key: cord-1003727-zdm1i65q authors: Parn, Simone; Savsani, Kush; Dakshanamurthy, Sivanesan title: SARS-CoV-2 Omicron (BA.1 and BA.2) specific novel CD8+ and CD4+ T cell epitopes targeting spike protein date: 2022-04-06 journal: bioRxiv DOI: 10.1101/2022.04.05.487186 sha: efd9c57f9970255aa88646a10b9d59b7b9bf982b doc_id: 1003727 cord_uid: zdm1i65q The Omicron (BA.1/B.1.1.529) variant of SARS-CoV-2 harbors an alarming 37 mutations on its spike protein, reducing the efficacy of current COVID-19 vaccines. This study identified CD8+ and CD4+ T cell epitopes from SARS-CoV-2 S protein mutants. To identify the highest quality CD8 and CD4 epitopes from the Omicron variant, we selected epitopes with a high binding affinity towards both MHC I and MHC II molecules and applied other clinical checkpoint predictors including immunogenicity, antigenicity, allergenicity, instability, and toxicity. Subsequently, we found eight Omicron (BA.1/B.1.1.529) specific CD8+ and eleven CD4+ T cell epitopes with a world population coverage of 76.16% and 97.46%, respectively. Additionally, we identified common epitopes across Omicron BA.1 and BA.2 lineages that target mutations critical to SARS-CoV-2 virulence. Further, we identified common epitopes across B.1.1.529 and other circulating SARS-CoV-2 variants, such as B.1.617.2 (Delta). We predicted CD8 epitopes’ binding affinity to murine MHC alleles to test the vaccine candidates in preclinical models. The CD8 epitopes were further validated using our previously developed software tool PCOptim. We then modeled the three-dimensional structures of our top CD8 epitopes to investigate the binding interaction between peptide-MHC and peptide-MHC-TCR complexes. Importantly, our identified epitopes are targeting the mutations on the RNA-binding domain and the fusion sites of S protein. This could potentially eliminate viral infections and form long-term immune responses compared to rather short-lived mRNA vaccines and maximize the efficacy of vaccine candidates against the current pandemic and potential future variants. Using IEDB recommended 2.22 based on rank, IFNEpitope, VaxiJen 2.0, AllerTop 2.0, and ToxinPred, only IFNγ inducing antigenic non-allergenic non-toxic CD4 epitopes were selected. Our analysis yielded 7 common S CD4+ T cell epitopes across Omicron (BA.1/B.1.1.529) and six other SARS-CoV-2 variants (Alpha, Beta, Delta, Gamma, US variants (S protein mutations), and Cluster 5 mink variants) reported by our other study ( Table 3, Supplementary Table S4 ) [43] . In addition, we identified 15 Omicron BA.1 specific S CD4+ T cell epitopes (Supplementary Table S5 ), whereas 11 of them were predicted to be stable in vivo ( Table 4) . We also predicted CD4+ T cell epitopes from the BA.2 variant to check for any common epitopes among Omicron BA.1 specific CD4 peptides, and across other SARS-CoV-2 strains. Our analysis resulted in 8 common CD4+ T cell epitopes across Omicron BA.1 and BA.2 sub-lineages (Tables 4,5). Moreover, BA.2 shares identical CD4 peptides with other SARS-CoV-2 variants listed in Table 3 . the promotion of specific B cell antibody production. It is thought that together these T cells can elicit a strong immune response against an evading pathogen. Hence, we overlapped the CD8+ T cell epitopes on the CD4+ T cell epitopes for the identification of overlapping Omicron BA.1 specific epitopes. The analysis from the previous section revealed 3 S CD4+ T cell epitopes (KSHRRARSVASQSII, SVLYNLAPFFTFKCY, and VLYNLAPFFTFKCYG) overlapping with our top CD8+ T cell epitopes ( Table 5 ). All identified CD4+ T cell epitopes are antigenic, IFNγ inducing, non-allergenic, and non-toxic. Moreover, we found that CD4+ T cell (KSHRRARSVASQSII) that overlaps with a CD8+ T cell epitope (KSHRRARSV) is common to both Omicron BA.1 and BA.2 sub-lineages. To assess the efficacy of the proposed vaccine candidates, we predicted their binding affinity to murine H2 alleles. Since the SYFPEITHI server is mostly validated for MHC Class I epitope prediction and has low reliability (50%) for MHC Class II peptides, we only identified CD8 peptides that bind to murine H2 alleles. The comparison of reference epitopes to the identified CD8 peptides yielded potential strong, intermediate, and weak binders. Our analysis revealed 7 common S CD8+ T cell epitopes among SARS-CoV-2 variants ( Table 1 ) and 4 Omicron BA.1 specific S CD8+ T cell epitopes ( Table 2) presented by murine MHC restriction. Moreover, we identified one BA.2 lineagespecific CD8+ T cell epitope (TPINLGRDL) that is presented by murine MHC restriction (Supplementary Table S3 ). To determine the world population coverage of the selected T cell epitopes, corresponding HLA associations were considered. Identified epitopes that bind several MHC (HLA) alleles are generally regarded as the best probable epitopes as they have a higher potential to show good coverage by approaching 100%. Subsequently, we found that the 12 common CD8 peptides among Omicron and other SARS-CoV-2 variants cover 84.0% of the world population (Table 1, Supplementary Fig. 1A ) whereas the 8 Omicron BA.1 specific CD8 peptides cover 76.16% ( Table 2, Supplementary Fig. 1B) . World population coverage was also computed for the MHC Class II peptides. We found that the 7 common CD4+ T cell epitopes across Omicron and other circulating variants including Delta can elicit an immune response that covers 96.65% of the world population ( Table 3 , Supplementary Fig. 1C) . The 11 top quality Omicron BA.1 variant-specific CD4 peptides were computed to cover 97.46% of the world population ( Table 4 , Supplementary Fig. 1D ) whereas the three overlapping CD4 peptides were found to cover 92.66% of the world population ( Table 5 , Supplementary Fig. 1E ). The top 300 epitopes found from IEDB NetMHCpan EL 4.1 have a rank range of 0-0.30. We filtered through all the raw epitope data to obtain a dataset of epitopes with a rank less than 0.30 and immunogenicity greater than 0. We entered this dataset into PCOptim to obtain a list of optimized epitopes which we can use as a basis of comparison to our 8 immunogenic, antigenic, nonallergenic, non-toxic, and stable epitopes. The optimized list of epitopes included 8 unique epitopes and returned 98.55% population coverage (Figure 6 ). Of the 8 optimized epitopes, 2 are present in our "Omicron BA.1 variant-specific S CD8 epitope" dataset (Supplementary Table S2) (RSYSFRPTY and YNLAPFFTF) and 1 is present in our "Top common S CD8 epitopes across Omicron and other SARS-CoV-2 variants" dataset (Supplementary Table S1 ) (VVFLHVTYV). The HLA alleles that were present in the optimized dataset but not the clinically relevant dataset are HLA-B*07:02, HLA-B*08:01, HLA-A*31:01, HLA-A*33:01, and HLA-B*51:01. These alleles account for the difference in population coverage between the optimal epitope dataset and datasets obtained after passing epitopes through several clinical checkpoints. Protein 3D structures offer useful insights into their molecular activity and provide a wide variety of applications in bioscience. We selected two immunogenic antigenic non-allergenic and non-toxic CD8+ T cell epitopes (KSHRRARSV and YNLAPFFTF) for 3D visualization and molecular docking analysis. KSHRRARSV and YNLAPFFTF are both strong MHC I binders and presented by murine MHC restriction. Epitope KSHRRARSV was docked with HLA-A*30:01 ( Figure 7A ) and epitope YNLAPFFTF with HLA-A*24:02 ( Figure 7B) . The structures were also energy minimized to reduce the overall potential energy between the epitope and MHC molecule. Furthermore, to get a better insight into the TCR interaction with the pMHC complex, we modeled the YNLAPFFTF-HLA-A*24:02 complex with the C1-28 TCR (RCSB PDB: 3VXM) structure specific to HLA-A24 ( Figure 7C ). Although several COVID-19 vaccines are currently available, the excessive mutations observed in the spike protein of Omicron can escape immune response, raising concerns over the efficacy of To design an efficient and safe vaccine that would provide broad protection across many ethnicities, we carefully selected the prediction tools based on their accuracy. The IEDB prediction servers are free and widely accepted in literature [50] . However, to select the most reliable T cell prediction tool, we used our previous study that analyzed the specificity and sensitivity of several MHCI and MHCII peptide prediction tools [43] . We used NetMHCpan EL 4.1 for CD8 epitope predictions and IEDB Recommended 2.22 for CD4 epitope predictions. Due to the lack of reliable immunogenicity prediction tools for CD4 peptides on IEDB.org, the IFNEpitope server was used to predict IFNγ inducing epitopes. IFNEpitope is predominantly designed to predict MHCII peptides with the capacity to induce IFNγ release and has an accuracy of 82.10%. The VaxiJen and AllerTOP servers have been regarded as highly reliable antigenicity and allergenicity prediction servers with an accuracy of 87% and 89%, respectively. For toxicity evaluation, ToxinPred was selected due to its accuracy of 94.50%. We revealed CD8+ and CD4+ T cell epitopes from SARS-CoV-2 S protein targeting the Omicron (BA.1/B.1.1.529) variant while considering many of its significant mutations such as A67V, del69-70, T95I, ins214EPE, S371L, S373P, S375F, Q493K, G496S, Q498R, N501Y, N679K, and P681H. We identified 8 immunogenic antigenic non-allergenic non-toxic CD8+ T cell epitopes (Figure 3 ) and 11 IFNγ inducing antigenic non-allergenic non-toxic stable CD4+ T cell epitopes that can provide a robust immune response and cover 76.16% and 97.46% of the world population, respectively (Tables 1 and 2) . Among the identified epitopes, 5 CD8 peptides (RSYSFRPTY, NLAPFFTFK, YNLAPFFTF, APFFTFKCY, and VLYNLAPFF) are located on the RBD region, suggesting their potential to target the impaired interaction between S protein and the neutralizing antibodies. To give a better insight into how our epitopes target Omicron S protein RBD mutations, we obtained the crystal structure of Fab fragment in complex with SARS-CoV-2 S protein (RCSB PDB 6XCM) ( Figure 4) . Moreover, our predicted CD8 epitope RSYSFRPTY targets the two important mutations, Q498R and N501Y, correlated with an enhanced binding affinity to the ACE2 and decreased binding affinity to neutralizing antibodies [51] . We predicted RSYSFRPTY to be immunogenic, highly antigenic, non-allergenic, and non-toxic with a world population coverage of 49.04% alone, suggesting its importance in epitope-based vaccine design. To further enhance the immune response, we identified two 15-mer peptides (SVLYNLAPFFTFKCY, VLYNLAPFFTFKCYG) that overlap with the identified CD8+ T cell epitopes on the RBD region. Together these epitopes can provide a strong and long-lasting immune response against the Omicron variant. We used our software tool PCOptim to validate the accuracy of the CD8 top epitope selection. Additionally, the PCOptim tool allowed us to determine the caveats in population coverage and thus identify the populations that may receive less coverage. The PCOptim tool provided us with an optimal set of epitopes that reaches the maximum possible population coverage given our raw dataset. Comparing this optimal dataset to our set of top epitopes revealed some overlap, confirming the success of our selection procedure. A total of three epitopes were found in both the optimal dataset and our top epitope datasets. Out of the 27 most common HLA alleles in the human population, we identified 5 alleles that are unaccounted for in the immunogenic, antigenic, nonallergenic, non-toxic, and stable epitope dataset (HLA-B*07:02, HLA-B*08:01, HLA-A*31:01, HLA-A*33:01, and HLA-B*51:01). The difference in population coverage between that optimal dataset and our set of 8 CD8 Omicron-related epitopes can be attributed to the five aforementioned HLA alleles. The only way to increase population coverage further is to include new epitopes that are predicted to be strong binders to the desired HLA alleles and pass the clinical checkpoint filters. While PCOptim does not account for the clinical checkpoint parameters we address in this study, it is a useful tool in determining several epitopes that are useful in obtaining high population coverage. The population coverage that we obtained from our top epitope dataset is substantially high, so adding new epitopes will make a very small impact on total population coverage. Previous studies have shown that SARS-CoV-2 may use mutations on the NTD region of S protein to escape potent polyclonal neutralizing responses [52] . Subsequently, we identified an immunogenic antigenic non-allergenic and non-toxic CD8+ T cell epitope REPEDLPQGF which consists of a novel insertion mutation ins214EPE, not observed in any other SARS-CoV-2 strains [53] . Although the mutation is believed to arise from co-infection of SARS-CoV-2 with HCoV-229E as the two share similar nucleotide sequences, we were unable to detect any common epitopes between the two human-coronavirus strains after running a multiple sequence alignment analysis. To create a robust immune response that targets the NTD region of SARS-CoV-2 Omicron, we identified an additional CD8+ T cell epitope (GVYFASIEK) and two high-quality CD4+ T cell epitopes (FLPFFSNVTWFHVIS and LPFFSNVTWFHVISG) which consist of mutation A67V and a pair of deletions at residues 69 and 70. We predicted another important immunogenic antigenic non-allergenic non-toxic CD8 peptide (KSHRRARSV) which is located at the furin cleavage site of SARS-CoV-2. The cleavage of the S protein into S1 and S2 is an essential step in viral entry into a host cell A large number of mutations on the RBD of S protein is speculated to help the virus escape neutralizing antibodies from natural and vaccine-induced immunity. Despite the loss of binding affinity to human ACE2 due to mutations such as K417N, Omicron has restored its strong binding to the ACE2 with mutations at residues 493, 496, 498, and 501. Previously we showed how epitopes can target the S protein interaction with neutralizing antibodies (Figure 4) . While completing the manuscript, Omicron-specific S protein structure in complex with human ACE2 (PDB 7T9K) became available which we used to visualize our predicted epitopes and their interaction with human ACE2 and Omicron S protein complex. We show that CD8+ T cell epitopes (RSYSFRPTY, NLAPFFTFK, YNLAPFFTF, APFFTFKCY, and VLYNLAPFF) on the RBD target the ACE2 binding site (Figure 5) . Together with the predicted CD4+ T cell epitopes, a robust immune response can be created which facilitates the production of antibodies, therefore blocking cell entry. Moreover, NTD and furin cleavage sites are critical to virus attachment to membrane protein and other sites, and we have identified epitopes that can target these sites (Figure 5) . With the emergence of new SARS-CoV-2 variants and their high likelihood to adapt to mutations on the RBD, NTD, and furin cleavage sites, we speculate that our identified epitopes can effectively target any future variants by blocking their binding to human ACE2. We have already observed that the BA. Tables 2,5) . Moreover, these epitopes are located on the S1/S2 cleavage site, associated with increased ACE2 binding affinity and impaired antibody recognition. Therefore, we propose that these epitopes can be used to target the critical site of SARS-CoV-2 S protein in both BA.1 and BA.2 lineages. As it is highly important to design an effective booster shot against the novel SARS-CoV-2 Omicron variant, we also attempted to design a protective vaccine against other SARS-CoV-2 strains. Subsequently, we identified 12 immunogenic CD8+ and 7 IFNγ inducing CD4+ T cell epitopes across Alpha, Beta, Delta, Gamma, Omicron (BA.1, BA.2), US variants (S protein mutations), and Cluster 5 mink variants that are predicted to protect 84.0% and 96.65% of the world population, respectively (Tables 1 and 3) . Moreover, all of the CD8 and CD4 epitopes were predicted to be antigenic, non-allergenic, and non-toxic for safe vaccine development. These epitopes will likely reactivate B and T cells against the original epitopes without creating a variant-specific immune response. The findings through computational analysis indicate that the designed epitopes could be vastly Available studies propose that CD8 and CD4 epitopes could be linked together using AYY linker and GPGPG linker, respectively [56] . It is detrimental to assess the safety and efficacy of proposed multi-epitope vaccine candidates in animal models. Murine models have low-cost maintenance and reflect the clinical signs, viral replication, and pathology of SARS-CoV-2 in humans [57] . We used the SYFPEITHI prediction server to predict murine MHC binding affinity to MHC I epitopes. The prediction server utilizes published T cell epitopes and motifs and is validated to be 80% accurate. We identified four Omicron BA.1 specific and one BA.2 specific S CD8+ T cell epitopes, as well as 7 common S CD8+ T cell epitopes among SARS-CoV-2 variants presented by murine MHC restriction. Ultimately, the predicted strong, intermediate, and weak MHCI binders can be used in further preclinical trials. The in-silico prediction analysis of the SARS-CoV-2 S protein reported in our study requires further laboratory validation to select the most efficient vaccine candidate. The recent emergence of the highly mutated Omicron variant of SARS-CoV-2 has disrupted confidence around whether the current vaccines and antibody therapies will provide long-term protection against the novel coronavirus. Subsequently, the possibility of escape from natural and vaccine-induced immunity has prompted an urgent need for new vaccine constructs which target the most concerning mutations on the variant. Using immunoinformatic methods and bioinformatics tools, we identified immunogenic antigenic non-allergenic non-toxic CD8+ T cell epitopes and IFNγ inducing antigenic non-allergenic non-toxic CD4+ T cell epitopes on the RBD, NTD, and furin cleavage sites of S protein which target Omicron specific mutations by blocking the binding of S protein to ACE2. Moreover, our analysis yielded common high-quality CD8+ and CD4+ T cell epitopes across Omicron (BA.1, BA.2) and other circulating SARS-CoV-2 variants including Delta, Alpha, Beta, Gamma, US variants (S protein mutations), and Cluster 5 mink variants with a world population coverage of 84.0% and 96.65%, respectively. We validated the findings using our software PCOptim which enabled us to identify an optimal set of epitopes reaching the maximum possible population coverage. Our in-silico epitope prediction together with murine MHC affinity prediction enables scientists to validate the efficacy of the proposed multi-peptide vaccine model through further preclinical studies. To understand the structural validation of the vaccine candidate and the binding affinity to MHC and TCR molecules, we performed a molecular docking analysis. To conclude, the multi-epitope vaccine constructs designed against S protein of SARS-CoV-2 by utilizing immunoinformatic methods may be considered as a new, safe, and efficient approach to control the Omicron variant as well as any future SARS-CoV-2 variants. and CD4+ T cell epitopes were predicted using freely available prediction tools on Immune Epitope Database (IEDB). MHCI binders were refined by immunogenicity and MHCII binders by IFNγ inducer capability. All CD8 and CD4 epitopes were then evaluated and selected based on their antigenicity, allergenicity, toxicity, and physicochemical properties. Population coverage was computed for both CD8 and CD4 epitopes with known MHC restriction. CD8 peptides were further validated using our software PCOptim. Murine MHC binding affinity to top CD8 epitopes was predicted and the peptide-MHC structures were modeled in three-dimensional analysis. Table 6 . Optimized list of epitopes returned by PCOptim. Input data included all raw epitopes with a rank less than 0.30 and immunogenicity greater than 0. HLA alleles that are bolded are also present in the HLA restrictions of epitopes that passed all clinical checkpoint parameters. RSYSFRPTY and YNLAPFFTF are present in Omicronvariant-specific S CD8 epitopes and VVFLHVTYV is present in the Top common epitopes across Omicron and other SARS-CoV-2 variants. Classification of omicron (B.1.1.529): SARS-COV-2 variant of concern. World Health Organization Omicron (B.1.1.529) Variant. 2021 Broadly neutralizing antibodies overcome SARS-CoV-2 Omicron antigenic shift Structural and functional properties of SARS-CoV-2 spike protein: potential antivirus drug development for COVID-19 The SARS-CoV-2 Spike Glycoprotein Biosynthesis, Structure, Function, and Antigenicity: Implications for the Design of Spike-Based Vaccine Immunogens SARS-CoV-2: Structure, Biology, and Structure-Based Therapeutics Development Receptor-binding domain-specific human neutralizing monoclonal antibodies against SARS-CoV and SARS-CoV-2 Striking Antibody Evasion Manifested by the Omicron Variant of SARS-CoV-2. 2021 Ancestral SARS-CoV-2-specific T cells cross-recognize Omicron (B.1.1.529). bioRxiv T-cell epitope vaccine design by immunoinformatics Epitope-Based Immunome-Derived Vaccines: A Strategy for Improved Design and Safety Memory T cell responses targeting the SARS coronavirus persist up to 11 years post-infection Convergent antibody responses to SARS-CoV-2 in convalescent individuals Definition of a human T cell epitope from influenza A non-structural protein 1 using HLA-A2.1 transgenic mice CTL responses of HLA-A2.1-transgenic mice specific for hepatitis C viral peptides predict epitopes for CTL of humans carrying HLA-A2.1 Contriving Multi-Epitope Subunit of Vaccine for COVID-19: Immunoinformatics Approaches SARS-CoV-2 Rapidly Adapts in Aged BALB/c Mice and Induces Typical Pneumonia We wish to acknowledge Lombardi Comprehensive Cancer Center, and Georgetown University Medical Center for their support.