Identification of immunodominant T cell epitopes in SARS-CoV-2 spike protein in Syrian hamsters

Abstract

T cell immunity has a crucial role in vaccine-induced protection against respiratory viruses, yet a detailed characterization of T cell responses and epitopes in Syrian hamsters, a highly utilized pre-clinical, small animal model for SARS-CoV-2 research, is lacking. In this study, using an intranasal Chimpanzee adenoviral vectored vaccine (ChAd-SARS-CoV-2-S), we characterized the T cell response to the spike protein of SARS-CoV-2 in Syrian hamsters and identified immunogenic CD4⁺ and CD8⁺ T cell epitopes using IFN-γ ELISpot assays and cell depletions. The mucosal ChAd-SARS-CoV-2-S vaccine elicited strong T cell responses, with evidence of CD4⁺ and CD8⁺ T cell activation in both lymphoid and mucosal tissues. Responses were directed toward the non-receptor-binding domain (RBD) regions of the spike protein, indicating that dominant T cell epitopes for hamsters reside elsewhere in this structural protein. Six different T cell epitopes (4 for CD4 and 2 for CD8) were identified in the spike protein, and epitope specific responses were detected in hamsters from two vendors, suggesting genetic similarity in terms of major histocompatibility complex (MHC) allele expression. Identifying T cell epitopes and characterizing T cell responses in lymphoid and mucosal compartments enhances the utility of Syrian hamsters as a preclinical model for SARS-CoV-2 vaccine studies.

INTRODUCTION

Cellular immunity has a pivotal role in controlling SARS-CoV-2 infection and enhancing the effectiveness of vaccines. CD4⁺ T cells support B cell activation and antibody production and have direct antiviral effects through cytokine production. Also, cytotoxic CD8⁺ T cells eliminate infected cells, thereby reducing disease severity and hospitalization risk. Unlike neutralizing antibodies, which can be readily evaded by viral mutations, T cell responses are less susceptible to single point mutations and generally encompass multiple epitopes, enabling cross-variant protection (1,2). Thus, understanding spike-specific T cell immunity is essential for developing vaccines that confer durable and broad protection against emerging variants (3,4). Among vaccine approaches, intranasal immunization has gained attention due to its ability to stimulate immunity in both systemic and mucosal compartments and prevent or limit infection at the portal of viral entry. Adenoviral-vectored intranasal vaccines, including ChAd-SARS-CoV-2-S (and its clinical derivative, iNCOVACC^®), have shown promising results in both preclinical and human studies (3–5). In mice, a single intranasal dose elicited robust and long-lasting mucosal T cell immunity, even in the absence of strong neutralizing antibody responses against immune-evasive Omicron subvariants like XBB.1.5 (5). Notably, CD8⁺ T cell depletion prior to viral challenge reduced protection, highlighting a central role of T cells in mediating cross-variant immunity (3,4).

The Syrian hamster is a well-established and highly relevant preclinical model for respiratory viral infections due to its physiological and immunological similarities to humans (6). Syrian hamsters are naturally susceptible to SARS-CoV-2 (7–9), as well as other respiratory viruses such as influenza (10) and respiratory syncytial virus (11–13), and also have been used to model hemorrhagic fever caused by Ebola and Marburg viruses (14–17). Following SARS-CoV-2 infection, hamsters exhibit hallmark features of Coronavirus disease 2019 (COVID-19), including weight loss, lung pathology, and prolonged immune activation, making them a valuable model for studying both acute and long-term disease mechanisms (18).

Our group has focused on testing different therapeutic interventions to understand both protection and viral dissemination in the context of transmission studies, which is a key advantage of the Syrian hamster model. In hamsters, intramuscular mRNA vaccines protect against disease, but do not fully block SARS-CoV-2 transmission. In contrast, intranasal mucosal vaccination with a ChAd-vectored vaccine significantly reduced viral burden in upper and lower respiratory tract tissues and prevented onward transmission (19,20). Potently neutralizing monoclonal antibodies also prevented infection and aerosol transmission (21). Despite their widespread use in COVID-19 studies, hamster research has primarily focused on humoral immune responses, with relatively few studies investigating T cell responses (22–27), perhaps because the T cell epitopes targeted in Syrian hamsters following infection or vaccination remain undefined. Given the critical role of T cells in controlling SARS-CoV-2 infection and providing cross-variant immunity (28–31), this gap limits the utility of this model in evaluating next-generation vaccines. Mapping these epitopes, especially those that are conserved and immunodominant, is vital for advancing vaccine design and the development of T cell-based immunotherapies (32–37). Here, we identify immunodominant CD4⁺ and CD8⁺ T cell epitopes in the Syrian hamster model and validate their use in evaluating vaccine-induced immunity.

RESULTS

Single-dose intranasal ChAd-SARS-VoV-2-S vaccination elicits a robust T cell response in Syrian hamsters.

To assess the cellular immune response induced by an intranasal (IN) SARS-CoV-2 vaccine, we immunized Syrian hamsters with a chimpanzee adenoviral vectored vaccine encoding the prefusion-stabilized spike (S) protein of the Wuhan-1 strain of SARS-CoV-2 (ChAd-SARS-CoV-2-S) or empty vector (38). The S-specific T cell responses in the spleen, cervical draining lymph nodes (DLN), and lungs were quantified at 7- and 30-days post-immunization using IFN-ɣ ELISpot assays and ex-vivo re-stimulation with mega pools of overlapping 15-mer peptides corresponding to the S1 and S2 subunits of the S protein. S-specific T cells were present in both systemic and mucosal sites following immunization with ChAd-SARS-CoV-2-S but not the empty vector control (Fig. 1A and Fig. S1). In both the spleen and DLN at day 30 post vaccination, ChAd-SARS-CoV-2-S induced ~220 S1-specific T cells and ~140 S2-specific spot-forming cells/million cells (SFC/10⁶) for a total of ~360 SFC/10⁶ cells. A greater response was detected in the lungs, with ~540 S1-specific SFC/10⁶ and ~450 S2-specific SFC/10⁶ observed (Fig. 1A), amounting to a total of ~1,000 SFC/10⁶ cells. A similar pattern was seen at day 7 post-vaccination, with ~350, ~650, and ~800 total S-specific SFC/10⁶ in the spleen, DLN, and lungs, respectively (Fig. S1B).

Figure 1. Single-dose ChAd-SARS-CoV-2-S vaccination in Syrian hamsters induces a T cell response dominated by CD4⁺ T cells in lymphoid organs and the lung.

(A) IFN-γ ELISpot assay on cells obtained from the spleen (left), DLNs (middle) and lungs (right) of Syrian hamsters 30 days after intranasal immunization with 10¹⁰ virus particles of a ChAd-SARS-CoV-2-S using a mega pool of 25 overlapping peptide pools covering the S protein of the WA1/2020 strain. Thirty days after immunization, tissues were collected, processed by IFN-γ ELISpot assay to quantify the number of subunit 1 (S1), subunit 2 (S2), or full-length spike (S1+S2) specific T cells. Results are presented as the number of SFC/10⁶ in each tissue (n=12) across 3 independent experiments. (B-D) Impact of ex-vivo CD4⁺ T cell depletion on S1, S2, or full-length S-specific responses in the spleen (B), DLN (C), and lungs (D) of six Syrian hamsters from two independent experiments. Results are expressed as SFC/10⁶ cells, and the fold reduction in average spot count is depicted above each graph. Fold change in responses to S1, S2, and S1+S2 peptides was calculated by comparing values with and without CD4⁺ T cell depletion. Statistical analysis: one-way ANOVA with Tukey post-test (A), and paired t-test (B-D): ns = not significant, * p < 0.05, ** p < 0.01, **** p < 0.0001.

To assess the contribution of CD4⁺ T cells to the T cell response after vaccination, we ex-vivo depleted the CD4⁺ T cells with an anti-CD4 antibody (clone GK1.5, BioLegend), quantified the remaining S-specific T cell response by ELISpot, and confirmed depletion by flow cytometry (Fig. S1D). At day 30 post vaccination, the ex-vivo depletion of CD4⁺ T cells led to a 9- and 18-fold reduction in S-specific T cells in the spleen and DLN, respectively, indicating that the response in lymphoid organs was mediated primarily by CD4⁺ T cells (Fig. 1B–C). Similar results were observed at 7 days post-immunization, with a 13- and 16-fold decrease in number of S-specific T cells in the spleen and DLN after CD4⁺ T cell depletion (Fig. S1C). The fold reduction was greater for S1-specific T cells than S2-specific T cells, suggesting the presence of one or more immunodominant CD4⁺ T cell epitopes in the S1 subunit of the S protein. In contrast, ex-vivo depletion of CD4⁺ T cells from cells obtained from the lung reduced the S-specific T cell response 1.7-fold on day 30 (Fig. 1D) and 5-fold on day 7 (Fig. S1C), indicating a greater contribution of CD8⁺ T cells to the S-specific T cell response in the lungs. Thus, ChAd-SARS-CoV-2-S induces a robust T cell response in hamsters, which is detectable as early as day 7 post-vaccination (Fig. S1) and sustained through day 30 (Fig. 1).

Hamster CD4⁺ and CD8⁺ T cells target multiple epitopes in the S protein of SARS-CoV-2.

To identify immunodominant CD4⁺ and CD8⁺ T cell epitopes in the S protein for hamsters, we immunized eight 6- to 10-week-old animals with ChAd-SARS-CoV-2-S, and 7 days later performed ELISpot assays on cells from the spleen, DLN, and lungs using 25 different mini-pools of overlapping 15-mer peptides spanning the full-length S protein (Table S1). Pools 1, 3, 13, 23, and 25 were the only ones that elicited detectable T cell responses in all the tested animals (8/8) (Fig. 2A). Among these, pool 13 induced the strongest T cell response, with an average of ~200 SFC/10⁶ cells. Pool 25 was the second most dominant, with an average of ~130 SFC/10⁶ while pools 1, 3, and 23 induced 87, 56, and 57 specific SFC/10⁶ cells respectively (Fig. 2A). The receptor binding domain (RBD) (pools 7–11) of the S protein, which correspond to the most variable regions among variants (39), did not contain any immunodominant T cell epitopes (Fig. 2A). Weak inconsistent responses were observed in cells stimulated with pools 7, 11, 12, 16, 17, 18, 20, 22, and 24 (< 4/8 animals responded), whereas all the remaining pools were consistently negative. These findings suggest that specific regions of S are immunogenic and may serve as targets for SARS-CoV-2 vaccine-induced T cell immunity. To determine whether these responses were specific to intranasal ChAd-SARS-CoV-2-S vaccination, two Syrian hamsters were immunized intramuscularly with two doses of an mRNA spike vaccine (BNT162b2, 5 μg/dose), administered 21 days apart. Animals were sacrificed on day 7 post-boost. Consistent with the ChAd-SARS-CoV-2-S findings, peptide pools 1, 13, 23, and 25 elicited spike-specific IFN-γ responses, with the strongest activity observed in pool 13. Additional reactivity was observed with peptide pools 18 and 20, whereas a response was not detected with pool 3 (Fig. S2).

Figure 2: Hamster CD4⁺ and CD8⁺ T cells target multiple epitopes in the S protein of SARS-CoV-2.

(A) IFN-γ ELISpot assay on cells obtained from the draining DLN of ChAd-SARS-CoV-2-S immunized Syrian hamsters using 25 overlapping peptide pools covering the SARS-CoV-2 (WA1/2020) spike protein. Results are presented as SFC/10⁶ cells. Each color represents a different hamster, and the data are from two independent experiments with four hamsters per experiment. (B) Identification of the immunodominant peptides in pools 1 (peptide 10, S_46–60), 3 (peptide 4, S_116–130), 13 (peptide 7, S_631–645), 23 (peptide 4, S_1116–1130) and 25 (peptides 9 and 10, S_1251–1265 and S_1256–1270). IFN-γ ELISpot assay were performed on the spleen, DLN, and lungs of five hamsters using the individual peptides that comprised each of these peptide pools. Results are expressed as SFC/10⁶ cells, and each shade of color represents a different hamster. (C) Impact of ex-vivo CD4⁺ T cell depletion on peptide specific responses in the spleen, DLN, and lungs of five animals from two independent experiments. Results are expressed as SFC/10⁶ cells, and all three tissues are presented in one graph. The fold difference is depicted above each graph. Statistical analysis: paired t-test (C): * p < 0.05, ** p < 0.01, *** p < 0.001.

Identification of individual hamster T cell epitopes in the spike protein of SARS-CoV-2.

To further delineate the dominant T cell epitopes, we screened individual peptides from the five most immunogenic pools (1, 3, 13, 23, and 25). The isolated cells from the spleen, DLN, or lungs were stimulated with 10 μg/mL of each peptide. In pool 1, peptide 10 (S_46–60, SVLHSTQDLFLPFFS) elicited T cell responses in all three tissues of all five hamsters tested (Fig. 2B). Ex-vivo CD4⁺ T cell depletion prior to peptide stimulation reduced the S_46–60 peptide-specific response 60-fold in the spleen, DLN, and lungs (p < 0.05, Fig. 2C), confirming it as a CD4⁺ T cell epitope. In pool 3, peptide 4 (S_116–130, SLLIVNNATNVVIKV) was the only peptide that re-stimulated S-specific T cells in all three tissues of all five hamsters (Fig. 2B). CD4⁺ T cell depletion reduced this response 30-fold (p <0.01, Fig. 2C), indicating that S_116–130 is a CD4⁺ T cell epitope. For pool 13, peptide 7 (S_631–645, PTWRVYSTGSNVFQT) was the only peptide that induced a detectable response across all five immunized hamsters (Fig. 2B). Ex-vivo CD4⁺ T cell depletion abrogated the response (30-fold, p < 0.001, Fig. 2C), also confirming its CD4⁺ T cell dependency. Collectively, these results suggest that S_46–60, S_116–130, and S_631–645 contain dominant CD4⁺ T cell epitopes in Syrian hamsters. In pool 23, peptide 4 (S_1116–1130, TTDNTFVSGNCDVVI) re-stimulated T cells from all ChAd-SARS-CoV-2-S vaccinated hamsters (Fig. 2B). Depletion of CD4⁺ T cells reduced the peptide 4 specific response 3-fold (p < 0.05, suggesting that S_1116–1130 may contain both CD4⁺ and CD8⁺ T cell epitopes (Fig. 2C). Finally, in pool 25, peptides 9 (S_1251–1265, GSCCKFDEDDSEPVL) and 10 (S_1256–1270, FDEDDSEPVLKGVKL) re-stimulated a T cell response across all tissues and hamsters tested (Fig. 2B). CD4⁺ T cell depletion reduced the peptide-specific response by just 2-fold (p < 0.01) indicating that these peptides also may contain both CD4⁺ and CD8⁺ epitopes (Fig. 2C). Collectively, our results identify immunodominant T cell epitopes in Syrian hamsters following intranasal ChAd-SARS-CoV-2-S vaccination, highlight the differential immunodominance of CD4⁺ and CD8⁺ T cells across distinct anatomical compartments, and demonstrate a high degree of similarity in peptide specific T cell responses among hamsters.

Characterization of the hamster SARS-CoV-2 CD4⁺ T cell epitope S_631–645.

To more precisely fine map the CD4⁺ T cell epitope contained within the S_631–645 peptide, we generated overlapping peptides flanking the peptide sequence (±1 amino acid) (Table S2). T cells from ChAd-SARS-CoV-2-S immunized hamsters were restimulated with 10 μg/mL of each peptide in an ELISpot assay; this analysis revealed S_633–644 (WRVYSTGSNVFQ) as the optimal peptide for CD4⁺ T cell activation (Fig. 3A). To further define the epitope boundaries, we tested overlapping peptides with incremental deletions of 1, 2, 3, or 4 amino acids from the N- or C-terminus of S_631–645 (Table S2). These experiments confirmed S_633–644 as an immunodominant CD4⁺ T cell epitope, as peptides lacking key residues at the N- or C-terminus exhibited reduced responses (Fig. S3A–B). To assess the binding affinity of the peptide to MHC class I molecules, a serial dilution binding assay was performed. The effective peptide concentration at which 50% of maximal T cell response was detected (EC₅₀) was determined to be 1 μg/mL, indicating moderate binding affinity (Fig. 3B). To further confirm that the S_631–645 peptide (WRVYSTGSNVFQ) is a CD4 epitope, we performed IFN-γ ELISpot assays following CD4+ or CD8+ cell depletion. Stimulation with the S_631–645 peptide showed that CD4 depletion resulted in a 7-fold reduction in IFN-γ production compared to non-depleted controls (p < 0.01), whereas CD8 depletion had no significant effect (Fig. 3D). These results confirm that the S_631–645 peptide elicits a CD4⁺ T cell–dependent response.

Figure 3: Identification of SARS-CoV-2 CD4⁺ T cell epitope S_631–645 in Syrian hamsters.

(A) Syrian hamsters (n=9) were obtained from CRL and received a single intranasal dose of ChAd-SARS-CoV-2-S (10¹⁰ virus particles). At day 7 post-vaccination, spleen, draining lymph node (DLN), and lung tissues were processed to identify the optimal peptide required for CD4⁺ T cell activation using an IFN-γ ELISpot assay with overlapping peptides flanking the S_631–645 sequence. Results are presented as SFC/10⁶ cells and representative of two independent experiments. The pale gray line corresponds to the response from individual hamsters, and the red line corresponds to the average response between all the tested animals. (B) The EC₅₀ values of the S_631–645 peptide were determined in lung (red), DLN (green), and spleen (blue) of 4–5 hamsters by serial dilution of the peptide and quantifying the number of SFC by IFN-γ ELISpot. Results are expressed as SFC/10⁶ cells, and all three tissues are presented in one graph. (C) Cross-reactivity of T cell responses to S_631–645 peptides derived from SARS-CoV-2 and SARS-CoV-1 was evaluated in ChAd-SARS-CoV-2-S vaccinated hamsters. Results are expressed as SFC/10⁶ cells, and all three tissues are presented in one graph. The fold difference is depicted above the graph. (D) Cells from the spleen, DLN, and lungs of ChAd-SARS-CoV-2-S vaccinated hamsters were depleted of CD4⁺ or CD8⁺ T cells, stimulated with S_631–645 peptide, and the number of IFN-γ secreting cells was quantified. Results are expressed as SFC/10⁶ cells. All three tissues are presented in one graph, and the fold difference is depicted above each graph. (E) T cell reactivity to S_631–645 was quantified by IFN-γ ELISpot in spleen, lung, and DLN of ChAd-SARS-CoV-2-S immunized (10¹⁰ virus particles) Syrian hamsters from Inotiv. The results are from one experiment with five animals. (F) The optimal peptide sequence within the S_631–645 peptide was defined in spleen, lungs, and DLN of ChAd-SARS-CoV-2-S immunized hamsters obtained from Inotiv (n=2). The pale gray line corresponds to the response from individual hamsters, and the red line corresponds to the average response between all the tested animals. (G) To determine the MHC class II restriction of the S_631–645 epitope, naïve hamster splenocytes were pulsed with a 10-fold serial dilution of S_631–645 for 1 h at 37°C, washed, and treated with blocking antibodies against I-Ek (14–4-4S), I-A (Y3JP), HLA-DR (L243), or left untreated. The cells were co-cultured with splenocytes or lung cells from ChAd-SARS-CoV-2-S immunized hamsters at a 1:1 ratio. IFN-γ ELISpot was used to quantify S_631–645-specific responses. Peptide-pulsed splenocytes from naive and immunized hamsters served as negative and positive controls, respectively. The left panel is a representative result for the lungs of one hamster. The right panel is the number SFC/10⁶ cells from four hamsters (spleen and lungs) at the 10 μg/mL peptide concentration. Statistical analysis: paired t-test (C-D), one-way ANOVA with a Tukey post-test (E), and one-way ANOVA with a Dunnett post-test comparing each condition against the naïve cells + S_631–645 (green) (G): ns= not significant, ** p < 0.01.

To determine to what degree the epitope was conserved across SARS-CoV-2 variants, we retrieved full-length spike glycoprotein sequences from representative lineages and performed multiple sequence alignment using Clustal Omega. Sequence alignment of the S protein region spanning residues 633–644 from multiple SARS-CoV-2 variants revealed a strong conservation of the S_633–644 epitope (Fig. S4). This 12-residue motif was 100% identical in all major SARS-CoV-2 lineages, including the ancestral Wuhan strain, WA/2020, B.1.1.7, P.1, BA.1, BA.2.12.1, BA.5, XBB.1.5, EG.5.1, JN.1., and NB.1.8.1. In a related Sarbecovirus, RaTG13, the epitope was also fully preserved, indicating strong evolutionary conservation to SARS-CoV-2. However, clade 1b human SARS-CoV-1 strains (e.g., Urbani, CUHK-W1, GZ02, A031, A022) and clade 1b bat-derived SARS-like coronaviruses (e.g., WIV1, WIV16, and SHC014) had several substitutions in the peptide with 83–91% identity conservation (Fig. S4). Previous experiments have shown that sequence conservation above 67% is often associated with CD4⁺ T cell cross-reactivity (40). To evaluate potential T cell cross-reactivity between SARS-CoV-2 and SARS-CoV-1, we stimulated T cells from ChAd-SARS-CoV-2-S immunized hamsters with the corresponding SARS-CoV-1 peptide (SARS-CoV-1-S_633–644, PAWRIYSTGNNVFQT) and compared the number of spots to the SARS-CoV-2 peptide (S_631–645, PTWRVYSTGSNVFQT). The T cell response to SARS-CoV-1-S_633–644 was only marginally reduced (1.9 -fold, p < 0.01, Fig. 3C), confirming the cross-reactivity between SARS-CoV-1 and SARS-CoV-2 epitopes.

While the similarity in T cell responses between individual hamsters suggests that the Syrian hamsters at Charles River Laboratories (CRL) are inbred, the breeding status of hamsters at other vendors is not known. To assess this, we measured the S_631–645 specific response in animals from Inotiv (previously Harlan and Envigo) that were immunized with ChAd-SARS-CoV-2-S. Like Syrian hamsters from CRL, hamsters from Inotiv exhibited strong T cell responses to the S_631–645 peptide in the three tested tissues (Fig. 3E). Additionally, epitope mapping using overlapping peptides flanking S_631–645 in these hamsters identified the same optimal epitope, WRVYSTGSNVFQ (Fig. 3F).

Identification and functional evidence of I-Ek-Like MHC Class II restriction of the S_631–645 epitope in Syrian hamsters.

To identify the putative hamster MHC class II allele that presents the S_631–645 peptide, as a first step, we determined whether commonly used anti-MHC antibodies that bind class I and II MHC molecules expressed in other species (i.e., humans, mice, or macaques) would cross-react with Syrian hamster MHC. A panel of different anti-MHC antibodies was tested for reactivity with Syrian hamster-derived splenocytes, utilizing murine and human cells as specificity controls and further validated using a SV40-immortalized Syrian hamster B cell line (41) (Table S3 and Fig. S5). MHC class I specific antibodies 34–5-8S, B8–24-3, and 2G5, and MHC class II specific antibodies 14–4-4S and Y3PJ bound to hamster splenocytes and the immortalized cell line. To identify the putative hamster MHC class II allele that presents the S_631–645 peptide, we tested anti-mouse H2-E^k (clone 14–4-4S) and anti-mouse H2-A (clone Y3JP) blocking antibodies. In the presence of 14–4-4S, but not Y3JP or an anti-human MHC class II antibody, we saw a significant reduction (3-fold, p < 0.05) in the number of S_631–645-specific T cells (Fig. 3G), suggesting that the S_631–645 peptide is restricted by a H2-E^k-like MHC class II molecule in Syrian hamsters. Computational analysis using NetMHCIIpan (version 4.1) via the IEDB Analysis Resource predicted 'VYSTGSNVF' as the core binding sequence within S_631–645 for the H2-E^k MHC class II molecule. In this core, residues V, T, S, and F at positions P1, P4, P6, and P9, respectively, are predicted to serve as anchor residues, with P1 and P9 acting as the primary anchors (42).

Identification and characterization of SARS-CoV-2 CD8⁺ T cell epitopes in Syrian hamsters.

Our results, with peptides S_1251–1265 and S_1256–1270, suggested the presence of an overlapping and shared immunodominant CD8⁺ T cell epitope near the C-terminal end of the S protein. To identify the optimal peptide and responding T cell type at greater resolution, we performed IFN-γ ELISpot assays on spleens, DLN, and lungs 7 days after intranasal ChAd-SARS-CoV-2-S immunization using overlapping 15-mer peptides spanning the two sequences (Table S4). Peptides 15 through 21, containing the central 9-mer peptide DEDDSEPVL (S_1257–1265), preferentially restimulated hamster T cells (Fig. 4A). We also screened a panel of 8 to 10-mer peptides computationally predicted to bind to hamster MHC (Table S4). Among these, the 9-mer DEDDSEPVL induced the strongest recall response, establishing it as the optimal minimal epitope (Fig. 4B). An additional peptide, SEPVLKGV (S_1261–1268), also triggered a strong IFN-γ response in some animals, although this was not supported by overlapping peptide mapping, suggesting it might represent a subdominant epitope or a result of differential antigen processing (Fig. 4B). We also detected a low, yet consistent, response to peptide S_1242–1256 (SCLKGCCSCGSCCKF), suggesting the presence of lower-frequency or tissue-specific CD8⁺ T cell targets near the C-terminus of the S protein. To evaluate the relative antigen sensitivity of the T cell recognizing these epitopes, we performed titration experiments across a range of peptide concentrations. The calculated EC₅₀ values for DEDDSEPVL and SEPVLKGV were 110 ng/mL (range 9–694 ng/mL) and 600 ng/mL (range 60–1400 ng/mL), respectively, with DEDDSEPVL exhibiting a lower EC₅₀, indicating that T cells specific for DEDDSEPVL exhibit greater antigen sensitivity and higher functional avidity (Fig. 4C). S_1257–1265 and S_1261–1268 are in the cytoplasmic tail of the S protein. This region contains di-acidic and di-hydrophobic endoplasmic reticulum (ER) export motifs that are crucial for COPII-dependent export from the ER (43). Due to their structural and functional importance, these sequences are 100% conserved among SARS-CoV-2 variants and related sarbecoviruses (Fig. S4). Like the immunodominant CD4⁺ T cell epitope S_633–644, the CD8⁺ T cell epitope S_1257–1265 elicited a strong and specific T cell response in Syrian hamsters from both CRL and Inotiv vendors (Fig. 4D). The comparable magnitude of the response, and similarity in epitope boundaries across these vendor-sourced populations suggests conserved CD8⁺ T cell recognition, further supporting the hypothesis that available Syrian hamster populations share MHC haplotypes (Fig. 4D). To confirm CD8⁺ T cell restriction, depletion experiments were conducted prior to stimulation with peptides. CD8⁺ T cell depletion significantly reduced the IFN-γ response to S_1257–1265 (DEDDSEPVL, p < 0.01) and S_1261–1268 (SEPVLKGV, p < 0.05), whereas CD4⁺ T cell depletion had no significant effect, validating that this epitope is recognized by CD8⁺ T cells (Fig. 4E).

Figure 4: Identification of SARS-CoV-2 CD8⁺ T cell epitopes S_1267–1266 and S_1261–1268 in Syrian hamsters.

(A) Syrian hamsters (n=4) received a single intranasal dose of ChAd-SARS-CoV-2-S (10¹⁰ virus particles). At day 7 post-vaccination, spleen, draining lymph node (DLN), and lung were processed to identify optimal T cell epitopes using an IFN-γ ELISpot assay with overlapping 15-mer peptides flanking the S_1251–1265 and S_1256–1270 epitope sequences. The results are reported as SFC/10⁶ cells. The pale gray line corresponds to the response from individual hamsters, and the red line corresponds to the average response between all the tested animals. (B) Predicted CD8⁺ T cell epitopes (8–10-mers) were screened using an IFN-γ ELISpot assay across different tissues (spleen, DLN, and lung) in ChAd-SARS-CoV-2-S vaccinated hamsters. The results are reported as SFC/10⁶ cells, and each color is a different tissue from a different hamster. (C) Titration experiments were conducted using 10-fold serial dilutions of 9-mer peptides S_1257–1265, S_1261–1268, and S_1242–1256 to determine the relative affinity of the epitope in lungs (red), DLN (green), and spleens (blue) of three hamsters. The results are reported as SFC/10⁶ cells. (D) Overlapping peptides spanning the region containing S_1251–1265 and S_1256–1270 epitopes were used to determine the optimal T cell epitope in Syrian hamsters (n = 3) from Inotiv immunized 7 days earlier with ChAd-SARS-CoV-2-S (10¹⁰ virus particles). The pale gray line corresponds to the response from individual hamsters, and the red line corresponds to the average response between all the tested animals. (E) Cells from the spleen, DLN, and lungs of ChAd-SARS-CoV-2-S vaccinated hamsters were depleted of CD4⁺ or CD8⁺ T cells and stimulated with S_1257–1265 or S_1261–1268. The number of IFN-γ secreting cells was enumerated and reported as SFC/10⁶ cells and the fold difference in SFC is depicted above each graph. Statistical analysis: paired t-test (E): ns= not significant, * p < 0.05, ** p < 0.01.

DISCUSSION

In this study, we characterized the T cell response in Syrian hamsters following intranasal immunization with the ChAd-SARS-CoV-2-S vaccine and mapped immunogenic T cell epitopes. A single intranasal dose induced a rapid and robust T cell response detectable in both mucosal and lymphoid compartments. Using overlapping peptide pools spanning the SARS-CoV-2 spike protein, we identified five regions containing immunodominant T cell epitopes. Further analysis revealed multiple CD4⁺ and CD8⁺ T cell epitopes recognized by hamsters sourced from different commercial vendors. These findings add to the growing body of preclinical work by providing a detailed functional analysis of T cell immunity in Syrian hamsters, a model widely used in SARS-CoV-2 research due to its susceptibility to infection and its ability to recapitulate key aspects of human disease. This work enhances the utility of the hamster model for vaccine research, particularly in the face of limited species-specific immunological tools and reagents.

Although it remains unclear whether Syrian hamsters represent inbred or outbred populations, historical records indicate that current colonies such as those maintained by CRL and Inotiv descend from just two littermates captured in Syria in 1930 (44). This shared ancestry likely explains the apparent lack of MHC polymorphism in these animals (45–47). For instance, tumor cell lines derived from hamsters are transplantable between individuals without rejection (48,49), and tissue grafts generally are not rejected (47). Similarly, adoptive T cell transfer between animals does not result in graft-versus-host disease. In our study, we observed similar CD4⁺ and CD8⁺ T cell responses between hamsters from the two vendors, supporting the idea that these animals share the same MHC class I and II alleles. Together, these findings suggest that Syrian hamsters used commonly in research laboratories are genetically similar at MHC loci and may be partially or fully inbred.

In humans, the RBD is a central target for the adaptive immune response, particularly for potently neutralizing antibodies, many of which bind to ACE2-binding region (50–52). Yet, this region is a mutational hotspot that represents an immune pressure resulting from RBD mutations in variants (53). Despite its immunodominance in antibody responses, the RBD contains relatively few T cell epitopes compared to other regions of the Spike protein and non-Spike viral proteins (50,54). In Syrian hamsters, prior studies demonstrated that vaccination or infection with full-length S triggers T cell responses; however, these studies did not investigate RBD as a source of T cell epitopes (55). Our data reveal that the RBD region lacks dominant T cell epitopes in Syrian hamsters and thus, predominantly induces B cell rather than T cell immunity. These findings highlight species-specific differences in T cell recognition of the spike protein, suggesting that while the RBD is highly immunogenic for mouse T cells, it elicits limited T cell responses in humans and hamsters. This is an important point and underlines the relevance of the hamster model to evaluate T cell immunity to vaccines and infections. Our epitope mapping efforts revealed that S-specific T cell responses in hamsters are focused on discrete regions, with pools 13 and 25 being particularly immunodominant. The strongest CD4⁺ T cell response was directed against peptide S_631–645 (PTWRVYSTGSNVFQT), located outside the RBD, with high frequencies of specific T cells observed across the spleen, DLN, and lungs. Using deletion scanning, we refined the minimal epitope to S_631–645 (WRVYSTGSNVFQ), a sequence that is highly conserved among global SARS-CoV-2 isolates and maintains cross-reactivity with SARS-CoV-1, thus suggesting a key functional role and reduced susceptibility to mutation. Furthermore, compared to related sarbecoviruses like RaTG13, WIV1, and SARS-CoV strains, WRVYSTGSNVFQ is largely conserved with only minor regional differences in the computationally predicted anchor residues (V, T, S and F).

Peptides S_1116–1130, S_1251–1265, and S_1256–1270 elicited robust CD8⁺ T cell responses in the lung. This finding reinforces the importance of mucosal vaccine delivery for generating local effector cells (56–58), a feature that systemically delivered mRNA or protein-based platforms do not match. Through overlapping peptide screening of the SARS-CoV-2 S protein, we identified a highly immunodominant CD8⁺ T cell epitope, S_1257–1265, located within the conserved S2 domain. This 9-mer epitope, present in peptides S_1251–1265 and S_1256–1270, elicited strong CD8⁺ T cell restricted IFN-γ responses, as confirmed by depletion assays. This response was localized to the lungs of vaccinated animals, further supporting a tissue-resident memory phenotype, which ultimately will require tetramer generation and flow cytometry analysis to confirm. DEDDSEPVL is conserved across all major Omicron subvariants, including XBB.1.5, EG.5.1, JN.1, and NB.1.8.1, human SARS-CoV strains, and bat SARS-related coronaviruses, suggesting it might serve as a vaccine-specific, cross-variant CD8⁺ T cell target in hamsters. When comparing our data to previous studies, we found that several T cell epitopes identified in Syrian hamsters are conserved and share cross-species similarity with epitopes recognized by human MHC alleles (59). This also supports the relevance of the hamster model for preclinical vaccine evaluation and highlights its translational potential in studying SARS-CoV-2 immunity. For instance, the peptide S_633–644, which triggered a robust CD4⁺ T cell response in hamsters, is predicted to bind multiple human MHC class II alleles, including HLA-DRB1*07:01, HLA-DRB1*15:02, and HLA-DP molecules. Notably, this same peptide also overlaps with epitopes reported to bind human class I alleles such as HLA-A*24:02, HLA-B*15:01, and HLA-A*30:01 in previous studies (59–65). This potential cross-recognition highlights its possible candidate use for broad-coverage vaccine designs. Similarly, the S_1257–1265 epitope, associated with CD8⁺ T cell responses in hamsters, overlaps with epitopes known to bind human HLA-DRB3*01:01, HLA-B*40:01, and other class II molecules (59,60,62,66–68). Furthermore, S_1261–1268, part of the longer immunogenic epitope S_1261–1270, is a known binder of HLA-B*40:01 and HLA-B*40:03, and is functionally relevant in human CD8⁺ T cell responses (60,62,69). Another epitope of interest, S_1242–1256, showed predicted binding to HLA-DRA01:01/DRB104:01 (70), further supporting the presence of conserved CD4⁺ T cell epitopes between hamsters and humans. This analysis supports the possible use of the hamster model for dissecting the T cell landscape in response to SARS-CoV-2 infection and vaccination. The conservation of these epitopes across species strengthens the rationale for including such epitopes in the design of universal SARS-CoV-2 vaccines and for evaluating T cell-based correlates of protection. Future work using transgenic hamster models expressing human MHC alleles may allow for more precise epitope validation, tracking of antigen-specific T cells in vivo, and greater insight into the correlations of protection.

Limitations of the study.

(a) The peptide-specific responses were not evaluated in the context of a SARS-CoV-2 infection; relative responses may differ due to the presence of more immunodominant epitopes in other viral proteins. (b) We were unable to determine the exact MHC class I or II restriction of T cell epitopes beyond S_631–645. The Syrian hamster MHC region contains multiple class I and II candidates, and functional evaluation of each would require substantial resources. Accordingly, we did not generate MHC class I or II tetramers to track T cell responses over time or across tissues. (c) Only IFN-γ production was measured, potentially underestimating the magnitude and functional diversity of T cell responses. Additional readouts such as TNF, IL-2, and granzyme B could provide a more comprehensive assessment; however, the limited availability of validated immunological reagents for Syrian hamsters remains a major constraint. (d) We did not compare epitopes induced by intranasal vaccination to those elicited by natural SARS-CoV-2 infection, which may differ in breadth and immunodominance. (e) Our study focused on primary immunization; the impact of repeated exposures or boosting on the quality, hierarchy, or breadth of T cell responses was not assessed. (f) The durability of these T cell responses, including persistence of immunodominant epitopes, was not evaluated.

In summary, this study provides a detailed epitope-level analysis of S-specific CD4⁺ and CD8⁺ T cell responses in Syrian hamsters, highlighting their value for preclinical vaccine evaluation. By identifying immunodominant peptides, particularly in the context of mucosal immunity, it establishes a framework for studying T cell responses in SARS-CoV-2 and other infectious or disease models.

MATERIALS AND METHODS

Hamster Experiments.

All animal experiments were conducted in compliance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Washington University School of Medicine (assurance number A3381–01). Five-week-old male Syrian hamsters were obtained from CRL or Envigo/Inotiv and housed in a controlled environment at Washington University. Animals were acclimated for five days before the start of the experiments. On the day of vaccination, hamsters were anesthetized with isoflurane and immunized with a single-dose, IN-delivered replication-defective ChAd-SARS-CoV-2-S encoding a pre-fusion stabilized Wuhan-1 SARS-CoV-2 S protein, GenBank MN908947.3 (71), or empty vector (ChAd). A total of 10¹⁰ virus particles were administered in a 100 μL volume of phosphate-buffered saline (PBS). Additional animals were immunized twice intramuscularly with 5 μg of an mRNA vaccine encoding a pre-fusion stabilized S protein (BNT162b2) (72). To assess immune responses, animals were euthanized at either 7- or 30-days post-vaccination, perfused with 15 mL of PBS, and spleen, DLNs, and lungs were collected. Cervical lymph nodes (DLN) were collected as the primary regional lymph nodes draining the nasal cavity and upper respiratory tract following intranasal immunization (73–75). The spleen was analyzed to evaluate systemic immune responses. In addition, lungs were harvested to assess local mucosal immune responses at the dominant site of SARS-CoV-2 infection. The latter were transferred to 15 mL conical tubes containing 5 mL of ice-cold RPMI-1640 medium supplemented with 2% fetal bovine serum (FBS) (R2). Tissue digestion was initiated by cutting the lungs into smaller pieces, followed by transferring the latter to 15 mL conical tubes containing 5 mL of Hank's Balanced Salt Solution (HBSS, calcium- and magnesium-free) supplemented with 50 μL of 5 mg/mL Liberase (Roche) and 12.5 μL of 10 mg/mL DNase I. Lung tissues were incubated at 37°C in a shaking incubator for 45 min for enzymatic digestion (76). Single-cell suspensions were prepared for the spleen, DLN and lungs by mechanically dissociating tissues using the plunger of a 1 mL syringe, followed by filtration through a sterile 70 μm cell strainer. The resulting cell suspensions were centrifuged at 300 × g for 5 minutes at 4°C, and erythrocytes were lysed using 500 μL of RBC lysis buffer (BioLegend) for 1 min at room temperature. The lysis reaction was quenched with 10 mL of ice-cold R2 medium, followed by an additional centrifugation step. Cell pellets were resuspended in 1 mL of ice-cold RPMI-1640 supplemented with 10% FBS and 1% penicillin-streptomycin (R10). Cell viability was assessed using acridine orange (AO) and propidium iodide (PI) staining, and the proportion of live and dead cells was quantified using an automated cell counter (Nexcelom Bioscience). The final cell concentration was adjusted to 10⁷ cells/mL in R10 medium (77).

T cell ELISpot assay.

To evaluate antigen-specific T cell responses, an IFN-γ enzyme-linked immunospot (ELISpot) assay was conducted using the ELISpot Flex: Hamster IFN-γ kit (MABTECH), following the manufacturer’s instructions to enhance assay sensitivity and reproducibility. Polyvinylidene difluoride (PVDF)-lined 96-well microplates (Millipore) were pre-activated by incubation with 35% ethanol for 1 min to ensure uniform membrane activation, followed by extensive rinsing with PBS to eliminate residual ethanol. Wells were coated overnight at 4°C with a sterile PBS-diluted IFN-γ capture antibody (15 μg/mL). The following day, unbound antibodies were removed by five successive washes with PBS. To minimize background staining and non-specific binding, plates were blocked with complete R10 medium for at least one hour at 37°C prior to the addition of cells. Freshly isolated cells (500,000 cells per well) were harvested from spleen, DLNs, or lungs and resuspended in R10 medium. Cells were stimulated for 18–24 h under the following conditions: (a) peptide mega pools spanning the SARS-CoV-2 S protein, including the S1 subunit (residues 1–668) and the S2 subunit (residues 659–1273) at a final concentration of 1 μg/mL (BEI Resources), (b) 25 smaller mesopools of overlapping peptides covering the full-length spike protein, each containing 10–11 peptides of 15-mers at a final concentration of 1 μg/mL per pool, or (c) individual peptides (8–10-mers or 15-mers) at a final concentration of 10 μg/mL. Positive control wells were stimulated with phorbol myristate acetate (PMA, 0.5 μg/mL) and ionomycin (1 μg/mL), and negative control wells received 1% dimethyl sulfoxide (DMSO) alone. Following antigen stimulation, the plates were washed 5 times with PBS to eliminate the cells and any residual cytokines. Plates were then incubated at room temperature for 2 to 3 h with a biotinylated anti-IFN-γ detection antibody (1 μg/mL in PBS-0.5% FBS). After five additional washes with PBS, plates were further incubated for 1 h with streptavidin-conjugated alkaline phosphatase (ALP) (1:1000 dilution in PBS-0.5% FBS). To prevent interference from phosphate residues in the final substrate reaction, a final wash was performed using Tris buffer (0.05 M) before spot development. Colorimetric detection was initiated by adding BCIP/NBT substrate, and spot formation was visually monitored under controlled lighting conditions. The reaction was terminated by extensive washing with distilled water, followed by air-drying of the plates in the dark to prevent oxidative artifacts that could impact signal clarity. ELISpot plates were analyzed using an automated BioSpot reader (Cellular Technology Limited). Background responses in unstimulated wells were consistently low, averaging fewer than 20 spots per million cells. To ensure accurate quantification, background values were subtracted from the final SFC readouts for each experimental condition. Data was expressed as the number of spot-forming cells per million live cells (SFC/10⁶ live cells). The MHC II restriction of the S_631–645 peptide was assessed using blocking antibodies against MHC class II alleles I-E^k (clone 14–4-4S) and I-A (clone Y3JP). A 10-fold serial concentration of P13.7 peptide was added to splenocytes of a naïve hamster for 1 h at 37°C before the cells were washed and treated with 50 μg/mL of mAb 14–4-4S, Y3JP, L243, or left untreated for 1 h at 37°C. Next the cells were washed and added to splenocytes or cells isolated from the lungs of ChAd-SARS-CoV-2-S immunized hamsters at a 1:1 ratio and the number of S_631–645 specific cells were quantified by IFN-γ ELISpot. As negative and positive controls, we added S_631–645 peptide to splenocytes of a naive or immunized hamster, respectively.

CD4⁺ and CD8⁺ T cell depletion.

To deplete CD4⁺ or CD8⁺ T cells ex-vivo, single-cell suspensions were prepared from spleens, LNs, or lungs. Cells were subjected to magnetic bead-based negative selection to remove either CD4⁺ or CD8⁺ T cell populations while preserving the remaining immune cell repertoire. For CD4⁺ T cell depletion, Dynabeads^™ Biotin Binder Kit (Invitrogen) was used. Beads were first washed twice with 2% FBS in PBS (P2 buffer). The washed beads were then incubated with biotinylated anti-CD4 monoclonal antibody (clone GK1.5, BioLegend) at a final concentration of 10 μg/mL for 45 min at room temperature with continuous gentle mixing to facilitate antibody binding. For CD8⁺ T cell depletion, MagnaBind^™ Goat Anti-Mouse Magnetic Beads (Invitrogen) were coated with anti-CD8β monoclonal antibody (clone 341, Invitrogen) using the same protocol as CD4⁺ T cell depletion. Beads were incubated with anti-CD8β at a final concentration of 10 μg/mL for 45 min at room temperature with shaking every 10 min. Following antibody conjugation, beads were washed five times with P2 buffer to remove excess unbound antibodies, using a magnetic separation stand to facilitate washes. One million freshly isolated splenocytes, DLN, or lung cells were then incubated with the pre-coated beads for 60 min on ice with intermittent gentle shaking to maximize cell-bead interactions and ensure efficient depletion. After incubation, a magnetic separation stand was used to remove the CD4⁺ or CD8⁺ T cells bound to the beads. The remaining T cell-depleted fractions were collected, resuspended in complete culture medium (RPMI 1640, 10% FBS, 1%PS), and processed for downstream analyses. The efficiency of depletion was validated by flow cytometric analysis using fluorochrome-conjugated anti-CD4 and anti-CD8 antibodies.

Flow cytometry.

Staining was conducted on one million freshly isolated cells from the spleen and DLNs. Cells were incubated on ice for 30 min with fluorophore-conjugated monoclonal antibodies specific to T and B cell markers, including CD4-PE (GK1.5, 1:100, BioLegend), CD8b-BB700 (341, 1:100, BD Biosciences), and B220-PE/Cyanine7 (RA3–6B2, 1:100, BioLegend). To exclude dead cells, samples were stained with Zombie Aqua viability dye (1:200, BioLegend). Following staining, cells were washed with cold P2 buffer, fixed with 2% paraformaldehyde, and resuspended in P2 buffer before flow cytometric acquisition. Data acquisition was performed using a Cytek Aurora spectral flow cytometer with SpectroFlo v2.2 software. Compensation controls were included for each fluorophore, and fluorescence-minus-one (FMO) controls were used to establish accurate gating thresholds. Flow cytometry data were analyzed using FlowJo v10 (BD Biosciences). Gating strategies were designed to first exclude doublets and dead cells, followed by selection of live, singlet T cells while excluding B220⁺ B cells. The final analysis focused on quantifying CD4⁺ and CD8⁺ T cell populations in both anti-CD4 and anti-CD8 depleted and non-depleted samples to assess residual T cell populations.

Flow cytometric analysis of MHC Class I and Class II molecule expression.

To identify suitable antibodies for MHC analysis, we tested in-house purified anti-MHC antibodies (derived from hybridomas obtained from the ATCC and purified as described elsewhere (78)) (Table S3) against Syrian hamster-derived splenocytes and a SV40-immortalized hamster B cell line (1109) (79). These antibodies had been previously validated for MHC class I and II peptide motif studies and binding assays in other species, including humans, mice, and macaques. Staining was conducted on one million cells from freshly thawed splenocytes or the 1109 immortalized cell line on ice for 30 minutes. After incubation, cells were centrifuged at 1,400 rpm, washed, and incubated on ice for an additional 30 min with a goat anti-mouse IgG FITC-conjugated secondary antibody and the Live/Dead Viability dye eFluor506 (Invitrogen) to exclude dead cells. Each sample was left unstained or incubated with an isotype control or with the secondary antibody alone as controls for each mAb tested. Positive controls, representing cell lines expressing alleles that correspond to the known specificity of each mAb, were included in each assay. Samples were immediately acquired on a ZE5 cell analyzer (Bio-Rad Laboratories, Hercules, CA) and data were analyzed with FlowJo software v10.7.2 (BD Biosciences). The relative expression of each antibody was assessed by calculating the ratio of mean fluorescence intensity (MFI) between the test antibody and its corresponding isotype control.

Preparation of peptide pools and individual peptides for SARS-CoV-2 T cell epitope identification.

15-mer peptides overlapping by 10 amino acids were synthesized to span the entire SARS-CoV-2 S proteome, providing a comprehensive and unbiased approach independent of predictive algorithms. Peptides (15-mers and 8–10-mers) were synthesized using solid-phase peptide synthesis (SPPS) with Fmoc chemistry on Wang or Rink amide resins, depending on the desired C-terminal functionality. Coupling was carried out for 45 to 60 min at room temperature using HBTU/Oxyma Pure or HATU/DIEA, followed by Fmoc deprotection with 20% piperidine in DMF and thorough washing. After synthesis, peptides were cleaved from the resin using TFA/TIS/water (95:2.5:2.5, v/v/v) for 2–3 h, precipitated in cold diethyl ether, centrifuged, and purified by reverse-phase HPLC (C18 column). Peptide purity (>70%) was confirmed by ESI-MS or MALDI-TOF, then lyophilized and stored at −80°C. Crude peptides (A&A, San Diego, CA) were individually resuspended in DMSO (10 mg/mL) and pooled into antigen-specific mega pools, which were sequentially lyophilized and stored in single-use aliquots at −80°C. A subset of 15-mers was also organized into mesopools of ten peptides each, resuspended in DMSO at 1 mg/mL. CD8⁺ (8–10-mers) and CD4⁺ (13–15-mers) peptides were designed to optimize MHC class I and II binding, respectively, and used individually at a final concentration of 10 μg/mL. For T cell stimulation assays, megapools were reconstituted in PBS + 5% DMSO, and mesopools in DMSO, and used at a final peptide concentration of 1 μg/mL.

Sequence retrieval.

Full-length amino acid sequences of the SARS-CoV-2 S protein from representative variants including Wuhan-Hu-1,WA/2020, Alpha (B.1.1.7), Gamma (P.1), and Omicron sublineages (BA.1, BA.2.12.1, BA.5, XBB.1.5, EG.5.1, JN.1, and NB.1.8.1) were retrieved from the National Center for Biotechnology Information (NCBI) Virus database and the Global Initiative on Sharing Avian Influenza Data (GISAID). Sequences were downloaded in FASTA format and verified for completeness and annotation accuracy.

Multiple sequence alignment.

This was performed using Clustal Omega (v1.2.4) accessed through the EMBL-EBI web server (https://www.ebi.ac.uk/Tools/msa/clustalo/). Default parameters were employed, and sequences were aligned based on full-length S proteins to assess conservation and variability among variants. Alignments were manually inspected and adjusted as necessary to ensure alignment accuracy, particularly within functional domains.

Domain annotation and motif mapping.

Domain boundaries of the S protein, including the N-terminal domain (NTD), RBD, and C-terminal domains 1 and 2 (CTD1 and CTD2 were defined according to previously published structural structures (80–83). Sequence motifs of interest, such as WRVYSTGSNVFQ, were mapped to their corresponding domains using the Wuhan-Hu-1 numbering scheme.

Statistical analysis.

Statistical analyses were performed using GraphPad Prism 10.1 software. All experiments were performed with multiple different hamsters, and most experiments were repeated once or twice. The number of animals used and the number of experiments performed are indicated in the figure legends. One-way ANOVA, followed by post hoc comparisons (Tukey’s or Dunnett’s), and paired t-tests were used to compare groups. A p-value of < 0.05 was considered significant.

Data availability.

All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials or available online (https://doi.org/10.17632/6vbrr8ctkg).