key: cord-0833641-em02pqtp
authors: Shah, Masaud; Ahmad, Bilal; Choi, Sangdun; Goo Woo, Hyun
title: Mutations in the SARS-CoV-2 Spike RBD are responsible for stronger ACE2 binding and poor anti-SARS-CoV mAbs cross-neutralization
date: 2020-11-12
journal: Comput Struct Biotechnol J
DOI: 10.1016/j.csbj.2020.11.002
sha: d2abbc5e65b924fc59bcf6e1918dc1193210397c
doc_id: 833641
cord_uid: em02pqtp

Severe acute respiratory syndrome coronavirus 2(SARS-CoV-2), which causes coronavirus disease (COVID-19), is a novel beta coronavirus. SARS-CoV-2 uses spike glycoprotein to interact with host angiotensin-converting enzyme 2 (ACE2) and ensure cell recognition. High infectivity of SARS-CoV-2 raises questions on spike-ACE2 binding affinity and its neutralization by anti-SARS-CoV monoclonal antibodies (mAbs). Here, we observed Val-to-Lys417 mutation in the receptor-binding domains (RBD) of SARS-CoV-2, which established a Lys-Asp electrostatic interaction enhancing its ACE2-binding. Pro-to-Ala475 substitution and Gly482 insertion in the AGSTPCNGV-loop of RBD possibly hinders neutralization of SARS-CoV-2 by anti-SARS-CoV mAbs. In addition, we identified unique and structurally conserved conformational-epitopes on RBDs, which can be potential therapeutic targets. Collectively, we provide new insights into the mechanisms underlying the high infectivity of SARS-CoV-2 and development of effective neutralizing agents.

Introduction 122 respectively. The temperature and pressure were coupled with V-rescale and Parrinello-Rahman 123 barostat methods (30) , respectively. The long-range electrostatic interactions were computed using 124 the particle mesh Ewald algorithm (31) , while LINCS algorithm was applied to constrain the bond 125 lengths (32) . After temperature and pressure equilibration, MD simulations were carried out for 126 30 ns for each system. A detailed procedure has been described in our previous studies (16, 18) .

127 Binding free energy analysis 128 Since both the cRBD and sRBD bind to the same ACE2 protein, we used Molecular mechanics 129 Poisson−Boltzmann surface area (MM-PBSA) approach (33) to calculate the relative binding free 130 energies of both complexes. In GROMACS, the built-in tool g_mmpbsa and APBSA were called 131 for the MMPBSA calculations. For g_mmpbsa analysis, the dielectric constant of the aqueous 132 solvent was set to 80, and the interior dielectric constant was set to 4; the surface tension constant 133 g was set to 0.022 kJ/mol. As g_mmpbsa tool is compatible with older versions of GROMACS 134 (versions 5 or lower), the "tpr" files created by GROMACS 2019.3 were recreated through 135 GROMACS 5.1 and used for binding energies calculations, as described previously (34) . The 136 relative binding energies of the complexes were approximated according to the following energies 137 terms. 

155 Structural modeling of the SARS-CoV-2 spike and ACE2 interaction 156 A full-length spike protein is composed of S1 and S2 subunits, which further contains sub-157 domains and motifs with distinct functions (Figure 1A) . Based on the hinge-like motion of the 158 RBD of S1 subunit, the trimeric S protein exists a transiently symmetric or asymmetric 159 conformation. Recent studies revealed that the cRBD, like other coronaviruses, exhibit stochastic 160 breathing-like movement, facilitating receptor binding to the exposed RBD and subsequent 161 shedding of the S1 subunit (11, 12) . This hinge-like motion of the RBD is considered as one of the 162 immune evasion strategy of the SARS-CoV-2 where the virus masks its RBD from host antibodies 163 in down conformation and makes it available for the ACE2 in up position (12, 13 Figure 1B) . On the contrary, ACE2 dock well onto the 212 interface of ACE2 with a chimeric RBD through X-ray crystallography (38) . Overall, this chimeric 213 structure shares structural folds with our model and recently reported Cryo-EM structure (37) and 214 validates the cRBD-ACE2 interface residues in our model; nonetheless, the chimeric structure 215 lacks the crucial mutation of Val-to-Lys417. This is because the core (scaffold) RBD in this 216 structure was taken from SARS-CoV, which is considerably conserved between the two viruses 217 with few mutations including Val-to-Lys417 (Supplementary Figure 1C) . As the high resolution 218 crystal structure of the SARS-CoV-2 RBD with ACE2 has been reported during the course of this 219 study, we validated our ACE2-RBD model by comparing the structure and interface with recently 220 reported structure (6LZG) resolved through X-ray diffraction (39) . As expected, there was 221 negligible differences in the overall RBDs (backbone atoms RMSD = 1.26 Å) structures and the 222 ACE2-RBD interface residues were same in both complexes, including Lys417 ( Figure 2D ).

223 Therefore, we suggest that our computational model provides detailed understanding about the 224 structural variation in cRBD and its interface with ACE2.

227 Differing from SARS-CoV and SARS-related CoVs, the S protein of SARS-CoV-2 has furin 228 cleavage site at the S1/S2 boundary as well as exhibits similar or more binding affinity towards 229 ACE2, which might be responsible for the efficient spread of SARS-CoV-2 (40) . In addition to 230 these two points, we next sought to identify mutations in cRBD that play critical roles in the 231 stronger binding-tendency towards ACE2 as compared to sRBD. As the information is limited 232 about the static conformation of a protein complex considering the changes of the binding interface 233 in physiological condition, we simulated the complex structure of cRBD-ACE2 and compared this 234 with the sRBD-ACE2 complex. The distances between interface residues were monitored as a 235 function of time to trace the shifting, breaking, or formation of new bonds. Surface Plasmon 236 resonance (SPR) and bio-layer interferometry (BLI) analyses have shown that cRBD-ACE2 237 interaction is stronger than sRBD-ACE2 interaction (11, 40, 41) . Supporting this, we also observed 238 that the total number of hydrogen bonds remained similar throughout the simulation time in both 239 sRBD-ACE2 and cRBD-ACE2 models ( Figure 3A ). This result may imply that the stronger 240 binding affinity of cRBD toward ACE2 might be attributed to the stronger interaction of Lys417-241 Asp30 compared to Arg426-Glu329. Interestingly, when the minimum interaction distances with 242 respect to the simulation time was monitored, we observed that Lys417-Asp30 pair was more 243 compact as compared to Arg426-Glu329 pair. Initially the residues in both pairs were ~1.4 Å apart;

244 however, the Arg426-Glu329 pair separated by 2.6 Å, but the Lys417-Asp30 pair remained intact 245 until the midpoint of the simulation. The bonds between both pairs broke at the same time point 246 and remained separated by ~5 Å till the end of simulation ( Figure 3A) . The relative strength and 247 variation in the bond distance between these two pairs have been recently evaluated 248 computationally, which were substantially in line with our findings (42) . These strong yet transient 249 electrostatic contacts can partly explain the phenomena of receptor recognition and S1 shedding 250 upon enzymatic cleavage of the S1-S2 junction. S protein transiently utilizes the RBD of S1 251 subunit for receptor recognition and sheds them during cell internalization. Thus, faster SARS-252 CoV-2 transmission as compared to SARS-CoV is, at least in part, might be facilitated by the 253 robust Lys417-Asp30 interaction.

In addition, we observed that Tyr449, Tyr489, Gln493, and Asn501 in cRBD established strong 255 hydrogen bonds with the interface residues of ACE2 and remained intact throughout the simulation 256 ( Figure 3A) . These results indicate that these residues are equally responsible for the relatively 257 stronger interaction of cRBD with ACE2. To demonstrate our results more clearly, we captured 258 the motions of these interface residues in animations (Supplementary movies 1 and 2), and 259 calculated binding free energies for each complex along the simulation time. The vdW, 260 electrostatic, and SASA energies of the cRBD-ACE2 were relatively stronger than that of sRBD-261 ACE2. Besides, the polar solvation energy of cRBD-ACE2 was relatively higher than sRBD-262 ACE2, which may compensate the differences in the other energies of these complexes, resulting 263 in overall slightly higher total binding free energies for the cRBD-ACE2 (see Figure 3B ). These 264 data are in agreement with recently reported study who estimated the binding-free energies of 265 sRBD and cRBD-ACE2 complexes through AMBER tool (43) . Collectively, our structural 266 modeling analyses could demonstrate stronger cRBD-ACE2 interaction compared to sRBD-ACE2 267 interaction. In addition, we demonstrated that Lys417 mutation may allow cRBD-ACE2 contact 268 more readily, which may facilitate the rapid transmission of SARS-CoV-2 compared to SARS-269 CoV.

With the help of structural information provided by the binding analysis of SARS-CoV-2 spike 273 protein and ACE2, many groups have been actively involved in designing peptide antidotes that 274 are able to block the receptor binding of the virus. A peptide, S 471-503 , derived from the ACE2 275 binding region of the sRBD has been able to hinder ACE2-sRBD interaction and thus SARS-CoV 276 entry into the cell, as confirmed in vitro (44) . By comparing the cRBD region corresponding to the 277 S 471-503 (ALNCYWPLNDYGFYTTTGIGYQPYRVVVLSFEL) peptide, we found that the N-278 terminus (bold letters) of S 471-503 and the corresponding cRBD region were considerably different;

279 however, the C-terminus portion (non-bold letters) was 100% identical to cRBD (see Figure 2A , 293 (45) . Others and we found that cRBD and sRBD interact with the overall same helical peptide (α1 294 helix) of the ACE2 with some differing interface residues from the RBDs (see Table 1 ).

295 Particularly, the two glutamic acids, EE, at the N-terminus of ACE2 did not make stable contacts 296 with both s-and cRBDs and were exposed to the solvent. Lysine 353 (bold letter in P6) is crucial 297 for RBD binding, establishing multiple stable hydrogen bonds with both s-and cRBDs ( Figure   298 3A, Table 1 ). Thus, we suggest that a peptide HW2, QAKTFLDKFNHEAEDLFYQSS-linker-

LGKGDFR (Figure 3C ) may hold the optimum capacity to engage RBD and halt its binding to 300 ACE2. More recently, a short peptide SBP1derived from the α1 helix of ACE2 was shown to bind 301 SARS-CoV-2 RBD that was expressed in insect; however, the results were not reproduced with 302 human and other insect-derived RBDs (46) . This finding suggests that either the α1 helix of ACE2 303 is not sufficient to bind RBD or it loses helicity and RBD-binding ability in solution state. A new 304 study has shown that stabilizing the helical fold of α1 helix (LCB1 & AHB1-2, Figure 3C ) retains 305 its RBD-binding capacity and effectively inhibit the SARS-CoV-2 cell entry (47) . Conclusively, 306 we suggest that merely α1 helix derived peptides, in the absence of structural constrains or 307 LGKGDFR motif, may poorly bind RBD and hinder its binding to ACE2. Figure 4A) . Differences in the sequence and length of the CDRs indicate that these mAbs 322 recognize distinct epitopes on the RBD and may not overlap thoroughly. Over a short period, more 323 than two dozen of SARS-CoV-2 S protein neutralizing mAbs have been identified and structurally 324 elucidated. We compared the sequence and structures of CDR regions of these mAbs with that of 325 CR3022 and F26G19. We found that the immunoglobulin G heavy-chain variable region 3 (in 326 other words the VH-CDR3) and to some extent the VL-CDR1 in these mAbs are diversified and 327 utilized to target the RBD of spike protein (Figure 4B, Supplementary Table 1 & 2) .

Next we sought to predict conformational epitopes of cRBD using structural information of the 329 mAbs. To ensure the authenticity of the epitope prediction, the co-crystal structure of sRBD-330 F26G19 was used as control. We observed that epitope 1 completely overlapped with the 331 experimental result, supporting the reliability of our analysis ( Figure 4C) . Among the predicted 332 cRBD epitopes, the residues in epitope 2 were mainly composed with highly variable regions 333 between sRBD and cRBD (cyan color arrows in the aligned sequences). In contrast, the residues 334 of the epitope 1 and 3 were significantly conserved between sRBD and cRBD (epitope 1, 93%; 335 epitope 3, 100%, Figure 4D ). This result implies that the anti-SARS-CoV sRBD mAbs 336 recognizing epitope 1 or epitope 3 could bind the cRBD and may hinder its receptor binding.

337 However, the epitope 2 region was highly variable between cRBD and sRBD, therefore the anti-338 sRBD mAbs recognizing epitope 2 may not be able to bind or neutralize cRBD.

Recent studies comprising SPR and BLI analyses have demonstrated that the sRBD mAbs 342 including m396, 80R, s230, and CR3014 are not able to recognize cRBD (11, 41) , although the 343 reason for failure was not understood. To evaluate the reason, we placed or docked the scFv regions 344 of these sRBD mAbs onto cRBD revealing their interface residues ( Table 2) . s230 and 80R 345 interacted with a part of the overlapping residues at the hypervariable RBDR region (epitope 2) of 346 cRBD; this could possibly explain their negative binding in the previous SPR and BLI experiments 347 (11, 41) . m396 and F26G19 were partly overlapped onto the residues at non-epitope regions 348 (Figure 5A) , suggesting that these mAbs may not bind cRBD. The binding affinity of F26G19 349 with cRBD has not been studied yet requiring further evaluation in future. Taken together, we 350 suggest that these m396, 80R, s230, and F26G19 mAbs recognize non-conserved or non-epitope 351 regions of cRBD, and therefore might not be able to block the cRBD interaction with ACE2.

352 Interestingly, cRBD escapes from the anti-sRBD mAbs even though cRBD can bind to ACE2 with 353 high affinity. This could be partly explained by the structural differences of binding regions 354 between them. Anti-sRBD mAbs have CDR that are very specific and recognize conformational 355 epitopes on RBD, while ACE2 utilize a long helix that binds longitudinally to RBD.

In addition, we observed a mutation at Ala475 in cRBD, which corresponded to Pro462 in 357 sRBD (see Figure 1A) . A previous study has shown that CR3014 mAb was not effective on the 358 mutant Pro462Leu viruses, although it could prevent lung damage and SARS-CoV shedding in 359 ferrets (48) . We found a glycine insertion mutation in the same loop (475-AGSTPCNGV-483), 360 lengthening the loop RMSD to 2-3 Å (see Figure 1C ). This might be the reason why the previous Residues participating in epitopes are indicated with arrows in the aligned cRBD and sRBD a.a.

sequences (the arrow colors correspond to their respective epitopes).

A) The binding interface of SARS-CoV mAbs are displayed with respect to the predicted epitopes on cRBD (Green = Epitope 1, Cyan = Epitope 2, Red = Epitope 3; explained in Figure   4 ). B) PLIF analysis of the CR3014 and CR3022 mAbs with cRBD. C) CR3014 cluster around epitope 2, which is highly variable between sRBD (left) and cRBD (right). CR3022 cluster near epitope 3, which is conserved between cRBD and sRBD. Figure 1B . 

Depending 184 on the state of ACE2, which exist in monomeric (solution form) as well as membrane-bound 185 dimeric form (23), the one spike trimer-two ACE2 notion could be possible in two ways. First, 186 two soluble ACE2 bind to two up RBDs on the same spike trimer (shown in figure 1B); second, 187 two ACE2 in the membrane-bound dimeric form bind to two up RBD on two separate spike trimers 188 (supplementary Figure 1B). As a dimeric ACE2 is known to accommodate two RBD 189 simultaneously (23), we suggest that soluble ACE2 may bind to the spike RBD more readily and 190 stoichiometrically. This is because the binding of membrane-bound ACE2 is highly likely 191 dependent on the spatial arrangement of the spike trimers on the virus surface, whereas soluble 192 ACE2 can bind to any RBD available in up conformation

Next, we sought to analyze and compare the interface cRBD-ACE2 with that of sRBD-ACE2

Interface analyses revealed that the electrostatic 201 contact between Arg426 and Glu329 in sRBD-ACE2 was analogous to that of Lys417 and Asp30 202 contact in cRBD-ACE2 (Figure 2B). However, this interaction was transient and break after the 203 Asp30 of ACE2 established an intrachain contact with the nearby His34 (explained later). By 204 performing protein patch analysis, we demonstrated that the standing cRBD exposes Lys417 that 205 establishes strong electrostatic interaction with Asp30 of the ACE2

With this concern, we 362 performed epitope mapping and protein ligand interaction fingerprints (PLIF) analyses to further 363 evaluate the binding of CR3014. CR3014 was clustered around the same AGSTPCNGV-loop, a 364 part of epitope 2, implying that the epitope 2-targeting sRBD mAbs may not be able to bind cRBD 365 (Figure 5B). Likewise, the sRBD mAbs recognizing the epitopes in variable RBDR regions may 366 also not be able to bind cRBD. Thus, we suggest that mAbs or therapeutic peptides that bind to a 367 conserved epitope on RBD may hinder the interaction of both SARS-CoV and SARS-CoV-2 spike 368 with ACE2. 369 CR3022 has been reported to completely neutralize the CR3014 escape SARS-CoV mutants 370 (i.e., Pro462Leu) and synergize the neutralizing effect of CR3014 without competing with its 371 epitopes (48). A recent study has also demonstrated that CR3022 does not compete with ACE2-372 binidng site of cRBD and exhibit binding to cRBD in the BLI analysis; in contrast, other mAbs, 373 such as CR3014, m396, and MERS-CoV neutralizing mAb m336 were not able to bind to cRBD 374 (41)

In their docking analysis, the binding of CR3022 and s230 overlapped on the same 384 interface of cRBD. In contrast, we and Tian et al. (41) have demonstrated that there is no overlap 385 between CR3022 and the ACE2-binding region of cRBD (see Supplementary Figure 2A)

To validate these results, we superimposed the structures of s230-sRBD and ACE2-sRBD 388 complexes, which revealed that ACE2 and s230 were overlapped with the same interface of sRBD 389 (Supplementary Figure 2B)

CoV (48) and the recent SARS-CoV-2 (41) (Supplementary Figure 2A). Notably, we found that 392 CR3022 recognized a highly conserved region partly overlapping with epitope 3 (Figure 5C)

Full-length CR3022 IgG also exhibited similar binding affinity towards both cRBD 397 and sRBD, but not able to neutralize SARS-CoV-2. The difference in the binding affinities of 398 CR3022 Fab and IgG with cRBD and its inability to neutralize SARS-CoV-2 needs further 399 investigation. We further speculated the binding mechanism of CR3022 IgG with SARS-CoV-2 400 trimeric spike. Yuan et al. suggested that CR3022 could bind to a trimeric spike in two or three 401 RBD up states, but not single up conformation due to steric hindrance. We observed that the light 402 chain constant region of CR3022 in their suggested model clash with CTD of adjacent or same 403 spike protomer, regardless of the RBD up position (Figure 6A). Conversely, we suggest that 404 epitope 3 can accommodate the Fab and CR3022 IgG differently without any clash with the 405 surrounding protomers. We further suggest that a single CR3022 can bind to two RBDs of two 406 different protomers in nearby trimers and an RBD up can be assessed by the Fab regardless of the 407 up or down conformation of the adjacent RBD (Figure 6B)

Nineteen mAbs bound to the cRBD were obtained from 412 PDB database and their CDRs were compared with those of sRBD binding mAbs. Most of the 413 SARS-CoV-2 spike neutralizing antibodies, with some exceptions, were bound to the RBDR 414 region of cRBD (Figure 6C). Surprisingly, all the three CDRs in the VH region of these mAbs 415 make contacts with the epitope 2. Whereas, the CDR3 in VL was docked into the epitope 1. This 416 suggests that the RBDR region in cRBD is highly immunogenic due to the differences in the RBDR 417 regions of SARS-CoV and SARS-CoV-2, producing highly specific and

CoV mAbs are not effective against SARS-CoV-2. The cRBD binding mAbs can hinder its 420 interaction with ACE2 even if they minimally bind or do not bind to the RBDR region. In fact, 421 some mAbs can simultaneously bind to cRBD at distinct epitopes and abrogate its biding to ACE2

generated nine cRBD-bindng mAbs from genetically humanized mice and 423 COVID-19 convalescent patients and identified their epitopes on cRBD through 424 hydrogendeuterium exchange mass spectrometry (HDX-MS)

Similar antibody cocktail strategy has also been reported by other group, where two RBD-429 neutralizing antibody bind two non-overlapping epitopes on cRBD (53)

and BD-368-2 binds to a non-overlapping epitope (epitope 431 not suggested in our study) on the opposite side of epitope 2 (Figure 6C). Overall, these findings 432 suggest that the RBD of SARS-CoV-2 is highly immunogenic and harbor multiple but overlapping 433 epitopes

This interaction 436 is further strengthened by electrostatic and hydrophobic contacts at the cRBD-ACE2 interface

Overview of Current Therapeutics and Novel Candidates Against Influenza, Respiratory Syncytial Virus, and Middle East Respiratory Syndrome Coronavirus Infections

The emergence of SARS-CoV-2 in Europe and North America

A new coronavirus associated with human respiratory disease in China

A pneumonia outbreak associated with a new coronavirus of probable bat origin

A Novel Coronavirus from Patients with Pneumonia in China

Full-genome evolutionary analysis of the novel corona virus (2019-nCoV) rejects the hypothesis of emergence as a result of a recent recombination event

Current Status of Multiple Drug Molecules, and Vaccines: An Update in SARS-CoV-2 Therapeutics

Coronavirus spike proteins in viral entry and pathogenesis

Functional assessment of cell entry and receptor usage for SARS-CoV-2 and other lineage B betacoronaviruses

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

Cryo-EM structures of MERS-CoV and SARS-CoV spike glycoproteins reveal the dynamic receptor binding domains

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

The SARS-CoV-2 Spike Variant D614G Favors an Open Conformational State

The D614G mutation in the SARS-CoV-2 spike protein reduces S1 shedding and increases infectivity

In silico mechanistic analysis of IRF3 inactivation and high-risk HPV E6 species-dependent drug response

TLR4/MD2 specific peptides stalled in vivo LPS-induced immune exacerbation

Extraction of molecular features for the drug discovery targeting protein-protein interaction of Helicobacter pylori CagA and tumor suppressor protein ASSP2

Linear and Rationally Designed Stapled Peptides Abrogate TLR4 Pathway and Relieve Inflammatory Symptoms in Rheumatoid Arthritis Rat Model

Energy-based analysis and prediction of the orientation between light-and heavy-chain antibody variable domains

Canonical structures for the hypervariable regions of immunoglobulins

SCOP: a structural classification of proteins database for the investigation of sequences and structures

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

Structural insights into immune recognition of the severe acute respiratory syndrome coronavirus S protein receptor binding domain

Structure of severe acute respiratory syndrome coronavirus receptor-binding domain complexed with neutralizing antibody

Unexpected Receptor Functional Mimicry Elucidates Activation of Coronavirus Fusion

SAbPred: a structurebased antibody prediction server

Improving B-cell epitope prediction and its application to global antibody-antigen docking

CHARMM36m: an improved force field for folded and intrinsically disordered proteins

GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit

Optimizing working parameters of the smooth particle mesh Ewald algorithm in terms of accuracy and efficiency

LINCS: a linear constraint solver for molecular simulations

Open Source Drug Discovery C, Lynn A. g_mmpbsa--a GROMACS tool for high-throughput MM-PBSA calculations

Insights into the species-specific TLR4 signaling mechanism in response to Rhodobacter sphaeroides lipid A detection

Inhibition of SARS-CoV-2 Infections in Engineered Human Tissues Using Clinical-Grade Soluble Human ACE2

Peptide Antidotes to SARS-CoV-2 (COVID-19)

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

Structural basis for receptor recognition by the novel coronavirus from Wuhan

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

Structure, function and antigenicity of the SARS-CoV-2 spike glycoprotein

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

Enhanced receptor binding of SARS-CoV-2 through networks of hydrogen-bonding and hydrophobic interactions

Molecular Mechanism of Evolution and Human Infection with SARS-CoV-2

Screening and identification of linear B-cell epitopes and entry-blocking peptide of severe acute respiratory syndrome (SARS)-associated coronavirus using synthetic overlapping peptide library

Identification of critical determinants on ACE2 for SARS-CoV entry and development of a potent entry inhibitor

Investigation of ACE2 N-terminal fragments binding to SARS-CoV-2 Spike RBD. biorxiv

De novo design of picomolar SARS-CoV-2 miniprotein inhibitors

Human monoclonal antibody combination against SARS coronavirus: synergy and coverage of escape mutants

Spike protein binding prediction with neutralizing antibodies of SARS-CoV-2. bioRxiv

Crystal structure of the receptor-binding domain from newly emerged Middle East respiratory syndrome coronavirus

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

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

Antibody Shows High Efficacy in Severely Infected Hamsters and Provides a Potent Cocktail Pairing Strategy

Sequence variation of SARS-CoV-2 spike protein may facilitate stronger interaction with ACE2 promoting high infectivity

Sangdun Choi supervised the study: Hyun Goo Woo was responsible for the funding acquisition and project administration: All the authors contributed to editing and reviewing the manuscript

All authors declare that there is no competing interest.