key: cord-0942662-6g7eeyjd
authors: Gattinger, Pia; Niespodziana, Katarzyna; Stiasny, Karin; Sahanic, Sabina; Tulaeva, Inna; Borochova, Kristina; Dorofeeva, Yulia; Schlederer, Thomas; Sonnweber, Thomas; Hofer, Gerhard; Kiss, Renata; Kratzer, Bernhard; Trapin, Doris; Tauber, Peter A.; Rottal, Arno; Körmöczi, Ulrike; Feichter, Melanie; Weber, Milena; Focke‐Tejkl, Margarete; Löffler‐Ragg, Judith; Mühl, Bernhard; Kropfmüller, Anna; Keller, Walter; Stolz, Frank; Henning, Rainer; Tancevski, Ivan; Puchhammer‐Stöckl, Elisabeth; Pickl, Winfried F.; Valenta, Rudolf
title: Neutralization of SARS‐CoV‐2 requires antibodies against conformational receptor‐binding domain epitopes
date: 2021-09-22
journal: Allergy
DOI: 10.1111/all.15066
sha: df8e15e70a285dbfff54bbb32f833f185d1c6d2d
doc_id: 942662
cord_uid: 6g7eeyjd

BACKGROUND: The determinants of successful humoral immune response to the severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) are of critical importance for the design of effective vaccines and the evaluation of the degree of protective immunity conferred by exposure to the virus. As novel variants emerge, understanding their likelihood of suppression by population antibody repertoires has become increasingly important. METHODS: In this study, we analyzed the SARS‐CoV‐2 polyclonal antibody response in a large population of clinically well‐characterized patients after mild and severe COVID‐19 using a panel of microarrayed structurally folded and unfolded SARS‐CoV‐2 proteins, as well as sequential peptides, spanning the surface spike protein (S) and the receptor‐binding domain (RBD) of the virus. RESULTS: S‐ and RBD‐specific antibody responses were dominated by immunoglobulin G (IgG), mainly IgG(1), and directed against structurally folded S and RBD and three distinct peptide epitopes in S2. The virus neutralization activity of patients´ sera was highly correlated with IgG antibodies specific for conformational but not sequential RBD epitopes and their ability to prevent RBD binding to its human receptor angiotensin‐converting enzyme 2 (ACE2). Twenty percent of patients selectively lacked RBD‐specific IgG. Only immunization with folded, but not with unfolded RBD, induced antibodies against conformational epitopes with high virus‐neutralizing activity. Conformational RBD epitopes required for protection do not seem to be altered in the currently emerging virus variants. CONCLUSION: These results are fundamental for estimating the protective activity of antibody responses after natural infection or vaccination and for the design of vaccines, which can induce high levels of SARS‐CoV‐2–neutralizing antibodies conferring sterilizing immunity.

The causative agent of coronavirus disease 2019 (COVID- 19) , SARS-CoV-2, is a ß-coronavirus closely related phylogenetically to previously identified pathogenic human coronaviruses that cause fatal respiratory disease in humans, severe acute respiratory syndrome coronavirus, SARS-CoV, and Middle East respiratory syndrome coronavirus, MERS-CoV. 1 Coronaviruses in general are responsible for substantial human and animal morbidity and mortality, and the potential for continued emergence of novel pathogenic coronaviruses from this class is highlighted by the relatively rapid appearance of three highly severe human diseases within two decades. 1,2 Wien, Grant numbers: COVID001 and COVID006. ITa was awarded an Investigator-Initiated Study (IIS) grant by Boehringer Ingelheim (IIS 1199-0424). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript proteins, as well as sequential peptides, spanning the surface spike protein (S) and the receptor-binding domain (RBD) of the virus.

Results: S-and RBD-specific antibody responses were dominated by immunoglobulin G (IgG), mainly IgG 1 , and directed against structurally folded S and RBD and three distinct peptide epitopes in S2. The virus neutralization activity of patients´ sera was highly correlated with IgG antibodies specific for conformational but not sequential RBD epitopes and their ability to prevent RBD binding to its human receptor angiotensin-converting enzyme 2 (ACE2). Twenty percent of patients selectively lacked RBD-specific IgG. Only immunization with folded, but not with unfolded RBD, induced antibodies against conformational epitopes with high virus-neutralizing activity. Conformational RBD epitopes required for protection do not seem to be altered in the currently emerging virus variants. 

IgG response in convalescent COVID-19 patients is directed to folded but not to unfolded RBD or RBD peptides. IgGs to folded RBD are required for virus neutralization. Twenty percent of convalescent COVID-19 patients selectively lack RBD-specific IgG. Only immunization with folded, but not with unfolded RBD, induces antibodies with virus-neutralizing activity.

SARS-CoV-2 binds to and enters human cells through an interaction between the receptor-binding domain (RBD) of S protein to angiotensin-converting enzyme 2 (ACE2). 3, 4 Potent neutralizing monoclonal antibodies against multiple epitopes on S have been isolated from convalescent patients, 5 and recent studies have shown that human antibodies can be effective for the treatment of COVID-19. 6, 7 Features and duration of patients´ antibody responses have been studied 8, 9 but a comprehensive characterization of the attributes of a protective polyclonal antibody response to SARS-CoV-2 and the prerequisites for the induction of such an antibody response by vaccination have not been reported to our knowledge. Understanding the natural polyclonal antibody response after COVID-19 may guide the design of additional vaccines capable of eliciting a SARS-CoV-2-specific sterilizing immunity for creating herd immunity 10, 11 and may help to identify antigenic features that may, if varied, allow viral escape from immune surveillance. This study reports the mapping of the polyclonal antibody responses in a large number of clinically well-characterized convalescent COVID-19 patients with a comprehensive panel of microarrayed folded and unfolded SARS-CoV-2 proteins and S-derived peptides in relation to their virus neutralization activity and ability to inhibit the RBD-ACE2 interaction. Experimental antibody responses induced by immunization with folded or unfolded RBD were then analyzed for neutralization potential. A polyclonal antibody response against conformational RBD epitopes is required for highly effective neutralization of SARS-CoV-2, and induction of this response requires immunization with folded RBD.

Between 29 April 2020 and 30 July 2020, 253 COVID-19convalescent patients were enrolled in this study. During the same period, 235 control subjects (CS), who according to self-assessment had no COVID-19-or common cold-like symptoms 10 weeks before and had a negative RT-PCR test at the time of their visit, were enrolled. All subjects gave their written informed consent. Visits and collection of blood samples in COVID-19 patients were performed 8 weeks (mean 61 days, SD ± 13.7, 19-98 days) after positive SARS-CoV-2 RT-PCR test. COVID-19 patients were grouped according to mild and severe symptoms ( Figure 1A ). Patients with mild COVID-19 had recovered at home, whereas the severe symptom group had been hospitalized during the acute phase, where they received oxygen supply or were treated in the intensive care unit (ICU).

COVID-19 symptoms and comorbidities were assessed by questionnaire 17, 18 (Table S1 , Supporting information Methods S1 

Immunoglobulin (Ig) response of human or rabbit serum samples of COVID-19-convalescent patients and healthy control sera to SARS-CoV-2-derived proteins was determined by enzyme-linked immunosorbent assay (ELISA) as previously described. 19, 20 Details can be found in the articles' Supporting information . Slides were again washed, dried, and subsequently scanned using a confocal laser scanner (Tecan, Männedorf, Switzerland). Image analysis was performed by MAPIX microarray image acquisition and analysis software (Innopsys, Carbonne, France), and conversion of measured fluorescence units to ISAC standardized units (ISU) was performed as described. 21, 22 For microarrayed inhibition experiments, human sera were diluted 1:100, rabbit sera 1:8000 in sample diluent and pre-incubated overnight with either folded RBD, unfolded RBD, unfolded S1 (10 µg/ml) or with an equimolar amount of a RBD-derived peptide mix comprising peptides 13-21 (Table S4, Microarrayed measurements and analysis were performed as described above. Details on rabbit immunization can be found in the articles' Supporting information.

The molecular interaction assay to detect inhibition of RBD to ACE2 receptor binding by patients´ sera was performed as described. 19 Shortly, 1:2 diluted serum was incubated for 3 h with HEK cell-expressed His-tagged RBD followed by a 3 h overlay onto plate-bound ACE2. Bound RBD was then detected with a mouse monoclonal anti-His antibody followed by a HRP-labeled anti-mouse IgG 1 antibody and detected with ABTS. All measurements were performed in duplicates with a variation of <5%. The SARS-CoV-2 virus neutralization test was carried out as described. 23 Twofold serial dilutions of heat-inactivated serum samples were incubated with 50-100 TCID50 SARS-CoV-2 for 1 h at 37℃. The mixture was added to Vero E6 cell (ATCC ® CRL-1586) monolayers, and incubation was continued for three days at 37℃. Virus neutralization titers (VNTs) were expressed as the reciprocal of the serum dilution that protected against virus-induced cytopathic effects. VNT titers ≥10

were considered positive.

Surface representation of SARS-CoV-2 spike protein was generated in PyMOL (PyMOL Molecular Graphics System, Version 2.5.0a0, Schrödinger, LLC) based on the PDB entry 6XR8. Mutations in RBD known at the date of submission were derived from https://spike mutan ts.exsca late4 cov.eu/.

No statistical methods were used to predetermine sample sizes. The experiments were not randomized. Investigators were blinded during experiments regarding demographic and clinical data with the samples having de-identified subjects IDs that did not contain any of this information. All statistical analyses were performed using GraphPad Prism Version 5.00 (La Jolla, CA, USA).

patients and healthy controls or mild and severe symptom groups were tested with chi-squared test. Differences in immunoglobulin reactivity to proteins or peptides were determined using two-tailed

Mann-Whitney U test. Correlations of immunoglobulin reactivity and virus neutralization titers were assessed by Spearman´s rank correlation coefficient. P values of <.05 were considered as significant.

From 29 April 2020 to 30 July 2020, 253 COVID-19-convalescent patients with positive SARS-CoV-2 RT-PCR test and/or positive antibody tests and 235 age and gender-matched control subjects who had no signs of COVID-19 or common-cold-like symptoms were enrolled in this study ( Figure 1A , Methods S1). Fatigue, myalgia, and anosmia were significantly more frequent in the mild group (59.7%) than in the severe group (42.2%).

The percentages of patients suffering from IgE-associated allergy were similar among patients with mild and severe COVID-19 and the control group (Table S1 Supporting information Methods S1).

The 235 control subjects had a negative SARS-CoV-2 RT-PCR test at the time of investigation and no common cold-like symptoms in the 10 weeks before the visit. Overall, the prevalence of malignancies, endocrine or circulatory comorbidities were significantly higher in COVID-19 patients as compared to control individuals (Table   S1 ). Blood samples were collected from COVID-19-convalescent patients approximately 8 weeks (mean 61 days, SD ± 13.7, min.

19 days, max. 98 days) after the positive SARS-CoV-2 RT-PCR test, which ensured that they had seroconverted and were already in the plateau phase of antibody production. 8, 25 To discriminate between antibodies specific for SARS-CoV-2 and those acquired through earlier infections by common cold-inducing coronaviruses, sera obtained before the occurrence of COVID-19 (ie, 1996-summer 2019, historic controls) from 38 age-matched control subjects were included in the analyses ( Figure 1A , Table S2 Supporting information Methods S1).

In (Table S3 Supporting information Methods S1). S-derived peptides of approximately 30 amino acids in length spanning the S-protein and in particular RBD were included ( Figure 1B , Table S4 Supporting information Methods S1). The analysis of the surface exposure of RBD-derived peptides indicated that the percentages of surface-exposed amino acids were highest for peptides 13-15 and 18-20 (Table S4 Supporting information Methods S1).

Peptides that were not adjacent in the RBD sequence (eg, peptides 18 and 20) could appear in close vicinity on the RBD surface ( Figure 1C ).

In a first set of experiments, we measured IgG, IgG subclasses, IgM, and IgA levels specific for folded S and RBD in the complete population of COVID-19-convalescent patients (mild: n = 139; severe; n = 114) and in the 235 control subjects by ELISA ( Figure 1A , Figure   S1A -B). Severe COVID-19 patients had significantly higher IgG, IgM,

and IgA levels to S and RBD compared to mild COVID-19 patients ( Figure S1A ). S-and RBD-specific IgG levels were higher than IgM levels, whereas few patients mounted low IgA responses ( Figure   S1A ). No significant correlations between S-and RBD-specific IgG, IgM, and IgA responses were found ( Figure S2A ).

In the control group, 7.6% (n = 18) had IgG to either S and/or RBD ( Figure S1 ). Eleven subjects had COVID-19-like symptoms longer than 10 weeks before the visit, whereas in 7 subjects (ie, 2.9%), no symptoms at all had been reported indicating a previous asymptomatic infection.

IgG subclass analysis revealed a predominant IgG 1 response to S and RBD with significantly higher IgG 1 levels in patients with severe COVID-19, compared to mild COVID-19 patients ( Figure S1B ). In 23 COVID-19 patients, a weak S-specific IgG 2 response was found but no S-specific IgG 3 or IgG 4 could be detected ( Figure S1B ). S-and RBD-specific IgG levels were significantly correlated with IgG 1 but not IgG 2 levels ( Figure S2B ).

Out of the 253 COVID-19 patients 53 (ie, 20.9%) lacked RBD-specific IgG antibodies ( Figure S1 ). Among these RBD 

For the assessment of antibody reactivity to a comprehensive panel of SARS-CoV-2 proteins and S-derived peptides, we used microarray technology 21 ( Figure 1D ). Figure Figure 2 shows that VNTs were associated with IgG titers to folded S, S1 and in particular to folded RBD. IgG levels significantly increased with VNTs and were as follows: VNT 10-80: mean S-specific IgG: 21.1 ISU; mean S1-specific IgG: 3.7 ISU; mean *** *** *** * ** *** *** *** *** *** *** *** *** *** *** * ** *** *** *** *** *** *** *** *** ***

RBD-specific IgG: 54.4 ISU; VNT 120-240: mean S-specific IgG: 42.1 ISU; mean S1-specific IgG: 10.1 ISU, mean RBD-specific IgG: 84.8 ISU; VNT 320-640: mean S-specific IgG: 54.4 ISU; mean S1specific IgG: 15.4 ISU: mean RBD-specific IgG: 93.1 ISU (Figure 2A ).

High and significant correlations between VNTs and IgG levels to folded S, S1, S2, and RBD but not with IgG levels to unfolded S1, S2, or RBD were found ( Figure 3A , Figure S5 ). For RBD-derived pep- IgG levels ( Figure 3A ).

Specific IgG levels greater than 15 ISU and associations of specific IgG levels with VNTs were also found for NP which does not play a role in virus neutralization (Figure 2A , Figure S5 ) and for three S2-derived peptides (ie, peptides 25, 33, and 46) ( Figure 2B ) which are outside RBD and hence are not directly involved in binding of RBD to ACE2.

Since VNTs were significantly correlated with levels of IgG antibodies to folded RBD in COVID-19 patients, we analyzed whether VNTs are associated with the ability of patients' sera to inhibit the binding of RBD to ACE2. Figure 3B shows that there is indeed a highly significant correlation of VNTs with the inhibition of the binding of RBD to ACE2 in sera from COVID-19

patients. Figure 3C shows the analysis of the ability of sera from 233 COVID-19 patients ( Figure 1A) to block the binding of RBD to ACE2. We found a median inhibition of RBD binding to ACE2 of 24% for this population. For 19.2% of patients, a greater than 50%

inhibition was found, in 38.4% of patients inhibitions ranged from 20 to 50% and a less than 20% inhibition occurred in 42.4% of the patients.

Together, these results demonstrate that neutralization of SARS-CoV-2 is associated with high levels of IgG antibodies against conformational epitopes of folded RBD and their ability to inhibit the binding of RBD to ACE2. However, the ability of patients'

antibodies to inhibit RBD binding to ACE2 varied considerably ( Figure 3C ). 

In order to further investigate the importance of conformational versus sequential RBD epitopes for the binding of patients´ IgG to RBD, inhibition experiments were performed. Patients' sera were pre-incubated either with folded RBD containing conformational epitopes, with unfolded S1 or a mix of RBD-derived peptides containing sequential epitopes. For control purposes, an unrelated protein (bovine serum albumin, BSA) was used. Then, IgG binding of pre-adsorbed sera to folded RBD, folded S, unfolded S1, unfolded RBD, and the RBD-derived peptides 13-21 was measured (Figure 4 ).

Only pre-incubation of sera with folded RBD, but not with unfolded S1 or the RBD-derived peptide mix significantly inhibited IgG binding to conformational epitopes on RBD and reduced IgG binding to folded S ( Figure 4A ). Some non-significant reduction in the low IgG binding to unfolded RBD was observed after pre-incubation of sera with folded RBD, unfolded S1, and the RBD peptide mix ( Figure 4A ).

Also, a non-significant reduction in the low IgG binding to unfolded S1 was observed by pre-adsorption with unfolded S1 and the RBD peptide mix ( Figure 4A ). Pre-incubation of sera with the RBD peptide mix reduced the low IgG binding to the individual RBD-derived peptides 13-21 with significant reductions observed for peptides 13, 17, 18, 20, and 21 ( Figure 4B ).

Immunization with denatured, synthetic or recombinant unfolded antigens can be used to induce antibodies recognizing the corresponding folded antigen to prevent and/or treat infectious diseases and allergy. [26] [27] [28] [29] We were therefore interested to study whether immunization with unfolded RBD could induce IgG antibodies against folded RBD, which exhibit high virus-neutralizing activity. Groups of rabbits were immunized with three doses (20, 40, or 80 microgram) of adjuvanted unfolded or folded RBD and for control purposes with adjuvant alone. Immunization with unfolded RBD induced IgG reactivity to unfolded S1 and unfolded RBD but almost no IgG responses against folded RBD ( Figure S6A ), whereas immunization with folded RBD induced strong IgG production against folded RBD but almost no IgG antibodies against unfolded RBD and unfolded S1 ( Figure S6B ). No relevant IgG responses were observed in rabbits immunized with folded or unfolded RBD to an unrelated control antigen (HHM0) (Figure S6A -B) and no IgG response to any of the tested antigens was observed in rabbits immunized with adjuvant alone ( Figure S6C ). The IgG reactivity of rabbits immunized with folded RBD to conformational epitopes on folded RBD and folded S was only inhibited by pre-adsorption with folded RBD but not with unfolded S1 or RBD-derived synthetic peptides containing only sequential epitopes ( Figure S6D ). The low IgG binding of rabbits immunized with unfolded RBD to unfolded proteins and RBD-derived peptides was only inhibited with unfolded S1 and/or RBD-derived peptides ( Figure S6E ).

We then tested rabbit antisera obtained after the second and third immunization with folded or unfolded RBD for their VNTs (Table S6 Supporting information Methods S1). With folded RBD (40 and 80 μg), VNTs between 240->1280 were obtained after the third immunization whereas unfolded RBD failed to induce any VNT (Table S6 Supporting information Methods S1). These results demonstrate that folded RBD containing conformational epitopes is required to induce high VNTs upon immunization.

The We collected blood samples approximately 8 weeks after acute infection when specific antibodies are in the plateau phase. 25, 30, 31 The analysis of sera from 253 COVID-19-convalescent patients showed that the antibody response against the spike protein and RBD is dominated by the IgG isotype, in particular by the IgG 1 subclass which is in agreement with an earlier report 32 and is similar as was found for other respiratory viruses (eg, RV and RSV) which did not induce IgG 3 and IgG 4 responses. 20,33 S-and RBD-specific IgG antibodies indicative of an asymptomatic infection were observed in 2.9% of our control cohort, and their levels were lower than those in the patients with mild or severe symptoms (data not shown). We did not have access to mucosal secretions but several studies have already investigated the specific immunoglobulin response in nasal secretions, tears, and stool indicating that SARS-CoV-2-specific IgA antibodies may occur in such samples in addition to IgG antibodies. [34] [35] [36] Using microarrayed folded S, folded and unfolded portions of the spike protein (S1, S2, and RBD), and synthetic peptides spanning S, the present work shows that VNTs in patients´ sera are highly correlated with the levels of IgG antibodies directed against conformational but not sequential RBD epitopes. In fact, the localization of the RBD-derived peptides in the three-dimensional structure of RBD shows that non-adjacent RBD-derived peptides appear in close vicinity on the RBD surface, which is required for the formation of conformational epitopes of the discontinuous type.

The finding that the majority of the SARS-CoV-2-neutralizing activity of the polyclonal antibody response in COVID-19convalescent patients can be attributed to IgG antibodies directed against conformational but not against sequential RBD epitopes is important because so far only 3 mutations (E484K, N501Y, and K417N) have been observed in the RBD of currently reported SARS-CoV-2 variants (https://spike mutan ts.exsca late4 cov.eu/) of which only one (ie, E484K) appears on the RBD surface ( Figure   S7 ) but does not seem to be involved in the ACE2 interaction. It is thus likely that IgG antibodies from COVID-19-convalescent patients directed to the conformational RBD epitopes will crossreact with the currently emerging SARS-CoV-2 variants and confer cross-protection.

Another interesting result of our study is that 20% of patients Lack of RBD-specific IgG in the non-responders did not seem to be a factor for severe disease because we found that the majority of the RBD non-responders (ie, 75.5%) had mild COVID-19. This may be due to low virus exposure of these subjects, sufficient early RBD-specific IgM responses, and/or a highly potent specific cellular immunity. Since we observed that, significantly more women than men are RBD non-responders, it will be of great interest to conduct further studies regarding the underlying mechanisms (eg, genetically determined differences regarding antigen presentation 19 Accordingly, antibodies against conformational RBD epitopes capable of inhibiting the RBD-ACE2 interaction seem to be an important parameter for the assessment of a protective SARS-CoV-2-specific immunity after disease or vaccination.

Other reports have shown that monoclonal antibodies or enriched antibody fractions specific for epitopes outside RBD or sequential epitopes may have SARS-CoV-2-neutralizing activity. 7, 37 Although this information is valuable for the creation of therapeutic antagonists of the virus, it is of less certain relevance for the development of effective vaccine strategies, for which an understanding of the natural pattern of neutralizing responses and their therapeutic implications is important. In this context, it is worth mentioning that our analysis also confirmed the presence of low antibody responses against certain SARS-CoV-2 peptides in sera from historic controls obtained before the COVID-19 pandemic. 38 Our result that the majority of virus-neutralizing activity in sera of SARS-CoV-2 patients can be attributed to antibodies against conformational RBD epitopes is supported by a study performed for SARS-CoV 39 which, similar to SARS-CoV-2, also binds with RBD to ACE2. Also for SARS-CoV, it has been demonstrated that the spike protein contains conformational epitopes which induce highly potent-neutralizing antibodies. 39 Furthermore, it has been shown that vaccines targeting the RBD of SARS-CoV-2 induce protective immunity. 40 However, for SARS-CoV, it has been reported that potent-neutralizing antibodies and protective immunity can be obtained by immunization with RBD expressed in a folded form in eukaryotic cells as well as with unfolded RBD, Escherichia coli-expressed RBD. 41 These results were consistent with data obtained for several vaccines for other infectious diseases and therapeutic vaccines for allergy demonstrating that one can induce protective antibody responses against the corresponding natural, folded antigen resembling conformational epitopes with the denatured antigens, the unfolded recombinant antigen, or sequential peptides thereof. 15, 16, [26] [27] [28] [29] Conversely, it has been suggested for certain viral diseases that immunization with correctly folded antigens is required for obtaining protective antibody responses. 12, 13 In order to compare antibody responses obtained by immunization with folded versus unfolded RBD and their virus-neutralizing activity, rabbits were immunized with a folded and unfolded recombinant RBD protein. Only immunization with folded but not with unfolded RBD induced antibodies against conformational RBD epitopes and high VNTs.

Collectively, our data demonstrate that the virus-neutralizing activity of antibodies in COVID-19 patients depends on the presence of antibodies directed to conformational epitopes of RBD, which do not seem to be altered in currently known mutated SARS-CoV-2 variants ( Figure S7 ). However, not all COVID-19 patients develop these antibodies. Importantly, the induction of such antibodies by vaccination requires folded RBD. Thus, our results suggest that antibodies against conformational RBD epitopes are a surrogate marker for a SARS-CoV-2-neutralizing antibody response and are important for the development of SARS-CoV-2-specific vaccines capable of inducing sterilizing immunity.

We wish to thank Jutta Hutecek, Paul Ettel, Fritz Tuppy, Katharina Grabmeier-Pfistershammer, Lisabeth Pimenov for excellent technical assistance. Furthermore, we are grateful to Brian Seed (Harvard) and Theodore Jardetzky (Stanford) for critically reading the manuscript and useful suggestions. We wish to acknowledge the help of Doris Werjant-Locmele and Anna

Guentcheva regarding the recruitment and administration of study subjects. We are grateful to all individuals who participated in our study. 

The hallmarks of COVID-19 disease

The origin of COVID-19 and why it matters

Structure of SARS coronavirus spike receptor-binding domain complexed with receptor

Coronavirus biology and replication: implications for SARS-CoV-2

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

Convalescent plasma treatment of severe COVID-19: a propensity score-matched control study

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

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

Mapping neutralizing and immunodominant sites on the SARS-CoV-2 spike receptorbinding domain by structure-guided high-resolution serology

SARS-CoV-2 vaccines in development

A guide to vaccinology: from basic principles to new developments

Structure-based design of a fusion glycoprotein vaccine for respiratory syncytial virus

De novo protein design enables the precise induction of RSV-neutralizing antibodies

Medical virology of hepatitis B: how it began and where we are now

Quantification, epitope mapping and genotype cross-reactivity of hepatitis B preSspecific antibodies in subjects vaccinated with different dosage regimens of BM32

Hepatitis B and D viruses exploit sodium taurocholate co-transporting polypeptide for species-specific entry into hepatocytes

Immunological imprint of COVID-19 on human peripheral blood leukocyte populations

Cardiopulmonary recovery after COVID-19 -an observational prospective multi-center trial

Antibodies in serum of convalescent patients following mild COVID-19 do not always prevent virus-receptor binding

Features of the human antibody response against the respiratory syncytial virus surface glycoprotein G. Vaccines (Basel)

PreDicta chip-based high resolution diagnosis of rhinovirus-induced wheeze

HIV microarray for the mapping and characterization of HIV-specific antibody responses

Dynamics of CD4 T cell and antibody responses in COVID-19 patients with different disease severity

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

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

Antigen-induced in vitro antibody production in humans: tetanus toxoid-specific antibody synthesis

Synthetic approaches to vaccines for infectious and autoimmune diseases

Studies on "allergoids" prepared from naturally occurring allergens. I. Assay of allergenicity and antigenicity of formalinized rye group I component

The future of antigen-specific immunotherapy of allergy

Evaluation of serum IgM and IgG antibodies in COVID-19 patients by enzyme linked immunosorbent assay

Persistence of SARS-CoV-2-specific antibodies in COVID-19 patients

A comprehensive, longitudinal analysis of humoral responses specific to four recombinant antigens of SARS-CoV-2 in severe and non-severe COVID-19 patients

Rhinovirusinduced VP1-specific antibodies are group-specific and associated with severity of respiratory symptoms

SARS-CoV-2 RNA and antibodies in tear fluid

Systemic and mucosal antibody responses specific to SARS-CoV-2 during mild versus severe COVID-19

Limited intestinal inflammation despite diarrhea, fecal viral RNA and SARS-CoV-2-specific IgA in patients with acute COVID-19

Linear epitopes of SARS-CoV-2 spike protein elicit neutralizing antibodies in COVID-19 patients

Viral epitope profiling of COVID-19 patients reveals cross-reactivity and correlates of severity

Receptor-binding domain of severe acute respiratory syndrome coronavirus spike protein contains multiple conformation-dependent epitopes that induce highly potent neutralizing antibodies

A vaccine targeting the RBD of the S protein of SARS-CoV-2 induces protective immunity

Recombinant receptor-binding domain of SARS-CoV spike protein expressed in mammalian, insect and E. coli cells elicits potent neutralizing antibody and protective immunity

Additional supporting information may be found in the online version of the article at the publisher's website.