key: cord-1042283-3rytky4d
authors: Jennewein, Madeleine F.; MacCamy, Anna J.; Akins, Nicholas R.; Feng, Junli; Homad, Leah J.; Hurlburt, Nicholas K.; Seydoux, Emilie; Wan, Yu-Hsin; Stuart, Andrew B.; Edara, Venkata Viswanadh; Floyd, Katharine; Vanderheiden, Abigail; Mascola, John R.; Doria-Rose, Nicole; Wang, Lingshu; Yang, Eun Sung; Chu, Helen Y.; Torres, Jonathan L.; Ozorowski, Gabriel; Ward, Andrew B.; Whaley, Rachael E.; Cohen, Kristen W.; Pancera, Marie; McElrath, M. Juliana; Englund, Janet A.; Finzi, Andrés; Suthar, Mehul S.; McGuire, Andrew T.; Stamatatos, Leonidas
title: Isolation and Characterization of Cross-Neutralizing Coronavirus Antibodies from COVID-19+ Subjects
date: 2021-06-22
journal: Cell Rep
DOI: 10.1016/j.celrep.2021.109353
sha: 18d461fa0349459ded9cc7341e8e8b443305439f
doc_id: 1042283
cord_uid: 3rytky4d

SARS-CoV-2 is one of three coronaviruses that have crossed the animal-to-human barrier and caused widespread disease in the past two decades. The development of a universal human coronavirus vaccine could prevent future pandemics. We characterized 198 antibodies isolated from four COVID19+ subjects and identified 14 SARS-CoV-2 neutralizing antibodies. One targeted the NTD, one recognized an epitope in S2 and eleven bound the RBD. Three anti-RBD neutralizing antibodies cross-neutralized SARS-CoV-1 by effectively blocking binding of both the SARS-CoV-1 and SARS-CoV-2 RBDs to the ACE2 receptor. Using the K18-hACE transgenic mouse model, we demonstrate that the neutralization potency and antibody epitope specificity regulates the in vivo protective potential of anti-SARS-CoV-2 antibodies. All four cross-neutralizing antibodies neutralized the B.1.351 mutant strain. Thus, our study reveals that epitopes in S2 can serve as blueprints for the design of immunogens capable of eliciting cross-neutralizing coronavirus antibodies.

In the past 2 decades there have been three zoonotic transmissions of highly pathogenic coronaviruses; SARS-CoV-1, MERS-CoV and SARS-CoV-2, which have caused widespread human disease. The most recent one, 54 SARS-CoV-2, has been rapidly spreading globally since late 2019/early 2020, infecting over 160 million people and killing almost 3.4 million people by May 2021 Patel et al., 2020) . Studies conducted in 56 mice, hamsters and non-human primates strongly suggest that neutralizing antibodies (nAbs) isolated from infected patients can protect from infection, and in the case of established infection, can reduce viremia and 58 mitigate the development of clinical symptoms (Baum et al., 2020b; Cao et al., 2020a; Mercado et al., 2020; Rogers et al., 2020b; Schafer et al., 2021; Shi et al., 2020; Tortorici et al., 2020; Wu et al., 2020; Yu et al., 2020) . 60

Cocktails of neutralizing monoclonal antibodies have been approved by the FDA for the treatment of infection (Baum et al., 2020a; Weinreich et al., 2020) . Thus, nAbs are believed to be an important component 62 of the protective immune responses elicited by effective vaccines. Indeed, both the mRNA-based Pfizer and Moderna vaccines elicit potent serum neutralizing antibody responses against SARS-CoV-2 (Jackson et al., 64 2020; Walsh et al., 2020) .

characterization led to the identification of vulnerable sites on the viral spike protein (S) (Cao et al., 2020b; Ju et al., 2020; Kreer et al., 2020; Liu et al., 2020; Nielsen et al., 2020; Robbiani et al., 2020; Seydoux et al., 2020; 68 Wan et al., 2020a; Zost et al., 2020) .

Many known SARS-CoV-2 nAbs bind the receptor-binding domain (RBD) and block its interaction with its 70 cellular receptor, Angiotensin converting enzyme 2 (ACE2), thus preventing viral attachment and cell fusion Yan et al., 2020) . However, some RBD-binding mAbs prevent infection without 72

interfering with the RBD-ACE2 interaction (Pinto et al., 2020; Tai et al., 2020; . Other mAbs such as the betacoronaviruses OC43 and HKU1 or the alphacoronaviruses 229E and NL63 have not yet been identified. 84

Here, we report on the isolation and full characterization of 198 S-specific mAbs from four SARS-CoV-2infected individuals. Although a number of these mAbs recognized both SARS-CoV-2 and SARS-CoV-1, we 86 observed minimal cross-reactivity with MERS-CoV, betacoronaviruses (OC43 and HKU1) or alphacoronaviruses (NL63 and 229E) . A significant fraction of cross-reactive antibodies bound the SARS-CoV-2 S2 domain of the 88 Spike protein. 14 mAbs neutralized SARS-CoV-2. One neutralizing mAb bound NTD, another bound the S2 subunit, one bound an unidentified site on S and the remaining 11 bound RBD. Some competed with the RBD-90 ACE-2 interaction while others did not. Although seven of the SARS-CoV-2 neutralizing mAbs bound SARS-CoV-1, only four mAbs neutralized both viruses. Three targeted the RBD and one targeted the S2 subunit. Using the 92 K18-hACE transgenic mouse model, therapeutic treatment with CV1-30, a potent RBD-binding antibody, reduced lung viral loads and protected mice from SARS-CoV-2 infection. In contrast, a weaker anti-RBD 94 neutralizing mAbs, CV2-75, and the anti-NTD neutralizing mAb, CV1-1, displayed minimal protective efficacies.

These observations strongly suggest that neutralization potency along with antibody epitope-specificity 96 regulate the in vivo protective potential of anti-SARS-CoV-2 antibodies. Interestingly, the anti-S2 mAb, CV3-25, was the only one that was unaffected by mutations found in the recently emerged B.1.351 variant. These 98 mAbs, especially CV3-25, can serve as starting points for the development of immunogens to elicit protective neutralizing antibody responses against multiple coronaviruses. 100

Peripheral blood mononuclear cells (PBMCs) and serum or plasma were collected from four SARS-CoV-2infected adults (CV1 -previously discussed in Seydoux et al. 2020, CV2, CV3 and PCV1) at 3, 3.5, 6 and 7 weeks 104 after the onset of symptoms, respectively (Table S1 ). Sera from PCV1 had the highest anti-stabilized spike (S-2P) IgG and IgM titers, while the anti-S-2P IgA titers were higher in CV1 (Figure 1 A-C) . In contrast, to the 106 higher anti-S-2P IgG titers in the PCV1 sera, all four sera displayed similar anti-receptor binding domain (RBD)

IgG titers (Figure 1 D-F ). PCV1 and CV1 had higher levels of anti-RBD IgA than the other two donors and CV1 108

showed slightly lower anti-RBD IgM than the three other sera.

While all sera neutralized SARS-CoV-2 ( Figure 1G ), serum from PCV1 was significantly more potent ( Figure 1H ). 110

The serum neutralizing differences track with timepoint in infection, with the samples collected at later timepoints show greater potency, potentially indicating maturation of the humoral response. Thus, though all 112 J o u r n a l P r e -p r o o f four patients had similar anti-RBD binding antibody titers, PCV1 developed higher anti-S-2P binding antibody titers and higher neutralizing titers than the other three patients examined here. 114

Monoclonal antibodies (mAbs) have been isolated and characterized previously by us and others (Cao et al., 116 2020b; Ju et al., 2020; Kreer et al., 2020; Nielsen et al., 2020; Robbiani et al., 2020; Seydoux et al., 2020) . We isolated individual S-2P+ and RBD+ IgG+ B cells (Table S1 ) from all four subjects. The percentage of S-2P+ cells 118

in the four patients ranged from 0.23%-1.84% of IgG+ B cells, of which 5-12.7% targeted the RBD. In agreement with the above-discussed serum antibody observations, the frequency of S-2P+ IgG+ B cells in PCV1 120

was 3-8-fold higher than those in the other patients, while no major differences were observed in the frequencies of RBD+ IgG+ B cells among the four patients. As expected, the frequency of S-2P+ cells in a 122 healthy (pre-pandemic) control individual, CN, was lower than those found in the four patients (0.104% and 0.128%), as were the frequency of RBD+ IgG+ B cells (first sort: 0.015% and second sort: 0.019%). A total of 124 341 HC, 353 LCs and 303 LCs were successfully sequenced from the four SARS-CoV-2-positive donors (Table   S1 , Figure S1 ), from which 228 paired HC/LCs were generated, and 198 antibodies were successfully produced 126 and characterized. We isolate 59 paired mAb sequences from the healthy individuals and then successfully generate 36 mAbs. As discussed above we performed an initial characterization of the 48 mAbs from CV1 128 , here we performed a more in-depth characterization of these mAbs.

In agreement with previous reports, the antibodies isolated from the patients utilized diverse V regions (Cao 130 et al., 2020b; Nielsen et al., 2020; Robbiani et al., 2020; Seydoux et al., 2020) (Figure 2A -C, Figure S1 ).

Similarly, the S-specific mAbs isolated from the healthy donor originate from diverse V regions. To determine 132

whether anti-S-2P+ B cells that express certain VH and VL genes preferentially expand during infection, we compared the relative frequencies of each VH and VL sequence to those present in healthy individuals. For 134 this, we performed a 10x-based sequence analysis of total circulating B cells (i.e., not S-2P specific) from five SARS-CoV-2-unexposed adults ( Figure 2D -F, Figure S2 ). Significantly higher frequencies of S-2P+ IGHV3-30 and 136 IGHV1-18 antibody sequences were observed in the patients as compared to the relative frequencies of these two genes present in healthy adults ( Figure 2D ). Interestingly, lower frequencies of S-2P+ IGHV3-33 usage was 138 observed in the patients than in healthy donors. Differences were also observed in kappa ( Figure 2E ) and lambda ( Figure 2F ) gene usage between patients and healthy donors. Specifically, IGKV3-15, IGKV1-33/1D-33 140

and IGKV1-17 were significantly elevated in patients as compared to healthy donors while the expression of IGKV1-39/1D-39 was reduced. IGLV1-51 was significantly elevated in the patients as compared to healthy 142 donors, as was IGLV2-23, though this appears to be driven by a greatly elevated usage in patient CV3.

The above observations suggest that naïve B cell clones expressing the above IGHV, IGKV or IGLV genes 144 preferentially recognize the viral S protein at the initial stages of infection. To address this point IgD+ IgM+ S-2P+ and RBD+ B cells were isolated from a healthy donor, CN (following two independent B cell sorting 146 experiments from this donor) (Table S1) , their V genes sequenced, and their relative frequencies were again compared to those found in total B cells from healthy donors (Figure 2 G) . Although IGHV3-30 was present in 148

higher frequency in B cells sorted with S-2P from CN than in the total B cell population, the difference was not as large as in the infected patients. Similarly, no differences were observed for the other IGHV, no instances of the IGKV or IGLV genes that were predominant in the anti-S response after infection appeared in CN. Thus, it appears that the anti-spike B cell response that predominates at 3-7 weeks post infection is 152 dissimilar from the naïve B cells that preferentially bind to S-2P.

The length distribution for the CDRH3 and CDRL3 of antibodies were comparable to those present in the pre-154

infection, healthy B cell repertoires ( Figure 2H , I). Interestingly, the IGHVs and IGLVs sequences derived from samples collected at 6 (CV3) and 7 (PCV1) weeks after infection had significantly more amino acid mutations 156 than those derived from samples collected at 3 (CV1) or 3.5 (CV2) weeks after symptom-development, or than the CN mAbs ( Figure 2J, K) . These observations are suggestive of a continuous B cell evolution during SARS-158

CoV-2 infection as others recently reported (Gaebler et al., 2021) .

The binding specificities of the 198 monoclonal antibodies (mAbs) to S-subdomains were determined using recombinant proteins including, S1, RBD, N-terminal domain (NTD) and S2 ectodomain (ECD) monomer 162 subunits ( Figure 3A , Figure S3 A, B) . Only a small percentage of mAbs bound RBD, irrespective of the time of B cell isolation following the development of symptoms. However, the relative proportion of anti-S2 antibodies 164 was higher in samples collected at 3 and 3.5 weeks (51% in CV1 and 70% in CV2, respectively) than in samples collected 6-and 7-weeks post symptom onset (35% in CV3 and 27% in PCV1, respectively). PCV1 had a high 166

proportion (32%) of antibodies whose epitopes could not be mapped to S1 or S2, while such antibodies were rarer in the other three patients examined here (15% in CV1, 7% in CV2 and 0% in CV3). Out of the 36 mAbs 168 produced from healthy individuals, 27 (75%) bound S2P and of these 40.7% could not be mapped to S1 or S2

binding.

We also determined the abilities of these antibodies to recognize SARS-CoV-1, MERS, the two endemic human beta coronaviruses, OC43 and HKU1, and the two endemic human alpha coronaviruses, NL63 and 229E ( Figure  172 3B). 81 mAbs (41%) displayed SARS-CoV-1 reactivity (to varying degrees), approximately half of which recognized the SARS-CoV-1 RBD. In contrast, only 4 mAbs (2.3%) displayed cross reactivity towards MERS (and 174 J o u r n a l P r e -p r o o f none to the MERS RBD), 13 bound OC43 (7.1%), 12 bound HKU1 (6.6%), 2 bound NL63 (1.1%) and only 1 bound 229E (0.56%). Of these cross-reactive mAbs, the majority mapped to S2-binding ( Figure 3C ) and most 176 bound only one coronavirus type beyond SARS-CoV-2 ( Figure S3C ). There was no association between the number of amino acid mutations in the antibody V regions and cross-reactivity with divergent HCoVs ( Figure  178 S3D-J).

Only 14 mAbs (7%) neutralized SARS-CoV-2 ( Figure 4A) , with IC 50 s ranging from 0.007 g/ml to 15.1 g/ml (although, as we discuss below, we were unable to assign an IC50 to CV2-74) ( Figure 4B , Figure 5A , Table S2 , 182 Figure S4 ). 11 of 14 neutralizing mAbs bound RBD, in agreement with our and other reports that RBD is the major target of anti-SARS-CoV-2 nAbs (Barnes et al., 2020; Cao et al., 2020b; Ju et al., 184 2020; Liu et al., 2020; Rogers et al., 2020a) . Three of the nAbs, CV1-1 (from patient CV1), CV2-74 (from patient CV2) and CV3-25 (from patient CV3) bound epitopes outside the RBD. CV1-1 binds the S1 NTD, CV3-25 binds 186 the S2 subunit, CV2-74 bound neither the recombinant S1 or S2 proteins used here, and we were unable to define its specificity ( Figure 5B and Figure S4A , B).

The three most potent nAbs, all anti-RBD, were CV1-30 (IC 50 =0.044 g/ml) , CV3-1 (IC 50 =0.007 g/ml) and PCV19 (IC 50 =0.072 g/ml). The anti-NTD mAb (CV1-1) had lower neutralizing potency 190

(IC 50 =8.2 g/ml) and as we previously reported , its maximum level of neutralization was lower than 100% ( Figure S4C ), similar to other anti-NTD mAbs . CV1-1 displayed decreased 192

binding against more stable SARS-2-CoV S engineered proteins (S-6P), as shown by lower overall unit response and faster off-rate by BLI and does not bind like other published NTD-targeting antibodies by negative stain 194 EM ( Figure S5A ,B) . The IC 50 of anti-S2 mAb, CV3-25, was 0.34 g/ml, which is comparable to most anti-RBD nAbs with the exception of CV1-30, CV3-1 and PCV19.

Out of the 14 nAbs, seven (CV2-20, CV2-71, CV2-75, CV3-7, CV3-17, CV3-25 and CV3-43) also bound the S-2P of SARS-CoV-1 and four of the seven neutralized this virus ( Figure 4C , Figure S4D and Table S2 ). Three were 198

anti-RBD (CV2-71, CV2-75 and CV3-17), while the fourth, CV3-25, bound to S2 ( Figure 5B , Table S2 ).

Interestingly, while the IC 50 s of CV2-71, CV2-75 and CV3-25 against SARS-CoV-1 and SARS-CoV-2 were not 200 significantly different, CV3-17 neutralized SARS-CoV-1 more potently than SARS-CoV-2 ( Figure S4E ).

Furthermore, the two most potent anti-SARS-CoV-2 mAbs (CV1-30 and CV3-1) did not neutralize SARS-CoV-1.

CV1-30 has only two non-silent somatic mutations (both in VH) that we previously reported are important for 204 potent neutralization of SARS-CoV-2 . To examine if this is a general phenomenon among anti-RBD SARS-CoV-2 nAbs, we generated the inferred-germline (iGL) versions of 6 anti-RBD Abs (CV2-20, CV2-206 71, CV2-75, CV3-1, CV3-7, and CV3-43) and measured their neutralizing potencies ( Figure 4D and Fig S6A) .

Three of six anti-RBD iGL-nAbs, CV2-20 (3 amino acid mutations), CV2-75 (3 amino acid mutations) and CV3-43 208

(9 amino acid mutations) were non-neutralizing. However, no differences in neutralizing potency between the mutated and iGL-CV2-71 (3 amino acid mutations), iGL-CV3-1 (2 amino acid mutations) and iGL-CV3-7 (9 210 amino acid mutations) were observed. Reductions in neutralizing potency of the iGL mAbs correlated with faster dissociation rates from RBD ( Figure S6B ). The anti-NTD mAb CV1-1 has no amino acid mutations in its V 212 genes while the anti-S2 Ab CV3-25 has 5 mutations. Reversion of the anti-S2 mAb CV3-25 to its germline form also led to a significant reduction in its neutralizing potency. Thus, some anti-SARS-CoV-2 nAbs are capable of 214 potent neutralization in the absence of affinity maturation, while the neutralizing activity of others depends on the accumulation of a small number of mutations. Overall, however, there was no correlation between the 216 neutralization potency and the degree of SHM (data not shown).

We next examined whether the differences in neutralizing potencies of the anti-RBD nAbs ( Figure 4B ) were due to differences in their relative abilities to block the RBD-ACE2 interaction ( Figure 4E , Figure S4F ). While 220 CV2-71, CV2-75 and CV3-1 abolished ACE2 binding to RBD, suggesting that they either directly bound the receptor binding motif (RBM) like CV1-30 , or indirectly (sterically) hindered this binding, 222 the remaining 7 anti-RBD NAbs (CV2-20, CV2-66, CV3-7, CV3-17, CV3-43, CV3-45 and PCV19) did not inhibit the RBD-ACE2 interaction. Similar observations were made when the abilities of mAbs to block the interaction 224 of recombinant S-2P to cells expressing ACE2 were examined ( Figure S4G ). Indeed, a correlation between the potency of neutralization and the extent to which a mAb blocked the RBD-ACE2 interaction was observed 226

( Figure 4F and Figure S4H ) in agreement with previous reports (Baum et al., 2020a; Brouwer et al., 2020; Gavor et al., 2020; Hoffmann et al., 2020; Wan et al., 2020a; Yan et al., 2020) . 228

As mentioned above, three of the anti-RBD mAbs (CV2-71, CV2-75, and CV3-17) also neutralized SARS-CoV-1.

The abilities of these antibodies to block the ACE2 interaction with the SARS-CoV-1 RBD were similar to their 230 abilities to block the interaction of ACE2 with the SARS-CoV-2 RBD, with CV2-71 and CV2-75 blocking ACE2 SARS-CoV-1 RBD . In contracts, CV2-75 binds the RBD at an epitope distinct from the receptor binding motif (RBM) ( Figure S5C , Table S3 ) and is only accessible when the RBD is in the up 236

conformation. The residues that CV2-75 interacts with on the RBD are nearly completely conserved between SARS-CoV-1 and -2 explaining the cross-neutralizing ability ( Figure S5D ). An alignment with the structure of 238 ACE2-RBD, showed that the heavy chain of CV2-75 would clash with the glycan at Asn322 in ACE2, establishing a mechanism of competition ( Figure S5E ).

As mentioned above CV2-74 binds to an undefined epitope on S, that is present on S-2P but absent or not 242

properly presented on the recombinant S1 or S2 proteins used here ( Figure 5B ). We identified several mAbs sharing this binding property (especially in PCV1) and the majority (75%) of these mAbs did compete the 244 binding of CV2-74 to S-2P ( Figure 5D , E). The fact that among these mAbs only CV2-74 displayed neutralizing activity suggests that either the other mAbs bind distinct epitopes on S-2P and indirectly affect the binding of 246

CV2-74 to S-2P or that CV2-74 binds a unique but overlapping epitope. It is note-worthy that CV2-74 displays an unusual neutralization curve, where the mAb neutralizes only 50% of the virus across a thousand-fold 248 concentration range ( Figure 5A ). For that reason, we did not assign an IC50 value to CV2-74.

Out of the 14 anti-NTD mAbs we identified, 8 (57%) competed the binding of CV1-1 to S-2P ( Figure 5D , E) and 250

yet, CV-1-1 was the only neutralizing anti-NTD mAb ( Figure 4B ). Interestingly, CV1-1 displayed decreased binding to more stable SARS-CoV-2 engineered soluble proteins ( Figure S5 ). While BLI revealed binding of CV1-252 1 to recombinant NTD, the on-rate and maximal binding signal was lower than to the entire S1 domain, suggesting that secondary (or quaternary) contacts are important ( Figure 5B ). Indeed, negative-stain EM 254 analysis indicates that it recognizes NTD differently than other anti-NTD mAbs (such as COVA1-22 (Brouwer et al., 2020) , with a footprint that might also include an area just above the S1/S2 cleavage site ( Figure S5B ).

Out of 87 anti-S2 mAbs, CV3-25 was the only one capable of neutralizing SARS-CoV-2 and SARS-CoV-1 ( Figure   4B and C) and of binding the S proteins of the OC43 and HKU1 betacoronaviruses ( Figure 5C , Table S2 ). As 258 none of the other 86 anti-S2 mAbs competed the binding of CV3-25 to S2-P ( Figure 5D ,E) we expect that CV3-25 binds a unique epitope on the S2 subunit which is present not only on SARS-CoV-1 but also on the other 260 coronaviruses tested here. 2020). As discussed above, CV1-1 binds NTD and has an IC50 of 8.2 g/ml, while CV1-30 and CV2-75 bind the RBD and have IC50s of 0.044 and 1.7 g/ml, respectively. Thus, CV1-1 and CV2-75 have neutralizing potentials 266 in a similar range but recognize different regions of the viral spike.

Mice were given a dose of 10 mg/kg of CV1-1, CV2-75, CV1-30, or an isotype control anti-EBV antibody 268 AMMO1 (Snijder et al., 2018) , and then challenged intranasally with 10,000 plaque forming units (PFU) of SARS-CoV-2 ( Figure 6A ). Two days post challenge, half of the animals were euthanized to assess viral loads in 270 the lung, and the remaining 5 animals were monitored for survival for up to 14 days. Two days post-challenge, mice receiving AMMO1, CV1-1 and CV2-75 had high levels (1 x 10 8 PFU) of infectious virus and viral RNA in the 272 lung ( Figure 6B and C). 3 of 5 remaining animals in the CV1-1-and CV2-75 groups did not survive beyond 6 days post-challenge ( Figure 6C ,D). In contrast, CV1-30 significantly limited viral replication in the lungs at 2 274 days post-challenge ( Figure 6B ,C) and all remaining mice survived ( Figure 6D ).

Collectively these data suggest that both neutralizing potency and epitope specificity are the most influential 276 factors in defining the prophylactic efficacy of anti-SARS-CoV-2 antibodies.

Recently We recently reported that these mutations abrogated the neutralizing activity of CV1-1 and reduced the neutralizing activities of the two most potent nAbs CV1-30 and CV3-1, but not of CV2-75 (Stamatatos et al., 288 2021b). Here we evaluated the ability of the 4 cross-neutralizing mAbs (CV2-75, CV3-17, CV2-71 and CV3-25)

to neutralize the B.1.351242-243 mutant strain ( Figure 7B ) (Stamatatos et al., 2021b) . We found that all four 290 mAbs retained their neutralizing activities against B.1.351.

Our study reveals that naïve B cells expressing VH3-30 and VH1-18 preferentially recognize the SARS-CoV-2 envelope spike, but that nAbs are produced by B cells expressing diverse BCRs. Of the 198 mAbs characterized 294 J o u r n a l P r e -p r o o f here (isolated at 3-7 weeks post-symptom development), 14 (7%) displayed neutralizing activities and among them, only CV3-7 was derived from VH3-30. In fact, the 11 anti-RBD nAbs were derived from distinct B cell 296 clones, that cross-competed for binding, and 4 prevented the RBD-ACE2 interaction. These observations, combined with the fact that anti-RBD nAbs can neutralize the virus with no, or minimal somatic mutation, may 298 explain why potent anti-SARS-CoV-2 neutralizing antibody responses are rapidly generated within a few weeks of infection, or shortly following 2 immunizations with vaccines that express the viral spike (Jackson et al., 300 2020; Walsh et al., 2020) . The observation that 7 of 11 anti-RBD nAbs do not prevent the RBD-ACE2

interaction, indicates different mechanisms of neutralization by anti-RBD antibodies. The former nAbs may 302 prevent RBD-heparin interactions (Clausen et al., 2020) , stabilize the RBDs in their 'up' conformation and thus prematurely activate the fusion machinery (Koenig et al., 2021; Wrapp et al., 2020a) , or limit the 304 conformational changes, and particularly the RBD movement, that are required for cell fusion, allowing them to neutralize without directly blocking ACE2 binding. 306

The two most potent anti-SARS-CoV-2 nAbs, CV1-30 and CV3-1, which both bind SARS-CoV-2 RBD but not SARS-CoV-1 RBD, did not neutralize SARS-CoV-1 while CV2-75 and CV3-17, which bind not only SARS-CoV-2 308

RBD but also SARS-CoV-1 RBD and displayed weaker anti-SARS-CoV-2 neutralizing activities, were able to efficiently neutralize SARS-CoV-1. A comparison of the CV2-75-RBD and CV1-30-RBD (Hurlburt et al., 2020) 310 structures reveal that CV2-75 binds an area of SARS-CoV-2 RBD with higher sequence homology with SARS-CoV-1 RBD. In contrast, CV1-30 binds directly to the receptor binding motif which only has 50% sequence 312 homology among SARS-CoV-1 and SARS-CoV-2 (Finkelstein et al., 2021; Hurlburt et al., 2020; Wan et al., 2020b) . 314

The mechanisms of neutralization of the three non-RBD binding nAbs characterized here (CV1-1, CV2-74 and CV3-25) are presently unknown. As CV1-1, CV2-74 and CV3-25 do not interfere with the binding of ACE-2 to S-316 2P, we anticipate that they mediate neutralization by interfering with a step in the fusion process that follows attachment. The viral spike undergoes conformational changes, specifically in the S2 region, during virus-cell 318 binding and fusion (Cai et al., 2020; Gavor et al., 2020; Walls et al., 2020) . Potentially, these three mAbs may prevent these conformational changes from occurring, either by locking the spike in an intermediate 320 conformation, preventing cleavage or by stabilizing its pre-fusion conformation.

The fact that the binding of CV1-1 to S-2P was competed by the other anti-NTD mAbs (14 total), all of which 322

were non-neutralizing, suggests that CV1-1 recognizes the NTD in a distinct manner from the non-neutralizing anti-NTD mAbs. Similarly, the binding of CV2-74 to S-2P was competed by the other non-neutralizing non-324 S1/S2 mAbs (23 total), which strongly suggests that these mAbs all recognize the same immunogenic region, but that CV2-74 recognizes it in a unique manner. In contrast, none of the anti-S2 mAbs isolated here (65 326 total) competed the binding of CV3-25 to S-2P. These observations and the fact that CV3-25 potently neutralizes both SARS-CoV-1 (IC 50 2.1 g/mL) and SARS-CoV-2 (IC 50 0.34 g/mL) and the B.1.351 mutant strain 328

and binds the S proteins of HKU1 and OC43, strongly suggests that it recognizes a conserved epitope among diverse coronaviruses. As only two other anti-S2 antibodies that neutralize both SARS-CoV-1 and SARS-CoV-2, 330 but with weaker neutralizing activities than CV3-25, were reported so far Wang et al., 2020b) we expect the epitope of CV3-25 to be less immunogenic than those recognize by non-neutralizing 332

anti-S2 antibodies.

We propose that because of its cross-neutralizing activity, its ability to neutralize the B.1.351 and because it 334

binds the OC43 and HKU1 spikes, CV3-25 is a potential starting point for developing a pan-coronavirus vaccine.

We expect that the protective potentials of antibodies that bind the same region of S2 as CV3-25, could be 336

improved through the accumulation of amino acid mutations in their VH and VLs by sequential immunizations.

As a first step, the epitope of CV3-25 must be identified, and immunogens should be designed expressing it in 338 the most immunogenic form.

In summary, our study indicates that neutralization of SARS-CoV-2 and SARS-CoV-1 does not necessitate the 340 expansion of B cell lineages that express particular VH/VL pairings and that even the unmutated forms of some antibodies can potently neutralize SARS-CoV-2 and SARS-CoV-1. As these viruses are capable of tolerating 342 mutations in distinct regions of its viral spike, they will be able to escape the neutralizing activities of most We also would like to thank L. Kehoe, S. P. Canny, K. Nanda and J. Czartoski for the care and enrollment of 362 patients. One or more of the authors of this paper self-identifies as a member of the LGBTQ+ community. Serum from four patients infected with SARS-CoV-2 (Table S1) Tukey's multiple comparison test *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. 408 Figure 3 : Epitope-specificities and cross-reactivity of SARS-CoV-2 antibodies.

The percentage of mAbs from each donor specific for the SARS-CoV-2 spike subdomains and their cross-410 reactivity was determined by BLI. (A) Using S1 and S2 monomer proteins, mAbs were grouped into the antibodies that bound RBD in the S1 subunit (S1: RBD, blue), mAbs that bound S1 outside of RBD (S1: non-412 J o u r n a l P r e -p r o o f RBD, teal), mAbs that bound the S2 ECD (S2 ECD, yellow) or those that bound S2P but did not bind either S1 or S2 (S2P: Non-S1/Non-S2. (B) The percentage of mAbs that bind to SARS-CoV-1, MERS and the four common 414

human coronavirus was also measured by BLI using S2-P timers for SARS-CoV1 and MERS and S1+S2 monomers for the four human coronavirus antigens. (C). The percentage of mAbs that bound each subdomain 416 of the coronavirus spike for the mAbs cross reactive with SARS-CoV-1, MERS S-2P and the four common human coronaviruses. Only mAbs isolated from the four SARS-CoV-2 infected donors are included. Significant 418 differences were determined using two-way ANOVA with Tukey's multiple comparison test, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Additional BLI data and comparison to number of amino acid mutations 420

in Figure 3S . (G) The competition of mAbs for binding to SARS-CoV-1 RBD with ACE2 is compared on this graph performed as in E. Full ACE2-competition data in Figure S4F -J. Additional characterization of CV1-1 ad CV2-75 in Figure  440 S6. indicates what is considered true competition, dots below the line are considered competitive. For CV1-1, S1

NTD mAbs from CV1, CV2 and CV3 were tested. For CV2-74, all non-S1/S2 mAb in all four sorts were tested. 450

For CV3 

Resource availability

Further information and requests for resources and reagents should be directed to and will be fulfilled by the CoV-1 RBD fused to a monomeric Fc (pαH-SARS-CoV RBD-Fc) and MERS RBD (pαH-MERS RBD-Fc) fused to a monomeric Fc have been previously described and were a kind gift from Dr. Jason McLellan (Hsieh et al., 2020; 514 Pallesen et al., 2017; Wrapp et al., 2020b) .

Proteins were produced as described in . Briefly, 1L of 293 EBNA cells at 1 x 10 6 cells/mL 516 were transfected with 500 mg of pαH-SARS-CoV-2 S2P, pαH-SARS-CoV S2P, pαH-SARS-CoV-2 RBD-Fc, pαH-SARS-CoV RBD-Fc, pαH-MERS S2P, pαH-MERS RBD-Fc using 2 mg of polyethyleneimine. After 6 days of growth, 518

supernatants were harvested and filtered through a 0.22 mM filter. S2P supernatants were passed over a HisTrap FF affinity column and further purified using a 2 mL StrepTactin sepharose column and a Strep-Tactin 520

Purification Buffer Set. The S-2P variants were further purified using a Superose 6 10/300 GL column. RBD proteins were purified using protein A agarose resin, followed by on-column cleavage with HRV3C protease to 522 release the RBD from the Fc domain. The RBD containing flow through was further purified by SEC using a HiLoad 16/600 Superdex 200 pg column. Proteins were flash frozen and stored at -80°C until use. 524

HCoV-OC43, HCoV-HKU1, HCoV-NL63 and HCoV-229E S1+S2 ECTs, SARS-HCoV-2 S1 domain, SARS-CoV-2 S1 Nterminal domain, SARS-HCoV-2 S2 extra-cellular domain (CAT#: 40590-V08B) and SARS-CoV-1 RBD were 526 purchased from SinoBiologicals.

J o u r n a l P r e -p r o o f S-2P and RBD were coated onto 384-well nunclon plates at 0.5 mg/mL in 30 ml overnight at 4C. Plates were washed with PBS 0.02% Tween (wash buffer) using a Biotek 405 select plate washer and then blocked in 100 530 mL of 10% milk, 0.02% Tween (Blocking/Dilution buffer) for 1 hour at 37C. Plates were washed again, and sera was loaded at a starting dilution of 1:50 with 11 serial 1:3 dilutions in dilution buffer in a total volume of 30 532 mL. After another hour at 37C, plates were washed again, and IgG, IgA or IgM was detected with 30 mL of HRP secondary (Goat anti-human IgG HRP, Goat anti-human IgA HRP, Goat anti-human IgM HRP) at a 1:3000 534 dilution for 1 hour at 37C. After the last wash, plates were developed with 30 mL SureBlue TMB Microwell Peroxidase Substrate. The reaction was quenched with 30 mL of 1N sulfuric acid. Plates were read on a 536

SpectraMax M2 plate reader at 450 nM.

B cell sorting B cell sorting was performed as described in . Briefly, fluorescent probes were made 540 from SARS-CoV-2 S-2P and RBD. S-2P and RBD were biotinylated protein at a theoretical 1:1 ration using the Easylink NHS-biotin kit according to manufacturer's instructions. Excess biotin was removed via size exclusion 542 chromatography using an Enrich SEC 650 10 x 300 mm column. The S-2P probes were made at a ratio of 2 moles of trimer to 1 mole streptavidin, one labeled with streptavidin-phycoerythrin (PE), and one with (Table  554 S1). Cells were sorted into 96-well plates containing 16 μl lysis buffer (3.90% IGEPAL, 7.81 mM DTT, 1250 units/ml RNase Out).

PCR amplification and sequencing of VH and VL genes RNA was reverse transcribed to cDNA by adding 4 ul of iScript to sorted B-cells and cycling according to the 558 manufacturer's instructions. VH and VL genes were amplified using two rounds of PCR as previously described (Tiller et al., 2008) . First round reactions contained 5 ul cDNA, 1-unit HotStarTaq Plus, 190 nM 3' primer pool, 560 290 nM 5' primer pool, 300 M GeneAmp dNTP Blend, 2 ul 10x buffer, and 12.4 ul nuclease-free H2O. Second round PCR reactions used 5 ul first round PCR as template and 190 nM of both 5' and 3' primers. Second round 562 PCR products were subjected to electrophoresis on a 1.5% agarose gel containing 0.1% Gel Red Nucleic Acid

Stain. Positive wells were then purified using either ExoSAP-IT following manufacturer's instructions or using a 564 homemade enzyme mix of 0.5 units Exonuclease I, 0.25 units of rAPid Alkaline Phosphatase, and 9.725 ul 1x PCR buffer mixed with 5 ul of second round PCR product and cycled for 30 minutes at 37C followed by 5 566 minutes at 95C. Purified samples were Sanger sequenced. IMGT/V-QUEST was used to assign V, D, J gene identity, and CDRL3 length to the sequences (Brochet et al., 2008) . Sequences were included in analysis if V 568

and J gene identity could be assigned and the CDR3 was in-frame.

For sorts CV1, CV2 and CN, paired VH and VL sequences were optimized for human expression using the 572

Integrated DNA Technologies (IDT) codon optimization tool. Sequences were ordered as eBlocks (IDT) and

cloned into full-length pTT3 derived IgL and IgK expression vectors (Snijder et al., 2018) or subcloned into the 574 pT4-341 HC vector (Mouquet et al., 2010) using InFusion cloning.

Sorts CV3 and PCV1 were directly cloned using Gibson Assembly. Second round PCR primers were adapted to 576

include homology regions that corresponding to the leader sequence and constant regions on the expression vector. Cycling parameters and post-PCR clean-up remained the same. The backbone expression plasmid was 578 amplified using primers specific for the leader sequence and constant regions in 25 l reactions containing 2x

Platinum SuperFi II DNA polymerase, 100 nM 5' and 3' primers, 10 ng template DNA, and 21.5 l Nuclease-free 580

water. The reaction was cycled at 98C for 30 seconds, 30 cycles of 98C for 10 seconds, 60C for 10 seconds, and 72C for 3 minutes and 30 seconds, followed by 72C for 5 minutes. The reaction was treated with 20 units of 582 dpnI and incubated at 37C for 60 minutes. The reaction was purified using a PCR clean-up kit according to manufacturer's directions or using. The cloning reaction was performed using 100 ng of second round PCR 584 product, 25 ng of backbone, 1 l 5x InFusion HD Enzyme and nuclease-free water for a total reaction volume of 3 ul and incubated at 50C for 15 minutes. 586

The cloning reactions were used to transform OneShot DH5 Alpha cells ( according to manufacturer's directions and plated on agar plates containing ampicillin and grown overnight. Colonies were used to seed 5 588 mL LB broth cultures containing ampicillin. DNA was prepared using QIAprep Spin Miniprep Kit. Equal amounts of heavy and light chain expression plasmids and a 1:3 ratio of PEI was used to transfect 293-6E cells at a 590 J o u r n a l P r e -p r o o f density of 1x10^6 cells/mL in Freestyle 293 media. Supernatants were collected 6 days post transfection by centrifugation at 4,000g followed by filtration through a 0.22 M filter. Clarified supernatants were then 592 incubated with Protein A agarose beads overnight followed by extensive washing with 1x PBS. Antibodies were eluted using 0.1M Citric Acid into a tube containing 1M Tris then buffer exchange into 1xPBS using an 594

Amicon centrifugal filter. We recently reported an initial characterization of the anti-S antibody responses generated by CV1 .

10X sequencing:

PBMCs were thawed in a 37C water bath with pre-warmed RPMI + 10% cells were analyzed per donor yielding 5,000-7,000 clonotypes each. Fred Hutch Genomics core performed the sequencing and the Fred Hutch Bioinformatics core performed processing of the raw sequence data.

BLI 608

All BLI experiments were performed on an Octet Red instrument at 30°C with shaking at 500-1000 rpm. All loading steps were 300s, followed by a 60s baseline in KB buffer (1X PBS, 0.01% Tween 20, 001% BSA, and 610 0.005% NaN 3 , pH 7.4), and then a 300s association phase and a 300s dissociation phase in KB. For the binding BLI experiments, mAbs were loaded at a concentration of 20 mg/mL in PBS onto Anti-Human IgG Fc capture 612

(AHC) biosensors. After baseline, probes were dipped in either SARS-CoV2 proteins; SARS-CoV-2 RBD, S-2P, S1, S1 NTD orS2; SARS-CoV proteins; SARS-CoV-RBD or S-2P, or human coronavirus spike proteins; HCoV2-OC43, 614 HKU1, NL63 or 229, at a concentration of 2-0.5 mM for the association phase. The binding of mature VRC01 was used as negative control to subtract the baseline binding in all of these experiments.

ACE2 competition BLI 618

To measure competition between mAb and RBD for ACE2 binding, ACE2-Fc was biotinylated with EZ-Link NHS-PEG4-Biotin at a molar ratio of 1:2. Biotinylated protein was purified using a Zeba spin desalting column. ACE2-620 J o u r n a l P r e -p r o o f

Fc was then diluted to 20-83.3 mg/mL in PBS and loaded onto streptavidin biosensors. Following the baseline phase, association was recorded by dipping into a 0.5 mM solution of either SARS-CoV-2 RBD or 0.5 mM 622

solution of SARS-CoV-2 RBD plus mAb. The binding of RBD and mAb to uncoated sensors was used as background binding and was subtracted from each sample. The area under the curve (AUC) of competition 624 was compared to the AUC of the RBD-alone condition. Samples that showed reduced binding are considered competition. Some samples appear to show enhanced binding in the presence of ACE2, perhaps because ACE2 626

binding stabilizes and exposes their binding sites, these antibodies are considered not competitive with ACE2.

To measure competition between individual mAbs for binding to SARS-CoV-2 S-2P and RBD, S-2P and RBD 630

were biotinylated using EZ-Link NHS-PEG4 Biotin at a molar ratio of 1:2/ Biotinylated protein was purified using a Zeba spin desalting column. RBD was loaded onto streptavidin biosensors. For these experiments, 632

following the baseline in KB, the probe was dipped in the first mAb for a first association phase, with this mAb at a saturating concentration of 2 mM. This was followed by a second baseline in KB. The probe was then 634 dipped into the secondary mAb, at a concentration of 0.5 mM for a second association phase, followed by the standard dissociation phase. For a background control, one sample was run with the second mAb identical to 636 the first mAb, to show the residual binding capacity, and this was subtracted from all samples.

To calculate the competition percentage, the binding of the secondary antibodies to RBD or S-2P was also 638 assessed. Here, streptavidin probes were loaded with biotinylated S-2P or RBD, probes were then dipped in the secondary antibody at 0.5 mM for the association phase, before dissociation in KB as normal. As a 640 background control, the binding of mature VRC01 to the RBD or S-2P was assessed and subtracted from all samples. To calculate competition percentage, the area under the curve (AUC) of this binding curve was 642 calculated, along with the AUC of the competition curve. Percent competition was calculated as:

( ÷ ) × 100. Full competition was considered when less than 15% 644 binding capacity remained.

HIV-1 derived viral particles were pseudotyped with full length wild-type SARS CoV-2 S, SARS CoV-2 S or SARS-648

CoV-2 B.1.351 S242-243 (Crawford et al., 2020; Seydoux et al., 2020; Stamatatos et al., 2021b) . The B.1.351242-243 SARS-CoV-2 variant was produced as described previously (Stamatatos et al., 2021b) with 650 J o u r n a l P r e -p r o o f the D80A, D215G, K417N, E484K, N501Y, D614G and A701V mutations. Briefly, plasmids expressing the HIV-1

Gag and pol (pHDM-Hgpm2), HIV-1Rev (pRC-CMV-rev1b), HIV-1 Tat (pHDM-tat1b), the SARS CoV2 spike 652

(pHDM-SARS-CoV-2 Spike) and a luciferase/GFP reporter (pHAGE-CMV-Luc2-IRES-ZsGreen-W) were cotransfected into 293T cells at a 1:1 that neutralized 50% of infectivity (IC 50 ) , or serum dilution that neutralized 50% infectivity (ID 50 ) was interpolated from the neutralization curves determined using the log(-inhibitor) versus response-variable 672 slope (four parameters) fit using automatic outlier detection in GraphPad Prism software.

The neutralizing activities of CV1-1 and CV1-30 mAbs were also determined with a slightly different 674

pseudovirus-based neutralization assay as previously described (Bottcher et al., 2006; Naldini et al., 1996) .

Monitoring RBD-binding to 293-ACE2 cells by flow cytometry 8 pmol of biotinylated S-2P with strep tag peptide sequence on C terminus were mixed with 10 pmol of mAb 678

and incubated for 10 min at RT in a round-bottom tissue culture 96-well plate. 200,000 HEK293T-hACE2 cells in 50 µL of cDMEM were then added to each well and the mixture of cells + RBD or S-2P + mAb was incubated 680 J o u r n a l P r e -p r o o f for 20 min on ice. Samples were washed once with ice-cold FACS buffer (PBS + 2% FBS + 1 mM EDTA), before staining cells with DY-549-labeled strep-tactin (1:100 dilution) or Allphycocyanin-labeled streptavidin (1:200 682 dilution). Cells were washed once with FACS buffer, fixed with 10% formalin for 15 min on ice in the dark, and resuspended in 200 μl of FACS buffer to be analyzed by flow cytometry using a LSRII. Control wells were 684 included on each plate and either had no mAb, no RBD or no S-2P, or were unstained. The mean fluorescence intensity (MFI) for each sample was determined and each sample was normalized to the MFI of the no mAb 686

control.

Antigen binding fragment (Fab) was generated by incubating IgG with LysC at a ratio of 1 μg LysC per 10mg 690

IgG at 37°C for 18hrs. Fab was isolated by incubating cleavage product with Protein A resin for 1hr at RT.

Supernatant containing Fab was collected and further purified by SEC.

Crystal Screening and Structure Determination 694

The CV2-75 Fab and SARS-CoV-2 RBD complex was obtained my mixing Fab with a 2-fold molar excess of RBD and incubated for 90 min at RT with nutation followed by SEC. The complex was verified by SDS-PAGE analysis.

The complex was concentrated to 19 mg/mL for initial crystal screening by sitting-drop vapor-diffusion in the MCSG Suite using a NT8 drop setter. Initial crystal conditions were optimized using the Additive Screen.

Diffracting crystals were obtained in a mother liquor (ML) containing 0.1 M Tris, pH 7.5, 0.1 M Calcium Acetate, 15% (w/v) PEG 3350, and 4mM glutathione. The crystals were cryoprotected by soaking in ML 700

supplemented with 30% (v/v) ethylene glycol. Diffraction data was collected at Advanced Photon Source (APS) SBC 19-ID at a 12.662 keV. The data set was processed using XDS (Kabsch, 2010) to a resolution of 2.80Å. The 702

structure of the complex was solved by molecular replacement using Phaser (McCoy et al., 2007 ) with a search model of SARS-CoV-2 RBD (PDBid: 6xe1) and the Fab structure (PDBid: 4fqq) (Mouquet 704 et al., 2012) divided into Fv and Fc portions. Remaining model building was completed using COOT (Emsley and Cowtan, 2004) and refinement was performed in Phenix (Adams et al., 2004) . The data collection and 706 refinement statistics are summarized in Table S3 . Structural figures were made in Pymol.

400 mesh copper grids. For data collection, a Thermo Fisher Tecnai Spirit (120 kV) and an FEI Eagle (4k x4k) 712 CCD camera were used to produce 296 raw micrographs. Leginon (Suloway et al., 2005) was used for automated data collection and resulting micrographs were stored in Appion (Lander et al., 2009) . Particles 714

were picked with DogPicker (Voss et al., 2009 ) and processed in RELION 3.0 (Scheres, 2012) .

Infection of k18-hACE2 mice with SARS-CoV-2

icSARS-CoV-2 virus (Xie et al., 2020) was diluted in PBS to a working concentration of 2 x 10 5 pfu/mL. Mice 718

were anesthetized with isoflurane and infected intranasally with icSARS-CoV-2 (50 uL, 1 x 10 4 pfu/ mouse) in a ABSL-3 facility. Mice were monitored daily for weight loss. At the indicated day post infection, mice were 720 euthanized via isoflurane overdose and lung tissue was collected in Omni-Bead ruptor tubes filled with 1% plates and run on a QuantStudio5 qPCR system. SARS-CoV-2 viral RNA-dependent RNA polymerase levels were measured as previously described (Vanderheiden et al., 2020) . The following Taqman Primer/Probe sets 732 (Thermo Fisher Scientific) were used in this study: Gapdh (Mm99999915_g1).

Sequences were analyzed using Geneious software (Version 8.1.9). Identification and alignments to VH/VL genes, quantification of mutations and CDRH3 length were done using V Quest (Brochet et al., 2008) . 738

Mutations were counted beginning at the 5' end of the V-gene to the 3' end of the 428 FW3.

All graphs were completed using GraphPad Prism. For column analysis of multiple independent groups one-742

way ANOVA with Tukey's multiple comparison test or with Dunnett's multiple comparison test was used. For grouped analysis two-way ANOVA with Tukey's or Šídák's multiple comparison test. Correlations were 744 determined using nonparametric spearmen correlation and p values and nonlinear fit R squared values are reported. Specific details of statistical methods can be found in figure legends for all panels. Specific pe values 746 are reported in legends and p value ranges are reported for every figure in the legend. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. 748

Recent developments in the PHENIX software for automated 754 crystallographic structure determination

Structures of Human Antibodies Bound to SARS-CoV-2 Spike Reveal Common Epitopes and Recurrent Features of Antibodies

REGN-COV2 antibodies prevent and treat SARS-CoV-2 infection in rhesus 760 macaques and hamsters

Antibody cocktail to SARS-CoV-2 spike protein prevents rapid mutational escape seen with individual antibodies

Proteolytic activation of influenza viruses by serine proteases TMPRSS2 and HAT from human airway epithelium

IMGT/V-QUEST: the highly customized and integrated 768 system for IG and TR standardized V-J and V-D-J sequence analysis

Potent neutralizing antibodies from COVID-19 patients define multiple targets of vulnerability

Distinct conformational states of SARS-CoV-2 spike protein

Potent neutralizing antibodies against SARS-CoV-2 identified by high-throughput single-cell sequencing of 776 convalescent patients' B cells

Potent Neutralizing Antibodies against SARS-CoV-2 Identified by High-Throughput Single-Cell Sequencing of Convalescent Patients' B Cells

Potent SARS-CoV-2 neutralizing antibodies directed against spike N-terminal domain target a single 782 supersite

A neutralizing human antibody binds to the N-terminal domain of the Spike protein of SARS-CoV-2

SARS-CoV-2 Infection Depends on Cellular Heparan Sulfate 788 and ACE2

Protocol and Reagents for Pseudotyping Lentiviral Particles with SARS-CoV-2 Spike Protein for Neutralization Assays

Estimated transmissibility and severity of novel SARS-CoV-2 794 Variant of Concern 202012/01 in England. medRxiv

An interactive web-based dashboard to track COVID-19 in real time

Infection and mRNA-1273 vaccine antibodies neutralize SARS-CoV-2 UK variant. medRxiv

Coot: model-building tools for molecular graphics

Genomic characterisation of an emergent SARS-CoV-2

Structural Analysis of Neutralizing Epitopes of the SARS-CoV-2 Spike to Guide Therapy and Vaccine Design Strategies

Evolution of antibody immunity to SARS-CoV-2. Nature

Structural Basis of SARS-CoV-2 and SARS-CoV Antibody Interactions

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is 814 Blocked by a Clinically Proven Protease Inhibitor

Structure-based design of prefusion-stabilized SARS-CoV-2 spikes

Structural basis for potent neutralization of SARS-CoV-2 and role of 820 antibody affinity maturation

An mRNA Vaccine against SARS-CoV-2 -Preliminary Report

Human neutralizing antibodies elicited by SARS-CoV-2 infection

Xds

Structure-guided multivalent nanobodies block SARS-CoV-2 infection and suppress mutational escape

Longitudinal Isolation of Potent Near-Germline SARS-CoV-2-832 Neutralizing Antibodies from COVID-19 Patients

Appion: an integrated, database-driven pipeline to facilitate EM image processing

Potent neutralizing antibodies against multiple epitopes on SARS-CoV-2 spike

Neutralizing Activity of BNT162b2-Elicited Serum

Cross-reactive Antibody Response between SARS-CoV-2 and SARS-CoV Infections

N-terminal domain antigenic mapping reveals a site of vulnerability for SARS-CoV-2

Phaser crystallographic software

Single-shot Ad26 vaccine protects against SARS-CoV-2 in rhesus macaques

Complex-type N-glycan recognition by potent broadly neutralizing HIV antibodies

Efficient transfer, integration, and sustained long-term expression of the transgene in adult rat brains injected with a lentiviral vector

Human B Cell Clonal Expansion and Convergent Antibody Responses to SARS-CoV-2

CoV-2 lineages B.1.1.7 and B.1.351/501Y-V2 (Virological.org)

Immunogenicity and structures of a rationally designed prefusion MERS-CoV spike antigen

Initial Public Health Response and Interim Clinical Guidance for the 2019 Novel Coronavirus Outbreak -United States

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

Preliminary genomic characterisation of an emergent SARS-CoV-2 lineage in the 874 UK defined by a novel set of spike mutations

Convergent antibody responses to SARS-CoV-2 in convalescent individuals

Isolation of potent SARS-CoV-2 neutralizing antibodies and protection from disease in a small 880 animal model

Isolation of potent SARS-CoV-2 neutralizing antibodies and protection from disease in a small animal model

Resurgence of COVID-19 in Manaus, Brazil, 886 despite high seroprevalence

Antibody potency, effector function, and combinations in protection and therapy for SARS-CoV-2 infection in vivo

RELION: implementation of a Bayesian approach to cryo-EM structure determination

Analysis of a SARS-CoV-2-Infected Individual Reveals Development 894 of Potent Neutralizing Antibodies with Limited Somatic Mutation

A human 896 neutralizing antibody targets the receptor binding site of SARS-CoV-2

An Antibody Targeting the Fusion Machinery Neutralizes Dual-Tropic Infection and Defines a Site of Vulnerability on Epstein-Barr Virus

Cross-reactive serum and memory B cell responses to spike protein in SARS-CoV-2 902 and endemic coronavirus infection

Antibodies elicited by SARS-CoV-2 infection and boosted by vaccination neutralize an emerging variant and SARS-CoV-1. medRxiv

mRNA vaccination boosts cross-variant neutralizing 908 antibodies elicited by SARS-CoV-2 infection

Automated molecular microscopy: the new Leginon system

Identification of SARS-CoV RBD-targeting 912 monoclonal antibodies with cross-reactive or neutralizing activity against SARS-CoV-2

Emergence and rapid spread of a new severe acute respiratory syndrome-916 related coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in South Africa. medRxiv

Efficient generation of monoclonal antibodies from single human B cells by single cell RT-PCR and expression vector 920 cloning

Ultrapotent human antibodies protect against SARS-CoV-2 challenge via multiple mechanisms

Type I and Type III Interferons Restrict SARS-CoV-2 926 Infection of Human Airway Epithelial Cultures

Transmission of SARS-CoV-2 Lineage B.1.1.7 in England: Insights from linking epidemiological and genetic data. medRxiv

DoG Picker and TiltPicker: software tools to facilitate particle selection in single particle electron microscopy

Structure, Function, 934 and Antigenicity of the SARS-CoV-2 Spike Glycoprotein

Safety and Immunogenicity of Two RNA-Based Covid-19 Vaccine Candidates

Human-IgG-Neutralizing Monoclonal Antibodies Block the SARS-CoV-2 Infection

Receptor Recognition by the Novel Coronavirus from Wuhan: an Analysis Based on Decade-Long Structural Studies of SARS Coronavirus

A human monoclonal antibody blocking SARS-CoV-2 infection

Isolation of cross-reactive monoclonal antibodies against 948 divergent human coronaviruses that delineate a conserved and vulnerable site on the spike protein

Broad neutralization of SARS-related viruses by human monoclonal 952 antibodies

REGN-COV2, a Neutralizing Antibody Cocktail, in Outpatients with Covid-19

SARS-CoV-2 501Y.V2 escapes neutralization by 958 South African COVID-19 donor plasma

SARS-CoV-2 infection of human ACE2-transgenic mice causes severe lung inflammation and impaired function

Structural Basis for Potent Neutralization of 964 Betacoronaviruses by Single-Domain Camelid Antibodies

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

Serum Neutralizing Activity Elicited by mRNA-1273 Vaccine -970 Preliminary Report

A 972 noncompeting pair of human neutralizing antibodies block COVID-19 virus binding to its receptor ACE2

An Infectious cDNA Clone of SARS-CoV-2

Structural basis for the recognition of SARS-978

CoV-2 by full-length human ACE2

DNA vaccine protection against SARS-CoV-2 in rhesus macaques

Potently neutralizing and protective human antibodies against SARS-CoV-2

HIGHLIGHTS  14 anti-SARS-CoV-2 neutralizing mAbs isolated from four patients  3 anti-RBD and 1 anti-S2 mAb neutralized SARS-CoV-1 and the B.1.351 variant  Mouse studies show potential protective effect of anti-NTD mAbs eTOC Jennewein et al. isolated 14 anti-SARS-CoV-2 neutralizing antibodies from 4 patients. One anti-S2, one anti-NTD and 11 anti-RBD. Three anti-RBD and the anti-S2 cross-neutralized SARS-CoV-1 and B.1.351. The anti-NTD mAb conferred partial protection in mice

 Figure 3 and 4. This table shows the 14 neutralizing mAbs isolated and their binding epitopes, neutralization IC50, the VH and VL genes they are derived from, the number of mutations in these genes, the length of their CDRH3 and CDRL3, and whether they bind to SARS-CoV-2, SARS-CoV-1 and the other endemic human coronaviruses. IGKV3-20*01 2 0 12 9 a. An IC50 was not able to be assigned to CV2-74, as discussed in the text. b. These mAbs did not neutralize SARS-CoV-1 J o u r n a l P r e -p r o o f Table S3 . Data collection and refinement statistics for crystal structure. Related to Figure 4 and Figure S5 .