key: cord-0718519-898oxxqi
authors: Amanat, Fatima; Thapa, Mahima; Lei, Tinting; Sayed Ahmed, Shaza M.; Adelsberg, Daniel C.; Carreno, Juan Manuel; Strohmeier, Shirin; Schmitz, Aaron J.; Zafar, Sarah; Zhou, Julian Q.; Rijnink, Willemijn; Alshammary, Hala; Borcherding, Nicholas; Reiche, Ana Gonzalez; Srivastava, Komal; Sordillo, Emilia Mia; van Bakel, Harm; Turner, Jackson S.; Bajic, Goran; Simon, Viviana; Ellebedy, Ali H.; Krammer, Florian
title: SARS-CoV-2 mRNA vaccination induces functionally diverse antibodies to NTD, RBD and S2
date: 2021-06-08
journal: Cell
DOI: 10.1016/j.cell.2021.06.005
sha: 627a8e3197d73058daad59f8ec45c1295973731b
doc_id: 718519
cord_uid: 898oxxqi

In this study we profiled vaccine-induced polyclonal antibodies as well as plasmablast derived mAbs from individuals who received SARS-CoV-2 spike mRNA vaccine. Polyclonal antibody responses in vaccinees were robust and comparable to or exceeded those seen after natural infection. However, the ratio of binding to neutralizing antibodies after vaccination was greater than that after natural infection and, at the monoclonal level, we found that the majority of vaccine-induced antibodies did not have neutralizing activity. We also found a co-dominance of mAbs targeting the NTD and RBD of SARS-CoV-2 spike and an original antigenic-sin like backboost to seasonal human coronaviruses OC43 and HKU1. Neutralizing activity of NTD mAbs but not RBD mAbs against a clinical viral isolate carrying E484K as well as extensive changes in the NTD was abolished, suggesting that a proportion of vaccine induced RBD binding antibodies may provide substantial protection against viral variants carrying single E484K RBD mutations.

First, we document that RBD and NTD co-dominate as B-cell targets on the viral spike protein, 93

highlighting the importance of the NTD. We also report the first vaccine-induced NTD mAbs. In addition, 94

we show that the majority of mAbs isolated are non-neutralizing, which is reflective of the higher 95 binding to neutralization ratios found in serum after vaccination compared to natural infection. Finally, 96 data from plasmablasts suggest that, at least, some of the vaccine-induced response is biased by pre-97 existing immunity to human β-coronaviruses. 98 99 J o u r n a l P r e -p r o o f

The polyclonal antibody response to mRNA vaccination exceeds titers seen in convalescent individuals 101 but is characterized by a high ratio of non-neutralizing antibodies 102

In late 2020, six adult participants of an ongoing observational study received mRNA-based 103 SARS-CoV-2 vaccines (Suppl . Table 1 ). Blood from these individuals (termed V1-V6) was collected at 104 several time points including before vaccination (for 4/6), after the first vaccination and at several time 105 points after the second vaccination. We examined their immune responses to recombinant spike protein 106 and RBD in enzyme-linked immunosorbent assays (ELISA), in comparison to those of 30 COVID-19 107 survivors ( Figure 1A and 1B, Suppl. Table 1 ). The sera from convalescent individuals were selected 108

based on their anti-spike titers and grouped into three groups (low +: n=8; moderate ++: n=11; and high 109 +++: n=11, based on the antibody titer measured in the Mount Sinai's CLIA laboratory (Wajnberg et al., 110 2020), taken 111-273 days post symptom onset), in order to facilitate identifying different features that 111 may track with the strength of the antibody response. Five out of six vaccinees produced anti-spike and 112

anti-RBD responses that were, at the peak, markedly higher than responses observed even in the high 113

titer convalescent group while one vaccinee (V4) produced titers comparable to the high titer group. 114

Notably, the antibody response peaked one week after the second vaccine dose, followed by a decline 115 in titers over the following weeks as expected from an antibody response to vaccination. Interestingly, 116

anti-RBD antibody titers seemed to decline faster than anti-spike antibody titers, which appeared to be 117 more stable over time. We also measured neutralizing antibody titers using authentic SARS-CoV-2 and 118

found a similar trend with all vaccinees displaying high titers, even though V4 responded with delayed 119 kinetics ( Figure 1C) . Importantly, although at the peak response, the vaccine group mounted 120 neutralization titers that fell in the upper range for the high convalescent group, they did not exceed 121 that group markedly. This finding prompted us to calculate the proportions of spike binding to 122 neutralizing antibodies. For the convalescent group, we found that individuals with lower titers had a 123 higher proportion of binding to neutralizing antibodies than high responding convalescent individuals 124 ( Figure 1D ). When determined at the time of peak response, the vaccinees had the highest proportion 125 of binding to neutralizing antibody titers, indicating an immune response focused on non-neutralizing 126 antibodies or an induction of less potent neutralizing antibodies in general (or both). These proportions 127 remained stable over time with the ratio of binding to neutralizing antibodies in vaccinated individuals 128 being significantly higher than those observed for any of the three convalescent groups (p = 0.0004, 129 0.0002 and 0.0041 for the three groups respectively; Suppl. Figure 1) . We also investigated the spike 130 binding to RBD binding ratio and found no difference to convalescent individuals except a general trend 131 towards proportionally less RBD binding over time in the vaccinees (Suppl. Figure 1) . 132 mRNA vaccination induces a modest but measurable immune response to seasonal β-coronavirus 133 spike proteins 134

It has been reported that SARS-CoV-2 infection induces an original antigenic sin-type immune 135 response against human coronaviruses (hCoVs) to which the majority of the human population has pre-136 existing immunity (Aydillo et al., 2020; Song et al., 2020) . Here, we explored whether this phenomenon 137 is also induced by SARS-CoV-2 mRNA vaccination. Antibody titers in four vaccinees against spike protein 138 from α-coronaviruses 229E and NL63 were detectable at the pre-vaccination time point, but did not 139

increase substantially post-vaccination (Figure 1E-F; for V5 and V6 no pre-vaccination serum was 140 available). However, titers against the spike proteins of β-coronaviruses OC43 and HKU1 increased 141 J o u r n a l P r e -p r o o f substantially in these four vaccinees after vaccination (Figure 1G-H) . Thus, vaccination with mRNA SARS-142

CoV-2 spike also boosts immune responses against seasonal β-coronavirus spike proteins in a manner 143 reminiscent of that reported for natural infection with SARS-CoV-2. 144

Plasmablast response to SARS-CoV-2 mRNA vaccination targets both the RBD and the NTD 145

In order to characterize the B-cell response to vaccination in an unbiased manner, plasmablasts 146

were single-cell sorted from blood specimens obtained from three individuals (V3, V5 and V6) one week 147 after the booster immunization (Suppl. Figure 2) . All mAbs were generated from single-cell sorted 148

plasmablasts and probed for binding to recombinant SARS-CoV-2 spike protein. Twenty-one (40 mAbs  149 were screened, with 28 being clonally unique, Suppl. Table 2 ) spike-reactive mAbs were isolated from 150 V3, six (82 screened, 20 unique) from V5 and fifteen (84 screened, 24 unique) from V6 (Figure 2A) . Using 151 recombinant spike, RBD, NTD and S2 proteins, we mapped the domains to which these mAbs bind. 152

Interestingly, only a minority of these antibodies recognized RBD (24% for V3, 47% for V6 and no RBD 153 binders were identified for V5) (Figure 2B and 2E) . A substantial number of the isolated mAbs bound to 154 NTD including 14% for V3, 33% for V5 and 33% for V6 ( Figure 2C and 2E) . These data indicate that RBD 155

and NTD are co-dominant in the context of mRNA-induced plasmablast response. The epitopes for the 156 majority of the remaining spike binding mAbs, 52% for V3, 50% for V5 and 20% for V6, mapped to S2 157 ( Figure 2D and 2E). Only three mAbs were not accounted for in terms of binding target (two for V3 and 158 one for V5, Figure 2E ). 159

The majority of isolated mAbs from SARS-CoV2 vaccinees are non-neutralizing 160

All antibodies were tested for neutralizing activity against the USA-WA1/2020 strain of SARS-161

CoV-2. Only a minority of the binding antibodies, even those targeting the RBD, showed neutralizing 162 activity ( Figure 2F and 2G) . For V3, only one (an RBD binder) out of 21 mAbs (5%) displayed neutralizing 163 activity ( Figure 2G) . For V5, a single NTD antibody neutralized authentic SARS-CoV-2 (17%) (Figure 2G ).

The highest frequency of neutralizing antibodies was found in V6 (34%) with one RBD neutralizer and 165

four NTD neutralizers ( Figure 2G) . Interestingly, the highest neutralizing potency was found in mAb 166 PVI.V5-6, an NTD binder followed by PVI.V6-4, an RBD binder. 167

We also tested all the antibodies for reactivity to the spike proteins of the four hCoVs 229E, 168 NL63, HKU1 and OC43. No antibody binding to the spike proteins of α-coronaviruses 229E and NL63 was 169 found but we identified five mAbs (including three from V3, one from V5 and one from V6) that bound, 170

to varying degrees, to the spike of OC43, which, like SARS-CoV-2, is a β-coronavirus ( Figure 2H) Figure 2I) . 176

The spike-reactive plasmablast response is dominated by IgG1+ cells and is comprised of a mixture of 177 cells with low and high levels of somatic hypermutation (SHM) 178

Single-cell RNA sequencing (scRNAseq) was performed on bulk sorted plasmablasts from the 179 three vaccinees (V3, V5, V6) to comprehensively examine the transcriptional profile, isotype distribution 180 and somatic hypermutation (SHM) of vaccine-induced plasmablasts. We analyzed 4,584, 3,523 and 181 J o u r n a l P r e -p r o o f 4,461 single cells from subjects V3, V5, and V6, respectively. We first verified the identity of sequenced 182 cells as plasmablasts through the combined expression of B cell receptors (BCRs) ( Figure 3A ) and that of 183 the canonical transcription as well as other factors essential for plasma cell differentiation, such as 184 PRDM1, XBP1 and MZB1 ( Figure 3B ). To identify vaccine-responding B cell clones among the analyzed 185 plasmablasts, we used scRNAseq to also analyze gene expression and V(D)J libraries from the sorted 186 plasmablasts and clonally matched the BCR sequences to those from which spike-specific mAbs had 187 been made. Using this method, we recovered 332, 7 and 1,384 BCR sequences from the scRNAseq data 188 that are clonally related to the spike-binding mAbs derived from subjects V3, V5 and V6, respectively 189 ( Figure 3C ). It is important to note here that we were not able to recover clonally related sequences for 190

all of the mAbs that we cloned and expressed from each of the three vaccinees. 191

We next examined the isotype and IgG subclass distribution among the recovered sequences. 192

IgG1 was by far the most dominant isotype in the three vaccinees ( Figure 3D) plasmablasts was equivalent to those observed after seasonal influenza virus vaccination ( Figure 3E , left 199 panel). We reasoned that the high level of SHM among spike-reactive plasmablasts may be derived from 200 those targeting conserved epitopes that are shared with human β-coronaviruses. Indeed, we found that 201 the SHM level among clones that are related to cross-reactive mAbs was significantly higher than their 202 non-cross-reactive counterparts ( Figure 3E, right panel) . 203

Two mAbs were identified as neutralizing and binding to RBD. We wanted, therefore, to test if 205 they competed with ACE2 for RBD binding. Concentration-dependent competition was indeed observed 206

for both mAbs demonstrating that inhibition of ACE2 binding is the mechanism of action of the two 207 mAbs (Figure 4 ). Since we prepared RBD proteins of viral variants of concern for analysis of antibody 208

binding (see below), we also wanted to assess the affinity of each variant RBD for human ACE2. mutations in RBD tested increased affinity to human ACE2. Specifically, N501Y and Y453F combined with 215 N439K increased affinity for human ACE2 by 5-fold ( Figure 4D , Suppl. Figure 3) . In contrast, E484K on its 216 own decreased affinity by 4-fold. Of note, the B.1.351 RBD affinity for ACE2 was comparable to that of 217 the wild-type RBD. These data were confirmed using an ELISA-based method which showed the same 218 trends (Suppl. Figure 4) . 219

Binding profiles of polyclonal serum and mAbs to RBDs carrying mutations found in viral variants of 220 concern 221

Next, we assessed binding of sera from vaccinated individuals, COVID Figure 5A ). In general, single mutants E406Q, E484K and F490K exerted the biggest 228 impact on binding. However, complete loss of binding was rare and 2-4-fold reduction in binding was 229 more common. Interestingly, almost all sera bound better to N501Y RBD (B.1.1.7) than to wild-type RBD 230

(average 129% compared to wild type). Conversely, the B.1.351 RBD caused, on average, a 39% 231 reduction in binding. The impact was slightly lower for the P.1 RBD (average 70% binding compared to 232 wild-type). For sera from the six vaccinated individuals, however, the highest reduction seen was only 233 two-fold for E406Q, N440K, E484K and F490K ( Figure 5B ). Of note, the vaccinees' later samples (V1=d89, 234

V2=d102, V3=d47, V4=d48, V5=49 and V6=48) were assayed to allow for some affinity maturation. The 235

highest reduction observed for E484K, F484A, B.1.351 and P.1 were also approximately two-fold but this 236

did not apply to all six vaccinees. Some vaccinees maintained binding levels against these RBDs at levels 237 comparable to wild-type RBD. 238

RBD binding mAbs were also tested for binding to the same variants. In general, mAbs 239 maintained binding levels within 2-fold of the binding seen with the wild-type RBD with some 240 exceptions. In fact, for most mAbs, no impact on binding was observed ( Figure 5C ) with the exception of 241 PVI.V3-9, which lost binding to the RBD carrying F486A. Although there was a negative impact on 242 binding of several mAbs to the B.1.351 variant, binding was almost unaffected by the mutations in the 243 P.1 variant RBD. Only one mAb, PVI.V6-4, showed a drop in binding to P.1. 244

NTD mutations significantly impact the neutralizing activity of NTD binding mAbs 246

Through the Mount Sinai Hospital's Pathogen Surveillance Program, we had access to the SARS-247

CoV-2 isolate PV14252 (Clade 20C, Pango lineage B.1) that featured two mutations (W64R, L141Y) and 248 one deletion (Δ142-145) in the NTD as well as the E484K mutation in the RBD ( Figure 5D ). To determine 249 the susceptibility of this virus variant to neutralization by post-vaccination serum, we performed 250 microneutralization assays. Wild-type SARS-CoV-2 and PV14252 were tested in parallel to ensure that 251 the assay setup for both viruses allowed comparison. We found a relatively minor impact when testing 252 polyclonal sera from vaccinees for neutralizing activity ( Figure 5E ). The activity of sera from V2, V5 and 253 V6 slightly increased while the activity for V1, V3 and V4 decreased. Next, we tested the seven 254 neutralizing mAbs that we isolated from plasmablasts. Consistent with their binding profiles in the 255 variant RBD ELISA, the two RBD mAbs neutralized both viruses with comparable efficiency ( Figure 5F ). In 256 fact, the activity of PVI.V3-9 increased slightly ( Figure 5F ). In stark contrast, all five anti-NTD antibodies 257 completely lost neutralizing activity against PV14252 due to mutations present in the NTD of this viral 258 isolate. 259

We also tested the neutralizing activity of the two RBD and the five NTD antibodies against the 261 variants of concern B.1.1.7 and B.1.351 ( Figure 5G ) absolute antibody titers and not ratios is underscored by the fact that post-vaccination neutralization 293 titers were equal to or exceeded the titers found in the high responder convalescent group. 294

Of the four seasonal CoVs that are widely circulating in humans, β-coronaviruses OC43 and HKU1 have 295 higher homology to SARS-CoV-2 spike. Vaccinated individuals mounted a response to spike proteins 296 from OC43 and HKU1 but not to α-coronaviruses 229E and NL63. This phenomenon resembles the 297 immune imprinting described in influenza virus immunology and has already been shown for natural 298 infection with SARS-CoV-2 where a 'backboost' to β-coronaviruses was also found (Aydillo et al., 2020; 299 Song et al., 2020) . A few of the mAbs isolated in our study had, indeed, such a cross-reactive phenotype. 300

It remains unclear whether these antibodies, which target mostly S2 epitopes, contribute to protection 301 against SARS-CoV-2, OC43 or HKU1 infection. However, the cross-reactive epitopes of mAbs that do bind 302 SARS-CoV-2, HKU1 and OC43 spikes could provide the basis for future pan-β-coronavirus vaccines. While 303

it is likely the case that the B-cells producing these mAbs come from recall responses and were initially 304 induced by human β-coronaviruses (which is supported by serology and of course the extensive SHM 305 that the mAbs show), they could hypothetically also be de novo induced antibodies. While this is 306

probably not the case, we cannot exclude this possibility with our current data. 307 308

Another interesting point we noted is the co-dominance of RBD and NTD. Previous analyses of 309 B-cell responses to SARS-CoV-2 mRNA vaccination focused on cells baited by labeled RBD (Wang et al., 310 2021). We, in contrast, took an unbiased approach to sort and clone plasmablasts in an antigen-agnostic 311 manner. We found similar levels of NTD and RBD binders with many mAbs binding to epitopes outside 312 the RBD and the NTD. In one vaccinee not a single RBD binding mAbs was isolated with the caveats that 313 the overall number of mAbs derived from that individual were low and their polyclonal serum antibody 314 responses included RBD recognition. These data suggest that the NTD, which also harbors neutralizing 315 epitopes, is -at least -as important as the RBD and warrants as much attention. 64 targeting the NTD and 46 binding outside of RBD and NTD and the second study finding 10 RBD 320 mAbs, 13 non-RBD S1 binding mAbs (strongly suggesting NTD binding) and 9 mAbs targeting S2. 321

Interestingly, and in contrast to our findings, a recent deep mutational scanning paper with sera from 322 mRNA-1273 found a very strong RBD-focused response (Greaney et al., 2021a). Further characterization 323 of the mAbs obtained in our study showed a complete loss of neutralization against an authentic, 324

replication-competent variant virus that harbored extensive changes in the NTD. All NTD mAbs also lost 325 neutralizing activity against B.1.351 and all but one lost activity against B.1.1.7. These observations may 326 explain why a reduction in neutralization against the viral variant of concern B.1. increased, suggesting that changes that increase affinity for the receptor may also increase affinity of a 346 set of antibodies that may mimic the receptor. 347

We also noted that the two neutralizing antibodies against the RBD showed some reduced 348 binding to a mutant RBD carrying the E484K mutation while having similar or even increased neutralizing 349 potency against a variant virus carrying the E484K mutation as the only change in its RBD. The reduced 350 affinity of the E484K variant RBD for hACE2 could render the virus more susceptible to RBD binding 351 mAbs. Thus, an antibody binding to the RBD may just be more effective in interfering with a low affinity 352 as compared to a high affinity RBD-hACE2 interaction. Increased affinity as an escape mechanism for 353 viruses has been described in mechanism could be at play here. 355

Whether or not the current vaccines will provide effective protection against circulating and 356 emerging viral variants of concern is an important question which has gathered a lot of attention in early 357 2021. Our data indicate that reduction in binding to the E484K and B. , 2021) ). Although not 360 tested here, it is likely that the reduction in binding to full length spike is even lower, given the many 361 epitopes on the spike other than NTD and RBD. The maintenance of binding to a large degree observed 362

in this study suggests that viral variants will have a minor impact on serological assays which are 363 currently in wide use for medical, scientific and public health reasons. Binding, non-neutralizing 364

antibodies have also been shown to have a protective effect in many viral infections (Asthagiri 365

Arunkumar Finally, although some antibodies may lose neutralizing activity due to reduced affinity, they do still 371

bind. Furthermore, B cells with these specificities potentially could undergo affinity maturation after 372 exposure to a variant virus or a variant spike-containing vaccine, leading to high affinity antibodies to 373 variant viruses of concern. 374 In summary, we demonstrate that the antibody responses to SARS-CoV-2 mRNA vaccination comprise a 375 large proportion of non-neutralizing antibodies and are co-dominated by NTD and RBD antibodies. The 376 NTD portion of the spike represents, thus, an important vaccine target. Since all viral variants of concern 377 are heavily mutated in this region, these observations warrant further attention to optimize SARS-CoV-2 378

vaccines. Finally, broadly cross-reactive mAbs to β-coronavirus spike proteins are induced after 379 vaccination, and suggest a potential development path for a pan-β-coronavirus vaccine. 380

While our study characterizes the antibody response after SARS-CoV-2 mRNA vaccination in detail, it has 382 several limitations. The first limitation is the small number of study participants, which makes this study 383 a qualitative rather than a quantitative study. Another limitation is the lack of plasmablast analysis of 384 SARS-CoV-2 infected individuals. We have compared our data with published data from plasmablast 385 analysis after SARS-CoV-2 infection but a side-by-side comparison would have been more accurate. We 386

have also not included a longitudinal analysis of the convalescent sera iin the study but feel that 387 providing a wide range of time points and titer levels offsets this limitation to a certain degree. 

The Icahn School of Medicine at Mount Sinai has filed patent applications relating to SARS-CoV-2 412 serological assays and NDV-based SARS-CoV-2 vaccines which list Florian Krammer as co-inventor. 413

Viviana Simon is also listed on the serological assay patent application as co-inventors. Mount Sinai has 414 spun out a company, Kantaro, to market serological tests for SARS-CoV-2. Florian Krammer has 415 consulted for Merck and Pfizer (before 2020), and is currently consulting for Pfizer, Seqirus and Avimex. 416

The Krammer laboratory is also collaborating with Pfizer on animal models of SARS-CoV-2. Ali Ellebedy 417 has consulted for InBios and Fimbrion Therapeutics (before 2021) and is currently a consultant for 418

Mubadala Investment Company. The Ellebedy laboratory received funding under sponsored research 419 agreements that are unrelated to the data presented in the current study from Emergent BioSolutions 420 and from AbbVie. 

Orthogonal SARS-CoV-2 Serological Assays Enable 597 Surveillance of Low-Prevalence Communities and Reveal Durable Humoral Immunity

Convergent antibody responses to SARS-CoV-2 in convalescent 601 individuals

Functional SARS-CoV-2-Specific Immune 604 Memory Persists after Mild COVID-19

Systematic Analysis of Monoclonal Antibodies against

Ebola Virus GP Defines Features that Contribute to Protection

Antibody potency, effector function, and combinations in 611 protection and therapy for SARS-CoV-2 infection in vivo

Longitudinal observation and decline of neutralizing antibody 614 responses in the three months following SARS-CoV-2 infection in humans

Preliminary Efficacy of the NVX-CoV2373 Covid-19 Vaccine Against 618 the B.1.351 Variant. medRxiv

Rapid 620 generation of fully human monoclonal antibodies specific to a vaccinating antigen

Cross-reactive serum and memory B cell responses to spike protein in SARS-CoV-624 2 and endemic coronavirus infection

SARS-CoV-2 Seroconversion in Humans: A Detailed Protocol 627 for a Serological Assay, Antigen Production, and Test Setup

Comprehensive Integration of Single-Cell Data

Newcastle disease virus (NDV) expressing the spike protein of SARS-634

CoV-2 as a live virus vaccine candidate

Emergence and rapid spread of a new severe acute respiratory 637 syndrome-related coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in South Africa. 638 medRxiv

Circulating SARS-CoV-2 spike N439K variants maintain 641 fitness while evading antibody-mediated immunity

Human germinal centres engage memory and naive B cells after influenza 644 vaccination

Robust neutralizing antibodies to SARS-CoV-2 infection persist 647 for months

RNA-Based COVID-19 Vaccine BNT162b2 Selected for a Pivotal Efficacy 650 Study. medRxiv

mRNA vaccine-elicited antibodies to SARS-CoV-2 and 653 circulating variants

Escape from neutralizing antibodies by SARS-CoV-2 spike 656 protein variants

SARS-CoV-2 501Y.V2 escapes neutralization by

South African COVID-19 donor plasma. bioRxiv, 2021

Distinct B cell subsets give rise to antigen-specific antibody responses against 662 SARS-CoV-2

Hemagglutinin Stalk-and Neuraminidase-Specific Monoclonal Antibodies Protect against Lethal 665 H10N8 Influenza Virus Infection in Mice

Vaccination with adjuvanted recombinant neuraminidase induces broad 668 heterologous, but not heterosubtypic, cross-protection against influenza virus infection in mice

Broadly cross-reactive antibodies dominate the 672 human B cell response against 2009 pandemic H1N1 influenza virus infection

Rapid cloning of high-affinity human monoclonal antibodies against influenza 677 virus

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

mRNA-1273 vaccine induces neutralizing antibodies against spike 683 mutants from global SARS-CoV-2 variants. bioRxiv

IgBLAST: an immunoglobulin variable domain 686 sequence analysis tool

RBD mutants were generated in the pCAGGS RBD construct by 779 changing single residues using mutagenesis primers. All proteins were purified after transient 780 transfections with each respective plasmid. Six-hundred million Expi293F cells were transfected using 781 the ExpiFectamine 293 Transfection Kit and purified DNA

Immulon 4 HBX; Thermo Scientific) were coated overnight at 4°C with 794 recombinant proteins at a concentration of 2 ug/ml in PBS (Gibco; Life Technologies) and 50 uls/well

All serum dilutions were prepared in 1% non-fat 798 milk prepared in TPBS. All serum samples were diluted 3-fold starting at a dilution of 1:50. After the 799 blocking step, serum dilutions were added to the respective plates for two hours at RT. Next, plates 800 were washed thrice with 250 uls/well of TPBS to remove any residual primary antibody. Secondary 801 antibody solution was prepared in 1% non-fat milk in TPBS as well and 100 uls/well was added to the 802 plates for 1 hour at RT. For human samples, anti-human IgG conjugated to horseradish peroxidase (HRP) 803 was used at a dilution of 1:3000 (Millipore Sigma; catalog #A0293). For mouse samples, anti-mouse

Once the secondary incubation was done, plates were again 807 washed thrice with 250uls/well of TPBS. Developing solution was made in 0.05M phosphate-citrate 808 buffer at pH 5 using o-phenylenediamine dihydrochloride tablets (Sigma-Aldrich; OPD) at a final 809 concentration of 0.04 mg/ml. One hundred uls/well of developing solution was added to each plate for 810 exactly 10 minutes after which the reaction was halted with addition of 50 uls/well of 3M hydrochloric 811 acid (HCl)

Graphpad Prism 7. This protocol has been described in detail earlier

Antibody responses after SARS-CoV-2 mRNA vaccination target RBD, NTD and S2

CoV-2 mRNA vaccination induces a high rate of non-neutralizing antibodies • Crossreactive antibodies to seasonal β-coronaviruses are induced by vaccination • Variant mutation N501Y enhances affinity to human ACE2 while E484K reduces it

An analysis of mRNA vaccine-induced polyclonal antibodies and plasmablast derived monoclonal antibodies from individuals vaccinated against SARS-CoV-2 identifies a high proportion of nonneutralizing antibodies, the induction of cross-reactive antibodies to seasonal coronaviruses and also maps the regions in the spike protein that are targeted

00791; IRB-20-03374). All participants provided written informed consent prior to collection of specimen 767 and clinical information. All specimens were coded prior to processing and analysis. An overview of the 768 characteristics of the vaccinees as well as the study participants with and without COVID-19 is provided 769in Suppl. starting at a dilution of 1:20. All work with authentic SARS-CoV-2 (isolate USA-WA1/2020 and PV14252) 868 was done in a biosafety level 3 (BSL3) laboratory following institutional biosafety guidelines and has 869 been described in much greater detail earlier (Amanat et al., 2020b; Amanat et al., 2020c) . Six hundred 870 median cell culture infectious doses (TCID 50 s) of authentic virus (USA-WA1/2020 and PV14252) was 871 added to each serum dilution and virus-serum mixture was incubated together for 1 hour inside the 872 biosafety cabinet. Media from the cells was removed and 120 uls of the virus-serum mixture was added 873 onto the cells for 1 hour at 37°C. After one hour, the virus-serum mixture was removed and 100 uls of 874 each corresponding dilution was added to every well. In addition, 100uls of 1X MEM was also added to 875 every well. Cells were incubated for 48 hours at 37°C after which the media was removed and 150 uls of 876 10% formaldehyde (Polysciences) was added to inactivate the virus. For assay control, remdesivir was 877 used against both the wild type virus as well as the patient isolate. After 24 hours, cells were 878 permeabilized and stained using an anti-nucleoprotein antibody 1C7 as discussed in detail earlier 879( BioLegend. Cells were washed twice, and single plasmablasts (live singlet CD19 + CD3 -IgD lo CD38 + CD20 -885 CD71 + ) were sorted using a FACSAria II into 96-well plates containing 2 µL Lysis Buffer (Clontech) 886supplemented with 1 U/µL RNase inhibitor (NEB) and immediately frozen on dry ice, or bulk sorted into 887 PBS supplemented with 0.05% BSA and processed for single cell RNAseq. 888Monoclonal antibody (mAb) generation. Antibodies were cloned as described previously (