key: cord-265277-ymvrserl
authors: Crooke, Stephen N.; Ovsyannikova, Inna G.; Kennedy, Richard B.; Poland, Gregory A.
title: Immunoinformatic identification of B cell and T cell epitopes in the SARS-CoV-2 proteome
date: 2020-05-14
journal: bioRxiv
DOI: 10.1101/2020.05.14.093757
sha: 
doc_id: 265277
cord_uid: ymvrserl

A novel coronavirus (SARS-CoV-2) emerged from China in late 2019 and rapidly spread across the globe, infecting millions of people and generating societal disruption on a level not seen since the 1918 influenza pandemic. A safe and effective vaccine is desperately needed to prevent the continued spread of SARS-CoV-2; yet, rational vaccine design efforts are currently hampered by the lack of knowledge regarding viral epitopes targeted during an immune response, and the need for more in-depth knowledge on betacoronavirus immunology. To that end, we developed a computational workflow using a series of open-source algorithms and webtools to analyze the proteome of SARS-CoV-2 and identify putative T cell and B cell epitopes. Using increasingly stringent selection criteria to select peptides with significant HLA promiscuity and predicted antigenicity, we identified 41 potential T cell epitopes (5 HLA class I, 36 HLA class II) and 6 potential B cell epitopes, respectively. Docking analysis and binding predictions demonstrated enrichment for peptide binding to HLA-B (class I) and HLA-DRB1 (class II) molecules. Overlays of predicted B cell epitopes with the structure of the viral spike (S) glycoprotein revealed that 4 of 6 epitopes were located in the receptor-binding domain of the S protein. To our knowledge, this is the first study to comprehensively analyze all 10 (structural, non-structural and accessory) proteins from SARS-CoV-2 using predictive algorithms to identify potential targets for vaccine development. Significance Statement The novel coronavirus SARS-CoV-2 recently emerged from China, rapidly spreading and ushering in a global pandemic. Despite intensive research efforts, our knowledge of SARS-CoV-2 immunology and the proteins targeted by the immune response remains relatively limited, making it difficult to rationally design candidate vaccines. We employed a suite of bioinformatic tools, computational algorithms, and structural modeling to comprehensively analyze the entire SARS-CoV-2 proteome for potential T cell and B cell epitopes. Utilizing a set of stringent selection criteria to filter peptide epitopes, we identified 41 T cell epitopes (5 HLA class I, 36 HLA class II) and 6 B cell epitopes that could serve as promising targets for peptide-based vaccine development against this emerging global pathogen.

In December 2019, public health officials in Wuhan, China, reported the first case of severe 60 respiratory disease attributed to infection with the novel coronavirus SARS-CoV-2 (1). Since its 61 emergence, SARS-CoV-2 has spread rapidly via human-to-human transmission (2), threatening to 62 overwhelm healthcare systems around the world and resulting in the declaration of a pandemic by the 63 World Health Organization (3). The disease caused by the virus is characterized by fever, 64 pneumonia, and other respiratory and inflammatory symptoms that can result in severe inflammation of 65 lung tissue and ultimately death-particularly among older adults or individuals with underlying 66 comorbidities (4-6). As of this writing, the SARS-CoV-2 pandemic has resulted in 4 million confirmed 67 cases of COVID-19 and over 280,000 deaths worldwide (7). 68 SARS-CoV-2 is the third pathogenic coronavirus to cross the species barrier into humans in the 69 past two decades, preceded by severe acute respiratory syndrome coronavirus (SARS-CoV) (8, 9) and 70

Middle-East respiratory syndrome coronavirus (MERS-CoV) (10). All three of these viruses belong to the 71 β -coronavirus genus and have either been confirmed (SARS-CoV) or suggested (MERS-CoV, SARS-72

CoV-2) to originate in bats, with transmission to humans occurring through intermediary animal hosts 73 (11) (12) (13) (14) . While previous zoonotic spillovers of coronaviruses have been marked by high case fatality rates 74 (~10% for SARS-CoV; ~34% for MERS-CoV), widespread transmission of disease has been relatively 75 limited (8,098 cases of SARS; 2,494 cases of MERS) (15). In contrast, SARS-CoV-2 is estimated to have 76 a lower case fatality rate (~2-4%) but is far more infectious and has achieved world-wide spread in a 77 matter of months (16). 78

As the number of COVID-19 cases continues to grow, there is an urgent need for a safe and 79 effective vaccine to combat the spread of SARS-CoV-2 and reduce the burden on hospitals and healthcare 80 systems. No licensed vaccine or therapeutic is currently available for SARS-CoV-2, although there are 81 over 100 vaccine candidates reportedly in development worldwide. Seven vaccine candidates have 82 peptide was removed from the structure using Chimera 1.14 (University of California-San Francisco) (50) 158 prior to running simulations. Ten models of each peptide-HLA complex were generated on the basis of 159 minimized energy scores, and the top model for each complex was selected for comparative analysis. 160

Prediction and structural modeling of SARS-CoV-2 B cell epitopes 161 Linear B cell epitope predictions were performed on the three exposed SARS-CoV-2 structural 162 proteins: S (GenBank accession: QHD43416), M (QHD43419), and E (QHD43418) using the BepiPred 163 1.0 algorithm (51). Epitope probability scores were calculated for each amino acid residue using a 164 threshold of 0.35 (corresponding to > 0.75 specificity and sensitivity below 0.5), and only epitopes > 5 165 amino acid residues in length were further analyzed. The structure of the SARS-CoV-2 S protein was 166 accessed from the Protein Data Bank (PDB ID: 6VSB) (52). Discontinuous (i.e., structural) B cell epitope 167 predictions for the S protein structure were carried out using DiscoTope 1.1 (53) with a score threshold 168 greater than -7.7 (corresponding to > 0.75 specificity and sensitivity below 0.5). The main protein 169 structure was modeled in PyMOL (Schrödinger, LLC), with predicted B cell epitopes identified by both 170 BepiPred 1.0 and DiscoTope 1.1 highlighted as spheres. 171

Genetic similarity of SARS-CoV-2 isolates 173

The primary goal of our study was to identify peptide epitopes that would be broadly applicable 174 in vaccine development efforts against SARS-CoV-2. We identified 64 point mutations and 4 deletions 175 across the genomes of 44 clinical isolates, with all deletions and the majority of mutations (n=45) 176 occurring in the ORF1ab polyprotein (Supp. Figure S1 ). Single-point mutations were also found in the S 177 protein (n=5), N protein (n=5), ORF8 protein (n=3), ORF3a protein (n=2), ORF10 protein (n=2), E 178 protein (n=1), and M protein (n=1). Despite the genetic diversity introduced by these events (Figure 1D) , 179 matrix analysis determined that > 99% sequence identity was maintained across all viral genomes. Based 180 on these findings and for study feasibility, the genome from the original virus isolate 181 GenBank: MN908947) was selected as the consensus sequence for all further analyses. 182

We next identified potential CD8 + T cell epitopes from all proteins in the SARS-CoV-2 184 proteome. Using the NetCTL 1.2 predictive algorithm, we analyzed the complete amino acid sequence of 185 each viral protein to generate sets of 9-mer peptides predicted to be recognized across at least one of the 186 major HLA class I supertypes (Figure 2A, Supp. Figure S2) . This approach yielded a significant number 187 of potential epitopes from each viral protein (ORF10: 9, ORF6: 17, ORF8: 23, E: 25, ORF7: 39, N: 80, 188 M: 87, ORF3a: 87, S: 321, ORF1ab: 2814) , with the number directly related to the size of the parent protein. We used the NetMHCpan 4.0 server to further refine the list of potential CD8 + T cell epitopes by 190 predicting binding affinity across representative HLA class I alleles (see Methods) and assigning 191 percentile scores to quantify binding propensity. Peptides with percentile rank scores < 0.5% (i.e., strong 192 binders) were filtered using a 500 nM threshold for binding affinity to further delineate 740 candidate 193 HLA class I epitopes from the viral proteome (54). For feasibility reasons, we refined our selection to 83 194 candidate epitopes by excluding peptides predicted to bind only one HLA molecule (Supp . Table S1 ). 195

The resultant peptides were enriched for predicted binders to HLA-B molecules 196 HLA-B*58:01=32; HLA-B*08:01=31) ( Figure 2B) . A final round of selection on the basis of HLA 197 promiscuity (i.e., predicted binding to > 3 HLA molecules) and predicted antigenicity scoring using the 198 VaxiJen 2.0 server produced a subset of five candidate peptides (four ORF1ab, one S protein) as potential 199 targets for vaccine development (Table 1) with the hypothesis that increased HLA binding promiscuity 200 meant broader population base coverage by those peptides. These peptides were predicted to provide 74% 201 global population coverage and had higher predicted binding affinities for HLA-B molecules 202 (B*08:01=42.6 nM; B*15:01=67.7 nM; B*58:01=110.3 nM) compared to HLA-A molecules 203 (A*01:01=238.6 nM; A*24:02=142.9 nM), with the exception of one ORF1ab-derived peptide 204 (MMISAGFSL) that was predicted to bind HLA-A*02:01 with high affinity (IC 50 = 6.9 nM) ( Figure 2C) . 205

We also sought to identify potential HLA class II peptides from SARS-CoV-2, as the stimulation 207 of CD4 + T-helper cells is critical for robust vaccine-induced adaptive immune responses. Using the 208 NetMHCIIpan 3.2 server, we identified 801 candidate HLA class II peptides from the viral proteome 209 predicted to have high binding affinity (< 500 nM) and percentile rank scores < 2% across a reference 210 panel of HLA molecules covering > 97% of the population (33, 45). Similar to HLA class I epitope 211

predictions, the number of class II epitopes identified for each viral protein (ORF10: 4, E protein: 7, 212 ORF7: 8, ORF8: 10, ORF6: 14, N: 15, M: 29, ORF3a: 31, S: 96, ORF1ab: 587) was largely proportional 213 to protein size. After excluding peptides predicted to bind to only a single HLA molecule in our panel, we 214 refined our selection to 211 peptides (Supp . Table S2 ), which were enriched for binding to HLA-DRB1 215 molecules (n=142) ( Figure 2D ). Filtering on HLA promiscuity and predicted antigenicity scores yielded 216 a subset of 36 peptides (24 ORF1ab, 5 S protein, 2 M protein, 2 ORF7, 1 ORF3a, 1 ORF6, 1 ORF8) as 217 CD4 + T cell epitopes for further study ( Table 1) . These peptides were predicted to collectively provide 218 99% population coverage and have significantly higher average binding affinities for HLA-DR alleles 219 (DRB1=56.4 nM; DRB3=50.9 nM; DRB4=70.1 nM; DRB5=18 nM) compared to HLA-DP (155.9 nM) 220 or HLA-DQ (238.6 nM) molecules ( Figure 2E) . 221

Characterization of HLA class I peptide docking with HLA-B*15:01

The five candidate HLA class I peptides identified by our computational approach were predicted 223 to provide coverage across six HLA alleles (A*01:01, A*02:01, A*24:02, B*08:01, B*15:01, B*58:01). 224

The peptide FAMQMAYRF was the only candidate predicted to bind to A*24:02 molecules, whereas 225 MMISAGFSL was predicted to uniquely bind A*02:01 and B*08:01 molecules. Four of the five peptides 226 were predicted to bind A*01:01 and B*58:01 molecules, but all were predicted to bind with relatively 227 high affinity (average IC 50 = 67.7 nM) to HLA-B*15:01. Therefore, we performed molecular docking 228 studies of each peptide with the molecular structure of HLA-B*15:01 (PDB: 3C9N). 229

All peptides were predicted to bind within the peptide binding groove, forming hydrogen bond 230 contacts with numerous amino acid side chains ( Figure 3A) . The binding motif for HLA-B*15:01 is 231 highly selective for residues at the P2 and P9 anchor positions, with a preference for bulky hydrophobic 232 amino acids at the C-terminus ( Figure 3B ) (55). All candidate peptides possessed terminal residues (Phe, 233

Tyr, Leu) that fit into the hydrophobic binding pocket of the HLA groove, further supporting that these 234 peptides should be strong binders of HLA-B*15:01 and promising candidates for vaccine development 235 studies. 236

An effective vaccine should stimulate both cellular and humoral immune responses against the 238 target pathogen; therefore, we also sought to identify potential B cell epitopes from SARS-CoV-2 239 proteins. We limited our analysis to the primary structural proteins exposed on the virus capsid (S, N, M, 240 and E), as these are the most accessible antigens for engaging B cell receptors. Using the Bepipred 1.0 241 algorithm, we identified 26 potential linear B cell epitopes in the S protein, 14 potential epitopes in the N 242 protein, and 3 potential epitopes in the M protein ( Table 2) . No epitopes were identified in the E protein. 243

Studies have previously shown the S protein to be the predominant target of neutralizing antibodies 244 against coronaviruses (56, 57), and, as our findings indicate this to likely be the case for SARS-CoV-2, 245 we focused all subsequent analyses on the S protein. While the N protein is also a major target of the 246 antibody response (58), it is unlikely these antibodies have any neutralizing activity based on the viral 247 structure. As epitope conformation can significantly influence recognition by antibodies, we also 248 employed DiscoTope 1.1 to identify discontinuous B cell epitopes in the protein structure. Our analysis 249 identified 14 potential structural epitopes in the S protein (7 in the S1 domain, 7 in the S2 domain), with 250 six regions having significant overlap with our predicted linear epitopes ( Table 2) . Antigenic regions 251 identified in both analyses were modeled using the recently published structure of the SARS-CoV-2 S 252 protein (52) to examine their accessibility for antibody binding. Epitopes in the S2 domain (P792-D796; 253 Y1138-D1146) were clustered near the base of the spike protein, whereas regions in the S1 domain 254 (D405-D428; N440-N450; G496-P507; D568-T573) were exposed on the protein surface (Figure 4 ).

In the face of the COVID-19 pandemic, it is imperative that safe and effective vaccines be rapidly 257 developed in order to induce widespread herd immunity in the population and prevent the continued 258 spread of SARS-CoV-2. Our study identified probable peptide targets of both cellular and humoral 259 immune responses against SARS-CoV-2 using computational methodologies to investigate the entire viral 260 proteome a priori. Studies such as these are paramount during the early stages of pandemic vaccine 261 development given the relative scarcity of biological data available on the viral immune response, and we 262 employed an approach that allowed us to systematically refine our predictions using increasingly stringent 263 criteria to select a subset of the most promising epitopes for further study. The data we have curated could 264 inform the design of a candidate peptide-based vaccine or diagnostic against SARS-CoV-2. 265

As selective pressures are known to introduce viral mutations that promote fitness and can lead 266 to evasion of immune responses (59, 60), we first sought to investigate the genetic similarity of all 267 reported SARS-CoV-2 clinical isolates and identify a consensus sequence for use in our epitope 268 prediction studies. We identified 68 mutations/deletions across the 44 genomes of clinical isolates 269 reported as of 27 February 2020. Despite these variations, the viral genomic identity was > 99% 270 conserved across all isolates. As the protein coding sequences were largely conserved, the genome of the 271 original virus isolate (Wuhan-Hu-1) was deemed a representative consensus sequence for analysis of the 272 SARS-CoV-2 proteome. 273 CD4 + and CD8 + T cell responses will likely be directed against both structural and non-structural 274 proteins during antiviral immune responses, as all viral proteins are accessible for processing and 275 presentation on the HLA molecules of infected cells. Therefore, we sought to identify T cell epitopes 276 across the entire viral proteome. Our analysis identified 83 potential CD8 + T cell epitopes (Supp. Table  277 S1) and 211 potential CD4 + T cell epitopes (Supp . Table S2) , with stringent filtering for more 278 promiscuous peptides with high predicted antigenicity yielding a subset of 5 CD8 + T cell epitopes and 36 279 CD4 + T cell epitopes ( Table 1 ) as potential targets for vaccine development. A single study by Grifoni 280 and colleagues has recently reported the computational identification of 241 CD4 + T cell epitopes from 281 SARS-CoV-2 (35), and 22 peptides from our analysis shared sequence homology or were nested within 282 peptides identified in their study. Moreover, seven peptides from this initial report were replicated in our 283 final subset of HLA class II epitopes, supporting that these peptides may be promising vaccine targets. 284

An increasing number of studies have employed predictive algorithms to identify potential HLA 285 class I epitopes for SARS-CoV-2, although relatively few have comprehensively analyzed the entire viral 286 proteome. A report from Feng et al. recently outlined the identification of 499 potential class I epitopes in 287 the main structural proteins from SARS-CoV-2 but did not consider any non-structural proteins (38). 288

Grifoni and colleagues conducted a more rigorous analysis, identifying 628 unique CD8 + T cell epitopes across all SARS-CoV-2 proteins but focusing their analyses solely on peptides with sequence homology 290 to known SARS-CoV epitopes (35). Our approach initially identified ~ 3,500 potential CD8 + T cell 291 epitopes across all viral proteins, which we refined to a subset of 5 peptides (Table 1) . One peptide 292 derived from ORF1ab (MMISAGFSL) was predicted to bind HLA-A*02:01 with high affinity (IC 50 = 6.9 293 nM) ( Figure 2C) . Given the prevalence of this allele in the American and European populations (25-60% 294 frequency) (61), MMISAGFSL may represent a promising epitope capable of providing broad vaccine 295 population coverage. 296

We also observed a notable enrichment of epitopes predicted to bind HLA-B molecules-297 particularly HLA-B*15:01-as we imposed more stringent selection criteria ( Figure 2B ). All five peptides 298 identified by our approach were predicted to be relatively strong binders for this allele (IC 50 = 67.7 nM), 299 with molecular docking simulations illustrating strong contacts with amino acid residues in the peptide 300 binding groove (Figure 3 A, B) . A recent computational study identified another HLA-B allele (B*15:03) 301 as having a high capacity for presenting epitopes from SARS-CoV-2 that were conserved among other 302 pathogenic coronaviruses (62). These data collectively suggest the HLA-B locus may be significantly 303 associated with the immune response to SARS-CoV-2 (and potentially other coronaviruses), with further 304 biological studies warranted to determine the true role of host genetics in SARS-CoV-2 immunology. 305

Lastly, we analyzed the primary structural proteins of SARS-CoV-2 (S, N, M, E proteins) for 306 potential B cell epitopes, as an ideal vaccine would be designed to stimulate both cellular and humoral 307 immunity. Our analysis identified potential linear B cell epitopes in all proteins except for the E protein 308 ( Table 2 ). The greatest number of epitopes were predicted in the surface-exposed S protein (n=26), but a 309 significant number of epitopes were also predicted for the N protein (n=14). This is not surprising, as 310 previous reports identified the N protein as a significant target of the humoral response to SARS-CoV 311 (63, 64). As the S protein is the predominant surface protein and has been the primary target of 312 neutralizing antibody responses against other coronaviruses (56, 57), we elected to focus our subsequent 313 analyses solely on antigenic regions in the S protein. We identified 14 potential structural epitopes in the 314 S protein structure and referenced against our linear epitope predictions to identify six regions that were 315 independently identified by both analyses (Table 2, Figure 4) To further evaluate the potential of these six antigenic regions as targets for antibody binding, we 319 modeled their surface accessibility on the crystal structure of the SARS-Cov-2 spike protein (52). Four 320 regions in the S1 domain (D405-D428; N440-N450; G496-P507; D568-T573) were solvent exposed 321 (Figure 4 A, B) , with minimal steric hindrance for antibody accessibility. The S1 domain contains the 322 residues (N331-V524) important for virus binding to angiotensin converting enzyme 2 (ACE2) on the cell surface (65), and studies have shown that antibodies with potent neutralizing activity against SARS-CoV 324 target this domain (66-68). Indeed, three of the four S1 epitopes identified in our analyses are located in 325 the ACE2-binding region, supporting their potential utility in vaccine development against SARS-CoV-2. 326

Two regions were identified in the S2 "stalk" domain of the S protein (Figure 4 A, C) . While Y1138-327 D1146 is located at the base of the S protein and likely inaccessible to antibodies, P792-D796 is on the 328 outer face of the protein and has been previously identified as part of a larger B cell epitope that is 329 conserved with SARS-CoV (35). As SARS-CoV S2-specific antibodies have previously been shown to 330 possess antiviral activity (66), it is interesting to speculate whether a strategy similar to targeting the 331 influenza hemagglutinin protein stalk could be employed for developing a broadly reactive coronavirus 332 vaccine. 333

Our study possessed several strengths and limitations. Rather than restricting our analyses of 334 HLA class I and class II epitopes to specific proteins based on prior studies of SARS-CoV immunology, 335

we investigated the complete proteome of SARS-CoV-2 using an unbiased approach. Furthermore, we 336 employed a multi-tiered strategy for identifying putative B cell and T cell epitopes from all viral proteins 337 studied. Our initial analyses were performed with liberal thresholds for epitope identification, and at each 338 additional step, we imposed more stringent selection criteria to filter these peptides to a subset of B cell 339 and T cell epitopes for further study. Nevertheless, the results of this study are derived purely from 340 computational methods, and it should be noted that computational algorithms can fail to capture a 341 significant number of antigenic peptides (69). Experimental validation with biological samples will 342 ultimately be needed. 343

During the early stages of a pandemic, access to sufficient biological samples may be extremely 344 limited, so we must continue to utilize methodologies-such as computational predictive algorithms-345 that allow us to explore the epitope landscape for experimental vaccine development. Our approach in this 346 study allowed us to identify and refine a manageable subset of T cell and B cell epitopes for further 347 testing as components of a SARS-CoV-2 vaccine. Based on our results, our proposed SARS-CoV-2 348 vaccine formulation could contain the following: 1) one or more B cell peptide epitopes from the S 349 protein to generate protective neutralizing antibodies; and 2) multiple HLA class I and class II-derived 350 peptides from other viral proteins to stimulate robust CD8 + and CD4 + T cell responses. Based on global 351 allele frequencies, these class I and class II peptides would be expected to collectively provide 74% and 352 99% population coverage, respectively. While such a vaccine could be readily formulated as a synthetic 353 polypeptide or an adjuvanted peptide mixture, these strategies may not retain the epitope structural 354 features necessary to induce a robust antibody response. Recombinant nanoparticles and assembly into 355

VLPs represent promising alternative vaccine platforms, as they have been extensively used for the 356 controlled display and delivery of peptide-based vaccine components (70-73). By omitting whole viral 357 proteins from the vaccine formulation, a peptide-based SARS-CoV-2 vaccine should have a well-358 tolerated safety profile and avoid the adverse events previously observed with experimental SARS-CoV 359 vaccines (19) (20) (21) (22) . 360

In summary, we have identified 41 potential T cell epitopes (5 HLA class I, 36 HLA class II) and 361 6 potential B cell epitopes from across the SARS-CoV-2 proteome that are predicted to have broad 362 population coverage and could serve as the basis for designing investigational peptide-based vaccines. 363

Further study on the biological relevance and immunogenicity of these peptides is warranted in an effort 364 to develop a safe and effective vaccine to combat the SARS-CoV-2 pandemic. 365

The authors would like to thank Caroline L. Vitse for editorial assistance with this manuscript. 367

The research presented here was not supported by any specific funding source. 368 :2020.2003.2022.20040600. 532 63. Huang LR, et al. (2004) Evaluation of antibody responses against SARS coronaviral nucleocapsid 533 or spike proteins by immunoblotting or ELISA. identification workflow illustrating the algorithms used (33, 40-43, 45-47, 51, 53) and filtering criterion 580 applied to refine peptide selection. (D) Cladogram illustrating the genetic relationship of SARS-CoV-2 581

isolates. The original viral isolate and consensus sequence (Wuhan-Hu-1) is highlighted in red. 582 583 Figure 2 . Immunogenicity scoring of peptides in the SARS-CoV-2 proteome with predicted HLA class I 584 and II coverage and binding affinities. DiscoTope prediction algorithms highlighted on the trimeric structure of the S glycoprotein. Inset panels 604

show the S1 domain (upper) and S2 domain (lower 

A new coronavirus associated with human respiratory disease in China

A familial cluster of pneumonia associated with the 2019 novel 392 coronavirus indicating person-to-person transmission: a study of a family cluster

WHO Declares COVID-19 a Pandemic

Epidemiological and clinical characteristics of 99 cases of 2019 novel 397 coronavirus pneumonia in Wuhan, China: a descriptive study

Clinical Characteristics of 138 Hospitalized Patients With

Clinical features of patients infected with

Coronavirus disease (COVID-19) Situation Report -113

Identification of a novel coronavirus in patients with severe acute 407 respiratory syndrome

A novel coronavirus associated with severe acute respiratory syndrome. 409

Isolation of a 411 novel coronavirus from a man with pneumonia in Saudi Arabia

Bats are natural reservoirs of SARS-like coronaviruses

A pneumonia outbreak associated with a new coronavirus of probable bat 416 origin

Middle East respiratory syndrome coronavirus in bats

Middle East respiratory syndrome coronavirus in dromedary camels: 420 an outbreak investigation

Structure, Function, and Antigenicity of the SARS-CoV-2 Spike 422 Glycoprotein

COVID-19: Knowns, Unknowns, and Questions. mSphere 5(2). 424 17. World Health Organization (2020) Draft landscape of COVID-19 candidate vaccines

Tortoises, hares, and vaccines: A cautionary note for SARS-CoV-2 vaccine 428 development

Immunization with SARS coronavirus vaccines leads to pulmonary 430 immunopathology on challenge with the SARS virus

Vaccine efficacy in senescent mice challenged with recombinant SARS-432 CoV bearing epidemic and zoonotic spike variants

Prior immunization with severe acute respiratory syndrome (SARS)-434 associated coronavirus (SARS-CoV) nucleocapsid protein causes severe pneumonia in mice 435 infected with SARS-CoV

Characterization of the receptor-binding domain (RBD) of 2019 novel 537 coronavirus: implication for development of RBD protein as a viral attachment inhibitor and 538 vaccine

Quantitative comparison of the efficiency of antibodies against S1 and S2 540 subunit of SARS coronavirus spike protein in virus neutralization and blocking of receptor 541 binding: implications for the functional roles of S2 subunit

Neutralizing epitopes of the SARS-CoV S-protein cluster independent of 543 repertoire, antigen structure or mAb technology

Identification and characterization of novel neutralizing epitopes in the 545 receptor-binding domain of SARS-CoV spike protein: revealing the critical antigenic 546 determinants in inactivated SARS-CoV vaccine

Discovery of 548 naturally processed and HLA-presented class I peptides from vaccinia virus infection using mass 549 spectrometry for vaccine development

Development of Autologous C5 Vaccine Nanoparticles to Reduce 551 Intravascular Hemolysis in Vivo

Plug-and-Display: decoration of Virus-Like Particles via isopeptide 553 bonds for modular immunization

A novel candidate HPV vaccine: MS2 phage VLP displaying a tandem HPV 555 L2 peptide offers similar protection in mice to Gardasil-9

Targeted 557 immunomodulation using antigen-conjugated nanoparticles

Scoring function for automated assessment of protein structure 560 template quality

GalaxyPepDock: a protein-peptide docking tool based 562 on interaction similarity and energy optimization