Oral yeast-displayed SARS-CoV-2 spike protein vaccine enhanced by trimerization and flagellin peptide Authors: Tahereh Saveii, Sareh Arjmand, Ismaeil Haririan, Navid Pourzardosht, Reza H Sajedi Published: 2026 ABSTRACT AbstractThe global COVID-19 pandemic has underscored the need for innovative vaccine strategies, including oral vaccines, to improve accessibility and enhance immunity. In this study, we explored a novel oral vaccine approach by using yeast surface display to express the receptor-binding domain (RBD) of the SARS-CoV-2 spike protein on Pichia pastoris using SED1 as a carrier protein. The constructs included a fibritin foldon domain from bacteriophage T4 for trimerization and truncated flagellin from Salmonella typhimurium (FliC) as an adjuvant to boost immunogenicity. Three recombinant yeast constructs were designed and evaluated: construct I (RBD-SED1), construct II (RBD-foldon-SED1), and construct III (FliC-RBD-foldon-SED1). Structural integrity, proper folding, and trimerization (in construct III) were modeled and confirmed using ColabFold. These constructs were successfully expressed on the yeast surface and confirmed through fluorescence microscopy, flow cytometry, and ELISA. Oral immunization of BALB/c mice with the recombinant yeast significantly induced spike protein-specific IgG antibodies in serum and IgA in fecal samples, indicating strong humoral and mucosal immune responses. Notably, construct III induced the highest immune response. These findings suggest that yeast-based oral vaccines expressing trimerized SARS-CoV-2 spike protein RBD, combined with a peptide fragment from flagellin, as an immunomodulatory adjuvant, can elicit robust immunity, offering a promising alternative for future COVID-19 vaccination strategies.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-50942-z.Keywords: SARS-CoV-2, Oral vaccine, Yeast surface display, Pichia pastoris, Spike protein, Trimerization, FlagellinSubject terms: Biotechnology, Immunology, Microbiology Introduction The COVID-19 pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has profoundly impacted global health, economies, and daily life since its emergence in late 20191. In response, unprecedented efforts have led to the rapid development of various vaccines, including mRNA, viral vector, protein subunit, and inactivated virus vaccines2–4. These vaccines helped to reduce disease severity and mortality, yet challenges such as vaccine distribution, storage requirements, and the emergence of new variants underline the need for innovative vaccine platforms5,6. Exploration of alternative vaccine strategies remains critical to address gaps in accessibility, cost, and immune protection durability7,8. Oral vaccines represent a promising avenue, offering key advantages over traditional injection-based methods9. They eliminate the need for trained healthcare professionals, improve patient compliance, and simplify large-scale immunization campaigns, particularly in resource-limited settings. Furthermore, oral vaccines can stimulate both systemic and mucosal immunity, providing a frontline defense against pathogens entering through mucosal surfaces10. One of the primary challenges associated with oral vaccines is their limited capacity to elicit robust T-cell responses11. This limitation arises mainly from the adverse conditions present within the gastrointestinal (GI) tract, which can significantly hinder antigen bioavailability and may even promote immune tolerance instead of activation. The GI environment is characterized by several harsh factors, including low pH levels, bile salts, and a variety of digestive enzymes, all of which can degrade many antigens before they reach their target immune cells. Consequently, this degradation results in a diminished immune response, as the antigen-presenting cells (APCs) located in the gut often remain in a tolerogenic state, further impairing T-cell activation and leading to inadequate stimulation of the immune system12. Yeast cells have emerged as a promising candidate for oral vaccine development. Studies have shown that oral administration of yeast displaying antigens can elicit robust IgG and secretory IgA responses, indicating effective mucosal immunity activation. Their inherent ability to withstand the hostile conditions of the GI tract is attributed to their robust cell wall structure13–15. The cell wall not only provides physical protection but also possesses intrinsic adjuvant properties. Key components such as β-1,3-glucans and mannoproteins play crucial roles in enhancing immune activation16–19. These components facilitate the recognition and uptake of antigens by immune cells, thereby stimulating dendritic cells (DCs) to mature and produce cytokines like interleukin-12 (IL-12), which is essential for effective T-cell responses17,20. One innovative approach to developing a yeast-based oral vaccine is yeast surface display technology. This platform involves genetically engineering yeast cells to express specific antigens on their surfaces, effectively mimicking the structure of pathogens and facilitating direct interaction with the immune system21. One significant advantage of yeast surface display is its scalability and cost-effectiveness, making it suitable for large-scale vaccine production. This technology has been successfully applied in various vaccine candidates targeting a wide range of pathogens22,23. In the present study, we employed yeast surface display to express the receptor-binding domain (RBD) of the SARS-CoV-2 spike protein on the surface of the yeast species Pichia pastoris, utilizing the SED1 protein as an anchor for this display24,25. The spike protein is a transmembrane glycoprotein that protrudes from the viral envelope and is critical for viral entry into host cells, as it mediates binding to the human angiotensin-converting enzyme 2 (ACE2). This interaction is essential for viral infection, making the spike protein RBD a primary target for neutralizing antibodies that can block SARS-CoV-2 from attaching to ACE2 and entering cells26–28. Consequently, targeting the RBD in vaccine formulations could enhance their efficacy by promoting robust immune responses that specifically inhibit viral entry29. Transmembrane glycoproteins, particularly those located in the viral envelope, predominantly exist as trimeric structures. This trimerization is crucial for their functionality, as these proteins serve as the primary targets for neutralizing antibodies during viral infections30,31. Trimeric envelope glycoproteins are recognized as superior immunogens, effectively eliciting robust antibody responses that are capable of neutralizing various viruses. The structural integrity and proper assembly of these glycoproteins are essential for their immunogenic properties32. Recent studies have highlighted the use of the Fibritin foldon domain as an artificial trimerization domain in several viral vaccines, including those targeting influenza33, the Middle East Respiratory Syndrome Coronavirus (MERS-CoV)34, and producing and enhancing of SARS-CoV-2 vaccines35,36. The Fibritin foldon domain is a 29-amino acid sequence derived from the C-terminal region of the fibritin protein found in the T4 bacteriophage. This sequence forms a stable β-propeller structure that facilitates the trimerization of various proteins, thereby enhancing their stability and immunogenicity33,37. For stable and high-density attachment of the recombinant protein to the yeast cell wall, the SED1 GPI-anchored domain was selected as the surface display anchor. As a major structural protein in the S. cerevisiae cell wall, SED1 covalently links to β-1,6-glucan via its C-terminal GPI signal peptide24,38, ensuring robust cell wall integration. Comparative studies indicate that SED1, among other GPI anchors, can achieve surface protein densities 6–8 times higher than commonly used anchors like α-agglutinin39. Its prior successful application in P. pastoris for displaying Fab fragment and lipase, while maintaining accessibility at the outermost cell wall layer40,41, validates its suitability for maximizing antigen exposure in yeast-based vaccine platforms, especially for oral vaccination strategies requiring effective interaction with mucosal antigen-presenting cells. To enhance our construct, we incorporated a truncated form of flagellin, specifically including residues (85–111) of the mature flagellin FliC42,43. Flagellin is a structural protein that forms the filament of bacterial flagella, which are whip-like appendages responsible for bacterial motility. Beyond its mechanical function, flagellin acts as a potent immunomodulatory agent by stimulating the innate immune system through its interaction with Toll-like receptor 5 (TLR5). This interaction triggers a cascade of immune responses, including the activation of pro-inflammatory cytokines and the recruitment of immune cells, making flagellin an attractive candidate for vaccine constructs44–48. The inclusion of this truncated flagellin in our vaccine construct may enhance its immunogenicity by promoting better recognition by the immune system and facilitating a stronger antibody response against the RBD of the spike protein42,43,49. Materials and methods Plasmids, strains, cell lines and media The shuttle plasmid pPINKα-HC (Invitrogen, USA) was utilized for cloning and expression, with Escherichia coli DH5α serving as the host strain for recombinant plasmid propagation. Bacteria were cultured in Luria-Bertani (LB) medium (1% (w/v) peptone (Merck, Germany), 0.5% (w/v) yeast extract (Merck, Germany), and 1% (w/v) sodium chloride (Merck, Germany)) supplemented with 100 µg/mL ampicillin. For surface display expression, the PichiaPinkTM expression system strain 4 (ade2, pep2, pep4) (Invitrogen, USA) was employed as the host. Yeast cultures were grown in Buffered-Glycerol-Complex Medium (BMGY) including 1% (w/v) yeast extract, 2% (w/v) peptone, 100 mM potassium phosphate (Merck, Germany), 1.34% (w/v) yeast nitrogen base (YNB) (Bio Basic Inc, Canada), 4 × 10⁻5% (w/v) biotin (Merck, Germany), 2% (v/v) glycerol (Dr. Mojallali Co., Iran), pH 6.0 and induced in Buffered Methanol-Complex Medium (BMMY), which is similar to BMGY but lacks glycerol and contains 0.5% (v/v) methanol (Merck, Germany) on the first day and 1% (v/v) methanol for subsequent cultures. The HEK293T-hACE2 (kind gift from Renap Therapeutics Co., Tehran, Iran) and MT2 cell lines used in this study were obtained from the Cell Culture Facility of the Faculty of Biological Sciences, Tarbiat Modares University (Tehran, Iran). These cell lines are well-established laboratory cell lines and were used in accordance with institutional biosafety and laboratory regulations. Design and structural modeling of SARS-CoV-2 spike-based antigenic constructs A protein fragment corresponding to the RBD of the SARS-CoV-2 spike protein (residues 332–541), including the C538A mutation50, was derived based on the sequence obtained from the National Center for Biotechnology Information (NCBI) (GenBank: QHD43416.1) and used for construct design. The SED1 anchor (residues 19–338; GenBank: CAA47314.1) was used as a carrier protein for surface display on P. pastoris. To facilitate trimerization and mimic the native-like structure of the SARS-CoV-2 spike protein, a peptide fragment of the fibritin foldon domain derived from bacteriophage T4 (residues 457–483; GYIPEAPRDGQAYVRKDGEWVLLSTFL) was incorporated, and a peptide fragment of flagellin (FliC) from Salmonella typhimurium (residues 85–118; INNNLQRVRELAVQSANSTNSQSDLDSIQAEITQ) was included as an adjuvant to boost immunogenicity. To evaluate the structural integrity, folding and spatial feasibility of these fusion constructs, 3D structural predictions were performed using ColabFold, a user-friendly implementation of AlphaFold2 optimized for rapid and accurate protein structure prediction51. Each fusion protein sequence was input into ColabFold, with critical parameters to model potential multimeric interactions, especially for foldon-mediated trimerization in constructs II and III. The top-ranked predicted models, based on confidence metrics such as predicted Local Distance Difference Test (pLDDT) scores, were downloaded for further analysis. The predicted PDB structures were visualized and analyzed using PyMOL to assess folding integrity, domain orientation, and spatial feasibility of the fusion protein. Based on the structural evaluation, gene constructs were finalized for expression in P. pastoris. Three constructs were designed as follows: (I) RBD-SED1, in which SED1 was fused to the C-terminal of RBD via a flexible (GS)3G linker; (II) RBD-foldon-SED1, where the foldon domain was linked to the C- and N-terminals of RBD and SED1, respectively, using (GS)3 and (GS)3G linkers; and III) FliC-RBD-foldon-SED1, which included the truncated flagellin (FliC) at the N-terminal of construct II, connected by a (GS)3 linker. The constructs were designed to be cloned into the pPINKα-HC plasmid, under the control of the AOX1 promoter and the α-MF signal peptide, which facilitate the expression and secretion of the recombinant protein, respectively, as depicted in Fig. 1. Schematic representation of three designed constructs for the display of spike protein RBD using SED1 on the surface of P. pastoris. AOX1 promoter: Alcohol oxidase 1 promoter; α-MF signal: α-mating factor secretion signal; Flagellin fragment: flagellin-derived peptide from Salmonella typhimurium (FliC); RBD: receptor-binding domain of the SARS-CoV-2 spike protein; Foldon domain: Trimerization foldon domain of T4 phage fibritin; SED1: SED1 anchor protein. Cloning and surface expression of recombinant constructs in P. pastoris The constructs were ordered for synthesis (Shanghai Generay Biotech Co.) based on the codon usage preference of P. pastoris, incorporating XhoI and KpnI restriction sites for sequential cloning into the pPINKα-HC plasmid. The nucleotide sequence data were submitted to the Genbank under the following accession numbers: construct I (PV657109), construct II (PV693362), and construct III (PV750214). The recombinant plasmids were introduced into competent Escherichia coli cells via thermal shock (90 s at 42 °C), and positive colonies were selected based on resistance to ampicillin (100 µg/mL). The plasmids were extracted from positive clones, and the cloning accuracy was verified through PCR using primers specific for the recombinant gene and the plasmid backbone (Supplementary 1), followed by sequencing of each construct. Each recombinant vector was transformed into P. pastoris. To achieve this, the recombinant plasmids were linearized using the SpeI restriction enzyme and then introduced into competent P. pastoris via electroporation (Gene Pulser, Bio-Rad, USA), as previously described52. Positive clones were selected on PAD medium. Integration of each construct (I, II, and III) into the P. pastoris genome was confirmed by PCR on the extracted yeast genomic DNA. P. pastoris cells transformed with the empty vector, lacking any foreign gene insert, served as the negative control. The P. pastoris cells were cultured in 50 mL of BMGY medium for 2 days (OD600= 10). The culture was centrifuged at 1500 × g for 5 min, and the cell pellet was resuspended in 50 mL of BMMY medium containing 0.5% methanol to induce expression of the constructs under the control of the AOX1 promoter. The culture continued for an additional three days, with methanol induction maintained at 1% as the carbon source. SDS-PAGE analysis of the surface-expressed SARS-CoV-2 spike protein RBD Following induction, one mL of yeast culture was harvested by centrifugation at 4,000 × g for 5 min, and the supernatant was discarded. The pellet was resuspended in 200 µL of lysis buffer containing 2% Triton X-100, 1% SDS, 100 mM NaCl, 1 mM EDTA, and 100 mM Tris-HCl (pH 8.0), along with glass beads. The suspension was vortexed vigorously for eight cycles to lyse the cells. After centrifugation at 12,000 × g for 15 min, the pellet was treated with mannanase (Bonda Faravar, Iran) and exo-1,3- β-glucanase from the yeast β-glucan assay kit (Megazyme, Ireland) to release surface-displayed proteins. The mixture was incubated at 37 °C for 2 hours. Both the supernatant and pellet fractions were analyzed by 12.5% SDS-PAGE to evaluate surface protein expression. Assessment of spike protein RBD surface expression using immunofluorescence and flow cytometry For immunofluorescence assay and flow cytometry analyses, one yeast unit (YU) of P. pastoris cells—defined as approximately 10^7 cells (equivalent to OD600 = 1 and about 0.5 mg)53—was collected by centrifugation from each experimental group. The resulting cell pellets were washed three times with PBS. The cell pellets were then incubated with 200 µL mouse anti-RBD antibody, diluted 1:200, sourced from the PastoCovac vaccine (Pasteur Institute, Iran)54. This incubation occurred at 4 °C for 1 hour to facilitate specific binding. Cell pellets were washed three times with PBST. Subsequently, 200 µL fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG secondary antibody (Razi Biotech, Iran), diluted 1:50, was added and incubated at 4 °C for an additional hour. After this incubation, the cells were centrifuged and re-suspended in 500 µL of sterile PBS. A volume of 5 µL of the P. pastoris pellet was utilized for immunofluorescence assay using an Olympus IX70 microscope (Japan). For quantitative assessment, 300 µL of the resuspended P. pastoris pellet was analyzed using a BD FacsCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). Yeast-based immunoassay for SARS-CoV-2 spike antigen RBD surface display P. pastoris cells (expressing constructs I, II, III, and negative control) treated with anti-RBD antibody were subjected to centrifugation and subsequently washed with PBST to remove unbound antibodies. The pellets were then incubated with a horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibody, diluted 1:1000 (Abcam, USA), at 37 °C for an additional hour. After the next washing, the cell pellets were treated with 100 µL of 3,3′,5,5′-tetramethylbenzidine (TMB) substrate (Sigma, St. Louis, MO, USA) for 15 min to allow for color development. The enzymatic reaction was then halted by the addition of 100 µL of H₂SO₄ (1 M). Absorbance was measured at 450 nm using a microplate reader (μQuant, BioTek, Winooski, VT, USA). Microscopic analysis of RBD-expressing yeast binding to ACE2-positive cells Human embryonic kidney 293 T cells expressing human ACE2 (HEK293T-hACE2) and MT2 T-lymphocyte cells (as a negative control) were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 100 µg/mL streptomycin, and 100 IU/mL penicillin. Cells (5,000 per well) were seeded in 96-well flat-bottom plates and incubated for 12 hours at 37 °C with 5% CO₂ to allow adherence. The cells were washed three times with PBS and fixed with 10% formalin at 4 °C for one hour. After the next wash, wells were blocked with 3% BSA in PBS for 30 min at room temperature. Following blocking, 100 µL of yeast suspension (OD₆₀₀ = 2), displaying spike protein RBD on the surface, was added and incubated overnight at room temperature. After a final PBS wash, yeast-cell interactions were examined under an inverted light microscope (Nikon Eclipse TE2000, Japan) to assess binding specificity. Preparation of heat-inactivated yeast cell After confirming the presence of recombinant constructs on the surface of P. pastoris cells, the cells were harvested following a 72-hour methanol induction. To ensure the safety and efficacy of these cells as an oral vaccine, they underwent heat inactivation by heating at 60 °C for one hour. Following heat treatment, the cells were centrifuged and resuspended in sterilized PBS to achieve two final cell densities (OD600 = 10 and OD600 = 150). The prepared cells were then stored at 4 °C until use. To confirm the effectiveness of heat inactivation, heat-inactivated P. pastoris cells were cultured on BMGY agar plates and incubated at 30 °C for 4 days. Oral administration and sample collection in mice Female BALB/c mice, aged seven weeks, were procured from the Tehran University of Medical Sciences (Tehran, Iran). The mice were divided into nine groups (n = 5 per group) and orally administered either recombinant yeast suspension at OD600 = 10 (equivalent to approximately 10 YU) or OD600 = 150 (equivalent to approximately 150 YU). These recombinant yeast groups contained constructs I, II, and III, while a non-recombinant yeast group served as the control. The immunization schedule consisted of a prime vaccination on days 1 and 2, followed by booster immunizations on days 14 and 15. Blood and fecal samples were collected from the mice on days 1, 12, and 26 post-initial immunizations. Blood samples were processed by centrifugation at 2000 rpm for 10 min at room temperature to separate serum, which was subsequently stored at -20 °C until analysis. For fecal samples, a suspension was prepared using 30 mg of feces in 300 µL of sterile PBS. After centrifugation at 15,000 g for 8 min, the supernatants were collected and stored at -20 °C until analysis. All procedures were conducted in compliance with the ethical guidelines approved by the institutional Ethics Committee of Tehran University of Medical Science (approval code: IR.TUMS.TIPS.REC.1401.009) and are reported following the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines 2.0. ELISA for detection of anti-RBD antibodies in serum and feces IgG titers in serum and IgA titers in fecal samples were quantified on day 26 after oral administration using an indirect ELISA to detect RBD-specific antibodies. Flat-bottom 96-well plates were coated with recombinant antigen (final concentration of 2 µg/mL) in 0.05 M carbonate coating buffer (pH 9.6), and incubated overnight at 4 °C. To assess IgG levels in serum, serial dilutions of the sera were added to the coated plates and incubated for two hours at 37 °C. Following three washes with PBS with Tween (PBST), 100 µL of HRP-conjugated goat anti-mouse IgG (1:1000 dilution) was introduced into each well, followed by another hour incubation at 37 °C. The plates were then washed six times with PBST and developed using 100 µL of TMB substrate in the dark for 15 min. The enzymatic reaction was terminated by adding 100 µL of 1 M H₂SO₄, and absorbance was measured at 450 nm using an ELISA microplate reader. For the determination of IgA, the fecal supernatants were treated similarly to the serum samples. After washing, 100 µL of HRP-conjugated goat anti-mouse IgA (1:10,000 dilution; Sigma-Aldrich, USA) was added to the wells, followed by one hour incubation at 37 °C. The subsequent steps were performed as mentioned above. Statistical analyses All ELISA assays for both IgG and IgA were performed in triplicate, and absorbance was measured at 450 nm. For each experimental group, serum samples (for IgG) and pooled fecal extracts (for IgA) were collected from five mice and combined prior to analysis to ensure consistency. In addition, cell-based immunoassays evaluating surface expression were also performed in triplicate. Statistical analyses were conducted using GraphPad Prism version 10 (GraphPad Software, USA). Differences among groups—including PBS, negative control, and recombinant P. pastoris strains expressing Constructs I, II, and III at two cell densities were evaluated by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test for multiple comparisons. Data are presented as mean ± standard deviation (SD). Statistical significance was defined as p < 0.05. Significance levels are as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns = not significant. Results Molecular modeling of RBD-based antigenic constructs The predicted three-dimensional structures of the designed antigens (Fig. 1) were obtained using ColabFold and visualized in PyMOL. All three constructs displayed well-folded domains with spatially accessible spike protein RBD regions (Fig. 2). Notably, the foldon domain in constructs II and III formed a trimeric coiled-coil structure, suggesting successful trimerization, which may enhance immunogenic mimicry of the native spike protein. The structural models demonstrate proper domain organization and trimerization in constructs II and III, supporting their potential application in yeast surface display systems. Analysis of the pLDDT profiles showed that both the spike protein RBD and the foldon trimerization domain were predicted with consistently high confidence, whereas lower scores were confined to the linker and selected SED1 regions, an expected pattern that reflects the known AlphaFold/ColabFold distinction between well-structured domains and flexible or disordered segments55. Predicted three-dimensional structures of the designed antigenic constructs, modeled using ColabFold and visualized in PyMOL. For each construct, both ribbon and surface representations are presented in monomeric (A, B, C and F) and trimeric (D, E, G and H) forms to illustrate the domain organization and spatial arrangement. The domains are color-coded as follows: truncated flagellin (FliC) (green), linkers (yellow), RBD of SARS-CoV-2 spike protein (red), foldon trimerization domain (purple), and SED1 anchor (cyan). Construction and expression of yeast cell surface-expressing SARS-CoV-2 RBD-based antigens The three constructs (I, II, and III) described in the methodology were successfully cloned into the pPINKα-HC vector and transformed into PichiaPink™ expression system. After induction and expression in P. pastoris, yeast cells were enzymatically treated with mannanase and β-glucanase to facilitate the release of surface-displayed antigens. SDS-PAGE analysis confirmed the expression of recombinant proteins. As shown in Fig. 3, distinct bands corresponding to the expected molecular weights of the three designed constructs were detected in the pellet fraction, migrating below the 75 kDa marker. These bands were predominantly observed in the pellet fractions, indicating that the majority of the recombinant proteins remained associated with the yeast cell surface form. This suggests successful expression of the target proteins, although their presence in the pellet may reflect strong surface attachment. No corresponding bands were observed in the supernatant fractions, indicating that the majority of expressed proteins remained associated with the yeast cell wall, consistent with successful surface localization. SDS-PAGE analysis of yeast-expressed recombinant antigens. Lane 1, Negative control (pellet); Lane 2, Negative control (supernatant); lane 3, Construct III (pellet); lane 4, Construct III (supernatant); lane 5, Protein size marker (Bio-Rad, USA); lane 6, Construct II (supernatant); lane 7, Construct II (pellet); lane 8, Construct I (supernatant) and lane 9, Construct I (pellet). Protein bands are indicated by arrows. Surface-displayed RBD detection using immunofluorescence and flow cytometry Figure 4 presents images of yeast cells under bright-field (4A and 4C) and fluorescence (4B and 4D) illumination. In panels 4A and 4B, which represent the non-recombinant construct serving as the control group, no detectable fluorescence was observed under both bright field and fluorescent illumination. However, the recombinant yeast cells demonstrated fluorescence emission when excited at the appropriate wavelength (panel 4D), indicating successful surface display of the RBD of the SARS-CoV-2 spike protein. To avoid misinterpretation of the fluorescence pattern, it should be noted that the whole-cell signal observed in panel 4D results from the optical projection of the curved yeast cell wall during surface labeling, rather than intracellular staining. Immunofluorescence microscopy of P. pastoris cells displaying the RBD of the SARS-CoV-2 spike protein. Bright-field (A and C) and fluorescence (B and D) images of P. pastoris cells are shown. Panels A and B represent the non-recombinant yeast cells as control, and Panels C and D show the recombinant yeast cells (Magnification: × 400). The flow cytometry and cell-based immunoassay results shown in Fig. 5 confirm these findings by quantitatively confirming surface expression of the recombinant construct. As shown in Fig. 5B, compared to non-recombinant P. pastoris cells, recombinant yeast expressing spike protein RBD exhibited a strong fluorescence signal upon staining with anti-RBD antibody, with approximately 90% of the cells testing positive. In addition, cell-based immunoassay revealed a markedly higher detection signal for the surface-displayed RBD of the SARS-CoV-2 spike protein in P. pastoris cells expressing constructs I, II, and III, compared to non-recombinant yeast cells, which served as the negative control (Fig. 5C). These results confirm the successful surface expression of the SARS-CoV-2 spike protein RBD on P. pastoris. Flow cytometry analysis and cell-based immunoassay of recombinant P. pastoris cells displaying the spike protein RBD. (A) Flow cytometry profile of non-recombinant yeast cells as control, (B) Flow cytometry profile of recombinant construct, and (C) cell-based immunoassay for P. pastoris cells expressing different recombinant constructs. Results represent the mean OD₄₅₀ values from three independent measurements performed on yeast cell suspensions. PBS was used as the assay background control. Binding of recombinant yeast displaying spike protein RBD to ACE2-expressing mammalian cells To assess the binding specificity of surface-displayed spike protein RBD on P. pastoris, yeast cells were incubated with ACE2-expressing HEK293T cells and MT2 as a non-expressing cell line. Light microscopy revealed clear adhesion of yeast cells surrounding and attached to the ACE2-expressing cells. In contrast, when spike protein RBD-expressing yeast cells were incubated with MT2 cells (as a non-expressing cell line), no attachment was detected (Fig. 6). These findings confirm the specific interaction between the surface-expressed spike protein RBD and the ACE2 receptor and demonstrate the successful functional presentation of RBD on the yeast cell surface. Light microscopy images for evaluating the binding of P. pastoris cells displaying spike protein RBD to fixed mammalian cells in a 96-well plate. (A) ACE2-expressing HEK293T cells and (B) MT2 cells non-expressing cell line as a control (Magnification: × 400). P. pastoris cells were identified in co‑culture based on their characteristic size (3–5 µm) and round/oval morphology, which clearly distinguish them from ACE2-expressing HEK293T cells and MT2 cells. Detecting humoral and mucosal responses Humoral and mucosal immune responses were evaluated by measuring IgG in serum and secretory IgA in feces from orally vaccinated mice on day 26 post-immunization with recombinant yeast constructs at two cell densities. Serum samples from each group were pooled and subjected to ELISA using a serial dilution range (1:50 to 1:2000). ELISA results demonstrated a significant increase in IgG antibody titers against the RBD, particularly in mice orally immunized with P. pastoris cells expressing construct III. This group exhibited the highest OD450 value (2.06), indicating a strong humoral immune response elicited by the trimerized and adjuvanted construct. As depicted in Fig. 7, the OD₄₅₀ values at the 1:1000 dilution demonstrated a significant increase in anti-RBD of the spike protein IgG levels, particularly in response to construct III at 150 YU, indicating a robust humoral immune response. Statistical analysis using one-way ANOVA revealed significant differences among groups (p < 0.0001), with Tukey’s post hoc test confirming that construct III at 150 YU was significantly more immunogenic than other groups. These findings suggest that both the construct design and the amount of yeast contribute to the observed IgG responses. Other constructs also produced significantly higher responses compared to the PBS and control groups (p < 0.001). Serum IgG titers against SARS-CoV-2 RBD in mice vaccinated with different yeast constructs at two cell densities (10 YU (A), and 150 YU (B)). Each bar represents the mean OD450 at 1:1000 dilution. Insets show antibody titration curves (1:50 to 1:2000). Data are means from triplicate ELISA measurements using pooled sera. Similarly, mucosal immune responses, evaluated by measuring fecal IgA levels, revealed distinct differences among constructs and cell densities (Fig. 8). Construct III induced the highest levels of fecal IgA at both tested cell densities, with a marked increase observed at the higher dose (1.56 vs. 2.39). Construct II also elicited a strong mucosal immune response at the higher cell density. In contrast, Construct I induced only modest IgA production and was less effective than the control at the same density. Fecal IgA levels in mice vaccinated with yeast constructs at 10 and 150 YU (A, B). Bars represent mean OD450 values from triplicate ELISA tests of pooled fecal extracts. Discussion Whole yeast-based vaccines represent a promising innovation in the field of vaccinology, capitalizing on the natural adjuvant properties and delivery potential of microbial platforms10,17,56,57. The motivation for this research stems from the urgent need for accessible, cost-effective, and easy-to-administer vaccines, particularly in low-resource environments. Oral vaccines offer several advantages over traditional injectable counterparts, including simpler administration, higher patient compliance, and the potential to stimulate both systemic and mucosal immunity10,58. Despite these benefits, one of the primary challenges in developing effective oral vaccines is the harsh environment of the GI tract, which can lead to antigen degradation and reduced efficacy11,56,59. To address this, our strategy utilizes the resilient cell wall of P. pastoris to safeguard the RBD of the SARS-CoV-2 spike protein antigen from GI tract degradation, while simultaneously harnessing the yeast cell wall’s inherent adjuvant properties to enhance immune response19. This approach aims to overcome key obstacles in oral vaccine delivery and improve overall vaccine efficacy. In this study, we evaluated the immune responses in mice immunized orally with recombinant yeast cells displaying the SARS-CoV-2 RBD, engineered with a trimerization foldon domain to promote native-like antigen structure, and combined with immunostimulatory elements such as truncated flagellin (FliC) and the yeast cell wall itself as natural adjuvants. The design of the three antigenic constructs is illustrated in Fig. 1. To gain insight into the spatial conformation of these constructs, structural modeling was performed using ColabFold. The primary goal was to explore the potential of yeast-based oral vaccines for eliciting robust humoral and mucosal immunity, particularly in the context of the ongoing COVID-19 pandemic. Our findings provide valuable insights into the immunogenicity of yeast-displayed RBD, and the role of foldon domain and FliC in enhancing immune responses. Surface display of the recombinant constructs on P. pastoris cells was successfully validated based on SDS-PAGE, fluorescence microscopy, flow cytometry, cell-based immunoassay, and ACE2-binding assays (Figs. 3, 4, 5 and 6). In terms of immune response, both humoral and mucosal immunity were assessed. Our results revealed that immunization with P. pastoris cells expressing the RBD of the SARS-CoV-2 spike protein, particularly with the addition of foldon domain and FliC, led to significantly elevated serum IgG titers, indicating a strong systemic immune response. The IgG response was more pronounced at a higher cell density, particularly for construct III, indicating an enhanced humoral immune response (Figs. 7). These findings are consistent with previous studies demonstrating that oral and intranasal vaccines, especially those delivered via yeast systems, can elicit potent systemic antibody responses when formulated with effective adjuvant21,56,60–65. Interestingly, the mucosal immune response, as measured by secretory IgA levels in fecal samples, showed a similar trend (Fig. 8). The presence of foldon domain and FliC in the vaccine constructs led to significantly higher IgA titers, further confirming that the inclusion of these adjuvants enhances the ability of the yeast display system to activate mucosal immunity. This is particularly important for oral vaccines, as mucosal immunity serves as the first line of defense against pathogens entering through the gastrointestinal tract. Notably, while increasing the cell density boosted mucosal immunity, it did not further enhance serum IgG levels, highlighting the potential for optimizing yeast-based vaccines for targeting both systemic and mucosal immune responses effectively. The role of FliC as an adjuvant is particularly noteworthy. Flagellin has been shown to engage TLR5 on immune cells, leading to the activation of pro-inflammatory cytokines and the recruitment of immune cells, which may contribute to the observed enhancement in both humoral and mucosal immunity42,44,66. The addition of the truncated form of FliC in this study offers a more controlled immune activation, reducing the risk of excessive inflammatory responses while still stimulating effective immune response67. The incorporation of the T4 fibritin foldon domain into the constructs represents another important factor contributing to the enhanced immunogenicity of the spike protein RBD. By promoting trimerization of the RBD, the foldon domain helps to more closely resemble the native trimeric conformation of the SARS-CoV-2 spike protein, which is known to be critical for the induction of potent neutralizing antibody responses. Trimeric antigen formats have been widely reported to exhibit superior immunogenicity compared to monomeric forms, eliciting stronger and more functionally relevant antibody responses68,69. Notably, the application of the foldon domain as a trimerization motif has been previously validated in several studies, including its use in SARS-CoV-2 RBD-based diagnostic and vaccine-related constructs such as the RBD–foldon fusion developed for serological applications70 and the engineered SARS-CoV-2 HR1 trimer construct demonstrating improved potency and broad-spectrum activity against SARS-CoV-2 variants71. In agreement with these reports, our findings further support the functional relevance of foldon-mediated trimerization. In the present study, structural predictions obtained from AlphaFold, together with the observed functional differences between constructs lacking and containing the foldon domain, collectively support the likelihood of enhanced structural stability and a more ordered trimeric organization. These results are consistent with the expected role of foldon as a trimerization scaffold and further validate the rational design of the engineered constructs. Overall, the combined structural and immunological evidence suggests that foldon-mediated trimerization contributes meaningfully to the improved immunogenic performance of the RBD-based vaccine constructs. The robust immune responses observed in this study also underline the potential of yeast-based vaccines in overcoming the challenges associated with oral vaccine development, particularly the need to protect the antigen from degradation in the GI tract while promoting strong immune activation. The use of P. pastoris, a yeast species known for its ability to survive harsh GI condition15,56, combined with the immunomodulatory effects of FliC and foldon domain, makes this platform a promising candidate for oral vaccine development against COVID-19 and potentially other infectious diseases. To compare the current study with previously developed yeast-based SARS-CoV-2 vaccine platforms, different antigen expression strategies have been employed, including surface display, secretion, and intracellular expression followed by purification. The most commonly used approach relies on the Aga1/Aga2 anchoring mechanism to display monomeric spike RBD on the yeast surface in S. cerevisiae. These constructs have been shown to induce both humoral and cellular immune responses following oral administration. In S. cerevisiae systems, oral vaccine strategies using Aga1/Aga2-mediated surface display of RBD have been reported60,72, demonstrating the feasibility of mucosal immunization. In contrast, S. boulardii has been engineered to secrete RBD for oral delivery, where immunogenicity is driven by soluble antigen release rather than cell-associated presentation73. Additionally, systemic immunization approaches using Aga1/Aga2-based surface display in S. cerevisiae have also been reported for injectable vaccine formats61. P. pastoris has also been utilized in multiple vaccine design strategies. Intracellular expression followed by purification of RBD combined with aluminum hydroxide adjuvant has been widely applied for injectable subunit vaccines74. RBD-dimer vaccines against SARS-CoV-2 have also been developed for injectable immunization strategies75. Moreover, both oral and surface display-based approaches using α-agglutinin anchor protein have also been explored in P. pastoris76, highlighting its versatility as a vaccine production platform. In comparison with these previously reported systems, the present study introduces a more integrated design. The use of SED1 as a glycosylphosphatidylinositol (GPI)-anchored cell wall protein enables stable and efficient surface display of the antigen in P. pastoris, distinguishing it from Aga1/aga2 display system S. cerevisiae as well as secretory and purified protein approaches. Importantly, incorporation of the T4 fibritin foldon domain introduces a defined trimerization strategy that is largely absent in earlier yeast-based vaccine designs, with the aim of more closely approximating the native trimeric organization of the SARS-CoV-2 spike protein. In addition, fusion of truncated flagellin (FliC) provides an intrinsic immunostimulatory component, enhancing innate immune activation without the need for external adjuvants. Collectively, these design features contribute to the induction of both systemic IgG and mucosal IgA responses, with the enhanced immunogenicity observed in the construct containing both foldon and FliC, underscoring the benefit of combining antigen structural engineering with built-in immune modulation in a single yeast-based oral vaccine platform (Table 1). Table 1Comparative overview of yeast-based SARS-CoV-2 vaccine platforms.Yeast-based SARS-CoV-2 vaccine platformsAnchoring approachAdministration methodMultimerizationImmunological outcomesReferencesSurface display/Whole-cell vaccine (S. cerevisiae)Aga1/Aga2 anchor proteinOralNoneStrong systemic IgG + mucosal IgA, robust cellular immunity; Th1/Th2 mixed with Th1 bias after oral administration60Secreted oral yeast vaccine (S. boulardii)No surface anchoringOralNoneDetectable IgG + IgA, strong IFN-γ (Th1-biased response), minimal IL-4; systemic + mucosal immunity73Intracellular expression(P. pastoris)No anchoringInjection with aluminum hydroxide adjuvantNoneStrong neutralizing antibody responses, variant-dependent cross-neutralization; strong adjuvant dependence (Alum ± CpG greatly increases NAb titers)74Surface display/Whole-cell vaccine (P. pastoris)α-agglutinin (Agα) anchor proteinOralNoneInduced IgG (including IgG2a/IgG2), mucosal IgA, neutralizing antibodies; no toxicity or gut pathology76Surface display/Whole-cell vaccine (S. cerevisiae)Aga1/Aga2 anchor proteinOralNoneIncreased IgG + IgA, immune modulation + gut microbiota changes, systemic immunoregulation72Secreted soluble protein (subunit vaccine)(P. pastoris)No anchoringIntramuscularly immunizedEngineered RBD dimerizationStrong protective neutralizing antibodies, suitable booster immunogen; improved epitope exposure confirmed structurally75Surface display/Whole-cell vaccine (S. cerevisiae)Aga1/Aga2 anchor proteinInjected subcutaneously in the mouse back skinNoneinduced strong humoral and cellular immune responses in mice61Surface display/Whole-cell vaccine (P. pastoris)SED1 anchor proteinOral administration with FliC adjuvantYES – Foldon trimerization domain (T4 fibritin)Strong IgG (serum) + mucosal IgA responses after oral immunization; enhanced immunogenicity via trimerization + flagellin adjuvant effect (TLR5 agonism)Current study Comparative overview of yeast-based SARS-CoV-2 vaccine platforms. Intracellular expression (P. pastoris) Secreted soluble protein (subunit vaccine) (P. pastoris) In conclusion, the results of this study highlight the potential of yeast-based oral vaccines, particularly those displaying the RBD of the SARS-CoV-2 spike protein engineered with a foldon domain to promote trimeric assembly, in combination with the immunomodulatory adjuvants such as truncated flagellin and the intrinsic adjuvant properties of the yeast cell wall. The successful induction of both humoral and mucosal immunity suggests that this approach could provide a valuable alternative to traditional vaccine platforms, offering advantages in terms of ease of administration, cost-effectiveness, and scalability. Future studies will need to further evaluate the long-term efficacy of this oral vaccine, including its ability to provide protection against viral challenges, and explore the potential for scaling up production for human use. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Acknowledgements The authors gratefully acknowledge the Pasteur Institute of Iran for providing the recombinant SARS-CoV-2 RBD antigen and the anti-RBD antibodies used in this study. We are also thankful to Dr. Mohammad Javad Rasaee (Department of Clinical Biochemistry, Tarbiat Modares University) for generously providing the FITC-conjugated secondary antibody. The authors would like to thank Dr. Mehdi Gholami (Faculty of Pharmacy, University of Tehran) for his valuable assistance with the animal experiments. Generative AI and AI-assisted technologies in the writing process During the preparation of this work the authors used ChatGPT (Open AI) in order to improve the clarity, grammar, and academic style of the manuscript text. After using this tool/service, the authors reviewed and edited the content as needed and takes full responsibility for the content of the published article. Author contributions Tahereh Saveii: Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Sareh Arjmand: Conceptualization, Methodology, Validation, Visualization, Writing—review & editing, Ismaeil Haririan: Methodology, Project administration, Resources, Supervision, Navid Pourzardosht: Methodology, Validation, Visualization, Reza H. Sajedi: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing—review & editing. Funding This work was supported by Tarbiat Modares University and Tehran University of Medical Sciences (TUMS) for financial support under grant number 56588–104-4–1402. Data availability The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Ethical approval All animal experiments in this study were conducted in accordance with the ethical guidelines and regulations approved by the Research Ethics Committee of Tehran University of Medical Sciences (TUMS). Ethical approval was granted under the approval code IR.TUMS.TIPS.REC.1401.009. The study is reported in accordance with the ARRIVE (Animal Research: Reporting of in vivo Experiments) guidelines 2.0 to ensure transparent and comprehensive reporting of in vivo animal research. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Data Availability Statement The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.