Cross-Neutralization of a SARS-CoV-2 Antibody to a Functionally Conserved Site Is Mediated by Avidity

1The ongoing coronavirus infectious disease 2019 (COVID-19) pandemic of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is unlikely to end anytime soon. Given the current lack of protective vaccines and antivirals, virus clearance and recovery of SARS-CoV-2 patients have to rely mainly on the generation of a neutralizing antibody response. To date, most neutralizing antibodies from convalescent patients target the receptor-binding domain (RBD) on the trimeric spike (S) glycoprotein, whose natural function is to mediate viral entry by first attaching to the human receptor angiotensin-converting enzyme 2 (ACE2) and then fusing its viral membrane with the host cell. SARS-CoV-2 is phylogenetically closely related to SARS-CoV, which caused the 2002–2003 human epidemic. However, SARS-CoV-2 and SARS-CoV only share 73% amino-acid sequence identity in their RBD, compared to 90% in their S2 fusion domain. Nevertheless, a highly conserved epitope on the SARS-CoV-2 RBD was previously identified from studies of a SARS-CoV neutralizing antibody CR3022, which was originally isolated almost 15 years ago. Many human monoclonal antibodies have now been shown to target the SARS-CoV-2 S protein, but cross-neutralizing antibodies are relatively uncommon in COVID-19 patients. To date, the only structurally characterized cross-neutralizing human antibodies are S309 and ADI-56046 from SARS-CoV survivors, as well as EY6A from a COVID-19 patient. Such structural and molecular characterization of cross-neutralizing antibodies is extremely valuable for therapeutic and vaccine design to confer broader protection against human SARS-like viruses that include the extensive reservoir of zoonotic coronaviruses in bats, camels, pangolins etc.

2Antibody COVA1–16 was recently isolated from a convalescent COVID-19 patient and can cross-neutralize both SARS-CoV-2 (IC50 0.13 μg/mL) and SARS-CoV (IC50 2.5 μg/mL) pseudovirus. The heavy and light chains of COVA1–16 are encoded by IGHV1–46, IGHD3–22, IGHJ1, and by IGKV1–33, IGKJ4, with a relatively long complementarity determining region (CDR) H3 of 20 amino acids (Figure S1). IGHV of COVA1–16 is only 1% somatically mutated at the nucleotide sequence level (one amino-acid change) from the germline gene, whereas its IGKV is 1.4% somatically mutated (three amino-acid changes). Here we determined the crystal structure of COVA1–16 in complex with SARS-CoV-2 RBD at 2.89 Å resolution to identify its binding site (epitope) and mechanism of cross-neutralization (Figure 1A, Table S1). The epitope of COVA1–16 overlaps extensively with that of CR3022, but also extends towards the periphery of the ACE2 binding site (Figure 1B). Seventeen out of 25 residues in the COVA1–16 epitope overlap with the CR3022 binding site (17 of 28 residues) (Figure 1C). Consistent with structural identification of its epitope, COVA1–16 can compete with CR3022 for RBD binding (Figure S2). COVA1–16 appears to have some resemblance to SARS-CoV cross-neutralizing antibody ADI-56046, whose epitope appears to span both the CR3022 epitope and ACE2-binding site, as indicated by negative-stain electron microscopy (nsEM). Interestingly, COVA1–16 also competes with ACE2 for RBD binding (Figure S2), although its epitope does not overlap the ACE2 binding site (Figure 1B). Therefore, COVA1–16 inhibits ACE2 binding due to steric hindrance with its light chain rather than by direct interaction with the receptor binding site (Figure 1D).

3The RBD can adopt up and down conformations on the S trimer. While the ACE2 receptor only binds the RBD in the up conformation, previously characterized cross-neutralizing antibodies S309 from a convalescent SARS-CoV patient and COVA2–15 from a SARS-CoV-2 patient, can bind the RBD in both up and down conformations. However, unlike S309, the COVA1–16 epitope is completely buried when the RBD is in the down conformation (Figure 2A), akin to the CR3022 epitope. Even in the up conformation of the RBD on an unliganded SARS-CoV-2 S trimer, the epitope of COVA1–16 would not be fully exposed (Figure 2A). We thus performed nsEM analysis of COVA1–16 in complex with the SARS-CoV-2 S trimer (Figure 2B). Three-dimensional (3D) reconstructions revealed that COVA1–16 can bind to a range of RBD orientations on the S protein when in the up position, indicating its rotational flexibility (Figure 2C). COVA1–16 can bind the S trimer either from the top (i.e. perpendicular to the trimer apex, Figure 2C, yellow, blue and pink) or from the side (i.e. more tilted, Figure 2C, brown). Model fitting of the COVA1–19/RBD crystal structure into the nsEM map indicates that the RBD on the S trimer is more open around the apex when COVA1–16 binds compared to unliganded trimers (Figure S3A–B). Bivalent binding of the COVA1–16 IgG between adjacent S trimers also appears to be plausible (Figure S3C). A recent cryo-electron tomography (cryo-ET) analysis demonstrated that the average distance between prefusion S on the viral surface is around 150 Å, which is comparable to the distance spanned between the tip of the two Fabs on an IgG (typically around 100 Å to 150 Å, although longer distances have been observed). Indeed, COVA1–16 IgG binds much more tightly than Fab to SARS-CoV-2 RBD, with dissociation constants (KD) of 0.2 nM and 46 nM, respectively (Figure S4A), reflecting bivalent binding in the assay format. Similarly, COVA1–16 IgG binds more strongly than Fab to SARS-CoV RBD (KD of 125 nM vs 405 nM) (Figure S4B). Moreover, the apparent affinity of COVA1–16 IgG decreased to approximately the Fab value when the amount of SARS-CoV-2 RBD loaded on the biosensor was decreased, substantiating the notion that COVA1–16 can bind bivalently in this assay (Figure S4C).

4Bivalent IgG binding is also important for the neutralization activity of COVA1–16 (Figure 2D–E). COVA1–16 IgG neutralizes SARS-CoV-2 pseudovirus with a half maximal inhibitory concentration (IC50) of 0.02 μg/mL, which is similar to that previously measured for SARS-CoV-2 pseudovirus (IC50 of 0.13 μg/mL). In contrast, COVA1–16 Fab does not neutralize SARS-CoV-2 pseudovirus even up to 13 μg/mL. A similar effect is also observed for SARS-CoV pseudovirus, which is neutralized by COVA1–16 IgG at an IC50 of 29 μg/mL, but not by COVA1–16 Fab even up to 67 μg/mL (Figure 2E). Of note, COVA1–16 is less potent against authentic SARS-CoV-2 (IC50 = 0.75 μg/mL). Whether such a difference is due to variation in S protein density on the viral surface versus pseudovirus or to other factors deserves future investigation. It will also be informative to compare the number, density and conformational states of the S proteins on SARS-CoV-2 and SARS-CoV virions. Overall, our findings support the importance of bivalent binding for SARS-CoV-2 neutralizing antibodies, and especially for cross-neutralization of SARS-CoV. Such a contribution of bivalent IgG (avidity) to SARS-CoV-2 neutralization has also been suggested in a recent study that compared binding of polyclonal IgGs and Fabs. Furthermore, a single-domain camelid antibody VHH-72 dramatically improved its neutralization activity to SARS-CoV-2 when expressed as a bivalent Fc-fusion. These observations are similar to some influenza broadly neutralizing antibodies to the hemagglutinin (HA) receptor binding site, where bivalent binding can increase avidity and neutralization breadth.

5Next we examined the molecular details of the interactions between COVA1–16 and SARS-CoV-2. COVA1–16 binding to the RBD is dominated by the heavy chain, which accounts for 82% of its total buried surface area (BSA, 694 Å2 out of a total of 844 Å2). Most of the interactions are mediated by CDR H3 (Figure 3A), which contributes 70% (594 Å2) of the total BSA. CDR H3 forms a beta-hairpin with a type I beta-turn at its tip and is largely encoded by IGHD3–22 (from VH N98 to VH Y100f, Figure S1C and Figure 3B). The beta-hairpin conformation is stabilized by four main chain-main chain hydrogen-bonds (H-bonds) and a side chain-side chain H-bond between VH N98 and VH Y100f at either end of the IGHD3–22-encoded region (Figure 3B). Four H-bonds between the tip of CDR H3 and the RBD are formed from two main chain-main chain interactions with RBD C379, and two with VH R100b (Table S2). The positively charged guanidinium of VH R100b also interacts with the negative dipole at the C-terminus of a short α-helix in the RBD (residues Y365 to Y369). Interestingly, VH R100b is a somatically mutated residue (codon = AGG in the IGHD3–22-encoded region, where the germline residue is a Ser (codon = AGT, Figure S1C). The short Ser side chain would likely not contact the RBD nor provide electrostatic complementarity. Interestingly, a somatic revertant VH R100bS actually improved binding affinity of COVA1–16 to the RBD, mostly due to an increased on-rate (Figure S5). Nevertheless, COVA1–16 has a much slower off-rate than its VH R100bS mutant, which may have led to its selection. The CDR H3 tip also interacts with the RBD through hydrophobic interactions between VH Y99 and the aliphatic portion of RBD K378, as well as a π−π interaction between VH Y100 and the RBD V382-S383 peptide backbone (Figure 3B). CDR H3 forms an additional four H-bonds with the RBD, involving the side chains of VH R97 and Q101 (Figure 3B). We further determined the unliganded structure of COVA1–16 Fab to 2.53 Å resolution and found that the CDR H3 distal region was not resolved due to lack of electron density indicating its inherent flexibility (Figure S6). CDR H1 and CDR L2 of COVA1–16 also interact with the RBD, but much less so compared to CDR H3. The VH T28 main chain and VH Y32 side chain in CDR H1 H-bond with D427 (Figure 3C, Table S2), whereas VL N53 in CDR L2 H-bonds with RBD R408 (Figure 3D, Table S2).

6CDR H3-dominant antibodies have been seen in the human immune response to other viral pathogens. Striking examples are antibodies PG9 and PG16, whose CDR H3s interact extensively along their length with the apex of the HIV-1 Envelope protein. Another example is C05, which is essentially a single loop binder that inserts its very long CDR H3 (24 residues) into the RBD of influenza HA, thereby providing a template for design of a high-avidity protein inhibitor of influenza virus, where the H3 loop was fused to a scaffold protein. The long CDR H3 of COVA1–16 may similarly facilitate therapeutic designs that could also include peptide-based antivirals, as exemplified by a potent cyclic peptide fusion inhibitor of influenza HA.

7Compared to the ACE2-binding site, the COVA1–16 epitope is much more highly conserved among SARS-CoV-2, SARS-CoV, and other SARS-related coronaviruses (SARSr-CoV) (Figure 4A–D, Figure S7 and Figure S8). To investigate possible structural and functional reasons for this sequence conservation, we analyzed the epitope location in the context of the SARS-CoV-2 trimeric S protein with all RBDs in the “down” conformation (Figure 4E and Figure S7). The COVA1–16 epitope is completely buried at the center of the trimer in the interface between the S1 and S2 domains and is largely hydrophilic (Figure S9). The polar side chains of K378, Q414, R408, and D427, which are involved in binding to COVA1–16, are all very close to the interface with adjacent protomers in the S trimer. Interestingly, the R408 side chain, which is positioned by Q414 via a H-bond, points towards a region in the adjacent protomer 2 with a positive electrostatic potential. Similarly, D427 is juxtaposed to a region in protomer 2 with a negative electrostatic potential. These repulsive charges would help favor the metastability required for transient opening and closing of the RBD in “up” and “down” conformations prior to ACE2 receptor binding. In contrast, the K378 side chain points towards a region in protomer 3 with negative electrostatic potential, thus favoring the “down” RBD conformation. Furthermore, in the “down” conformation, part of the COVA1–16 epitope interacts with the long helices formed from the heptad repeat motifs of S2 fusion domain (Figure 4E–F). Notably, S383 and T385 in the COVA1–16 epitope make three H-bonds with the tops of the helices and their connecting regions (Figure 4F). This mixture of attractive and repulsive forces would seem to be important for control of the dynamics of the RBD and, hence, for the biological function of the metastable pre-fusion S protein in receptor binding and fusion. The complementarity of fit of the epitope interface with the other RBDs and the S2 domain in the S trimer further explains the epitope conservation (Figure S10). Therefore, the high sequence conservation of the COVA1–16 epitope appears related to the functional requirement for this component of the RBD surface to be deeply buried within the S trimer in the “down” conformation.

8From the SARS-CoV-2 RBD/antibody complex structures to date, a significant portion of the RBD surface can be targeted by antibodies (Figure 5). One surface not yet observed to be targeted is partially covered by N-glycans at residues N165 on the N-terminal domain (NTD) and N343 on the RBD, which may hinder B cell receptor access and create a “silent face” (Figure S11), although the N343 glycan is incorporated in the S309 epitope. While antibodies that target the ACE2-binding site, such as BD23, CB6, B38, P2B-2F6, CC12.1, and CC12.3, do not show cross-neutralization activity to SARS-CoV, the conserved epitopes further from the ACE2-binding site seem to be more able to support cross-neutralization. It is also interesting that these so far rare cross-neutralizing antibodies, including COVA1–16, often seem to bind to epitopes that are not readily accessible in the pre-fusion native structure. This finding is similar to a recent discovery in influenza virus, where a class of cross-protective antibodies target a conserved epitope in the trimeric interface of the HA. Due to the inaccessibility of the COVA1–16 epitope on the S protein, it is possible that an RBD-based rather than S-based immunogen can elicit larger numbers of COVA1–16-like antibodies. Cross-neutralizing antibodies have also provided important insights into therapeutic and vaccine design, as for influenza virus and HIV. As SARS-CoV-2 continues to circulate in the human population and other zoonotic coronaviruses constitute future pandemic threats, it is certainly worth considering the development of more universal coronavirus vaccines and therapeutics that can cross-neutralize antigenically drifted SARS-CoV-2 viruses, as well as zoonotic SARS-like coronaviruses.

9The receptor-binding domain (RBD) (residues 319–541) of the SARS-CoV-2 spike (S) protein (GenBank: QHD43416.1), and the RBD (residues 306–527) of the SARS-CoV S protein (GenBank: ABF65836.1), were cloned into a customized pFastBac vector, and fused with an N-terminal gp67 signal peptide and C-terminal His6 tag. For each RBD, we further cloned a construct with an AviTag inserted in front of the His6 tag. To express the RBD, a recombinant bacmid DNA was generated using the Bac-to-Bac system (Life Technologies). Baculovirus was generated by transfecting purified bacmid DNA into Sf9 cells using FuGENE HD (Promega), and subsequently used to infect suspension cultures of High Five cells (Life Technologies) at an MOI of 5 to 10. Infected High Five cells were incubated at 28 °C with shaking at 110 r.p.m. for 72 h for protein expression. The supernatant was then concentrated using a 10 kDa MW cutoff Centramate cassette (Pall Corporation). The RBD protein was purified by Ni-NTA, followed by size exclusion chromatography, and buffer exchanged into 20 mM Tris-HCl pH 7.4 and 150 mM NaCl. For binding experiments, RBD with AviTag was biotinylated as described previously and purified by size exclusion chromatography on a Hiload 16/90 Superdex 200 column (GE Healthcare) in 20 mM Tris-HCl pH 7.4 and 150 mM NaCl.

10Expression plasmids encoding the heavy and light chains of the COVA1–16 Fab were transiently co-transfected into ExpiCHO cells at a ratio of 2:1 (HC:LC) using ExpiFectamine™ CHO Reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. The supernatant was collected at 10 days post-transfection. The Fabs were purified with a CaptureSelect™ CH1-XL Affinity Matrix (Thermo Fisher Scientific) followed by size exclusion chromatography.

11The N-terminal peptidase domain of human ACE2 (residues 19 to 615, GenBank: BAB40370.1) was cloned into phCMV3 vector and fused with a C-terminal Fc tag. The plasmids were transiently transfected into Expi293F cells using ExpiFectamine™ 293 Reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. The supernatant was collected at 7 days post-transfection. Fc-tagged ACE2 protein was then purified with a Protein A column (GE Healthcare) followed by size exclusion chromatography.

12The COVA1–16 Fab complex with RBD was formed by mixing each of the protein components in an equimolar ratio and incubating overnight at 4°C. The COVA1–16 Fab/RBD complex and COVA1–16 Fab apo (unliganded) protein were adjusted to around 11 mg/mL and screened for crystallization using the 384 conditions of the JCSG Core Suite (Qiagen) on our custom-designed robotic CrystalMation system (Rigaku) at Scripps Research. Crystallization trials were set-up by the vapor diffusion method in sitting drops containing 0.1 μl of protein and 0.1 μl of reservoir solution. Crystals used for x-ray data collection were harvested from drops containing 0.2 M sodium iodide and 20% (w/v) polyethylene glycol 3350 for the COVA1–16 Fab/RBD complex and from drops containing 0.08 M acetate pH 4.6, 20% (w/v) polyethylene glycol 4000, 0.16 M ammonium sulfate and 20% (v/v) glycerol for the COVA1–16 Fab. Crystals appeared on day 3, were harvested on day 7, pre-equilibrated in cryoprotectant containing 20% glycerol, and then flash cooled and stored in liquid nitrogen until data collection. Diffraction data were collected at cryogenic temperature (100 K) at Stanford Synchrotron Radiation Lightsource (SSRL) on the new Scripps/Stanford beamline 12–1 with a beam wavelength of 0.97946 Å, and processed with HKL2000. Structures were solved by molecular replacement using PHASER. The models for molecular replacement of RBD and COVA1–16 were from PDB 6XC4, 4IMK and 2Q20. Iterative model building and refinement were carried out in COOT and PHENIX, respectively. Epitope and paratope residues, as well as their interactions, were identified by accessing PISA at the European Bioinformatics Institute (http://www.ebi.ac.uk/pdbe/prot_int/pistart.html).

13The SARS-CoV-2 S construct used for negative stain EM contains the mammalian-codon-optimized gene encoding residues 1–1208 of the S protein (GenBank: QHD43416.1), followed by a C-terminal T4 fibritin trimerization domain, an HRV3C cleavage site, 8x-His tag and a Twin-strep tags subcloned into the eukaryotic-expression vector pcDNA3.4. Three amino-acid mutations were introduced into the S1/S2 cleavage site (RRAR to GSAS) to prevent cleavage and two stabilizing proline mutations (K986P and V987P) to the HR1 domain. For additional S stabilization, residues T883 and V705 were mutated to cysteines to introduce a disulphide bond. The S plasmid was transfected into 293F cells and supernatant was harvested at 6 days post transfection. S protein was purified by running the supernatant through a streptactin column and then by size exclusion chromatography using a Superose 6 increase 10/300 column (GE Healthcare Biosciences). Protein fractions corresponding to the trimeric S protein were collected and concentrated.

14SARS-COV-2 S protein was complexed with 3x molar excess of Fab at 30 minutes prior to direct deposition onto carbon-coated 400-mesh copper grids. The grids were stained with 2 % (w/v) uranyl-formate for 90 seconds immediately following sample application. Grids were either imaged at 200 KeV or at 120 KeV on a Tecnai T12 Spirit using a 4kx4k Eagle CCD. Micrographs were collected using Leginon and the images were transferred to Appion for processing. Particle stacks were generated in Appion with particles picked using a difference-of-Gaussians picker (DoG-picker). Particle stacks were then transferred to Relion for 2D classification followed by 3D classification to sort well-behaved classes. Selected 3D classes were auto-refined on Relion and used to make figures with UCSF Chimera.

15All constructs were expressed transiently in HEK293F (Invitrogen, cat no. R79009) cells maintained in Freestyle medium (Life Technologies). For soluble RBD proteins, cells were transfected at a density of 0.8–1.2 million cells/mL by addition of a mix of PEImax (1 μg/μL) with expression plasmids (312.5 μg/L) in a 3:1 ratio in OptiMEM. Supernatants of the soluble RBD proteins were harvested six days post transfection, centrifuged for 30 min at 4000 rpm and filtered using 0.22 μm Steritop filters (Merck Millipore). Constructs with a His6-tag were purified by affinity purification using Ni-NTA agarose beads. Protein eluates were concentrated, and buffer exchanged to PBS using Vivaspin filters with a 10 kDa molecular weight cutoff (GE Healthcare). Protein concentrations were determined by Nanodrop using the proteins peptidic molecular weight and extinction coefficient as determined by the online ExPASy software (ProtParam). For the COVA1–16 IgG1 antibody, suspension HEK293F cells (Invitrogen, cat no. R79007) were cultured in FreeStyle medium (Gibco) and co-transfected with the two IgG plasmids expressing the corresponding HC and LC in a 1:1 ratio at a density of 0.8–1.2 million cells/mL in a 1:3 ratio with 1 mg/L PEImax (Polysciences). The recombinant IgG antibodies were isolated from the cell supernatant after five days as described previously (20, 48). In short, the cell suspension was centrifuged 25 min at 4000 rpm, and the supernatant was filtered using 0.22 μm pore size SteriTop filters (Millipore). The filtered supernatant was run over a 10 mL protein A/G column (Pierce) followed by two column volumes of PBS wash. The antibodies were eluted with 0.1 M glycine pH 2.5, into the neutralization buffer of 1 M TRIS pH 8.7 in a 1:9 ratio. The purified antibodies were buffer exchanged to PBS using 100 kDa VivaSpin20 columns (Sartorius). The IgG concentration was determined on the NanoDrop 2000 and the antibodies were stored at 4°C until further analyses.

16To determine the binding affinity of COVA1–16 IgG and His-tagged Fabs, 20 μg/mL of His-tagged SARS-CoV or SARS-CoV-2 RBD protein in running buffer (PBS, 0.02% Tween-20, 0.1% BSA) was loaded on Ni-NTA biosensors (ForteBio) for 300 s. Streptavidin biosensors (ForteBio) were used if the RBD was biotinylated. Next, the biosensors were transferred to running buffer containing IgG or Fab to determine the association rate, after which the sensor was transferred to a well containing running buffer to allow dissociation. As negative control, an anti-HIV-1 His-tagged Fab was tested at the highest concentration used for COVA1–16 Fab (400 nM). After each cycle, the sensors were regenerated by alternating 20 mM glycine in PBS and running buffer three times, followed by reactivation in 20 mM NiCl2 for 120 s. All steps were performed at 1000 rpm shaking speed. KDs were determined using ForteBio Octet CFR software. The avidity effects of IgG were investigated by titrating the SARS-CoV-2 RBD concentration (5, 1, 0.2 and 0.04 μg/mL) followed by loading on Ni-NTA biosensors for 480 s with an additional loading step with His-tagged HIV-1 gp41 for 480 s to minimize background binding of His-tagged Fabs to the biosensor. All other steps were performed as described above.

17For competition assays, COVA1–16 IgG, CR3022 IgG, and human ACE2-Fc were all diluted to 250 nM. Ni-NTA biosensors were used. In brief, the assay has five steps: 1) baseline: 60 s with 1× kinetics buffer; 2) loading: 180 s with 20 μg/mL, His6-tagged SARS-CoV-2 RBD proteins; 3) baseline: 150 s with 1× kinetics buffer; 4) first association: 300 s with CR3022 IgG or human ACE2-Fc; and 5) second association: 300 s with human ACE2-Fc, CR3022 IgG, or COVA1–16 IgG.

18Neutralization assays were performed using SARS-CoV and SARS-CoV-2 S-pseudotyped HIV-1 virus and HEK-293T/ACE2 cells as described previously. In brief, pseudotyped virus was produced by co-transfecting expression plasmids of SARS-CoV S and SARS-CoV-2Δ19 S proteins (GenBank; AAP33697.1 and MT449663.1, respectively) with an HIV backbone expressing NanoLuc luciferase (pHIV-1NL4–3 ΔEnv-NanoLuc) in HEK293T cells (ATCC, CRL-11268). After 3 days, the cell culture supernatants containing SARS-CoV and SARS-CoV-2 S-pseudotyped HIV-1 viruses were stored at −80°C. HEK-293T/ACE2 cells were seeded 10,000 cells/well in a 96-well plate one day prior to the start of the neutralization assay. To determine the neutralizing capacity of COVA1–16 IgG and His6-tagged Fab, 20 or 100 μg/mL COVA1–16 IgG and equal molar of COVA1–16 Fab were serially diluted in 3-fold steps and mixed with SARS-CoV or SARS-CoV-2 pseudotyped virus and incubated for 1 h at 37°C. The pseudotyped virus and COVA1–16 IgG/Fab mix were then added to the HEK-293T/ACE2 cells and incubated at 37°C. After 48 h, cells were washed twice with PBS (Dulbecco’s Phosphate-Buffered Saline, eBiosciences) and lysis buffer was added. Luciferase activity of cell lysate was measured using the Nano-Glo Luciferase Assay System (Promega) and GloMax Discover System. The inhibitory concentration (IC50) was determined as the concentration of IgG or Fab that neutralized 50% of the pseudotyped virus using GraphPad Prism software (version 8.3.0).

19RBD protein sequences from SARS-CoV and SARS-related coronavirus (SARSr-CoV) strains were retrieved from the following accession codes: GenBank ABF65836.1 (SARS-CoV)GenBank ALK02457.1 (Bat SARSr-CoV WIV16)GenBank AGZ48828.1 (Bat SARSr-CoV WIV1)GenBank ACU31032.1 (Bat SARSr-CoV Rs672)GenBank AIA62320.1 (Bat SARSr-CoV GX2013)GenBank AAZ67052.1 (Bat SARSr-CoV Rp3)GenBank AIA62300.1 (Bat SARSr-CoV SX2013)GenBank ABD75323.1 (Bat SARSr-CoV Rf1)GenBank AIA62310.1 (Bat SARSr-CoV HuB2013)GenBank AAY88866.1 (Bat SARSr-CoV HKU3–1)GenBank AID16716.1 (Bat SARSr-CoV Longquan-140)GenBank AVP78031.1 (Bat SARSr-CoV ZC45)GenBank AVP78042.1 (Bat SARSr-CoV ZXC21)GenBank QHR63300.2 (Bat CoV RaTG13)NCBI Reference Sequence YP_003858584.1 (Bat SARSr-CoV BM48-31)GISAID EPI_ISL_410721 (Pangolin BetaCoV Guandong2019)

20GenBank ABF65836.1 (SARS-CoV)

21GenBank ALK02457.1 (Bat SARSr-CoV WIV16)

22GenBank AGZ48828.1 (Bat SARSr-CoV WIV1)

23GenBank ACU31032.1 (Bat SARSr-CoV Rs672)

24GenBank AIA62320.1 (Bat SARSr-CoV GX2013)

25GenBank AAZ67052.1 (Bat SARSr-CoV Rp3)

26GenBank AIA62300.1 (Bat SARSr-CoV SX2013)

27GenBank ABD75323.1 (Bat SARSr-CoV Rf1)

28GenBank AIA62310.1 (Bat SARSr-CoV HuB2013)

29GenBank AAY88866.1 (Bat SARSr-CoV HKU3–1)

30GenBank AID16716.1 (Bat SARSr-CoV Longquan-140)

31GenBank AVP78031.1 (Bat SARSr-CoV ZC45)

32GenBank AVP78042.1 (Bat SARSr-CoV ZXC21)

33GenBank QHR63300.2 (Bat CoV RaTG13)

34NCBI Reference Sequence YP_003858584.1 (Bat SARSr-CoV BM48-31)

35GISAID EPI_ISL_410721 (Pangolin BetaCoV Guandong2019)

36Multiple sequence alignment of the RBD sequences was performed by MUSCLE version 3.8.31. Sequence logos were generated by WebLogo. The conservation score of each RBD residue was calculated and mapped onto the SARS-CoV-2 RBD x-ray structure with ConSurf.