key: cord-332134-88wfcc3y
authors: Li, Tingting; Cai, Hongmin; Yao, Hebang; Zhou, Bingjie; Zhang, Ning; Gong, Yuhuan; Zhao, Yapei; Shen, Quan; Qin, Wenming; Hutter, Cedric A.J.; Lai, Yanling; Kuo, Shu-Ming; Bao, Juan; Lan, Jiaming; Seeger, Markus A.; Wong, Gary; Bi, Yuhai; Lavillette, Dimitri; Li, Dianfan
title: A potent synthetic nanobody targets RBD and protects mice from SARS-CoV-2 infection
date: 2020-09-24
journal: bioRxiv
DOI: 10.1101/2020.06.09.143438
sha: 
doc_id: 332134
cord_uid: 88wfcc3y

SARS-CoV-2, the causative agent of COVID-191, recognizes host cells by attaching its receptor-binding domain (RBD) to the host receptor ACE22–7. Neutralizing antibodies that block RBD-ACE2 interaction have been a major focus for therapeutic development8–18. Llama-derived single-domain antibodies (nanobodies, ∼15 kDa) offer advantages including ease of production and possibility for direct delivery to the lungs by nebulization19, which are attractive features for bio-drugs against the global respiratory disease. Here, we generated 99 synthetic nanobodies (sybodies) by in vitro selection using three libraries. The best sybody, MR3 bound to RBD with high affinity (KD = 1.0 nM) and showed high neutralization activity against SARS-CoV-2 pseudoviruses (IC50 = 0.40 μg mL−1). Structural, biochemical, and biological characterization of sybodies suggest a common neutralizing mechanism, in which the RBD-ACE2 interaction is competitively inhibited by sybodies. Various forms of sybodies with improved potency were generated by structure-based design, biparatopic construction, and divalent engineering. Among these, a divalent MR3 conjugated with the albumin-binding domain for prolonged half-life displayed highest potency (IC50 = 12 ng mL−1) and protected mice from live SARS-CoV-2 challenge. Our results pave the way to the development of therapeutic nanobodies against COVID-19 and present a strategy for rapid responses for future outbreaks.

strategy for rapid responses for future outbreaks. the binding through hydrophobic interactions and H-bonding that involves both side 142 chains and main chains (Fig. 1D ). In addition, Tyr37, a framework residue, also 143 participated binding by forming an H-bond with the RBD Gly447 backbone. 144 145 MR17 also binds to the RBD at the 'seat' and 'backrest' regions but approaches 146 the RBD at an almost perfect opposite direction of SR4 (Fig. 1C, 1E) , indicating 147 divergent binding mode for these sybodies. The binding of MR17 to the RBD occurred 148 on an 853.94 Å 2 surface area with noticeable electrostatic complementarity (Extended 149 Data Fig. 5B) . Interestingly, this surface was largely shared with the SR4 binding surface 150 (Fig. 1F) . The interactions between MR17 and the RBD were mainly mediated by H- 151 bonding. Apart from the three CDRs, two framework residues, Lys65 and Tyr60, 152 interacted with the same RBD residue Glu484, via a salt bridge with its side chain, and 153 an H-bond with its main chain (Fig. 1G) . Molecular mechanism for neutralization 157 Structure alignment of SR4-, MR17-and ACE2-RBD 4 showed that both sybodies 158 engage with RBD at the receptor-binding motif (RBM) ( Fig. 2A, 2B) . Superposing SR4 159 and MR17 to the S trimer showed both sybodies could bind to the 'up' conformation 2 160 of RBD with no steric clashes (Fig. 2C, 2D) , and to the 'down' conformation with only 161 minor clashes (Extended Data Fig. 6 ) owing to their minute sizes. Consistent with the 162 structure observation, both SR4 and MR17 inhibited the binding of ACE2 to RBD, as 163 revealed by bio-layer interferometry (BLI) assays (Fig. 2E, 2F) . 164 165 To probe the epitope for MR3 without a structure, competitive BLI binding assays 166 were carried out. The results showed that MR3 could block ACE2 (Fig. 2G) , and SR4 167 and MR17 (Fig. 2H, 2I) , suggesting it also binds to at least part of the RBM, although 168 the possibility of allosteric inhibition remains to be investigated. Taken together, SR4 169 and MR17, and probably MR3, neutralize SARS-CoV-2 by competitively blocking the For biparatopic fusion, we first identified two sybodies, namely LR1 and LR5 (Fig. 208 3A, 3B), that could bind RBD in addition to MR3 using the BLI assay. As LR5 showed 209 higher affinity and neutralization activity than LR1 (Fig. 1A) , we fused this non-210 competing sybody to the N-terminal of MR3 with various length of GS linkers ranging 211 from 13 to 34 amino acids (Extended Data Table S1 ). Interestingly, the linker length 212 had little effect on neutralization activity and these biparatopic LR5-MR3 sybodies 213 were more potent than either sybodies alone ( Fig. 1A) with an IC50 of 0.11 g mL -1 (Fig. 214 3C). LR5-MR3 may be more tolerant to escape mutants 34-37 owing to its ability to 215 recognize two distinct epitopes. This decreased IC50 by 10 folds for Fc-MR3 (39 ng mL -1 ) and 25 folds Fc-MR17 (0.48 g 219 mL -1 ), respectively (Fig. 3D, 3E) . Consistently, the Fc fusion increased the apparent 220 binding affinity for both sybodies, with a KD of 0.22 nM for Fc-MR3 and less than 1 pM 221 for Fc-MR17 (Extended Data Fig. 4H, 4I) . Note, however, Fc-MR17 did not gain as much 222 neutralization potency as for the apparent binding affinity. Table   228 1). The optimal construct for MR17m-MR17m had the shortest linker (13-GS) (Fig. 3D , 229 3E). By contrast, optimal neutralization activity was observed with the longest linker 230 (34-GS) for MR3-MR3 (Fig. 3D, 3E) . Again, MR3-MR3 was superior compared to 231 MR17m-MR17m, showing a 2-fold higher neutralization activity with an IC50 of 12 ng 232 mL -1 (Fig. 3E) . Compared to the monovalent MR3 (IC50 of 0.40 g mL -1 ), the divalent The most potent divalent sybody (MR3-MR3) was chosen to investigate the 240 potential of nanobodies to protect mice from SARS-CoV-2 infection. Nanobodies have 241 very short serum half-lives of several minutes due to their minute size 38 . To circumvent 242 this, we fused MR3-MR3 to the N-terminus of an albumin-binding domain (ABD) 39 243 which has been known to extend the circulating half-life of its fusion partners by 244 increase in size and preventing intracellular degradation 31 . Conveniently, we expressed 245 MR3-MR3-ABD in Pichia pastoris, which is the preferred host to express nanobody 246 therapeutics owing to its robustness and its endotoxin-free production. Small-scale 247 expression of MR3-MR3-ABD showed a secretion level of ~250 mg L -1 with an apparent 248 purity of >80% without purification (Fig. 4A) . Note, this experiment was carried out 249 using a shaker which gave cell density of OD600 of 16. Given its ability to grow to OD600 250 of 500 without compromising yield, the expression level of MR3-MR3-ABD may reach 251 7.5 g L -1 in fermenters. The potential for simple and high-yield production is especially 252 attractive for the pandemic at a global scale. 353 The construct for the RBD with an Avi-tag for biotinylation was made by fusing 354 DNA, from 5'-to 3'-end, of the encoding sequence for the honey bee melittin signal 

Neutralization assay 718 results for SARS-CoV-2 pseudovirus. (B) Neutralization assay results for SARS-CoV 719 pseudovirus. VeroE6-hACE2 cells were infected with a premix of pseudotypes and 720 sybodies at two concentrations (1 M and 100 nM)

A-I) Biotinylated RBD 724 immobilized on a streptavidin-coated sensor was titrated with various concentrations 725 (nM) of sybodies as indicated

Open-book' view of molecular electrical potential surfaces of the 730 interface between the RBD and SR4 (A) and between the RBD and MR17 (B). The 731 electrical potential maps were calculated by Adaptive Poisson-Boltzmann Solver (APBS) 732 52 built-in in PyMol

Structure-based design of a MR17 mutant (MR17m) with 743 improved affinity and potency. (A,B) Neutralization assay for SARS-CoV-2 (A) or

Sybody concentrations were used at 1 M (green) and 746 100 nM (magenta) concentrations. Data are from three independent experiments

Electrostatic repel and hydrophobic mismatch would make Lys99 750 unfavorable at this position. According to the original library design, Lys99 was 751 unvaried 26 , meaning that Lys99 was not selected and hence opportunities for 752 optimization. (E) The K99Y mutation fits the hydrophobic microenvironment well, as 753 revealed by the crystal structure of MR17m (Extended Data Table 2). (F) Binding 754 kinetics of MR17m binding to RBD. BLI signals were recorded under

IC50 values (g mL -1 ) for SARS-Cov-2 are indicated in brackets. Data for MR17 are from 757

Data are from three independent experiments

Extended Data Fig. 8. Evaluation of in vivo stability and toxicity of nanobodies. (A)

For neutralization assay, sera were preincubated with SARS-CoV-763 2 pseudovirus for 1 h before infection at 1/200 dilution. The infection rates on VeroE6-764 hACE2 were measure by FACS 3 days post infection. (B) Body weight changes. The body 765 weight data are presented as means  the SD of mice in each group (n= 4). No 766 significant differences are observed. (C) Representative histopathology of the lungs, 767 heart, liver, spleen, lungs, kidney, and thymus for the different sybodies injected

The images and areas of interest are magnified 100 ×. Bars indicate 100 m

A novel coronavirus outbreak of global health 773 concern

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

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

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 779 receptor

Structural basis of receptor recognition by SARS-CoV-2

Structural and Functional Basis of SARS-CoV-2 Entry by Using Human ACE2

Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2

A noncompeting pair of human neutralizing antibodies block COVID-19 virus 787 binding to its receptor ACE2

Potent neutralizing antibodies against SARS-CoV-2 identified by high-throughput 789 single-cell sequencing of convalescent patients' B cells

Human neutralizing antibodies elicited by SARS-CoV-2 infection

General strategy to humanize a camelid single-domain antibody and 794 identification of a universal humanized nanobody scaffold

Isolation of potent SARS-CoV-2 neutralizing antibodies and protection from 797 disease in a small animal model

Studies in humanized mice and convalescent humans yield a SARS-CoV-2 799 antibody cocktail

An alpaca nanobody neutralizes SARS-CoV-2 by blocking receptor interaction. 801 bioRxiv

Neutralizing nanobodies bind SARS-CoV-2 spike RBD and block interaction with 803

Selection, biophysical and structural analysis of synthetic nanobodies that 805 effectively neutralize SARS-CoV-2

Synthetic nanobodies targeting the SARS-CoV-2 receptor-binding domain

An ultra-high affinity synthetic nanobody blocks SARS-CoV-2 infection by 809 locking Spike into an inactive conformation. bioRxiv

Nanobodies® as inhaled biotherapeutics for lung diseases

The socio-economic implications of the coronavirus pandemic (COVID-19): A 814 review

Cell entry mechanisms of SARS-CoV-2

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a 818 Clinically Proven Protease Inhibitor

Convergent antibody responses to SARS-CoV-2 in convalescent individuals

Natural Single-Domain Antibodies

Yeast surface display platform for rapid discovery of conformationally 825 selective nanobodies

Synthetic single domain antibodies for the conformational trapping of 827 membrane proteins

Generation of synthetic nanobodies against delicate proteins

NaLi-H1: A universal synthetic library of humanized nanobodies providing 831 highly functional antibodies and intrabodies

An improved yeast surface display platform for the screening of nanobody 833 immune libraries

Selection, biophysical and structural analysis of synthetic nanobodies that 835 effectively neutralize SARS-CoV-2. bioRxiv

The Therapeutic Potential of Nanobodies

Inference of Macromolecular Assemblies from Crystalline State

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Molecular cancer therapeutics

Fusion to a highly stable consensus albumin binding domain allows for tunable 860 pharmacokinetics. Protein engineering, design & selection : PEDS 28

Generation of a Broadly Useful Model for COVID-19 Pathogenesis, Vaccination, 863 and Treatment

A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2

A SARS-CoV-2 Infection Model in Mice Demonstrates Protection by 867 Neutralizing Antibodies

Enzymatic assembly of DNA molecules up to several hundred kilobases

A fluorescence-detection size-exclusion chromatography-872 based thermostability assay for membrane protein precrystallization screening

The protein complex crystallography beamline (BL19U1) at the Shanghai 875

Synchrotron Radiation Facility

How good are my data and what is the resolution?

Phaser crystallographic software

Features and development of Coot

PHENIX: a comprehensive Python-based system for macromolecular 885 structure solution

The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC

Electrostatics of nanosystems: 889 Application to microtubules and the ribosome