key: cord-0887600-1wfeatf0
authors: Noy-Porat, Tal; Makdasi, Efi; Alcalay, Ron; Mechaly, Adva; Levi, Yinon; Bercovich-Kinori, Adi; Zauberman, Ayelet; Tamir, Hadas; Yahalom-Ronen, Yfat; Israeli, Ma’ayan; Epstein, Eyal; Achdout, Hagit; Melamed, Sharon; Chitlaru, Theodor; Weiss, Shay; Peretz, Eldar; Rosen, Osnat; Paran, Nir; Yitzhaki, Shmuel; Shapira, Shmuel C.; Israely, Tomer; Mazor, Ohad; Rosenfeld, Ronit
title: Tiger team: a panel of human neutralizing mAbs targeting SARS-CoV-2 spike at multiple epitopes
date: 2020-05-20
journal: bioRxiv
DOI: 10.1101/2020.05.20.106609
sha: ce8819ef441658b6480cfee528f120fae2ba67d7
doc_id: 887600
cord_uid: 1wfeatf0

The novel highly transmissible human coronavirus SARS-CoV-2 is the causative agent of the COVID-19 pandemic. Thus far, there is no approved therapeutic drug, specifically targeting this emerging virus. Here we report the isolation and characterization of a panel of human neutralizing monoclonal antibodies targeting the SARS-CoV-2 receptor binding domain (RBD). These antibodies were selected from a phage display library constructed using peripheral circulatory lymphocytes collected from patients at the acute phase of the disease. These neutralizing antibodies are shown to recognize distinct epitopes on the viral spike RBD, therefore they represent a promising basis for the design of efficient combined post-exposure therapy for SARS-CoV-2 infection.

The present global pandemic of coronavirus induced disease , declared by the World Health Organization (WHO) as a public health emergency of international concern, is caused by the highly transmissible severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). To this date, about 5 million confirmed cases and over 300,000 deaths have been reported worldwide 1 . Yet, there is no approved therapeutic drug specifically targeting the SARS-CoV-2. The novel coronavirus SARS-CoV-2, emerged as the seventh type of coronavirus infecting humans and the third most pathogenic preceded by the Severe Acute Respiratory Syndrome coronavirus (SARS-CoV) and the Middle East Respiratory Syndrome coronavirus (MERS-CoV).

Due to their exceptional antigen specificity, therapeutic monoclonal antibodies (mAbs) are considered an attractive candidate to target exposed antigenic sites on viruses, and prevent their infectivity 2 . The therapeutic efficacy of mAbs, specifically targeting surface viral proteins was demonstrated before for several viruses including SARS-CoV, MERS and Ebola 3-6 .

SARS-CoV-2 utilizes the surface homotrimeric Spike glycoprotein (S) as a major mediator of cellular infection. The SARS-CoV-2 S protein is composed of two distinct subunits, S1 and S2. The S1 subunit contains the receptor binding domain (RBD), known to bind the Angiotensin-Converting Enzyme 2 (ACE2) receptor on host cell surfaces 7 . The S2 subunit mediates the fusion of the viral and cellular membranes, essential for viral entry into the cell. The receptor interaction site on S1 is considered the main target for efficient neutralization of cell infection and therefore a prime candidate for therapeutic antibody development [8] [9] [10] [11] [12] . Although the S protein of the SARS-CoV and SARS-CoV-2 share 77.5% identity, most of the mAbs isolated against SARS-CoV reportedly failed to cross-neutralize the novel virus [13] [14] [15] .

For designing optimal therapeutic strategies, there is an urgent need for the identification of neutralizing monoclonal antibodies that specifically target SARS-CoV-2 (such antibodies may be identified either in humans in the course of illness/recovery, or in immunized animals). Recently, first two mAbs, elicited against the SARS-CoV-2, were reported 16 . Furthermore, efficient post-exposure therapy in humans, may require integration of several noncompeting mAbs, ideally neutralizing the virus infectivity by different mechanisms. Such combined therapies are expected to provide superior control of potential neutralizing escape variants 9, 17 .

Here we describe the isolation of a panel of neutralizing mAbs selected against the SARS-CoV-2 RBD from phage display library constructed based on patient samples collected in the acute phase of the disease. These specific antibodies were found to recognize distinct epitopes and can potentially be used for therapy and immune prophylaxis.

Sera and whole blood samples collected from convalescent or severe COVID-19 patients were kindly provided under written inform consent and treated in accordance with the biosafety guidelines of the IIBR in BL3 facility. PBMCs were separated from whole blood using density centrifugation by Ficoll. Sera samples were heat-inactivated (30 min at 60°C) prior to use for binding or neutralizing assays.

Total RNA was purified from PBMCs using RNeasy mini kit (Qiagen GmbH, Germany).

CDNA synthesis was performed using Verso cDNA synthesis kit (Thermoscientific, USA) and used as a template for Abs variable region coding fragments amplification. Heavy and light Ig variable domains (VH and VL) were then amplified, using specific primer set. The VH and VL used in PCR overlap extension reaction, resulted in scFv repertoire cloned into pCC16 phagemid vector 22 using NcoI/NotI. Total of 9.2e6 independent clones obtained (2e6 and 7.2e6), representing the library complexity.

Panning performed against huFc-RBD directly absorbed to polystyrene plates and against biotinylated-huFc-RBD (biotinylation performed using a commercial kit: EZ-Link sulfo-NHS-biotin, Pierce-Thermoscientific, USA) attached to streptavidin coated magnetic beads (Dynabeads Invitrogen, USA). All routine phage display techniques, performed as described 23 . Screen of specific binders, perform using phage ELISA against huFc-RBD Vs huFc-NTD as control. In addition, huFc-RBD fused protein was expressed using previously designed Fc-fused protein expression vector 26 , giving rise to a protein comprising a homodimer of two RBD moieties (amino acids 318-542, GenPept: QHD43416 ORF) fused to a dimer of IgG1 human Fc (huFc). Expression performed using ExpiCHO TM Expression system (ThermoFisher scientific, Gibco TM ).

Phagemid DNA of the desired clones were isolated using QIAprep spin Miniprep kit (Qiagen, GmbH, Hilden, Germany), and the entire scFv was cloned into a pcDNA3.1+

based expression vector that was modified, providing the scFv with the human (IgG1) CH2-CH3 Fc fragments, resulting in scFv-Fc antibody format. ScFv-Fc were expressed using ExpiCHO TM Expression system (ThermoFisher scientific, Gibco TM ) and purified on HiTrap Protein-A column (GE healthcare, UK).

Routine ELISA protocol applied essentially as described 27 

Binding studies were carried out using were cultured at 37°C, 5% CO2 at 95% air atmosphere.

SARS-CoV-2 (GISAID accession EPI_ISL_406862) was kindly provided by Bundeswehr 

Several blood samples, either of COVID-19 recovered patients or of patients with severe ongoing disease, were evaluated for titers of RBD binding and viral neutralization activity.

Two samples, exhibiting the highest neutralizing ability (NT50 >5000), demonstrating significant binding to both S1 subunit [DIL50 (half-dilution value) of 494 and 473] and RBD (DIL50 of 252 and 226; Fig. 1a) , were subsequently selected for antibody library generation.

A phage display (PD) single chain Fv (scFv) library, representing approx. 10 million distinct antibodies, was constructed. With the objective of isolating neutralizing Abs, three consecutive enrichment steps of panning were performed against both S1 and RBD.

Resulting clones were tested for their ability to bind S1, and those found positive were expressed as full-length antibodies (in a scFv-Fc format) for further analysis. Subsequently, eight RBD-specific antibodies carrying unique sequences were selected (Fig 1b) .

Binding specificity assays of these eight antibodies confirmed their specificity to the spike protein, the S1 subunit and the RBD, while no reactivity was observed against the Spike N terminal domain (NTD) of the spike protein (Fig. 1c) . Evaluation of the Abs affinity toward S1 by ELISA evidenced apparent KD values of 1. (Fig. 1e) . It should be noted that the biotinylation of MD47 was found to significantly hinder its ability to bind RBD and therefore was not evaluated in this format.

Additional sequence analysis by IgBlast 18 (Fig 1b) -23) and was the only mAb, carrying VL (IGLV3-21).

Binding curves of polyclonal antibodies in serially diluted serum samples of COVID-19 patients obtained by ELISA using S1 or RBD as coating antigen. b. Amino acid sequences of the HCDR3 and LCDR3 of the selected antibodies and their respective germ line genes. c. Specificity of the selected antibodies determined by ELISA against the indicated SARS-CoV-2 proteins. d. Reactivity profile of antibodies determined by ELISA, using S1 as the coating antigen. Data is presented as binding percent of Bmax for each antibody. e. Binding characteristics of the monoclonal antibodies determined using biolayer interferometry. All antibodies (except MD47) were biotinylated, immobilized to the sensor and interacted with increasing amounts of RBD. Binding kinetics were fitted using the 1:1 binding model. The values shown represent average ± SEM of triplicates.

SARS-CoV-2 spike RBD is known to mediate the binding of the human ACE2 receptor and thus, this domain is considered as main target for neutralizing mAbs. However, direct blocking of the RBD-ACE2 interaction is not the exclusive modality by which neutralizing antibodies can exert their effect 21 . Consequently, the selected mAbs, were classified on the basis of their epitope specificity determined by BLI epitope binning. In this assay, each individual antibody was biotinylated, immobilized to a streptavidin sensor, loaded with RBD and then challenged with each of the other antibodies. Simultaneous binding of the second antibody to RBD induces a wavelength shift in the interference pattern, which indicates that the two antibodies bind to non-overlapping epitopes 19 . Conversely, if the two antibodies bind the same or partially-overlapping epitope on RBD, no or very low wavelength shift, respectively, is induced. As a representative example, sensograms of the various antibody interactions with a pre-complexed MD65-RBD is shown (Fig. 2a) .

Antibody MD65 used as a negative control, and did not elicit any wavelength shift, as expected. In contrast, antibodies MD29, MD47, MD62 and MD63 induced a marked wavelength shift indicating that they could bind to RBD simultaneously with MD65. The analysis revealed that antibodies MD45 and MD67 could not bind to RBD in the presence of MD65, indicating that these three antibodies target the same epitope. Analysis was then performed for the next seven antibodies and the ability of each pair to simultaneously bind RBD was determined (Fig. 2b) . Results ranged from full to no competition and enabled the classification of the mAbs into 4 groups recognizing distinct epitopes (Fig 2c) : I (MD17, MD29 and MD63); II (MD45, MD65 and MD67); III (MD62) and IV (MD47).

Most notably, the classification of the antibodies on the basis of their specific targeted antigenic epitopes, is strongly supported by their observed sequence similarity, as discussed above. Group I mAbs shared the same V germ lines and were found to target the same epitope. Similarly, mAbs MD45 and MD67 which shared the same V germ lines and the same epitope, were classified as group II. Although MD65 differ in its VH germ line, he was included in this group as well. On the other hand, MD62 sharing the same VH germ line with MD45 and MD67, appears to bind a distinct epitope (III). Finally, MD47

exhibited both a unique sequence and targeted a unique epitope (Fig. 2c) .

Recently, the RBD-located epitope recognized by the SARS-CoV specific antibody CR3022, was determined 14, 20 . We chose to use this antibody in order to further determine the epitopes recognized by the antibodies described in this report. Therefore, a recombinant in-house version of this antibody was generated (Supplementary data) and used in epitope binning assays, together with the selected set of novel mAbs. CR3022 IgG was immobilized to the BLI sensor, loaded with RBD and further challenged with each of the selected mAbs. Group I mAbs were found to compete with the CR3022 as evidenced by the lack of interaction with the mAb-RBD complex ( Supplementary Fig. 1 ). The remaining five tested mAbs did bind the RBD in the presence of the CR3022 antibody. Thus, we conclude that antibodies MD17, MD29 and MD63 bind to the RBD epitope that spans the RBD residues 369-386, previously defined as the CR3022 epitope 20 . Complete epitope binning of the eight selected MD monoclonal antibodies. Binding was evaluated by the ability of each pair of antibodies to simultaneously bind RBD, using biolayer interferometry. c. Four non-competing RBD binding epitopes were identified and accordingly classified into four groups: I (blue), II (green), III (pink) and IV (yellow). d. SARS-CoV-2 in vitro neutralization using plaque reduction neutralization test (PRNT). Neutralization potency was determined by the ability of each antibody (at indicated concentrations) to reduce plaque formation; results are expressed as percent inhibition of control without Ab.

The neutralization potency of the antibodies was evaluated by plaque reduction test (PRNT) using VeroE6 cells infected with the pathogenic SARS-CoV-2. A fixed amount of the virus was incubated with increasing concentrations of each antibody, the complex was then added to the cells, and the number of plaques was quantified 48 hours later. Antibodies MD45, MD67 and MD65 displayed the highest neutralization potency, with a neutralization dose needed to inhibit 50% of the plaques (NT50) of 4 g/ml (MD45 and MD67) and 0.9 g/ml (MD65). MD65 exhibited the highest neutralization capacity amongst the entire set of antibodies (Fig. 2d) . Antibody MD62 was also found to effectively neutralize SARS-CoV-2 with a NT50 value of 7.5 g/ml.

Interestingly, antibodies MD17, MD29 and MD63 shared DNA sequence homology indicative of a common germ line and competed for the same epitope, yet only the first two showed neutralizing activity (NT50 of 87 and 67 g/ml, respectively). This discrepancy may be explained by differences in the affinities of the antibodies in this group. A recent report similarly documented that antibody CR3022 (targeting an epitope overlapping with MD17 and MD29, see above) does not neutralize the novel SARS-CoV-2 presumably due to its low binding affinity to the RBD (115 nM) 20 . Additional antibody tested in this study, MD47, which represents a unique epitopic group, display low virus neutralization of about 30% at the highest tested concentration.

To conclude, we report the isolation and characterization of a set of fully human SARS-CoV-2, neutralizing antibodies that target four distinct epitopes on the spike RBD. As neutralizing antibodies are generally known to be useful as post-exposure therapy for viral infection and more specifically for treatment of human corona viral diseases 3,21 , we suggest that these antibodies might serve for efficient treatment or prophylaxis of COVID-19

patients. Furthermore, since these neutralizing antibodies target different epitopes, they can be combined to further improve treatment efficacy and to reduce the risk of the emergence of treatment-escaping viral mutants.

The potential danger of suboptimal antibody responses in COVID-19

Potent human monoclonal antibodies against SARS CoV, Nipah and Hendra viruses

Antibodymediated protection against Ebola virus

Exceptionally potent neutralization of Middle East respiratory syndrome coronavirus by human monoclonal antibodies

Structural definition of a neutralization epitope on the N-terminal domain of MERS-CoV spike glycoprotein

Structure, Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein

Structural basis for the neutralization of MERS-CoV by a human monoclonal antibody MERS-27

Escape from human monoclonal antibody neutralization affects in vitro and in vivo fitness of severe acute respiratory syndrome coronavirus

Structural basis for potent cross-neutralizing human monoclonal antibody protection against lethal human and zoonotic severe acute respiratory syndrome coronavirus challenge

Structural bases of coronavirus attachment to host aminopeptidase N and its inhibition by neutralizing antibodies

Structural basis of neutralization by a human anti-severe acute respiratory syndrome spike protein antibody, 80R

Publisher Correction: A human monoclonal antibody blocking SARS-CoV-2 infection

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

A SARS-like cluster of circulating bat coronaviruses shows potential for human emergence

A noncompeting pair of human neutralizing antibodies block COVID-19 virus binding to its receptor ACE2

A single amino acid change within antigenic domain II of the spike protein of bovine coronavirus confers resistance to virus neutralization

IgBLAST: an immunoglobulin variable domain sequence analysis tool

A biolayer interferometry-based assay for rapid and highly sensitive detection of biowarfare agents

A highly conserved cryptic epitope in the receptor binding domains of SARS-CoV-2 and SARS-CoV

Neutralizing Antibodies against SARS-CoV-2 and Other Human Coronaviruses

A human synthetic combinatorial library of arrayable single-chain antibodies based on shuffling in vivo formed CDRs into general framework regions

Phage display of single-chain antibody constructs

Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation

SARS-CoV-2 Seroconversion in Humans: A Detailed Protocol for a Serological Assay, Antigen Production, and Test Setup

Acetylcholinesterase-Fc Fusion Protein (AChE-Fc): A Novel Potential Organophosphate Bioscavenger with Extended Plasma Half-Life

Isolation of Anti-Ricin Protective Antibodies Exhibiting High Affinity from Immunized Non-Human Primates

Development of a tissue-culture-based enzyme-immunoassay method for the quantitation of anti-vaccinia-neutralizing antibodies in human sera

We greatly thank our department friends: Ofer Cohen, Emmanuel Mamrud, Moshe Mantzur, Moshe Aftalion, Noa Caspi, Yentel Yivgy, Hila Cohen, Liat Bar-On, RonitAloni-Grinstein, Didi Gur, Sarit Sterenberg and Tamar Aminov for fruitful discussions and support. We also deeply acknowledge the IIBR management board for being accommodating to our needs. We are appreciating Dalit Brener for graphical design. 

Patent application for the described antibodies was filed by the Israel Institute for Biological Research.