key: cord-0814187-iqyfsjnl
authors: Mannar, Dhiraj; Saville, James W.; Zhu, Xing; Srivastava, Shanti S.; Berezuk, Alison M.; Zhou, Steven; Tuttle, Katharine S.; Kim, Andrew; Li, Wei; Dimitrov, Dimiter S.; Subramaniam, Sriram
title: Structural Analysis of Receptor Binding Domain Mutations in SARS-CoV-2 Variants of Concern that Modulate ACE2 and Antibody Binding
date: 2021-12-04
journal: Cell Rep
DOI: 10.1016/j.celrep.2021.110156
sha: 7b420ad4eac95f0fc63714828fb2f87e3e3d87e5
doc_id: 814187
cord_uid: iqyfsjnl

The recently emerged SARS-CoV-2 Beta (B.1.351) and Gamma (P.1) variants of concern (VoCs) include a key mutation (N501Y) found in the Alpha (B.1.1.7) variant that enhances affinity of the spike protein for its receptor, ACE2. Additional mutations are found in these variants at residues 417 and 484 that appear to promote antibody evasion. In contrast, the Epsilon variants (B.1.427/429) lack the N501Y mutation, yet exhibit antibody evasion. We have engineered spike proteins to express these RBD VoC mutations either in isolation, or in different combinations, and analyze the effects using biochemical assays and cryo-EM structural analyses. Overall, our findings suggest that the emergence of new SARS-CoV-2 variant spikes can be rationalized as the result of mutations that confer either increased ACE2 affinity, increased antibody evasion, or both, providing a framework to dissect the molecular factors that drive VoC evolution.

Structural comparison of D614G-ACE2 and D614G+L452R-ACE2 complexes reveals no significant 128 changes at the RBD-ACE2 interface ( Figures 3F, S4P) , indicating the enhanced ACE2 affinity afforded by L452R is not due to modulation of direct ACE2 contacts. In contrast to L452, the side 130 chain of R452 extends farther away from the RBD core ( Figure S4Q ) and is better exposed to 131 solvent, suggesting that R452 may enhance the solvation of the RBD in the up position. In addition 132 to solvation effects, the L452R substitution introduces a positive charge at position 452 which may 133 increase the electrostatic complementarity between the RBD and ACE2. Figure S4R shows the 134 increase in electropositivity at position 452 upon L452R substitution, with position 452 135 approximately 13 Å away from the highly electronegative site on ACE2 centered at E35. Thus, in 136 contrast to the local rearrangements observed at the RBD-ACE2 interface for the N501Y, E484K 137 and K417N/T mutations, the binding effect of the L452R mutation is likely mediated by solvation 138 and/or electrostatic complementarity effects.

Mutations E484K, L452R and K417N/T facilitate decreased antibody binding 140 We next sought to evaluate the effect of VoC/VoI RBD mutations on antibody binding. We selected 141 a panel of previously reported antibodies which cover the four distinct anti-RBD antibody classes 142 (Barnes et al., 2020b, Table S3 ) and an ultrapotent antibody, S2M11, that uniquely binds two 143 neighbouring RBDs (Tortorici et al., 2020) . Antibody binding was quantified via enzyme-linked 144 immunosorbent assay (ELISA) (Figures 4, S1D ). As expected, class 3 (S309) and class 4 (CR3022) 145 antibodies, whose footprints did not span VoC/VoI mutations, exhibited relatively unchanged 146 binding across all variant spikes ( Figure 4B ). Mutations at position 417 of the S protein to either N 147 or T abolished or significantly reduced ab1 (Li et al., 2020a) (class 1 like) binding respectively, 148 demonstrating the importance of K417 within the molecular epitope of ab1. Similarly, the E484K 149 mutation resulted in loss of binding to ab8 (Li et al., 2020b) (class 2) and S2M11, highlighting the 150 critical nature of E484 within the epitopes of these antibodies. L452 sits peripherally within the 151 J o u r n a l P r e -p r o o f footprint of S2M11 and mutation of this residue to R452 reduced but did not abolish its binding, 152 possibly via steric or charge mediated effects, or by allosteric modulation of direct contacts. Taken 153 together, these results suggest the escape of antibody binding from the four major anti-RBD 154 classes is primarily mediated by modulation of direct contacts at mutational sites. (Table S2 ). Using this structural data, we evaluated the frequency of Barnes et al., 2020b Barnes et al., , 2020a Cao et al., 2020; Hurlburt et al., 2020; 2020a , 2020b Yuan et al., 2020; Zhou et al., 2020) . The majority of deposited human-derived 217 neutralizing antibodies bound the RBD with footprints spanning at least one of the positions 218 corresponding to RBD mutation in VoC/VoIs ( Figure 4E ). Of these antibodies, the majority 219 interacted with more than one position corresponding to RBD mutations in VoC/VoIs ( Figure 4F ).

Of these variants, B.1.351, P.1 and VOC 202102/02 possess mutations that are collectively 221 recognized by the majority of the antibodies selected, suggesting these variants may exhibit the 222 greatest RBD-directed antibody escape during human infection ( Figure 4G ). 223 We additionally generated novel combinations of RBD mutations by introducing L452R into 224 B.1.351 and P.1 constructs. While these mutational combinations enable enhanced ACE2 binding 225 compared to wild-type spikes, the increase in ACE2 binding affinity conferred by the L452R 226 mutation in isolation was not preserved. We demonstrate that these novel constructs retain The wild type SARS-CoV-2 S HexaPro expression plasmid was previously described (Hsieh et al. Enzyme-Linked Immunosorbent Assay (ELISA) 100 µl of wild-type or VoC RBD mutant SARS-CoV-2 S protein preparations were coated onto 96-611 well MaxiSorp™ plates at 2 µg/ml in PBS overnight at 4°C. All washing steps were performed 5 612 times with PBS + 0.05% Tween 20 (PBS-T). After washing, wells were either incubated with 613 blocking buffer (PBS-T + 2% BSA) for 1 hr at room temperature. After washing, wells were 614 incubated with dilutions of primary antibodies in PBS-T + 0.5% BSA buffer for 1 hr at room 615 temperature. After washing, wells were incubated with goat anti-human IgG (Jackson 616 ImmunoResearch) at a 1:8,000 dilution in PBS-T + 0.5% BSA buffer for 1 hr at room temperature.

After washing, the substrate solution (Pierce™ 1-Step™) was used for colour development 

at a cost to ACE2 affinity • Cryo-EM structures reveal details of ACE2 contacts with mutant spike proteins • Analysis of ACE2 binding and antibody evasion in unnatural RBD mutational combinations eTOC Blurb Mannar et al. use structural and biochemical techniques to dissect the role of SARS-CoV-2 spike glycoprotein mutations within the receptor binding domain, demonstrating modular mutational effects which combine to simultaneously enhance receptor engagement and decrease antibody binding in emerging variant spike proteins

 (Li et al. 2020b