B38 binding showed reduced 300 binding below pH=7.0 (Fig. 2E, left panel) . In contrast, all class members maintained strong 301 binding to mutant D614G S2P spike into the endosomal pH range (where one RBD likely remains 302 up), and the potent antibody class member 1-20 recognized D614G spike down to pH 4.0 ( Fig. 303 2E, right panel). These data suggested that the most potent antibodies can maintain the bound 304 state (and stabilize the RBD-up conformation) more effectively under endosomal pH conditions 305 for D614 S2P spike, whereas all antibody class members could effectively recognize the native 306 RBD-up conformation for D614G across a broad pH range. Our data support ACE2 competition 307 as a functional signature of IGHV3-53/3-66 public antibody class neutralization, and we show that 308 the RBD-up vs. RBD-down conformation substantially influenced the ability of IGHV3-53/3-66 309 class antibodies to recognize spike trimer. (lambda) residues in CDR-L1 that make key contributions to RBD recognition. We also note that 319 class member light chains use common aromatic/hydrophobic residues 28 Val, 29 Ile/Val, or 320 30/32 Tyr30/32 to achieve similar interactions with 505 Tyr in the RBD, which is part of the shared 321 ACE2 and IGHV3-53/3-66 class binding epitope. These shared light chain features illuminate the 322 structural rationale for broader light chain diversity among IGHV3-53/3-66 class members. 323 The frequency of anti-SARS-CoV-2 IGHV3-53/3-66 precursor antibodies in healthy donors 324 (around 1 in 44,000) was more common than the previously studied anti-HIV-1 VRC01-class 325 antibody precursors observed in 1 per 1-4 million antibodies (Zhou et al., 2013) . In addition, it has 326 been shown that anti-HIV-1 VRC01-class antibodies also require much higher levels of somatic 327 hypermutation (SHM) to achieve potent neutralization (Zhou et al., 2013) . The comparably limited 328 SHM required for anti-SARS-CoV-2 IGHV3-53/3-66 class antibodies appears to be a feature of 329 IGHV germline gene neutralizing interactions and the need to recognize highly conserved viral 330 variants, as compared to HIV-1 broadly neutralizing antibodies that must recognize broadly 331 diverse viral variants and show limited germline gene neutralization. These findings help explain 332 the observed reproducibility of public IGHV3-53/3-66 anti-RBD antibodies in convalescent 333 COVID-19 patients. 334 D614G S2P spike variant shows a greater prevalence of RBD-up than D614G, which may 335 enhance spike and the ACE2 host receptor recognition to confer higher D614G viral infectivity 336 (Hou et al., 2020; Mansbach et al., 2020; Yurkovetskiy et al., 2020; . 337 Conversely, a sustained RBD 'up' also could make the virus more sensitive to neutralization, as 338 the exposed 'up' RBD enhances exposure of vulnerable epitopes (Mansbach et or Spike divided by dimeric ACE2 affinity to RBD or Spike. Table Legends 527 528 Table S1 . Cryo-EM data collection and refinement statistics for 910-30 Fab in complex with 529 SARS-CoV-2 spike at pH 5.5. 530 531 A structure-based method was applied to define sequence signatures for the HV3-53/3-66 798 class COVID neutralizing antibody (Zhu et al., 2013) . Briefly, protein structures of IGHV3-53/3-66 799 antibodies complexed with RBD or spike were selected for analysis, and the buried surface area 800 (BSA) between antibody and RBD was calculated by the PDBePISA server 801 (https://www.ebi.ac.uk/pdbe/pisa/). We examined the BSA larger than 20 Å 2 , and residues making 802 contacting with the RBD projected surface that were encoded by the conserved germline 803 sequence were selected as initial class sequence signatures, and amino acids from somatic 804 hypermutations were used to refine the signature of the class antibody. For germline sequence 805 alignments, heavy and light chain germline sequences were downloaded from IMGT (Lefranc et 806 al., 2003) sample was applied to a glow-discharged carbon-coated copper grid. The grid was washed with 822 a buffer with the same pH as the sample buffer (10 mM HEPES with 150 mM NaCl for pH 7.4; 10 mM acetate with 150 mM NaCl for the lower pH values). Protein molecules adsorbed to the carbon 824 were negatively stained with 0.75% uranyl formate. Datasets were collected using a 825 ThermoFisher Talos F200C electron microscope equipped with a Ceta CCD camera. The 826 microscope was operated at 200 kV, the pixel size was 2.53 Å (nominal magnification: 57,000), 827 and the defocus was set at -1.2 µm. Particles were picked and extracted automatically using in-828 house written software (YT, unpublished). 2D classification was performed using Relion 1.4 829 (Scheres, 2012) . Binding of mAbs 4-3, B38, 910-30, and 1-20 to SAR-CoV-2 S2P D614 and D614G variants 866 was assessed an a FortéBio Octet HTX instrument (FortéBio). Experiments were run in tilted 867 black 384-well plates (Geiger Bio-One) at 30°C and 1,000 rpm agitation. Running buffer was 868 comprised of 10mM of the corresponding pH buffer plus 150mM NaCl, 0.02% Tween20, 0.1% 869 BSA and 0.05% sodium azide. The following buffers were used to achieve the range of pH: pH 9 870 (borate), pH 8.5 (Tris), pH 8 (Tris), pH 7.4 (PBS), pH 7 (HEPES), pH 6.5 (MES), pH 6 (MES), pH 871 5.5 (NaAc), pH 5 (NaAc), pH 4.5 (NaAc), pH 4.2 (NaAc), pH 4.0 (NaAc). 300nM IgG solution was 872 used for immobilization at pH 7.4 on anti-human IgG Fc capture biosensors (FortéBio) that were 873 pre-hydrated for 30 minutes. Sensors were then equilibrated in 7.4pH buffer for 30 seconds 874 1-20 IgGs were tested over the biotinylated S2P surfaces at four concentrations ranging from 1-889 27nM, while B38 and 4-3 were tested at four concentrations ranging from 3-81nM, to account for 890 higher binding KDs. Biotinylated RBD was captured over independent flow cells at 250-500 RU 891 and B38 was tested at four concentrations ranging from 3-81nM, HKU910-30 and 4-3 were tested 892 at four concentrations of 1-27 nM and 1-20 at four concentrations ranging from 0.333-9 nM, to 893 account for differences in their binding affinities. To avoid the need for surface regeneration that 894 arises with the slowly dissociating interactions, we used single-cycle kinetics binding experiments. 895 The four concentrations for each IgG were prepared in running buffers at each of pH, using a 896 three-fold dilution series. 897 Binding of HKU910-30, 4-3 and B38 over the S2P or RBD surface as well as over a 898 streptavidin reference surface was monitored for 120s, followed by a dissociation phase of 120s-899 1080s depending on the interaction at 50μL/min. For the interaction of 1-20 with the RBD, which 900 showed an unusually slow dissociation rate, an extended dissociation phase of 4500s was 901 necessary to extrapolate accurate apparent dissociation constants. Four blank buffer single 902 cycles were performed by injecting running buffer instead of Fab to remove systematic noise from 903 the binding signal. The data was processed and fit to 1:1 single cycle model using the Scrubber 904 2.0 (BioLogic Software). The results from these assays, are reported in terms of apparent kinetic 905 parameters and KDs to account for potential avidity effects arising from the binding of bivalent 906 IgGs to trivalent S2P. Supplementary Studies in humanized mice and convalescent humans yield a 992 SARS-CoV-2 antibody cocktail Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor SARS-CoV-2 D614G variant exhibits efficient 999 replication ex vivo and transmission in vivo Neutralization of SARS-CoV-2 by 1002 Destruction of the Prefusion Spike Structural basis for potent neutralization 1005 of SARS-CoV-2 and role of antibody affinity maturation A distinct name is 1007 needed for the new coronavirus Human neutralizing antibodies elicited by SARS-CoV-2 infection Tracking Changes in SARS-CoV-2 Spike: 1012 Evidence that D614G Increases Infectivity of the COVID-19 Virus A novel coronavirus associated with severe acute 1015 respiratory syndrome Ultrasonically-guided flow focusing generates precise emulsion droplets for high-1018 throughput single cell analyses IMGT unique numbering for immunoglobulin and T cell 1021 receptor variable domains and Ig superfamily V-like domains Potent neutralizing antibodies against multiple epitopes on SARS-1024 CoV-2 spike Aerodynamic analysis of SARS-CoV-2 in two Wuhan hospitals Molecular Architecture of Early 1030 Dissemination and Massive Second Wave of the SARS-CoV-2 Virus in a Major Metropolitan 1031 The SARS-CoV-2 Spike Variant D614G Favors an Open Conformational State Ultra-high-1036 throughput sequencing of the immune receptor repertoire from millions of lymphocytes Establishment and validation of a pseudovirus neutralization assay for Characterization of spike glycoprotein of SARS-CoV-2 on virus entry and its immune cross-1043 reactivity with SARS-CoV UCSF Chimera--a visualization system for exploratory research and 1046 analysis UCSF ChimeraX: Structure visualization for researchers, educators, and 1049 developers Spike mutation D614G alters SARS-CoV-2 1052 fitness and neutralization susceptibility cryoSPARC: algorithms 1054 for rapid unsupervised cryo-EM structure determination Convergent antibody responses CoV-2 in convalescent individuals Isolation of potent SARS-CoV-2 neutralizing antibodies and protection 1060 from disease in a small animal model RELION: implementation of a Bayesian approach to cryo-EM structure 1062 determination OLGA: fast 1064 computation of generation probabilities of B-and T-cell receptor amino acid sequences and 1065 motifs Analysis of a SARS-CoV-2-Infected 1068 Individual Reveals Development of Potent Neutralizing Antibodies with Limited Somatic 1069 Mutation A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2 The Panitumumab EGFR Complex Reveals a Binding Mechanism That Overcomes Cetuximab 1075 Induced Resistance High frequency of shared clonotypes in human B 1078 cell receptor repertoires Deep mutational scanning of SARS-1081 CoV-2 receptor binding domain reveals constraints on folding and ACE2 binding Automated molecular microscopy: the new Leginon system Coronavirus 1086 membrane fusion mechanism offers a potential target for antiviral development Antibody Combination against SARS Coronavirus: Synergy and Coverage of Escape Mutants Temporal profiles of viral load in posterior 1094 oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: 1095 an observational cohort study SARS-CoV-2 challenge via multiple mechanisms Evaluating the effects of SARS Spike mutation D614G on transmissibility and pathogenicity Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein Functional interrogation and mining of natively 1107 paired human V H :V L antibody repertoires A human 1110 monoclonal antibody blocking SARS-CoV-2 infection SARS-CoV-2 Neutralizing Antibody Responses Are More Robust in 1113 Patients with Severe Disease Escape from neutralizing 1116 antibodies by SARS-CoV-2 spike protein variants D614G Spike Mutation Increases 1119 SARS CoV-2 Susceptibility to Neutralization Optimization of affinity, specificity and 1122 function of designed influenza inhibitors using deep sequencing Generation of VSV Pseudotypes Using Recombinant ΔG-VSV for Studies 1124 on Virus Entry, Identification of Entry Inhibitors, and Immune Responses to Vaccines SARS and MERS: 1127 recent insights into emerging coronaviruses Structural Basis for Potent 1133 Neutralization of Betacoronaviruses by Single-Domain Camelid Antibodies An Alternative Binding Mode Antibodies to the SARS-CoV-2 Receptor Binding Domain A noncompeting pair of human neutralizing antibodies block COVID-19 virus binding 1140 to its receptor ACE2 IgBLAST: an immunoglobulin variable 1142 domain sequence analysis tool Structural basis of a shared antibody response to SARS-CoV-2 1147 (2020b). A highly conserved cryptic epitope in the receptor binding domains of SARS-CoV-2 1148 and SARS-CoV Structural and Functional Analysis of 1151 the D614G SARS-CoV-2 Spike Protein Variant The D614G mutation in the SARS-CoV-2 spike protein reduces S1 shedding and 1157 increases infectivity Structural basis for the neutralization 1160 of SARS-CoV-2 by an antibody from a convalescent patient A pneumonia outbreak associated with a new coronavirus of probable 1164 bat origin Multidonor Analysis Reveals Structural Elements Genetic Determinants, and Maturation Pathway for HIV-1 Neutralization by VRC01-Class 1168 Cryo-EM Structures of SARS-CoV-2 Spike 1171 without and with ACE2 Reveal a pH-Dependent Switch to Mediate Endosomal Positioning of 1172 Receptor-Binding Domains. Cell Host Microbe Advanced Online Publication De novo identification of 1176 VRC01 class HIV-1-neutralizing antibodies by next-generation sequencing of B-cell transcripts A novel coronavirus from patients with pneumonia in China Rapid isolation and profiling of a diverse panel of 1182 human monoclonal antibodies targeting the SARS-CoV-2 spike protein We thank Jennifer Hackett from the Genome Sequencing Core Lab at the