key: cord-314072-av3r7it7 authors: Liu, Zhuoming; VanBlargan, Laura A.; Rothlauf, Paul W.; Bloyet, Louis-Marie; Chen, Rita E.; Stumpf, Spencer; Zhao, Haiyan; Errico, John M.; Theel, Elitza S.; Ellebedy, Ali H.; Fremont, Daved H.; Diamond, Michael S.; Whelan, Sean P. J. title: Landscape analysis of escape variants identifies SARS-CoV-2 spike mutations that attenuate monoclonal and serum antibody neutralization date: 2020-11-08 journal: bioRxiv DOI: 10.1101/2020.11.06.372037 sha: doc_id: 314072 cord_uid: av3r7it7 Although neutralizing antibodies against the SARS-CoV-2 spike (S) protein are a goal of most COVID-19 vaccines and being developed as therapeutics, escape mutations could compromise such countermeasures. To define the immune-mediated mutational landscape in S protein, we used a VSV-eGFP-SARS-CoV-2-S chimeric virus and 19 neutralizing monoclonal antibodies (mAbs) against the receptor binding domain (RBD) to generate 48 escape mutants. These variants were mapped onto the RBD structure and evaluated for cross-resistance by convalescent human plasma. Although each mAb had unique resistance profiles, many shared residues within an epitope, as several variants were resistant to multiple mAbs. Remarkably, we identified mutants that escaped neutralization by convalescent human sera, suggesting that some humans induce a narrow repertoire of neutralizing antibodies. By comparing the antibody-mediated mutational landscape in S protein with sequence variation in circulating SARS-CoV-2 strains, we identified single amino acid substitutions that could attenuate neutralizing immune responses in some humans. Control of the SARS-CoV-2 pandemic likely will require the deployment of multiple This data and functional approach may be useful for monitoring and evaluating the emergence 100 of escape from antibody-based therapeutic and vaccine countermeasures as they are deployed. To select for SARS-CoV-2 S variants that escape neutralization, we used VSV-SARSas we isolated this mutation alone, and acquisition of the L517R substitution appeared to 141 increase viral fitness as judged by plaque morphology (Fig S2) . For SARS2-19, S477N was 142 isolated as a single variant suggesting that this substitution arose first, however acquisition of 143 the S514F did not alter plaque morphology (Fig S2) . As the L517R or S514F substitutions were 144 not identified alone, it remains unclear whether they cause resistance to 2H04 or SARS2-19 145 respectively. Collectively, these results show that escape mutational profiling can identify key 146 epitopes and dominant antigenic sites. Escape mutants confer cross-resistance to multiple mAbs. We next evaluated whether individual mutants could escape neutralization by the other 150 inhibitory mAbs in the panel. We tested the 29 identified escape mutants for neutralization by 151 ten different mAbs. We defined the degree of resistance as a percentage by expressing the 152 number of plaques formed by each mutant in the presence of antibody versus its absence. We plotted the degree of resistance to neutralization as a heat map and arbitrarily set 50% as the than one mAb, with substitutions at S477 and E484 exhibiting broad resistance (Fig 3) (Fig 4A and S4 ). As observed with 186 chimeric viruses expressing the wild-type S protein, all of the escape mutants were inhibited by 187 hACE2-Fc but not mACE2-Fc. However, the extent of neutralization by hACE2-Fc varied (Fig 188 4A) , with some mutants more sensitive to receptor inhibition and others exhibiting relative 189 resistance. Substitutions at residues R346, A352, N450, S477, S494 and P499 were more 190 sensitive to inhibition by soluble hACE2 than the wild-type S as evidenced by reduced IC 50 191 values (Fig 4A) and leftward shifts of the inhibition curves (Fig S4) . This effect was substitution-192 dependent as N450K was 6-fold more sensitive to hACE2 than N450Y (P < 0.001). Several 199 inhibition at the highest concentration (20 µg/ml) of hACE2-Fc tested (Fig 4A and B) . Residue initial dilution) of sera tested. All four of the substitutions at residue E484 were resistant to each 214 of the four sera, suggesting that this is a dominant neutralizing epitope. Indeed, change at E484 substitutions (K444E, G446V, L452R and F490S) resulted in resistance to neutralization of sera 217 13, 35 and 37 (Fig 5A and S5 ). Substitutions N450D and N450Y but not N450K were resistant 218 to sera 13 and 35. Sera 13 and 35 also did not efficiently neutralize S477G, L441R, and T478I. All four sera neutralized the single substitution S477N as well as wild-type virus (Fig 5A-B) . Substitution S477N was sensitive to neutralization by sera 13 and 35 except in the presence of 221 a second S514F substitution (Fig 5A and S5) . Additional amino acid substitutions that conferred 222 resistance to serum 13 include T345S and G446D. Substitution F486S, which altered sensitivity 223 to soluble ACE2, escaped neutralization by serum 35 but not 13, 29 or 37. Thus, individual 224 escape mutants can exhibit resistance to neutralization by polyclonal human convalescent sera. This observation suggests that the repertoire of antigenic sites on RBD that bind high titer 226 neutralizing antibodies is limited in some humans. Comparison of escape mutants with S sequence variants isolated in humans. To broaden our analysis, we performed a second campaign of escape mutant selection compiled all publically available genome sequences of SARS-CoV-2. Using 161,182 genomes 234 from GISAID, we calculated the substitution frequencies throughout RBD protein (Fig 7A) and 235 mapped the identified residues onto RBD structure (Fig 7B) . Of the 48 escape variants we 236 selected, 27 are present in circulating human isolates of SARS-CoV-2 (Fig 7A) . The most 237 frequent S sequence variant seen in clinical isolates is D614G which is present in 86% of 238 sequenced isolates. The second most frequent substitution is S477N, which is present in 5.1% The mutations we selected also inform the mechanism by which the different antibodies Structural studies on the mechanism by which 2B04 and 2H04 neutralize SARS-CoV-2 support 278 inhibition by directly competing with ACE2 binding and an indirect mechanism, respectively. Direct competition with ACE2 binding is consistent with the escape mutants we selected with 280 2B04, and an indirect mechanism fits with the escape mutants we identified to 2H04. The Viral mutation 612 rates Structural basis of receptor recognition by SARS-CoV-2 Coronaviruses 616 lacking exoribonuclease activity are susceptible to lethal mutagenesis: evidence for 617 proofreading and potential therapeutics Coronaviruses as DNA wannabes: a new model for the 619 regulation of RNA virus replication fidelity Escape from neutralizing 622 antibodies by SARS-CoV-2 spike protein variants Cryo-EM structure of the 2019-nCoV spike in the prefusion 625 conformation A noncompeting pair of human neutralizing antibodies block COVID-19 virus binding to 628 its receptor ACE2 Potently neutralizing and protective human 634 antibodies against SARS-CoV-2 This study was supported by NIH contracts and grants (75N93019C00062 and R01