key: cord-0812865-ssts6er4
authors: Nelson, Gard; Buzko, Oleksandr; Spilman, Patricia; Niazi, Kayvan; Rabizadeh, Shahrooz; Soon-Shiong, Patrick
title: Molecular dynamic simulation reveals E484K mutation enhances spike RBD-ACE2 affinity and the combination of E484K, K417N and N501Y mutations (501Y.V2 variant) induces conformational change greater than N501Y mutant alone, potentially resulting in an escape mutant
date: 2021-01-13
journal: bioRxiv
DOI: 10.1101/2021.01.13.426558
sha: 7832894f850f2f46c540d0ec2ba1e9ad9612769b
doc_id: 812865
cord_uid: ssts6er4

Rapidly spreading SARS-CoV-2 variants present not only an increased threat to human health due to the confirmed greater transmissibility of several of these new strains but, due to conformational changes induced by the mutations, may render first-wave SARS-CoV-2 convalescent sera, vaccine-induced antibodies, or recombinant neutralizing antibodies (nAbs) ineffective. To be able to assess the risk of viral escape from neutralization by first-wave antibodies, we leveraged our capability for Molecular Dynamic (MD) simulation of the spike receptor binding domain (S RBD) and its binding to human angiotensin-converting enzyme 2 (hACE2) to predict alterations in molecular interactions resulting from the presence of the E484K, K417N, and N501Y variants found in the South African 501Y.V2 strain – alone and in combination. We report here the combination of E484K, K417N and N501Y results in the highest degree of conformational alterations of S RBD when bound to hACE2, compared to either E484K or N501Y alone. Both E484K and N501Y increase affinity of S RBD for hACE2 and E484K in particular switches the charge on the flexible loop region of RBD which leads to the formation of novel favorable contacts. Enhanced affinity of S RBD for hACE2 very likely underpins the greater transmissibility conferred by the presence of either E484K or N501Y; while the induction of conformational changes may provide an explanation for evidence that the 501Y.V2 variant, distinguished from the B.1.1.7 UK variant by the presence of E484K, is able to escape neutralization by existing first-wave anti-SARS-CoV-2 antibodies and re-infect COVID-19 convalescent individuals.

As many SARS-CoV-2 variants emerge and displace first-wave viruses 1,2 , it is important not 31 only to assess their relative transmissibility, but also their ability to escape antibody neutralization 32 by convalescent antibodies in recovered COVID-19 patients 3 , recombinant neutralizing antibodies 33 (nAbs) developed as therapeutics, or antibodies elicited by first-generation vaccines. 34

Of great interest are variants that include mutations with the potential to affect the interaction 35 of the SARS-CoV-2 spike receptor binding domain (S RBD) with the host receptor, angiotensin-36 converting enzyme 2 (ACE2). The binding of the S RBD of SARS-CoV-2, like SARS-CoV before 37 it, to ACE2 initiates infection 4-7 , thus variants that have a greater binding affinity for ACE2 are 38 likely to be more readily transmissible 8 . Transmissibility goes hand-in-hand with mortality, 39 because even if a variant does not produce a higher rate of morbidity or mortality, the total number 40 of severe cases and death would be expected to increase due to what may be an exponential 41 increase in infections. 42

The dire consequences of more rapid and widespread infection can further be compounded by 43 a decrease in efficacy of available antibody-based therapeutics and vaccines; and by a loss of 44 protective immunity in persons previously infected with a 'first wave' virus. The efficacy of 45 vaccines may be altered if a specific mutation or combination of mutations in a variant results in 46 significant conformational changes that render key regions of S that participate in ACE2 binding 47 'unrecognizable' to antibodies generated in response to a first-generation vaccine. A similar 48 principle is in play for the efficacy of nAbs targeted to the receptor interface 9,10 and convalescent 49 sera. 50

Here, to better understand the risks posed by individual or combined mutations in the 'second-51

wave' variants, we leveraged our in silico Molecular Dynamic (MD) simulation capabilities to 52 perform computational analysis of interactions of the S RBD with human ACE2. In our first report, 53

Nelson et al. 11 (in preprint) "Millisecond-scale molecular dynamics simulation of spike RBD 54 structure reveals evolutionary adaption of SARS-CoV-2 to stably bind ACE2", we initially used 55 millisecond-scale MD simulation to simulate free SARS-CoV-2 S RBD based on previously 56

reported structures 12,13 as well as its molecular interactions with ACE2 and showed S adopts a 57 binding-ready conformation, incurring little entropic penalty during ACE2 interaction. We further 58 revealed areas of high-affinity interaction between S RBD and ACE2 that have a high likelihood 59 of determining binding kinetics. 60

Here, we utilized the MD simulation methods employed in our first study to investigate what 61 effects mutations found at the S RBD-ACE2 interface in the rapidly spreading South African 62 variant 501Y.V2 14 -E484K, K417N, and N501Y -have on RBD binding affinity and spike 63 conformation. 64

As shown in Figure 1a , the E484K, K417N, and N501Y mutants span the S RBD-ACE2 66 interface, with the E484K substitution occurring in a highly flexible loop region of the S RBD 67 (Fig. 1c) . The N501Y substitution is found in a second region of contact 11 , and the K417N mutation 68 in a region between the two that shows relatively little interaction with ACE2. 

To determine if it is likely the presence of the E484K or N501Y mutations alter the 91 conformation of S RBD, we performed Principal Component Analysis (PCA) for the triple mutant, 92 E484K and N501Y mutants alone, and compared them to the original cryo-EM 13 structure 93 representing the 'first wave' sequence. The PCA plots are shown in Figure 2 . While both mutants 94 sample similar conformations, S RBD with K484 alone (Fig. 2b) preferentially adopts 95 conformations similar to the cryo-EM structure described in Wrapp et al. 13 . This is similar to the 96 behavior observed for the first-wave strain as described in our initial report 11 . Y501 alone ( structure, the red circled region to those found for N501Y, the black dashed circle those found for 117 K484 alone, and the gray dashed circles those found for the triple mutant; with a greater density 118

(purple) indicating a greater probability other conformations occur. 119 120 E484K shows increased contact with hACE2 E75 121 E484K, whether in the presence of both K417N, and N501Y variants (Fig. 3a) , or as the only 122 variant in the presence of K417 and N501 (Fig. 3b) is associated with increased contact between 123 RBD residue 484 and ACE2 E75 compared to either the N501Y mutant alone or the 'first wave' 124 (WT) sequence ( Fig. 3c and d, respectively) . In the presence of either Y501 and WT, E484 shows 125 little contact with hACE2 E75. In addition, The RBD E484K mutant shows increased contact 126 between it and several residue pairs in addition to hACE2 E75. The strong contacts seen for only the E484K mutant here suggest that spike may adopt a 142 distinct 'binding pose' relative to other conformations. We further posit the differing PCA 143 densities seen for the triple mutant that seem to result in lesser loop interaction with hACE2 as 144 compared to E484K alone may be due to the other mutations. We will investigate this in depth in 145 future studies. Additional support for the threat posed by mutations at residue 484 is provided by Greaney et 155 al. 17 who undertook an impressive effort to map mutations that affect binding of ten human 156 monoclonal antibodies. By employing a deep mutational scanning method, they found that 157 mutations at residue 484 have a high probability of affecting antibody binding. They further 158 suggested their analytical method provides a tool for design of escape-resistant antibody cocktails 159 to overcome the threat posed by viral evolution. We believe the MD simulation approach used 160 here similarly represents a tool to be used in the arsenal against the continuing pandemic, as it 161 provides insight into the likelihood mutations alone or in combination may have effects that lessen 162 the efficacy of existing therapies or vaccines. 163

In vaccine design, it has been suggest that the makers of vaccines could keep up with viral 164 evolution by continual alteration of the 'payload' -almost all vaccines in development use the 165 spike sequence -to fit currently predominant strains. We suggest vaccines whose efficacies are 166 largely dependent upon humoral responses to the S antigen only are inherently limited by the 167 emergence of novel strains and dependent upon frequent re-design. In contrast, a vaccine that 168 elicits a vigorous T-cell response that is far less subject to changes due to accruing mutations 169 provides a better, more efficient approach to protection. The ideal vaccine would also deliver a 170 second, conserved antigen such as the SARS-CoV-2 nucleocapsid protein, that very likely will 171 elicit humoral and cell-mediated immune responses that will remain effective, even in the face of 172 a rapidly changing virus. 173

Although not the subject of the investigation described here, we are developing a dual-antigen 174 The backbones of residues at the RBD/hACE2 interface were used for all PCA calculations. 208

Structures were RMSD aligned to the cryo-EM structure. Eigenvectors for the PCA plots were 209 then calculated using the full set of simulations of the triple mutant, E484K and N501Y systems. 210

Simulation structures were projected onto the eigenvectors for each mutation system separately. 211

All calculations were run with cpptraj and plotted using gnuplot. 

Estimated transmissibility and severity of novel SARS-CoV-2 Variant 228 of Concern 202012/01 in England. medRxiv

Spike mutation D614G alters SARS-CoV-2 fitness

Escape from neutralizing antibodies by SARS-CoV-2 spike protein 233 variants

Characterization of severe acute respiratory syndrome-associated 235 coronavirus (SARS-CoV) spike glycoprotein-mediated viral entry

Structure, Function, and Evolution of Coronavirus Spike Proteins

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is 241 Blocked by a Clinically Proven Protease Inhibitor

Receptor binding and priming of the spike protein of SARS-CoV-2 for 244 membrane fusion

SARS-CoV-2 RBD in vitro evolution follows contagious mutation 246 spread, yet generates an able infection inhibitor. bioRxiv

Receptor-binding domain-specific human neutralizing monoclonal antibodies 249 against SARS-CoV and SARS-CoV-2

Cross-neutralization of SARS-CoV-2 by a human monoclonal SARS-CoV 252 antibody

Millisecond-scale molecular dynamics simulation of spike RBD structure 254 reveals evolutionary adaption of SARS-CoV-2 to stably bind ACE2. bioRxiv

Function, and Antigenicity of the SARS-CoV

Cryo-EM structure of the 2019-nCoV spike in the prefusion 259 conformation

Emergence and rapid spread of a new severe acute respiratory 261 syndrome-related coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in 262

SARS-CoV-2 escape <em>in vitro</em> from a highly 265 neutralizing COVID-19 convalescent plasma. bioRxiv

Neutralization of N501Y mutant SARS-CoV-2 by BNT162b2 vaccine-elicited 268 sera. bioRxiv

Complete Mapping of Mutations to the SARS-CoV-2 Spike Receptor-270

Binding Domain that Escape Antibody Recognition

Th1 Dominant Nucleocapsid and Spike Antigen-Specific CD4+ and CD8+ 273

Memory T Cell Recall Induced by hAd5 S-Fusion + N-ETSD Infection of Autologous

Dendritic Cells from Patients Previously Infected with SARS-CoV-2

A Next Generation Bivalent Human Ad5 COVID-19 Vaccine Delivering Both 278 Spike and Nucleocapsid Antigens Elicits Th1 Dominant CD4+

Complete Protection of Nasal and Lung Airways Against SARS-CoV-2 282 Challenge by Antibody Plus Th1 Dominant N-and S-Specific T-Cell Responses to 283

Subcutaneous Prime and Thermally-Stable Oral Boost Bivalent hAd5 Vaccination in an 284 NHP Study. bioRxiv

Structural basis for the recognition of SARS-CoV-2 by full-length human 286 ACE2

ff14SB: Improving the Accuracy of Protein Side Chain and Backbone 288 Parameters from ff99SB

University of California