key: cord-268894-amfv3z2y
authors: Nguyen-Contant, Phuong; Embong, A. Karim; Kanagaiah, Preshetha; Chaves, Francisco A.; Yang, Hongmei; Branche, Angela R.; Topham, David J.; Sangster, Mark Y.
title: S protein-reactive IgG and memory B cell production after human SARS-CoV-2 infection includes broad reactivity to the S2 subunit
date: 2020-07-21
journal: bioRxiv
DOI: 10.1101/2020.07.20.213298
sha: 
doc_id: 268894
cord_uid: amfv3z2y

The high susceptibility of humans to SARS-CoV-2 infection, the cause of COVID-19, reflects the novelty of the virus and limited preexisting B cell immunity. IgG against the SARS-CoV-2 spike (S) protein, which carries the novel receptor binding domain (RBD), is absent or at low levels in unexposed individuals. To better understand the B cell response to SARS-CoV-2 infection, we asked whether virus-reactive memory B cells (MBCs) were present in unexposed subjects and whether MBC generation accompanied virus-specific IgG production in infected subjects. We analyzed sera and PBMCs from non-SARS-CoV-2-exposed healthy donors and COVID-19 convalescent subjects. Serum IgG levels specific for SARS-CoV-2 proteins (S, including the RBD and S2 subunit, and nucleocapsid [N]) and non-SARS-CoV-2 proteins were related to measurements of circulating IgG MBCs. Anti-RBD IgG was absent in unexposed subjects. Most unexposed subjects had anti-S2 IgG and a minority had anti-N IgG, but IgG MBCs with these specificities were not detected, perhaps reflecting low frequencies. Convalescent subjects had high levels of IgG against the RBD, S2, and N, together with large populations of RBD- and S2-reactive IgG MBCs. Notably, IgG titers against the S protein of the human coronavirus OC43 in convalescent subjects were higher than in unexposed subjects and correlated strongly with anti-S2 titers. Our findings indicate cross-reactive B cell responses against the S2 subunit that might enhance broad coronavirus protection. Importantly, our demonstration of MBC induction by SARS-CoV-2 infection suggests that a durable form of B cell immunity is maintained even if circulating antibody levels wane. IMPORTANCE Recent rapid worldwide spread of SARS-CoV-2 has established a pandemic of potentially serious disease in the highly susceptible human population. Key questions are whether humans have preexisting immune memory that provides some protection against SARS-CoV-2 and whether SARS-CoV-2 infection generates lasting immune protection against reinfection. Our analysis focused on pre- and post-infection IgG and IgG memory B cells (MBCs) reactive to SARS-CoV-2 proteins. Most importantly, we demonstrate that infection generates both IgG and IgG MBCs against the novel receptor binding domain and the conserved S2 subunit of the SARS-CoV-2 spike protein. Thus, even if antibody levels wane, long-lived MBCs remain to mediate rapid antibody production. Our study also suggests that SARS-CoV-2 infection strengthens preexisting broad coronavirus protection through S2-reactive antibody and MBC formation.

The high susceptibility of humans to SARS-CoV-2 infection, the cause of COVID-19, reflects the 24 novelty of the virus and limited preexisting B cell immunity. IgG against the SARS-CoV-2 spike 25 (S) protein, which carries the novel receptor binding domain (RBD), is absent or at low levels in 26 unexposed individuals. To better understand the B cell response to SARS-CoV-2 infection, we 27 asked whether virus-reactive memory B cells (MBCs) were present in unexposed subjects and 28

whether MBC generation accompanied virus-specific IgG production in infected subjects. We 29 analyzed sera and PBMCs from non-SARS-CoV-2-exposed healthy donors and COVID-19 30 convalescent subjects. Serum IgG levels specific for SARS-CoV-2 proteins (S, including the RBD 31 and S2 subunit, and nucleocapsid [N] ) and non-SARS-CoV-2 proteins were related to 32 measurements of circulating IgG MBCs. Anti-RBD IgG was absent in unexposed subjects. Most 33 unexposed subjects had anti-S2 IgG and a minority had anti-N IgG, but IgG MBCs with these 34 specificities were not detected, perhaps reflecting low frequencies. Convalescent subjects had high 35 levels of IgG against the RBD, S2, and N, together with large populations of RBD-and S2-reactive 36 IgG MBCs. Notably, IgG titers against the S protein of the human coronavirus OC43 in 37 convalescent subjects were higher than in unexposed subjects and correlated strongly with anti-S2 38 titers. Our findings indicate cross-reactive B cell responses against the S2 subunit that might 39 enhance broad coronavirus protection. Importantly, our demonstration of MBC induction by 40 SARS-CoV-2 infection suggests that a durable form of B cell immunity is maintained even if 41 circulating antibody levels wane. 42 43 IMPORTANCE 44

The betacoronavirus SARS-CoV-2, the causative agent of a respiratory disease termed 56 COVID-19, emerged in China in late 2019 and rapidly spread worldwide (1). A pandemic was 57 declared in March 2020 and global deaths from COVID-19 now exceed 500,000. The rapid 58 increase in cases in many countries has challenged healthcare systems and shutdowns and 59 quarantine measures introduced to slow virus spread have caused major disruptions to society and 60 economies (2). SARS-CoV-2 infection produces a wide spectrum of outcomes. A proportion of 61 infections, likely more than 20%, remain asymptomatic. Most clinical cases develop mild to 62 moderate respiratory symptoms, but up to 20% progress to a more severe disease with extensive 63 pneumonia (3, 4). 64

When SARS-CoV-2 emerged and began to spread, the severity of the threat was primarily 65 attributed to the novelty of the virus to the human immune system and, consequently, a lack of 66 preexisting immune memory to quickly clear virus and limit disease progression. Four types of 67 common cold coronavirus are endemic in humans, the alphacoronaviruses 229E and NL63 and the 68 betacoronaviruses OC43 and HKU1. However, limited relatedness between key structural proteins 69 of these human coronaviruses (HCoVs) and those of SARS-CoV-2 suggested that significant 70 cross-reactive immunity was unlikely (5, 6). Initial studies of non-SARS-CoV-2-exposed 71 individuals found negligible levels of IgG against the SARS-CoV-2 spike (S) protein, the viral 72 attachment protein that binds the receptor angiotensin converting enzyme 2 (ACE2) on host cells 73 to initiate infection (7). More recently, however, studies have provided evidence of SARS-CoV-74 2-reactive B and T cell memory in unexposed subjects that could confer some protection against 75 SARS-CoV-2 or modulate disease pathogenesis. 76

Sera from non-SARS-CoV-2-exposed individuals have been screened for IgG binding to 77 the S1 and S2 subunits of the SARS-CoV-2 S protein. The membrane-distal S1 subunit contains 78 the receptor binding domain (RBD) for receptor recognition, and the membrane-proximal S2, 79 which has higher homology among coronaviruses than does S1 (6, 8), mediates membrane fusion 80

to release viral RNA into the host cell. In two large cohorts of unexposed subjects, approximately 81 10% had IgG that bound S2, but not S1 or the RBD. Approximately 4% of subjects had IgG against 82 the SARS-CoV-2 nucleocapsid (N) protein, which is highly conserved among coronaviruses (9, 83 10) . Although N is an internal viral protein and not a target of neutralizing antibodies (Abs) , 84 coronavirus infections typically elicit strong anti-N Ab production (11). The idea that circulating 85

HCoVs elicit IgG that cross-reacts with SARS-CoV-2 is supported by the finding that SARS-CoV-86 2 infection increases IgG titers against the S proteins of multiple HCoVs (12). In T cell studies, 87

CD4 + T cells in up to 50% of non-SARS-CoV-2-exposed donors responded to epitopes in S and 88 non-S proteins of SARS- 13) . Notably, S-reactive CD4 + T cells in unexposed subjects 89

were mostly reactive to the conserved S2 subunit, consistent with cross-reactivity to circulating 90 HCoVs (8). SARS-CoV-2-reactive CD8 + T cells were also detected in unexposed donors, but the 91 response was less marked than for CD4 + T cells (13). 92

are also likely to be present in non-SARS-CoV-2-exposed individuals. Indeed, MBCs might be 94 more important than preexisting cross-reactive Abs as a source of protection against SARS-CoV-95 2. IgG MBCs are more broadly reactive than Abs generated against the same antigen, they persist 96 after circulating Ab levels wane, and they are readily activated to generate strong Ab responses or 97 seed germinal centers for additional rounds of affinity maturation (14). Concurrent early 98 production of virus-specific IgM and IgG in the response to SARS-CoV-2 infection suggests a 99 response mediated by IgG MBCs as well as naïve B cells (9, (15) (16) (17) . This picture is supported by To extend our understanding of the B cell response to SARS-CoV-2 infection, the current 116 study compared Ab and MBC immunity to SARS-CoV-2 in unexposed individuals and individuals 117 in the convalescent phase of infection. In particular, we were interested in the presence of SARS-118

CoV-2-reactive MBCs in unexposed subjects that could confer some protection against SARS-119

CoV-2, and formation of MBCs by SARS-CoV-2 infection to provide durable protection against 120

IgG MBCs reactive to the novel RBD and the conserved S2 subunit of the S protein. MBCs are thus likely to be available to mediate rapid protective Ab responses if circulating Ab 123 levels wane and reinfection occurs. Our study also draws attention to preexisting SARS-CoV-2-124 cross-reactive B cell memory to the S2 subunit in SARS-CoV-2-naïve subjects. We speculate that 125 the strong response to S2 after SARS-CoV-2 infection reflects preexisting S2-reactive MBC 126 activation and strengthens broad coronavirus protection. convalescent subjects sampled 4-9 weeks after symptom onset. Reactivity was measured against 135 the S (including the RBD and S2 subunit) and N proteins of SARS-CoV-2 and the S proteins of 136 the human alphacoronavirus 229E and betacoronavirus OC43. The H1 influenza virus 137 hemagglutinin and tetanus toxoid (TTd) were included as control antigens that humans are 138 commonly exposed to through infection and vaccination. 139

Serum IgG levels were measured by ELISA. Approximately one-third of non-SARS-CoV-140 2-exposed subjects in the healthy donor cohort had low levels of serum IgG against the S and N 141 proteins of SARS-CoV-2, likely reflecting cross-reactivity with seasonal HCoVs ( Figure 1A ). 142

Notably, 86% of unexposed subjects had IgG against the highly conserved S2 subunit of the S 143 protein. It is possible that inherent features of the bulky S reagent used in our analysis reduced 144 binding by anti-S2 Abs. IgG that bound the highly novel RBD was not detected in unexposed 145 subjects. All non-SARS-CoV-2-exposed subjects had IgG against S proteins of the HCoVs 229E 146 and OC43, indicating previous infection, and against the control proteins H1 and TTd ( Figures 1C-147 1F). 148

response to the S2 subunit. Levels of IgG against S, RBD, S2 and N were markedly higher in 150 convalescent subjects than unexposed subjects, indicating strong induction of these Abs by SARS-151

CoV-2 infection ( Figure 1A) . In a small number of convalescent subjects, high anti-S IgG titers 152 were associated with low levels of anti-N IgG. Indeed, more than 40% of convalescent subjects 153 had anti-N IgG levels within the range in unexposed subjects, questioning the reliability of using 154 anti-N IgG measurement to identify previous SARS-CoV-2 infection. 155

Notably, serum IgG titers against S2 were consistently higher than against the RBD in 156 convalescent subjects, perhaps reflecting the novelty of the RBD and a response dependent on 157 naive B cell activation ( Figure 1B) . Interestingly, titers of IgG were higher against the S protein 158 of the HCoV OC43 in convalescent subjects than in unexposed subjects, but this was not the case 159

for the S protein of HCoV 229E (or for the control proteins H1 and TTd) ( Figures 1C-1F ). The CoV-2 infection ( Figure 1G ). The particularly strong correlation between IgG titers against OC43 163 S and the SARS-CoV-2 S2 suggests a cross-reactive response to the S2 subunit. 164

Since the healthy donor samples in our analysis were collected 6-10 years before the 165 emergence of SARS-CoV-2, we considered the possibility that a recently circulating HCoV could 166 have been responsible for the higher anti-OC43 S IgG titers in the convalescent subjects. To 167 exclude this possibility, we measured anti-OC43 S IgG titers in sera collected from 20 healthcare 168 workers in 2020. The healthcare workers cared for hospitalized SARS-CoV-2 patients, but all were 169 negative for IgG against SARS-CoV-2 S and RBD, consistent with the effectiveness of personal 170 protective equipment and appropriate work practices. OC43 S-reactive IgG levels in healthcare 171 worker sera were similar to those in non-SARS-CoV-2-exposed healthy donor sera and 172 significantly lower than those in sera from convalescent subjects ( Figure 1C ). Taken together, our 173 results indicate that SARS-CoV-2 infection generates a strong IgG response that cross-reacts with 174 the S2 of human betacoronaviruses. 175

reactivity to the RBD and S2 subunit. PBMCs from non-SARS-CoV-2-exposed subjects and 177 convalescent subjects were analyzed for MBCs reactive to SARS-CoV-2 proteins. Circulating proportion of unexposed subjects suggested that IgG MBCs with the same specificity had also 190 been formed. However, these MBCs were not detected, possibly because of very low frequencies 191 in the circulation. In contrast, IgG MBCs reactive to the S proteins of the HCoVs OC43 and 229E 192 and the control proteins H1 and TTd were detected in nearly 50% or more of non-SARS-CoV-2-193 exposed subjects, consistent with the higher levels of serum IgG against these antigens ( Figure 2E -194 2H) . As expected, SARS-CoV-2 RBD-reactive MBCs were not detected in unexposed subjects. 195

In marked contrast to non-SARS-CoV-2-exposed subjects, the vast majority of 196 convalescent subjects had circulating IgG MBCs reactive to the SARS-CoV-2 S, RBD, and S2, 197 indicating strong induction by SARS-CoV-2 infection of MBCs reactive to novel and conserved 198 regions of the S protein ( Figure 2A) . Notably, numbers of IgG MBCs reactive to the S protein of 199 the HCoV OC43 were higher in convalescent subjects than in unexposed subjects ( Figure 1E generates IgG MBCs reactive to the SARS-CoV-2 S2 that cross-react with the S2 of human 207 betacoronaviruses. Interestingly, only a small proportion of the convalescent subjects generated 208 detectable N-reactive IgG MBCs, even though most subjects produced high levels of anti-N IgG 209 in serum (Figures 2C, 2D) . It is unclear whether this reflects a real difference between S-and N-210 reactive MBC formation or an effect of the sampling time. Overall, we demonstrate that SARS-211

CoV-2 infection induces strong S-reactive MBC formation that would be expected to provide 212 lasting protection against reinfection and potentially broad protection against betacoronaviruses. 213

Our goals in this study were to investigate SARS-CoV-2-reactive B cell memory in 216 unexposed subjects that could provide some protection against SARS-CoV-2 infection, and the 217 generation of B cell memory by SARS-CoV-2 infection that could provide lasting protection 218 against re-infection. In particular, we were interested in IgG MBCs, which respond to cognate 219 antigens with rapid, vigorous, and high-affinity Ab production. Importantly, MBCs are long-lived 220 cells that continue to provide strong protection when circulating Ab levels wane. Our approach 221 was to analyze circulating IgG as well as IgG MBCs from the SARS-CoV-2-naïve and SARS-222

CoV-2-convalescent subject groups. Our key findings are as follows: (i) the presence of IgG 223 reactive to the S2 subunit of SARS-CoV-2 in most unexposed subjects, likely reflecting cross-224 reactivity to HCoVs, (ii) markedly increased levels of IgG against the SARS-CoV-2 S and N 225 proteins, including reactivity to the RBD and S2 subunit of S, in convalescent subjects, (iii) 226

increased IgG binding to the S protein of the OC43 HCoV, but not 229E HCoV, in convalescent 227 subjects, reflecting greater cross-reactivity between S2 subunits of betacoronaviruses, (iv) strong 228 formation of IgG MBCs reactive with the RBD and S2 subunit of the SARS-CoV-2 S protein in 229 convalescent subjects, and (v) formation of IgG MBCs reactive with the S protein of OC43, but 230 not 229E, in convalescent subjects, consistent with S2 subunit cross-reactivity between 231

Approximately one-third of our cohort of non-SARS-CoV-2-exposed subjects had low 233 levels of IgG against the SARS-CoV-2 S and N proteins. The anti-N IgG likely reflects infection 234

with HCoVs, which have low level (20-30%) homology with the SARS-CoV-2 N protein (10). 235

However, a protective function for anti-N Abs has not been established (21). Notably, 86% of 236 unexposed subjects had IgG against the S2 subunit, reflecting homology with HCoVs, but none 237 had IgG against the highly novel SARS-CoV-2 RBD (6, 8, 22) . Abs that target the S2 subunit have 238 been shown to have virus neutralizing activity, raising the possibility that preexisting anti-S2 IgG 239 confers some protection against SARS-CoV-2 (23). The processes that generate anti-S2 IgG are 240 also likely to generate S2-reactive IgG MBCs and these might provide more significant protection 241 than low levels of anti-S2 Abs. However, S2-reactive MBCs (or S-reactive and N-reactive MBCs) 242

were not detected in non-SARS-CoV-2-exposed subjects. Taken together with the identification 243 of S-reactive MBCs in unexposed healthy donors (19), it is likely that S2-reactive MBCs were 244 below the limit of detection in our assays. Most MBCs are resident in lymphoid tissues, except for 245

MBCs against frequently seen immunogenic antigens (for example, the influenza H1 or TTd in 246 this study), and are at very low frequencies in circulation in steady state (24, 25) . 247

Anti-RBD, -S, and -N IgG levels were markedly higher in the convalescent subjects than 248 in non-SARS-CoV-2-exposed subjects, indicating strong induction by SARS-CoV-2 infection. 249

Perhaps notably, the majority of convalescent subjects had higher IgG titers against the S2 than 250 against the RBD. This is particularly surprising because of the accessibility of the RBD to B cells 251 and the expected immunodominance over the S2 subunit (26, 27). Our demonstration of strong 252 anti-S2 IgG production is consistent with the activation of a preexisting population of IgG MBCs 253 against the conserved S2 subunit in the absence of MBCs reactive to the novel RBD. However, we 254 cannot exclude inherent differences in the stability or antigenicity of RBD and S2 reagents as an 255 explanation. In convalescent subjects, IgG levels against the S protein of HCoV OC43 (but not 256 229E) were significantly higher than in non-SARS-CoV-2-exposed subjects and correlated 257 strongly with anti-S2 IgG levels. These findings support stronger B cell cross-reactivity between 258 the S2 subunits of SARS-Cov-2 and human betacoronaviruses than alphacoronaviruses (8). 259

Importantly, we demonstrate that SARS-CoV-2 infection generates RBD-reactive and S2-260 reactive IgG MBCs. Recently, Long et al. (4) found that levels of SARS-CoV-2-reactive Abs, 261 including neutralizing Abs, start to decrease within 8-12 weeks of infection, especially when the 262 infection is asymptomatic. Since MBC populations are maintained for many years, perhaps 263 decades, our findings indicate that MBCs generated by SARS-CoV-2 infection will be available 264 to rapidly generate protective Abs if waning Ab levels allow re-infection to occur (28). Notably, 265 three convalescent subjects in our analysis had undetectable RBD-reactive IgG, but nevertheless 266 had RBD-reactive IgG MBCs. This might reflect MBC production by germinal centers that 267 remained active after recovery from infection (29). The proportion of subjects with MBCs reactive 268 to the HCoVs OC43 and 229E was greater for the convalescent group than the unexposed group, 269 likely reflecting the increase in S2-reactive MBCs in the convalescent group and cross-reactivity 270 with HCoVs. S2-reactive MBC expansion by SARS-CoV-2 infection could enhance protection 271 against a broad range of coronaviruses (23). N-reactive MBC formation in convalescent subjects 272 was less than expected given the large number of subjects with high titers of N-reactive IgG, but 273 additional sampling times are required to confirm this observation. 274

In conclusion, our analysis investigated Ab and MBC immunity to SARS-CoV-2 in 275 unexposed subjects and individuals soon after recovery from SARS-CoV-2 infection. Findings 276 emphasized the novelty of the SARS-CoV-2 S protein RBD in unexposed subjects. However, IgG 277 reactive to the S2 was widespread in unexposed subjects and likely resulted from exposure to 278

HCoVs. Although our approach was unable to directly identify S2-reactive MBCs in the 279 unexposed subjects, we suggest that these cells are present and strongly contribute S2-reactive IgG 280 early in the response to SARS-CoV-2 infection. The IgG response in SARS-CoV-2 convalescent 281 subjects was also strong against the RBD and, less consistently, against the N protein. Importantly, 282 SARS-CoV-2 convalescent subjects had generated RBD-reactive and S2-reactive IgG MBCs. The May, 2020 and consisted of 22 PCR-confirmed patients and 5 non-PCR-confirmed subjects who 300 were contacts of confirmed cases or displayed COVID-19-like symptoms. The convalescent 301 subjects were sampled 4-9 weeks after symptom onset. Symptoms reported (percent of subjects) 302

were fever (67%) cough (74%), sore throat (48%), stuffy/runny nose (56%), difficulty breathing 303 (52%), fatigue (85%), headache (67%), body aches (67%), nausea/vomiting (19%), and 304 diarrhea/loose stool (41%). 305

(isolate Wuhan-Hu-1) were expressed in-house in HEK293 cells using pCAGGS plasmid 307 constructs kindly provided by Florian Krammer (Icahn School of Medicine at Mount Sinai) (7). 308

Baculovirus-expressed S2 subdomain and HEK293 cell-expressed N protein were obtained from 309

Sino Biological (Chesterbrook, PA) and RayBiotech (Peachtree Corners, GA), respectively. 310

Baculovirus-expressed S proteins from seasonal HCoVs OC43 and 229E were obtained from Sino 311

Biological. In-house HEK293 cell-expressed hemagglutinin from egg-derived H1N1 312 Mabtech Stockholm, Sweden) and p-nitrophenyl phosphate substrate (Thermo Fisher) were 337 subsequently added to detect bound antigen-specific Abs. Absorbance was read at 405 nm after 338 color development. A weight-based concentration method was used to quantify antigen-specific 339

Ab levels in test samples as described previously (30, 31) . Sera from healthy donors and 340 convalescent subjects with high titers for test antigens were used to establish human serum 341 standards. The cutoff for assay positivity was set at approximately 2x the mean OD value for 342 negative wells. 343 Statistical analyses. The medians with (q1, q3) were summarized by subject group and 344 compared by the Wilcoxon rank-sum test. Spearman correlation analysis together with 345 corresponding robust regression models was used to assess monotonic associations among Ab 346 responses. Multiple test adjustment was not applied for this explorative study and thus a P value 347 < 0.05 was considered significant for all analyses. Statistical analyses were performed using 348

Software SAS 9.4 (SAS Institute Inc, Cary, NC). CoV-2-exposed and COVID-19 convalescent subjects. Sera were collected from ( 2 proteins in non-SARS-CoV-2-exposed and COVID-19 convalescent subjects. PBMCs for MBC 522 analysis were collected from (i) 21 healthy donors sampled from 2011-14 (HD) and (ii) 26 523 COVID-19 convalescent subjects sampled 4-9 weeks after symptom onset (CONV). PBMCs were 524 stimulated in vitro to induce MBC differentiation into Ab-secreting cells. Antigen-specific 525

A pneumonia outbreak associated with a new coronavirus of probable bat origin

SARS-CoV-2 vaccines: Status report

Clinical features of patients infected with 369 2019 novel coronavirus in Wuhan

Clinical and 372 immunological assessment of asymptomatic SARS-CoV-2 infections

Genome composition and 376 divergence of the novel coronavirus (2019-nCoV) originating in China

Phylogenetic analysis 379 and structural modeling of SARS-CoV-2 spike protein reveals an evolutionary distinct and 380 proteolytically sensitive activation loop

A serological assay to detect 386

SARS-CoV-2 seroconversion in humans

Reimer 390

Presence of SARS-CoV-2 reactive T cells in COVID-19 patients and healthy donors

Infectious Diseases (except HIV/AIDS)

Pre-existing and de novo humoral immunity to SARS-CoV-398 2 in humans

Characterization 404 of a novel coronavirus associated with severe acute respiratory syndrome

Antibody response of patients with severe acute respiratory syndrome (SARS) targets the 408 viral nucleocapsid

Virological assessment of hospitalized 412 patients with COVID-2019

Targets of T 416 cell responses to SARS-CoV-2 coronavirus in humans with COVID-19 disease and 417 unexposed individuals

B cell responses: Cell interaction dynamics and decisions

Antibody responses to SARS-CoV-2 in patients with COVID-19

Kinetics of SARS-CoV-2 specific IgM and IgG responses in COVID-19 430 patients

COVID-19 serology at population scale: SARS-CoV-2-specific 435 antibody responses in saliva. Infectious Diseases (except HIV/AIDS)

Deep 439 sequencing of B cell receptor repertoires from COVID-19 patients reveals strong 440 convergent immune signatures

Broad neutralization of SARS-related viruses by human monoclonal antibodies

Convergent antibody responses to

SARS-CoV-2 in convalescent individuals

Contributions of the structural proteins of severe acute respiratory syndrome 458 coronavirus to protective immunity

Human monoclonal 466 antibodies against highly conserved HR1 and HR2 domains of the SARS-CoV spike 467 protein are more broadly neutralizing

The Transcription Factor T-bet Resolves Memory B Cell 472

Subsets with Distinct Tissue Distributions and Antibody Specificities in Mice and Humans

Broad dispersion and lung localization of virus-475 specific memory B cells induced by influenza pneumonia

A Sequence 478

Approach Can Predict Candidate Targets for Immune 479 Responses to SARS-CoV-2

The receptor binding domain of the viral spike protein is 484 an immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

Cutting edge: 487 long-term B cell memory in humans after smallpox vaccination

Role of Memory B Cells in Hemagglutinin-489

Specific Antibody Production Following Human Influenza A Virus Infection

Broad hemagglutinin-specific memory B cell 495 expansion by seasonal influenza virus infection reflects early-life imprinting and 496 adaptation to the infecting virus

Assignment of weight-based antibody units to a human antipneumococcal 499 standard reference serum, lot 89-S

individual HD and CONV subjects in order of ascending titers against S. The assigned cutoff for 509 positivity is shown by the shaded bar. (B) Proportions of serum IgG against the SARS-CoV

C) Serum IgG concentrations against the S protein 511 of the HCoV OC43 in CONV, HD, and HCW subjects. (D-F) Serum IgG concentrations against 512 the S protein of the HCoV 229E (D), the influenza virus H1 hemagglutinin (E), and TTd (F) in 513 CONV and HD subjects. (G) Correlation between serum IgG concentrations against the S2 subunit 514 of SARS-CoV-2 and the S protein of the HCoV OC43

001; ns [not significant]) for comparisons of serum IgG concentrations between subject 516 groups was determined by the Wilcoxon rank-sum test. Correlations were tested by Spearman 517 correlation analysis with corresponding robust regression models

quantitation of MBC-derived Ab (IgG)-secreting cells (MASCs) or MBC-derived polyclonal

MPAbs) provided a measure of the abundance of specific IgG MBCs. (A) IgG MBCs reactive 527 to the SARS-CoV-2 spike (S), receptor binding domain (RBD), and nucleocapsid (N) in CONV 528 subjects. MBC numbers were determined by enumeration of IgG MASCs by ELISpot essay after 529 in vitro MBC stimulation. The assigned cutoff for positivity is shown by the shaded bar

MBCs reactive to the influenza virus H1 hemagglutinin and TTd in CONV subjects. MBC 531 numbers were determined by enumeration of IgG MASCs. (C) Proportions of IgG MBCs reactive 532 to the SARS-CoV-2 RBD, S2, and N for individual CONV subjects. (D) Comparison of serum 533 IgG concentrations (upper panels) and IgG MBC numbers

CoV-2 S (left-hand side) and N (right-hand side) proteins. Serum IgG was measured by ELISA

Dilution curves are shown for 536 individual CONV subjects; curves for 4 subjects are shown in different colors to identify particular 537 response patterns