key: cord-0777724-kdyw5xaf
authors: Guo, Youjia; Kawaguchi, Atsushi; Takeshita, Masaru; Sekiya, Takeshi; Hirohama, Mikako; Yamashita, Akio; Siomi, Haruhiko; Murano, Kensaku
title: Potent mouse monoclonal antibodies that block SARS-CoV-2 infection
date: 2021-01-30
journal: J Biol Chem
DOI: 10.1016/j.jbc.2021.100346
sha: 4267b304f55f99fe2ea5db034409bd8ae9fb9868
doc_id: 777724
cord_uid: kdyw5xaf

Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has developed into a global pandemic since its first outbreak in the winter of 2019. An extensive investigation of SARS-CoV-2 is critical for disease control. Various recombinant monoclonal antibodies of human origin that neutralize SARS-CoV-2 infection have been isolated from convalescent patients and will be applied as therapies and prophylaxis. However, the need for dedicated monoclonal antibodies suitable for molecular pathology research is not fully addressed. Here, we produced six mouse anti-SARS-CoV-2 spike monoclonal antibodies that exhibit not only robust performance in immunoassays including western blotting, ELISA, immunofluorescence, and immunoprecipitation, but also demonstrate neutralizing activity against SARS-CoV-2 infection to VeroE6/TMPRSS2 cells. Due to their mouse origin, our monoclonal antibodies are compatible with the experimental immunoassay setups commonly used in basic molecular biology research laboratories, providing a useful tool for future research. Furthermore, in the hope of applying the antibodies of clinical setting, we determined the variable regions of the antibodies and used them to produce recombinant human/mouse chimeric antibodies.

We adopted the design principle reported by Wrapp et al. (37) , in which the SARS-CoV-2 spike protein was engineered to form a stable homotrimer that was resistant to proteolysis during protein preparation. In our practice, recombinant spike protein RBD and ectodomain were constructed. A T4 fabritin trimerization motif (foldon) was incorporated into the C-terminal of the recombinant spike ectodomain to promote homotrimer formation (38) (Fig. 1A) . Recombinant RBD proteins tagged with GST or MBP were produced using an E. coli expression system (Fig. 1B) . Both recombinant spike protein RBD and ectodomain (S∆TM) were produced using a mammalian expression system that retained proper protein glycosylation equivalent to that observed during virus replication (Fig. 1C , S1A). Mice were immunized with these recombinant spike proteins to generate antibodies against the SARS-CoV-2 virus, followed by cell fusion to generate a hybridoma-producing antibody. Culture supernatants were pre-screened by enzyme-linked immunosorbent assay (ELISA), western blotting (WB), and immunoprecipitation (IP), and six monoclonal hybridomas were isolated and evaluated.

To characterize these antibodies in detail, they were first purified from the culture supernatant and examined in terms of ELISA and WB performance. Four 7 where RBD and S∆TM glycoproteins were pulled down in their native conformation.

Of note, we found that S1D7 and S3D8 could maintain intact IP efficiency under highly stringent experimental conditions where sodium dodecyl sulfate (SDS) was present (Fig.   S2B ).

Next, we examined whether our antibodies could be used in the immunofluorescence assay (IF). An antibody applicable for IP would also have activity in IF. Cellular localization of spike proteins is essential for elucidating the mechanism of packaging and maturation of virions during release from the cellular membrane. We tested our antibodies' performance in IF using HeLa cells overexpressing spike protein with the transmembrane domain. Consistent with their performance in the above-mentioned assays ( Fig. 2A and 2B) , both S1D7 and S3D8 could detect spike proteins expressed homogeneously on the apical side of HeLa cells with a high signal-to-noise ratio ( Fig. 2C and S2C ). However, their localization pattern is different from that observed for SARS-CoV-1 spike proteins, which are exclusively localized in the Golgi during infection (39) 

The manner in which antibodies bind and pull-down spike glycoproteins in an J o u r n a l P r e -p r o o f IP experiment resembles the process of antibody-mediated neutralization, where spike-ACE2 interaction is intercepted by competitive binding between neutralizing antibodies and spike glycoprotein. We then examined whether they were capable of inhibiting spike-ACE2 binding or even neutralizing SARS-CoV-2 infection. First, we performed a spike pull-down assay in which the spike glycoprotein was pulled down by ACE2 in the presence of monoclonal antibodies ( Fig. 3A and S3A ). Clones S1D7 and S3D8 clearly inhibited spike-ACE2 binding, as shown by the dimmed spike signal in WB (Fig. 3B ). To quantify the inhibition ability, we performed a bead-based neutralization assay by measuring the amount of ACE2 bound to RBD beads after blocking with monoclonal antibodies (Fig. 3C) . Antibodies R22 and R31 showed no disruption of ACE2-RBD interaction, whereas S1D7 and S3D8 showed robust hindrance of ACE2-RBD binding with IC 50 values of 248.2 ng/mL and 225.6 ng/mL, respectively ( Fig. 3D and 3E ). S1D7 and S3D8's abilities to inhibit spike-ACE2 binding was consistent with their superior performance in IP experiments. Four monoclonal antibodies derived from the antigen produced by E. coli (Clones R15, R22, R31, and R52) were found to recognize continuous epitope 549-TGVLTESNKKFLPFQQFGRD-568 of spike protein RBD (Figs. S3B-D). In contrast, an epitope of two antibodies from mammalian cells (S1D7 and S3D8) could not be determined (Figs. S3B-D). The fact that they fail to recognize segmented RBD suggests that they recognize an intact tertiary structure of the spike protein.

Next, we asked whether our antibodies inhibit SARS-CoV-2 infection in VeroE6/TMPRSS2 (TM2) cells, which is susceptible to SARS-CoV-2 infection J o u r n a l P r e -p r o o f compared with the parental VeroE6 cell line by expressing TMPRSS2 (41) . In WB, antibodies R52 and R22, but not S1D7 and S3D8, could detect spike glycoprotein along with the progression of SARS-CoV-2 infection in VeroE6/TM2 cells (Fig. 4A ). On the other hand, S1D7 and S3D8 were applicable to IF in infected VeroE6/TM2 cells. Spike showed a punctate distribution pattern in the perinuclear region resembling ER and ERGIC (42) (Fig. 4B) Recombinant human/mouse chimeric antibodies R52h and S1D7h are applicable

Our mouse antibodies would not be applicable for use in clinical treatment, if not chimeric and humanized, due to their immunogenicity (30, 43) . In the hope of applying the antibodies of clinical settings, the variable regions of the antibodies were determined (Table S2) , followed by the production of recombinant antibodies based on plasmid transfection to Expi293 or 293T cell lines. We selected three antibodies from among these and generated humanized chimeric antibodies designated as R52h, S1D7h, J o u r n a l P r e -p r o o f and S3D8h by fusing them with the constant region of human IgG1κ for R52, S1D7, and S3D8. R52h was capable of detecting artificial spike glycoprotein carrying T4 foldon, and native spike glycoprotein expressed in 293T cells on WB as well as R52

( Fig. 5A and S4A ). In IF, S1D7h could detect spike proteins expressed in HeLa cells ( Fig. 5B and S4B) . Notably, S1D7h and S3D8h showed robust hindrance of ACE2-RBD binding with IC 50 values of 116.3 ng/mL and 137.2 ng/mL, respectively (Fig. 5C ).

Emerging SARS-CoV-2 is a global public health threat to society, which is predicted to be long-term for several years (44 

Experimental procedures are provided as supporting information.

All data are contained within the manuscript. 

The authors declare no competing interests. H. Detection of spike proteins expressed in 293T cells. Lysates of 293T cells expressing artificial spikes carrying T4 foldon or wild-type spike glycoproteins were separated by SDS-PAGE, followed by WB using antibody R52.

A. Immunoprecipitation (IP) of trimeric glycosylated spike protein (S∆TM) using established monoclonal antibodies. S1, S1D7; S3, S3D8; ni, non-immune mouse Higher IP efficiency was observed in clone R22, R31, S1D7, and S3D8.

C. Immunofluorescence (IF) staining of spike glycoprotein expressed in HeLa cells with monoclonal antibodies S1D7 and S3D8. Spike protein localized on the apical surface of transfected HeLa cells. Scale bar, 30 µm.

A. A schematic of the spike pull-down assay designed to evaluate inhibition of ACE2-spike binding by monoclonal antibody. Spike glycoprotein lacking TM J o u r n a l P r e -p r o o f domain (S∆TM) was mixed with a monoclonal antibody. ACE2-SBP was applied to capture S∆TM onto streptavidin beads competitively. Captured S∆TM was detected by WB as a measurement of the antibody's inhibitory ability. S1, S1D7; S3, S3D8;

ni, non-immune mouse IgG.

B. WB of spike pull-down assay using antibody R52. In the presence of clones S1D7 and S3D8, ACE2 was not able to pull down S∆TM.

C. Schematic of bead-based neutralization assay designed to quantify inhibition of ACE2-RBD binding by monoclonal antibody. RBD-SBP glycoprotein immobilized on streptavidin beads was mixed with a monoclonal antibody. ACE2-FLAG was applied to bind competitively with RBD. ACE2-RBD binding was quantified by measuring the signal given by an anti-FLAG antibody conjugated with APC fluorophore using FACS.

D. One set of representative FACS results of a bead-based neutralization assay in the presence of 4 µg/mL monoclonal antibodies. Clones S1D7 and S3D8 significantly inhibited ACE2-RBD interaction, shown as lowered fluorescence intensity of APC.

E. Binding profiles of potent neutralizing antibodies. ni, non-immune mouse IgG.

Error bars indicate standard deviation (n=3). Clones R22 and R31 showed no inhibition of ACE2-RBD binding, while S1D7 and S3D8 inhibited ACE2-RBD interaction at lower ng/mL levels. A. Recombinant spike glycoproteins were treated with HRV3C protease to remove SBP-tag before immunizing mice.

B. Clone R52 showed the highest performance on western blotting among our antibodies and detected even 0.08 ng S∆TM glycoprotein.

A. Quantification of signal intensity of spike glycoprotein immunoprecipitated by monoclonal antibodies. S1, S1D7; S3, S3D8; ni, non-immune IgG. Error bars indicate standard deviation (n=3).

B. Monoclonal antibody clones S1D7 and S3D8 maintain high efficiency even in the presence of 0.1% SDS. S1, S1D7; S3, S3D8; ni, non-immune IgG; In, input. C. ELISA using monoclonal antibodies against 6×His-tagged RBD segments secreted in the culture medium of 293T cells. Clone R15, R22, R31, and R52 recognized spike protein RBD segment C, which contains continuous epitope a.a. 549-568.

Clone S1D7 and S3D8 failed to recognize all RBD segments, suggesting that they require an intact RBD tertiary structure to bind spike protein. Error bars indicate standard deviation (n=3).

D. 6×His-tagged RBD segments were secreted in the culture medium of 293T cells and detected by western blotting.

A. R52h is applicable for WB. Lysates of 293T cells expressing artificial spikes carrying T4 foldon or wild-type spike glycoproteins were separated by SDS-PAGE, followed by WB using human/mouse chimeric antibody R52h which was secreted by 293T cells.

B. S1D7h and S3D8h are applicable for IF. Spike glycoprotein expressed in HeLa cells was stained with human/mouse chimeric antibody S1D7h or S3D8h which 

Sequencing of antibody variable regions was carried out as described previously (56) . Total RNA was extracted from the hybridoma cells using ISOGEN 

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Passive immunotherapy of viral infections: 'super-antibodies' enter the fray

Antibodies to SARS-CoV-2 and their potential for therapeutic passive immunization

History of passive antibody administration for prevention and treatment of infectious diseases

Palivizumab, a humanized respiratory syncytial virus monoclonal antibody, reduces hospitalization from respiratory syncytial virus infection in hieh-risk infants. the IMpact-RSV study group

Broadly neutralizing anti-HIV-1 monoclonal antibodies in the clinic

Rapid generation of a human (2020) COVID-19, an emerging coronavirus infection: advances and prospects in designing and developing vaccines, immunotherapeutics, and therapeutics

Could Intravenous Immunoglobulin Collected from Recovered Coronavirus Patients Protect against COVID-19 and Strengthen the Immune System of New Patients?

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

Detection of SARS-CoV-2-Specific Humoral and Cellular Immunity in COVID-19 Convalescent Individuals

A SARS-CoV-2 Infection Model in Mice Demonstrates Protection by Neutralizing Antibodies

Potently neutralizing and protective human antibodies against SARS-CoV-2

The safety and side effects of monoclonal antibodies

The race is on for antibodies that stop the new coronavirus

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

Cryo-EM structure of the SARS coronavirus spike glycoprotein in complex with its host cell receptor ACE2

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

Engineering trimeric fibrous proteins based on bacteriophage T4 adhesins

The intracellular sites of early replication and budding of SARS-coronavirus

Envelope glycoprotein interactions in coronavirus assembly

Enhanced isolation of SARS-CoV-2 by TMPRSS2-expressing cells

Cytoplasmic tail of coronavirus spike protein has intracellular targeting signals

Monoclonal antibody successes in the clinic

Projecting the transmission dynamics of SARS-CoV-2 through the postpandemic period

The race for coronavirus vaccines: a graphical guide

The Current and Future State of Vaccines, Antivirals and Gene Therapies Against Emerging Coronaviruses

The pathogenicity of SARS-CoV-2 in hACE2 transgenic mice

Pathogenesis of SARS-CoV-2 in Transgenic Mice Expressing Human Angiotensin-Converting Enzyme 2

A mouse-adapted model of SARS-CoV-2 to test COVID-19 countermeasures

SARS-CoV-2 infection of human ACE2-transgenic mice causes severe lung inflammation and impaired function

Mouse model of SARS-CoV-2 reveals inflammatory role of type I interferon signaling

We thank A. Ishida, M. Kobayashi, and Y. Kurihara for technical assistance and