Identi cation of four linear B-cell epitopes on the SARS-CoV-2 spike protein able to elicit neutralizing antibodiesIdenti cation of four linear B-cell epitopes on the SARS-CoV-2 spike protein able to elicit neutralizing antibodies


 SARS-CoV-2 unprecedentedly threatens the public health at worldwide level. There is an urgent need to develop an effective vaccine within a highly accelerated time. Here, we present the most comprehensive S-protein-based linear B-cell epitope candidate list by combining epitopes predicted by eight widely-used immune-informatics methods with the epitopes curated from literature published between Feb 6, 2020 and July 10, 2020. We find four top prioritized linear B-cell epitopes in the hotspot regions of S protein can specifically bind with pooled serum antibodies from horses, mice, and monkeys inoculated with different SARS-CoV-2 vaccine candidates or five patients recovering from COVID-19. The four linear B-cell epitopes can induce neutralizing antibodies against both pseudo and live SARS-CoV-2 virus in immunized wild-type BALB/c mice. This study suggests that the four linear B-cell epitopes are potentially important candidates for serological assay or vaccine development.

retained for downstream analysis with intracellular epitopes eliminated. The linear B-cell epitopes consisting of less than six amino acids or more than 50 amino acids were further removed. The antigenicity of the remained linear B-cell epitopes was evaluated by VaxiJen 2.0 [52]. A stringent criterion was employed to have linear B-cell epitopes with an antigenicity score larger than 0.9 viewed adequate to initiate a defensive immune reaction.
Identi cation of hotspot regions enriched with highly antigenic linear B-cell epitopes The distribution of linear B-cell epitopes with high antigenicity scores (larger than 0.9) were drawn along the S protein. The regions with coverage of epitopes larger than 11 were determined as the peaks. The peaks locating on the non-outer surface were excluded, based on the transmembrane topology of SARS-CoV-2 S protein predicted by TMHMM v2.0 (outside:1-1213; transmembrane:1214-1236; inside: 1237-1273). The enrichment ratio was calculated as (The number of epitopes in the peaks/the total length of peaks)/(the number of epitopes in the non-peak regions/the total length of non-peak regions). A permutation-based method was adopted to calculate the signi cance of the enrichment ratio. Speci cally, for each round, the positions of all epitopes were randomly assigned along the S protein, followed by calculation of the enrichment ratio. A total of 1000 rounds were performed. The signi cance was determined as the number of rounds where the random enrichment ratio larger than the real enrichment ratio divided by the total number of rounds.

Characterization of the 18 selected linear B-cell epitopes
The target pro les of IgG or IgA antibodies from 232 COVID-19 patient sera (101 for hospitalized, 131 for non-hospitalized) and 190 pre-COVID-19 era controls were generated in duplicates by recent research [53], using coronavirus 20-mer libraries which included 20-mer peptides tiling every 5 amino acids across the SARS-CoV-2 proteome. The target pro les for IgG/IgA IPs using coronavirus 20-mer libraries were in the form of Z-score, directly obtained from the literature [53]. The Z-scores of the 20-mer peptides in the target pro les were rst averaged for the two duplicates of each sample. For each of the 18 linear B-cell epitopes, the Z-scores of 20-mer peptides with more than 75% overlapping with the linear B-cell epitope were compared between hospitalized, non-hospitalized COVID-19 patient groups and negative controls using Mann Whitney U test (two-sided). The signi cance was determined if the p value was less than or equal to 0.05.

Conservation analysis of selected B-cell epitopes
The conservation status for each residue of SARS-CoV-2 were investigated by ConSurf [66] using the amino acid sequences of S protein from seven known coronaviruses including SARS-CoV-2

Peptide Synthesis
The selected 18 linear B-cell epitopes were synthesized by Scilight-Peptide Inc., Beijing, China via a practical approach of Fmoc solid-phase peptide synthesis. The unsophisticated peptides were puri ed using a Varian ProStar 218 high-performance liquid chromatography (HPLC) instrument with an Agilent Venusil MP C18 reversed phase column. Peptides were eluted with a linear gradient of water, H2O, and acetonitrile, CAN, (both having 0.05% TFA) at a ow rate of 1 mL/min. The separation was monitored at 220 nm using UV detection. Then peptides were subjected to Voyager-DE STR mass spectrometric (MS) analysis. The solvents for gradient elution HPLC are: solvent A, CAN 2%, TFA 0.05% and solvent B, CAN 90%, TFA 0.05%. Peptides were dissolved in deionized H2O at a nal concentration of 20mg/ml and stored at -20 o C until further use.
The experiments with infectious SARS-CoV-2 were performed at the biosafety level 3 facilities in Beijing Institute of Microbiology and Epidemiology, Academy of Military Sciences, China.

Animal experiments
All animal experiments were approved by and carried out in accordance with the guidelines of the Institutional Experimental Animal Welfare and Ethics Committee. A total of 30 speci c-pathogen-free female BALB/c mice aged 6-8 weeks were obtained from Beijing Vital River Laboratory Animal Technologies Co., Ltd (Beijing, China) and were housed and bred in the temperature-, humidity, and light cycle-controlled animal facility (20 ± 2°C; 50 ± 10%; light, 7:00-19:00; dark,19:00-7:00). Five rhesus monkeys (2.0 ± 0.5 kg, 1.5-2.5 years) were provided by Laboratory Animal Center, Academy of Military Medical Science. The animals were housed under standard laboratory conditions. Five 4-6 year old, healthy brown horses (300-350kg in weight) that had no detectable antibodies against SARS-CoV-2, were provided by Chifeng Boen Pharmacy Co., LTD (InnerMongolia, China).
Ten Balb/C mice and the ve monkeys were intramuscularly (i.m.) immunized with the RBD-IgG1 Fc subunit vaccine candidate (10µg per mouse, 40µg per monkey) ( ve for mouse, three for monkey), or Al(OH)3 adjuvant as a control( ve for mouse, two for monkey). After the primary injection, all animals received two booster injections with 14 days intervals. The ve horses were inoculated subcutaneously with SARS-CoV-2 Spike Protein (S1 Subunit, His tag) (Sino Biological, China, cat no:40591-V08H) antigens containing 1.0mg, 1.5mg, 2.0mg, and 3.0mg on day 0, 14, 28, and 42, respectively (three horses), or Freund′s adjuvant as a control (two horses). For each inoculation, 5ml of antigen suspension was mixed with an equal volume of Freund′s adjuvant according to the directions and injected into the several sites near the submandibular and inguinal lumph nodes. Blood samples of were collected and pooled sera of each model animal were used for ELISA. SARS-CoV-2 S1-speci c IgG assay in horse, mouse, monkey, and human (Indirect ELISA) 96-well polystyrene microplates (Oriental Ocean Global Health, China) were coated with 2 µg/mL (50µL/well) SARS-CoV-2 Spike Protein (S1 Subunit, His tag) (Sino Biological, China, cat no:40591-V08H) in carbonate bicarbonate buffer pH9.6 and the plates were incubated at 4 o C overnight. The plates were then blocked at 37°C for one hour with PBS (Solarbio, China, cat no: A8020) pH 7.4 in 5% skim milk (blocking buffer) and washed with PBST (0. Vaccination of mice Linear B-cell epitopes ('YNSASFSTFKCYGVSPTKLNDLCFT', 'GDEVRQIAPGQTGKIADYNYKLP', 'YQPYRVVVLSFELLH', and 'CVNFNFNGL') were respectively used to immunize BALB/c mice (n = 5 per linear B-cell epitope) through ve consequent subcutaneous injections of ve microgram dose with 7 days interval between two consecutive injections. Sera were collected for SARS-CoV-2 live virus neutralization assay titration (NT 50 ) and SARS-CoV-2 pseudovirus neutralization assay titration (EC50) at different time point.

SARS-CoV-2 pseudovirus neutralization assay
A pseudotyped virus-based neutralization assay against SARS-CoV-2 in biosafety level 2 facilities was performed as previously described [61]. Brie y, serial dilutions of the samples to be tested were mixed with 325-1,300 TCID 50 /ml of pseudotyped virus. The target cells were incubated for 24 hours, and then the neutralizing antibody content of the sample was obtained by calculating the amount of pseudotyped virus entering the target cells which were detected by the expression of luciferase. The half maximal effective concentration (EC 50 ) was calculated for the tested samples using the Reed-Muench method in

SARS-CoV-2 neutralization assay
Microneutralization (MN) assay was performed to assess the neutralizing activity of sera from the mice. 50µl (100 CCID 50 /0.05ml) of SARS-CoV-2 IME-BJ01 strain was incubated with serial dilution of heatinactivated sera in 5% CO 2 environment at 37 o C for one hour. The complexes of antibody-virus (100TCID50/50µl) were added to pre-plated Vero cell monolayers in 96-well plates and incubated for 72 hours. The Reed-Muench method was applied to estimate the dilution of sera required for NT 50 . The initial dilution of sera (1:16) was set as the con dence limit of the assay. Seropositivity was de ned as a titre ≥ 16.

Results
Identi cation of linear B-cell epitopes on the SARS-CoV-2 Spike protein Spike protein is an important target for vaccine development due to its indispensable function in helping SARS-CoV-2 gain entry into host cells. B-cells can be guided through linear B-cell epitopes to recognize and activate defense responses against viral infection. To construct a comprehensive linear B-cell epitope candidate list, we rst performed in-silicon prediction of B-cell epitopes from S protein through eight methods, obtaining a total of 4044 linear B-cell epitopes (256 for Bepipred and Bepipred2.0 with default parameter settings, Kolaskar and Tongaonkar antigenicity, Parker hydrophilicity, Chou and Fasman betaturn, and Karplus and Schulz exibility provided by IEDB (Immune-Epitope-Database And Analysis-Resource) [44]; 128 for BcePred [45] using accessibility, antigenic propensity, exposed surface, exibility, hydrophilicity, polarity, and turns; 3007 for the ANNpred-based server ABCpred [46]; 44 for Ellipro [47]; 176 for BCPREDS [48]; 191 for AAP [49]; 215 for FBCPRED [50]; and 27 for COVIDep [51]). We additionally extracted 279 linear B-cell epitopes from 25 articles or preprints  published between Feb 6, 2020, and July 10, 2020. We nally established a full list of 3836 unique linear B-cell epitope candidates by combining predictions by the eight methods and those curated from literature (Supplementary Table 1).
To obtain linear B-cell epitopes with high potential to initiate a defensive immune reaction, we adopted a series of high stringent criteria to lter out epitopes with low antigenicity. 614 linear B-cell epitopes were estimated by VaxiJen 2.0 [52] to have high antigenicity scores (larger than 0.9 viewed adequate to initiate a defensive immune reaction) (Fig. 1A) with length varying from 6 amino acids to 30 amino acids ( (AAT98580.1), we discovered that RBD region (319-514) of the S protein was not conserved among the seven coronaviruses, consistent with the observation that more mutations were found in the RBD region by performing sequence alignment of S protein against 118,694 (20200927) sequences of SARS-CoV-2 in the NGDC database (Fig. 1C). The linear B-cell epitopes from the 14 hotspot regions were mapped to the 3D structure of the SARS-CoV-2 S protein (PDB ID:6VSB), suggesting that hotspot regions locating on the exposed area of spike stem or spike head would harbor good B-cell epitope candidates (Fig. 1D and E).
To evaluate the stability of the 18 linear B-cell epitopes, we estimated the number of peptide-digesting enzymes through the protein digest server with13 enzymes available (See Method). A linear B-cell epitope would have potentially higher stability if more enzymes were predicted to be unable to digest it. All the 18 linear B-cell epitopes were found having at least two non-digesting enzymes varying from 2 to 12 enzymes (Fig. 2B). The 18 linear B-cell epitopes were mapped to the 3D structure of the SARS-CoV-2 S protein (PDB ID: 6VSB), showing that 'YNSASFSTFKCYGVSPTKLNDLCFT', 'SFSTFKCYGVSPTKLNDL', 'KLNDLCFTNVYAD ', 'GDEVRQIAPGQTGKIADYNYKLP', 'DEVRQIAPGQTGKIADYNYKLPDDFT', 'VRQIAPGQTGKIAD', 'APGQTGK', 'APGQTGKIADYNYKL', 'APGQTGKIADYNYKLPDDFT', 'KIADYNYKLPDDFT', 'YQPYRVVVLSFELLH', and 'CVNFNFNGL' located in the most exposed RBD region of the spike head (Fig. 2C), and 'QCVNLTT', 'LDITPCSFGGVSVI', 'LGQSKR', 'GQSKRVDF', 'GQSKRVDFC', and 'VVFLHVTYV' were in the spike stem region (Fig. 2C). The 12 linear B-cell epitopes in the spike head were found to substantially overlap with the predicted interacting surface where ACE-2 binds to the RBD of SARS-CoV-2 S protein (Fig. 2D) [54][55][56][57], suggesting that an antibody binding to this surface may block viral entry into cells. Discontinuous B-cell epitopes were also predicted by Discotope 2.0 using A, B, and C chain of the 3D structure of S protein (PDB ID:6VSB). The positions of discontinuous B-cell epitopes mapped on the 3D structure of S protein, revealing that RBD region is enriched with discontinuous B-cell epitopes (Fig. 3E). Similarly, the discontinuous B-cell epitopes in the RBD regions overlapped with the interacting surface between ACE2 and the RBD of S protein (Fig. 3F). The linear B-cell epitopes locating in the RBD regions ('GDEVRQIAPGQTGKIADYNYKLP', 'DEVRQIAPGQTGKIADYNYKLPDDFT', 'VRQIAPGQTGKIAD', 'APGQTGK', 'APGQTGKIADYNYKL', 'APGQTGKIADYNYKLPDDFT', 'KIADYNYKLPDDFT', and 'YQPYRVVVLSFELLH') incorporated multiple discontinuous B-cell epitopes ( Fig. 2A, E). By comparing the mutations documented in 53,969 SARS-CoV-2 virus strains (NGDC), we discovered that eight linear B-cell epitopes in the RBD region incorporated one or two mutation sites in few virus strains (C mutated in 11 virus strain for 'QCVNLTT'; R mutated in one virus strain and Q mutated in 10 virus strain for 'GDEVRQIAPGQTGKIADYNYKLP', 'DEVRQIAPGQTGKIADYNYKLPDDFT', and 'VRQIAPGQTGKIAD'; Q mutated in 10 virus strain for 'APGQTGK', 'APGQTGKIADYNYKL', 'APGQTGKIADYNYKLPDDFT'; Y mutated in 4 virus strain and Q mutated in 2 virus strain for 'YQPYRVVVLSFELLH'; D mutated in one virus strain for 'GQSKRVDF' and 'GQSKRVDFC'; Y mutated in 11 virus strain for 'VVFLHVTYV') ( Fig. 2A). None of the 18 linear B-cell epitopes were found toxic.

Four linear B-cell epitopes speci cally binding with serum antibodies from the animals inoculated with different SARS-CoV-2 vaccine candidates or patients recovering from COVID-19
Animal models are necessary to demonstrate e cacy and safety in the development of vaccines against SARS-CoV-2 infection [58,59]. BALB/c mice are a good animal model for investigating SARS-CoV-2 infection in both upper and lower respiratory tracts [5]. Monkey, phylogenetically close to humans, has been used to test whether seroconversion provides protective immunity against SARS-CoV-2 [58]. To assess the binding of the 18 linear B-cell epitopes with serum IgG antibodies against SARS-CoV-2, we immunized three model animals including ve horses, ve monkeys, and ten mice (SARS-CoV-2 vaccine candidates: three for horse and monkey, respectively, and ve for mouse; adjuvants as control: two for horse and monkey, respectively, and ve for mouse). Indirect enzyme-linked immunosorbent assay (ELISA) was performed between pooled serum antibodies (from animals inoculated with SARS-CoV-2 vaccine candidates or control animal models), and the 18 linear B-cell epitopes and arbitrary control peptides ('RRRRRRRRRRRRRRRR' and 'RRRRRRR') ( Supplementary Fig. 2). We discovered that four linear B-cell epitopes reacted speci cally and dose dependently with pooled serum antibodies from the vaccinated horses ('YNSASFSTFKCYGVSPTKLNDLCFT', and three highly overlapped linear B-cell epitopes including 'GDEVRQIAPGQTGKIADYNYKLP', 'DEVRQIAPGQTGKIADYNYKLPDDFT' and 'APGQTGKIADYNYKLPDDFT' for horse (Fig. 3A); 'YQPYRVVVLSFELLH' and 'CVNFNFNGL' for mouse (Fig. 3B); 'CVNFNFNGL' and 'GDEVRQIAPGQTGKIADYNYKLP' for monkey (Fig. 3C)), whereas the arbitrary control peptides had no effects. The sera from control horse models, control mice, and control monkeys were not reactive. In addition, we also used pooled convalescent sera of ve COVID-19 patients and pooled sera of ve healthy people from Beijing Institute of Microbiology and Epidemiology, China, to test the binding a nity between IgG antibodies and linear B-cell epitopes. Indirect ELISA results demonstrated that 'CVNFNFNGL' speci cally bond to serum IgG antibodies in the patients recovering from COVID-19, but not in healthy humans (Fig. 3D). These results suggest that the four linear B-cell epitopes ('YNSASFSTFKCYGVSPTKLNDLCFT', 'GDEVRQIAPGQTGKIADYNYKLP', 'YQPYRVVVLSFELLH', and 'CVNFNFNGL') may be able to induce antibodies against SARS-CoV-2.
Antibodies against the four linear B-cell epitopes neutralize SARS-CoV-2 To con rm that the four linear B-cell epitopes generated antibodies against SARS-CoV-2, we immunized six-to eight-week-old female BALB/c mice through ve consequent subcutaneous injections of ve microgram dose of linear B-cell epitope-based synthetic peptides with 7 days interval between two consecutive injections ( ve mice per linear B-cell epitope). Serum samples collected from each synthetic peptide immunized mice were assessed for binding to corresponding linear B-cell epitope by ELISA, as well as for neutralization against SARS-CoV-2 pseudovirus and live virus. It was evident that a substantial percentage of antibodies generated in vaccinated mice against the linear B-cell epitope antigens as demonstrated in the ELISA results (Fig. 3E), suggesting that antibodies directed at the four linear B-cell epitopes could bind SARS-CoV-2.
In SARS-CoV and MERS-CoV [60], pseudovirus neutralization assay is a sensitive and quantitative method. We therefore tested the concentration of neutralizing antibodies in immune sera from wild type mice 7 days after the fth vaccination against SARS-CoV-2 pseudovirus using a pseudotyped virus-based neutralization assay developed recently for SARS-CoV-2 [61]. The neutralizing antibodies were detected in all mice immunized with each of the four linear B-cell epitopes (mean EC50 is 1963 for 'YNSASFSTFKCYGVSPTKLNDLCFT'; mean EC50 is 704 for 'YQPYRVVVLSFELLH'; mean EC50 is 230 for 'GDEVRQIAPGQTGKIADYNYKLP'; mean EC50 is 154 for 'CVNFNFNGL') (Fig. 3F), whereas no neutralizing antibodies were found in the two control group.

Discussion
SARS-CoV-2 has caused a serious pandemic all over the world. Safe and effective vaccines are urgently needed to be developed and deployed in a rapid but highly reliable manner. More than one hundred projects are being executed in the WHO draft landscape of COVID-19 candidate vaccines including a variety of vaccine types such as viral vector-based vaccines, mRNA and DNA vaccines, subunit vaccines, nanoparticle-based vaccines, to inactivated-whole virus vaccines [5][6][7][8][9]. However, even for vaccines in clinical trial phase III, some mild side effects are still observed and sporadically reported. Ongoing efforts are necessary for developing new vaccines. Rigorous selection of epitopes most likely to elicit neutralizing antibodies against virus can accelerate this process.
S protein is an important target in vaccine development for its critical functions for SARS-CoV-2 to fuse and enter into the host cells [15][16][17]. A large number of predicted epitopes are generated within a highly accelerated time frame , but the selection of in-silicon prediction methods is based on the experiences and preferences of each individual researcher. There is no comprehensive benchmark about the accuracy and speci city of these epitopes prediction methods yet. Lack of limited number of target epitopes with supports from biological validation experiments makes the epitope-based vaccine development in the middle of nowhere.
In this study, we predicted linear B-cell epitopes from S protein by eight widely-used immune-informatics methods including Bepipred and Bepipred2.0 with default parameter settings, Kolaskar and Tongaonkar antigenicity, Parker hydrophilicity, Chou and Fasman beta-turn, and Karplus and Schulz exibility provided by IEDB (Immune-Epitope-Database And Analysis-Resource) [44], BcePred [45] using accessibility, antigenic propensity, exposed surface, exibility, hydrophilicity, polarity, and turns, ANNpred-based server ABCpred [46], Ellipro [47], BCPREDS [48], AAP [49], FBCPRED [50] and COVIDep [51]. To generate a comprehensive linear B-cell epitope candidate list, we incorporated 279 linear B-cell epitopes curated from 25 articles or preprints  published between Feb 6, 2020 and July 10, 2020. Interestingly, the linear B-cell epitopes predicted by different methods converged to some hotspot regions in the S protein, suggesting the pivotal role of these regions in diagnostics assay and vaccine development. Integrating antigenicity, toxicity, stability, and physiochemical properties, the 3D structure of S protein, the 3D conformation structure between RBD of S protein and ACE2, we selected 18 top prioritized linear B-cell epitopes for further investigation. Four out of 18 linear B-cell epitope-based synthetic peptides were found to speci cally bind with serum antibodies from horses, mice, and monkeys inoculated with different SARS-CoV-2 vaccine candidates or patients recovering from COVID-19. The serum antibodies we used here were generated by different regions of SARS-CoV-2 spike protein (S1-based vaccine for horse, RBDbased vaccines for mouse and monkey, unknown regions for patients recovering from COVID-19), leading to the observation that the four peptides didn't consistently show speci c binding with all antibodies produced across horses, mice, monkeys and patients recovering from COVID-19. However, the four peptides were able to elicit neutralizing antibodies in immunized wild-type BALB/c mice against both pseudo and live SARS-CoV-2 virus.
To the best of our knowledge, the linear B-cell epitope candidate list we presented is one of the most comprehensive and valuable source for developing vaccines. Importantly, the four linear B-cell epitopes we identi ed able to elicit neutralizing antibodies against SARS-CoV-2 are promising and valuable candidates which have immediate usefulness for developing vaccines against SARS-CoV-2. Besides, more linear B-cell epitopes in the candidate list may also deserve being examined. The analysis work ow we adopted can be broadly applied to identify B-cell epitopes for a large repertoire of virus, not limited to coronavirus. In short, the results we presented here will be useful to guide the identi cation and prioritization of linear B-cell epitope-based diagnostics and vaccine designs during this unprecedented pandemic. China (IACUC-2020-030). All participants provided written informed consent for the collection of information, and their clinical samples were stored and used for research. Data generated from the research were agreed to be published.

Consent for publication
All the authors have approved the manuscript for publication and declare no potential con icts of interest. We also certify that the manuscript is our own research work with nothing taken from the published or unpublished work of others, and has not simultaneously been submitted elsewhere.

Availability of data and material
The data that support the ndings of this study are openly available. SARS-CoV-2 protein sequence (Accession number MN908947.3) [61] was extracted from the NCBI database. Experimentally solved 3D structure of SARS-CoV-2 S protein (PDB ID: 6VSB) [11] was retrieved from Protein Data Bank. The predicted 3D structure of protein interactions between SARS-CoV-2 S protein RBD domain and human ACE-2 was obtained from recent reports [54-57, 62, 63]. The target pro les of IgG or IgA antibodies from 232 COVID-19 patient sera (101 for hospitalized, 131 for non-hospitalized) and 190 pre-COVID-19 era controls were from a recent research [53]. The code and software that support the ndings of this study are openly available in the original literatures which were cited properly in the Methods section. The predicted linear B-cell epitopes in the Spike protein of SARS-CoV-2. a, The number of linear B-cell epitopes shared among the distinct methods and literature mining. The pink, green, and light blue represent epitopes with antigenicity scores >0.9, 0.4 and 0.9, and <0.4, respectively. b, The antigenicity score and peptide length distribution of the predicted linear B-cell epitopes. c, The positions of the linear B-cell epitopes with high antigenicity scores (>0.9) on the S protein amino acid sequence. The y-axis is the coverage of predicted linear B-cell epitopes. The bars below are the protein domain, mutation sites, conservative scores across seven known coronaviruses viruses, and the transmembrane domain predicted by TMHMM, respectively. d-e, The localizations of regions enriched with linear B-cell epitopes (antigenicity score>0.9) on SARS-CoV-2 S (PDB: 6VSB) protein. The grey is protein, the green is the regions enriched with linear B-cell epitopes (antigenicity score>0.9), and the wheat color is the RBD regions.

Figure 2
The characteristics of the 18 selected linear B cell epitopes. a, The sequences of 18 selected linear B-cell epitopes. The bold is the mutated site in less than ten of 118,694 virus strains; The red is the predicted discontinuous residues. The bars on the right side are the Wilcoxon test p value for the comparisons of IgG or IgA antibody enrichment scores associated with each linear B-cell epitope between COVID-19 patients and negative controls. b, The digesting enzymes pro le of the epitope sequence. Red indicated not digest, blue indicated digest. c-d, The localization of the 18 selected epitopes mapped on SARS-CoV-2 S (PDB: 6VSB) protein (c) and ACE-RBD complex (d). e-f, The localizations of B cell discontinuous epitopes on SARS-CoV-2 S (PDB: 6VSB) protein (e) and ACE-RBD complex (f). The spike protein is grey, the RBD region is wheat color, the selected epitopes are green, the mutation sites are red, the human ACE domain is blue, and the discontinuous B-cell epitopes are purple.

Figure 3
Measurements of the selected Linear B cell epitope binding to antibody and neutralization e ciency of selected epitopes against SARS-CoV-2. a-d, The binding a nity assessed by ELISA between linear B-cell epitopes and serum antibodies from immunized horse with S1-based vaccines (a), immunized mouse with RBD-based vaccines (b), immunized monkey with RBD-based vaccines (c), and a patient recovering from COVID-19 (d). e, The binding a nity assessed by ELISA between the linear B-cell epitopes and serum antibodies from immunized mice with corresponding epitopes of 'YNSASFSTFKCYGVSPTKLNDLCFT', 'GDEVRQIAPGQTGKIADYNYKLP', 'YQPYRVVVLSFELLH', and 'CVNFNFNGL'. f, Neutralization assay against SARS-CoV-2 pseudovirus in 'YNSASFSTFKCYGVSPTKLNDLCFT', 'GDEVRQIAPGQTGKIADYNYKLP', 'YQPYRVVVLSFELLH', and