Young (2–3 months) male and female C57BL/6J mice were purchased from Janvier (Le Genest-Saint-Isle, France). Furthermore, middle-aged (16 months) and old (23–24 months) male and female C57BL/6J mice (initially obtained from Janvier) were obtained from our in-house breeding colonies (SFB 1506 “Aging at interfaces”; project Z02) at the Tierforschungszentrum of Ulm University. Mice were routinely housed in our animal facility under standardized pathogen-free (SPF) conditions. Aging mice were routinely checked for overall appearance, and only healthy aged mice were used in the described experiments.

The original SARS-CoV-2 spike sequence (S), encoding aa1-1273 of the Wuhan-Hu-1 strain (NCBI reference sequence: NC_045512.2), was codon optimized and synthesized by GeneArt (Thermo Fisher Scientific) and cloned into the pCI expression vector (cat. no. E1731, Promega) using NheI/NotI sites. This sequence was used to generate the other S-constructs described in this study by site-directed mutagenesis or Gibson Assembly. The respective primers for each mutagenesis PCR were ordered at Biomers.net GmbH (Ulm, Germany). The PCRs were conducted using 12.5 μl of the Phusion High-Fidelity PCR Master Mix (New England Biolabs, USA), 12.5 ng of the respective template, and 0.5 μM of the respective forward and reverse primer. PCR protocols, especially the annealing temperatures and the extension time, were adjusted to the respective primer pairs and constructs. The general PCR protocol was as follows: initial denaturation (98°C, 30 s), 20× denaturation (98°C, 10 s), annealing (62°C–72°C, 30 s), extension (72°C, 180 s–420 s) and final extension (72°C, 600 s). All antigen constructs used for protein- and DNA-based vaccination studies were fused with a COOH-terminal Strep-tag sequence (SWSHPQFEKGGGSGGGSGGGSWSHPQFEK). Furthermore, we designed a vector encoding a S6-P_ΔTM/EPEA detection antigen for S-specific ELISA (see below) by adding a COOH-terminal linker sequence and a glutamic acid–proline–glutamic acid–alanine tag sequence (GYQDY-EPEA). Plasmids were produced in transformed E. coli DH5α (Thermo Fisher Scientific, USA) and purified using the endotoxin-free Plasmid Mega Kit (#12183; Qiagen). All plasmids were sequenced before use by Eurofins Genomics (Germany) to ensure the correct sequence of the respective construct.

S proteins were expressed in HEK293T cells (ATCC CRL-3216) transiently transfected for 48 h with the indicated plasmid DNAs using the calcium phosphate method (42). Usually, approximately 5×10⁸ cells were used for large-scale production of the indicated cell-associated and secreted antigens. Cells were cultured in Dulbecco´s Modified Eagle Medium (DMEM) high-glucose medium supplemented with 10% heat-inactivated Fetal calf serum (FCS), 1% Penicillin/Streptomycin, 1% L-glutamine, and 0.1% β-mercaptoethanol at 37°C/5% CO₂. At 24 h post-transfection, the medium was changed, and at 48 h post-transfection, cell culture supernatants were collected, cleared by centrifugation (450×g, 5 min), and subsequently passed through a 0.45-µm filter. Notably, the expression conditions for the production of the secreted S6-P_ΔTM were further optimized. At 24 h post-transfection, the culture medium was additionally supplemented with 3.75 mM valproic acid and 3 g/L D-(+)-glucose (43), and cells were cultured for additional 5 days before harvesting the supernatant.

For purification of cell-associated/membrane-anchored S-antigens, cells were lysed for 30 min at 4°C with 5 ml of lysis buffer [100 mM Tris–HCl, 150 mM NaCl, 0.5% NP40 detergent, and Protease Inhibitor Cocktail Tablets (cat. no. 11836145001, Roche Applied Science; Penzberg, Germany) pH 8.0] and cleared by centrifugation (8,000×g, 30 min). Recombinant antigens in cell lysates and supernatants were purified with Strep-Tactin (IBA Lifesciences, Germany) as described previously (42). Briefly, cell extracts or supernatants were passed over StrepTactin Sepharose (#2-1201-025, IBA, Göttingen, Germany) packed polypropylene columns (cat. no. 29922, Pierce, Rockford, USA). Sepharose-bound antigens were purified with five column volumes of wash buffer (100 mM Tris–HCl, 150 mM NaCl, 1mM EDTA; pH 8.0) and eluted in eight 500-µl fractions in elution buffer (20 mM Tris–HCl, 30 mM NaCl, 0.2mM EDTA; pH 8.0) supplemented with 2.5 mM desthiobiotin (cat. no. 2-1000-002, IBA, Göttingen, Germany). S-protein-containing fractions were determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie Brillant Blue staining of the gels (see also Supplementary Figure S1 ). S-protein-containing samples were pooled and concentrated in Vivaspin 500 columns (30,000 or 100,000 MW cutoff) (cat. no. VS0141; Sartorius; UK) to a volume of approximately 100–200μl. Furthermore, an S6-P_ΔTM/EPEA detection antigen used for S-specific ELISA (see below), modified with a COOH-terminal linker sequence and a glutamic acid–proline–glutamic acid–alanine tag sequence (GYQDY-EPEA), was purified from cell culture supernatants using CaptureSelect™ C-tagXL Affinity Matrix (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions (https://assets.thermofisher.com/TFS-Assets/LSG/manuals/MAN0017302_CapSelCtagXLAffMatrix_PI.pdf). We also purchased a recombinant, carrier-free trimeric S2-P_ΔTM/GCN4-IZ SARS-CoV-2 S protein (accession no. YP_009724390.1) (cat. no. 10561-CV; R&D Systems). This S protein was produced in HEK293 cells, contains the ectodomain up to aa1211, a silenced furin cleavage site (S_R682S,R685S), a prefusion stabilization motive (S_K986P,V987P), a COOH-terminal GCN4-IZ trimerization domain, and a 6x-His-tag.

Samples of recombinant S proteins were either mixed with a fourfold concentrated SDS loading buffer (250 mM Tris–HCl (pH 6.8), 4% SDS, 40% glycerol, 0.02% bromphenol blue) either with (reducing) or without (non-reducing) 200 mM β-mercaptoethanol. Samples were mixed with the indicated loading buffer at a ratio of 3:1 (v/v). Only the samples analyzed under reducing conditions were additionally boiled (95°C for 5 min). Afterwards, the samples were loaded onto a 10% Bis–Tris gel using the Mini-PROTEAN electrophoresis chamber (Bio-Rad, USA). Gels were stained with Coomassie Brilliant Blue (44.8% methanol, 9.2% acetic acid, and 2 mM Coomassie R-250) for 30 min and destained in a mixture of 10% methanol and 10% isopropanol for 60–120 min at room temperature.

For Western blotting, the gels were blotted onto a 0.2µm nitrocellulose membrane in transfer buffer (25 mM tris, 200 mM glycine, and 15% methanol) at 60 V for 60 min using the Mini Trans-Blot Cell system (Bio-Rad, USA). Thereafter, the membranes were incubated with blocking buffer for 60 min at room temperature (1% milk powder in PBS), washed twice with washing buffer (15 mM Tris–HCl, 150 mM NaCl, 5 mM sodium azide, 1 mM EDTA, and 0.1% Tween-20; pH 7.8) and incubated with primary antibodies for 60 min at room temperature: anti-Strep-tag (1:33.333; StrepMAB-Classic-HRP, IBA Lifesciences, Germany), anti-Grp78/BiP (1:5,000, Proteintech, USA), anti-Hsc70/Hsp73 (1:1,000, Enzo Life Sciences, USA), or mouse anti-S sera (1:500) derived from mice vaccinated twice with pCI/S2-P_ΔTM DNA, and detected with the respective secondary antibodies, namely, anti-mouse IgG-HRP (1:2,500, Amersham, UK) or anti-rabbit IgG-HRP (1:2,000, Amersham, UK). Blots were developed for 15–30 s using the Immobilon Western Chemiluminescent HRP Substrate (Merck Millipore, USA) and documented using the Fusion FX (Vilber Lourmat, France).

Size exclusion chromatography (SEC) experiments were performed with Superose 6 Increase 10/300 GL column (Cytiva 29-0915-96). All runs were performed with PBS buffer at 0.5 ml/min after injecting 250-µl spike constructs (0.5 mg/ml). The column was calibrated using a mixture of dextran blue (MW ≥ 2MDa) and globular proteins of known molecular weights of 670 kDa, 158 kDa, 44 kDa, 17 kDa, and 1.35 kDa (cat no. 151-1901, Biorad gel-filtration calibration kit). The calibration was used to estimate the apparent molecular weight of the different S-species analyzed by SEC (Kav method).

Mice were immunized i.m. into both tibialis anterior muscles at day 0 (prime) and day 22 (boost) with 100 µg DNA diluted in 50 µl PBS and/or 10 µg of protein mixed with 10 µg Quil-A adjuvant in 50 µl PBS (provided by Dr. Eric Lindblad, Brenntag Biosector, Frederikssund, Denmark). Blood samples were collected at day 36 (day 14 post-boost).

S-RBD-specific antibody titers were routinely determined using the commercially available Elecsys anti-SARS-CoV-2 S immunoassay (Roche, Switzerland) applied on a cobas pro e 801 module (Roche, Switzerland) in the Institute of Clinical Chemistry, Ulm University Medical Center. Briefly, serum samples were diluted 1:10 in PBS to a final volume of 150 µl and analyzed according to the manufacturer’s instructions. The raw values were corrected with the dilution factor and depicted as S-RBD-specific IgG/IgM antibody titers (U/ml). The dotted lines in the respective graphs showing S-RBD-specific IgG/IgM antibody titers in (U/ml) represent the limit of quantification (0.8 U/ml) as described by the manufacturer of the test. The assay was validated by the “Clinical Chemistry” of the Ulm University Medical Center. Confirmatory, all negative control sera derived from unimmunized young and old control mice resulted in a value below 0.8 U/ml.

Where indicated, S-antigen-specific IgG antibody titers were measured with the secreted recombinant S6-P_ΔTM/EPEA detection antigen by endpoint ELISA. This protein was used to exclude any cross-reactivities against the strep-tag. Briefly, Nunc MaxiSorp™ 96-well plates (Thermo Fisher Scientific, USA) were coated with 0.1 µg of S6-P_ΔTM/EPEA/well diluted in 0.1 M sodium carbonate buffer (pH 9.5) and incubated at 4°C overnight, washed (0.05% Tween-20 in PBS), and incubated with blocking buffer (3% BSA in PBS). Serial dilutions (1:2) of serum samples derived from unimmunized young or old control and vaccinated mice (starting at 1:30 to a dilution of maximal 409,600) were added for 2 h at room temperature, washed, incubated for 1 h at 37°C with anti-mouse IgG-HRP (BD Biosciences, USA), developed with 0.4 mg/ml OPD (Merck) and 0.012% hydrogen peroxide dissolved in 0.05 M citrate-phosphate buffer (pH 5.0). The samples were then analyzed at 492 nm using a 96-well plate reader (MWG Biotech, Germany). The S-specific endpoint titers were defined as the highest serum dilution that resulted in an absorbance value three times greater than that of control sera from unimmunized young or old mice.

The production of pseudotyped particles has been previously described (40, 41). In brief, HEK293T cells were transfected with plasmids encoding SARS-CoV-2 S variants B.1, B.1.1.7 (Alpha), B.1.351 (Beta), B.1.617.2 (Delta), or B.1.1.529.1 (Omicron BA.1) (44–46) by Transit LT-1 (Mirus, USA). After 24 h, cells were inoculated with a replication-deficient vesicular stomatitis virus (VSV) vector in which the genetic information for its native glycoprotein (VSV-G) is replaced by genes encoding enhanced green fluorescent protein and firefly luciferase (kindly provided by Gert Zimmer, Institute of Virology and Immunology, Mittelhäusern, Switzerland) and incubated at 37°C for 2 h. Then, the inoculum was removed, cells were washed with PBS, and fresh medium containing anti-VSV-G antibody (from I1-hybridoma cells; ATCC no. CRL-2700) was added to block residual VSV-G carrying particles. After 16–18 h, supernatants were collected and cleared by centrifugation. Samples were aliquoted and stored at −80°C.

The antibody-mediated SARS-CoV-2 spike-specific pseudovirus neutralization assay was performed as described previously (40, 41). Briefly, Vero E6 cells (6,000 cells/well) were seeded in 96-well plates 1 day prior in DMEM supplemented with 2.5% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate, and 1× non-essential amino acids. Heat-inactivated (56°C, 30 min) serum samples were serially titrated in PBS (fourfold titration series with seven steps plus buffer only control) and mixed with VSV∗ΔG-FLuc pseudovirus stocks (1:1, v/v) in a total volume of 48 µl. Serum-pseudovirus mixes were incubated for 30 min at 37°C before being added to the Vero E6 target cells in duplicates (final on-cell dilution of sera: 100-, 400-, 1,600-, 6,400, 25,600-, 102,400-, 409,600-fold). After 16–18 h, the transduction efficiency was analyzed. Cells were lysed with Cell Culture Lysis Reagent (cat. no. E1531; Promega) at room temperature. Lysates were then transferred into white 96-well plates (cat. no. 136101; Thermo Fisher), and luciferase activity was measured using a Luciferase Assay System (cat. no. E1501; Promega) and a plate luminometer (Orion II Microplate Luminometer, Berthold). For quantitative analysis of raw values, the background signal of untreated control cells (in relative light units per second; RLU/s) was subtracted, and values were normalized to pseudovirus-infected control cells in the absence of serum. Results are given as serum dilution factors resulting in 50% pseudovirus neutralization (50% pseudovirus neutralizing titer; PVNT50) on cells, calculated by nonlinear regression ([inhibitor] vs. normalized response–variable slope) in GraphPad Prism Version 9.1.1. As the tested serum dilutions range from 100- to 409,600-fold, we could detect PVNT50 titers within this range, defining the lower and upper cutoff values as 100 and 409,600, respectively. The dotted lines in the respective graphs showing PVNT50 values represent the limit of quantification (100).

The GraphPad Prism 8.4.3 software (GraphPad, San Diego, CA, USA) was used for statistical analyses and creation of graphs. Statistically significant differences between two indicated groups were determined using Student’s unpaired t-test. Where indicated, we also used Kruskal–Wallis followed by Dunn’s multiple comparisons test. p-values smaller than 0.05 were considered as statistically significant and indicated with asterisks in the graphs (p<0.05*, p<0.01**, and p<0.001***). Non-significant (ns) differences between individual groups were not marked in the graphs and described in the respective figure legends. Data are depicted as geometric mean ± geometric SD, and group sizes are stated in the figure legends.

To investigate the antigenicity of cell-associated trimeric and secreted monomeric S proteins through DNA- and protein-based immunization, we generated mammalian expression vectors that were used for both the production of recombinant S proteins in transiently transfected HEK-293T cells and DNA-based vaccination studies. We generated vectors encoding the TM-anchored full-length S protein from the original Wuhan-Hu-1 strain (pCI/S) and a mutant S2-P with two proline substitutions at positions S_K986P,_V987P that convey a superior conformation and antigenicity by stabilization of the prefusion conformation (pCI/S2-P) (47, 48) ( Figure 1A ). Considering that TM-deficient S proteins were efficiently secreted into the cell culture supernatant, but did not form higher molecular trimeric structures unless modified with a specific trimerization domain (28, 49, 50), we designed vector DNAs encoding a prefusion stabilized (S_K986P,_V987P) TM-deficient and secreted S2-P_ΔTM (pCI/S2-P_ΔTM) protein and a derivative thereof with four additional proline substitutions (S_{K986P,V987P,F817P,A892P,A899P,A942P}) (29) and a silenced furin cleavage site (S_{R682G,R683G,R685S}) ensuring structural stability (pCI/S6-P_ΔTM) ( Figure 1A ). Furthermore, to enable standardized affinity purification of the respective S proteins and detection of intrinsic structures or possible interactions with cellular proteins, we cloned a Strep-tag sequence at the COOH-termini of the S-sequences ( Figure 1A ). The Strep-tag purification system was used because we had already confirmed its capability to produce recombinant antigens, like RNA-binding proteins or protein particles, under very mild conditions with high purity (42). We thus expected to express and produce different monomeric and trimeric S proteins and to reveal if de novo expressed TM⁺ and TM^- S proteins interact with cellular proteins, an issue that could mimic antigen expression in antigen-presenting cells upon DNA-based vaccination and therefore impact the efficacy of the respective vaccines (51–53).

For large-scale production of recombinant proteins, usually 5–10×10⁸ HEK-293T cells were transiently transfected with the different vector DNAs, followed by Strep-Tactin-specific purification of the respective S proteins from cell lysates or cell culture supernatants. A representative SDS-PAGE analysis of the purification of the secreted S6-P_ΔTM is shown in Supplementary Figure S1 . Samples were subjected to SDS-PAGE under reducing conditions followed by Coomassie Brilliant Blue staining of the gels ( Figure 1B ) or Western blot analyses using S-protein-specific antisera from vaccinated mice ( Figure 1C ). These analyses showed prominent steady-state levels of protein bands of approximately 180 kDa present in the cell lysates of all four constructs, depicting the S, S2-P, and the slightly smaller TM-deficient S2-P_ΔTM and S6-P_ΔTM proteins ( Figures 1B, C ). Western blot analysis but not Coomassie Brillant Blue staining of the gel also revealed a protein band of approximately 120 kDa that could not yet be identified ( Figures 1B, C ). As expected, the S and S2-P antigens were exclusively found in the cell lysates, whereas the S2-P_ΔTM and S6-P_ΔTM antigens were also detected in the cell culture supernatant ( Figures 1B, C ). However, the S6-P_ΔTM showed a much better secretion efficacy into the culture supernatant than the S2-P_ΔTM ( Figures 1B, C ). Therefore, we only used the S6-P_ΔTM in our vaccination studies. After optimizing the antigen expression conditions in transfected HEK-293T cells, we could gain approximately 600 μg of affinity-purified S6-P_ΔTM from the supernatant of 5×10⁸ transfected cells. Notably, it was difficult to reproducibly produce the cell-associated S and S-2P proteins, and on average, we could gain approximately 15–30 µg S and 30 µg S-2P proteins from cellular extracts of 10×10⁸ HEK-293T cells. Interestingly, Coomassie Brilliant Blue staining of the gels also revealed a prominent co-purification of additional 70–80 kDa protein bands with cellular S, S2-P, S2-P_ΔTM, and S6-P_ΔTM, but not secreted S2-P_ΔTM and S6-P_ΔTM proteins. ( Figure 1B ). These bands were identified by commercial NH₂-terminal protein sequencing (TopLab; Munich, Germany) or Western blot analyses as ER-associated Grp78 and constitutively expressed cytosolic Hsp73 ( Figures 1B, D ). The cell-associated, but not the secreted S proteins thus stably bound Hsp73 and Grp78 ( Figures 1B, D ).

When non-reducing conditions were applied for SDS-PAGE, the 180-kDa full-length S and S2-P protein bands were no longer detectable and instead appeared as a high molecular band ( Figure 2A , lanes 2 and 4), suggesting that these proteins maintained a high molecular multimeric structure. Furthermore, we observed in these non-reduced samples that anionic SDS-PAGE conditions were sufficient to strip off the stress proteins from the S-multimers appearing in the gels as separate bands ( Figure 2A , lanes 2 and 4). In contrast, the prominent 180 kDa protein band of the secreted S6-P_ΔTM formed under both reducing and non-reducing conditions of SDS-PAGE ( Figure 2B ), indicating that the secreted S6-P_ΔTM was mainly composed of monomers (34, 49, 50). A very low amount of S6-P_ΔTM assembled into high molecular structures detectable in SDS-PAGE under non-reducing conditions ( Figure 2B , lanes 3 and 4). Notably, an almost identical pattern of protein migration in SDS-PAGE has been shown for full-length, trimeric (TM⁺), and monomeric (TM⁻) coronavirus S proteins (34).

The affinity-purified secreted S6-P_ΔTM was further analyzed by size exclusion chromatography (SEC). As expected from SDS-PAGE analyses, we observed a SEC elution profile with one main protein peak of approximately 370 kDa as determined with globular standard proteins of known molecular weight ( Figure 2C ). Furthermore, minor protein peaks corresponding to a high molecular protein of approximately 650 kDa and a small protein of approximately 17 kDa were detectable ( Figure 2C ). To relate the individual protein peaks to S protein trimers, we used a commercially available, trimeric S2-P_ΔTM protein modified with a GCN4-IZ trimerization motif (S2-P_ΔTM/GCN4-IZ). Analytical SEC showed a single peak of recombinant S2-P_ΔTM/GCN4-IZ corresponding to a S-trimer of approximately 670 kDa ( Figure 2C ). As expected, the trimeric S2-P_ΔTM/GCN4-IZ protein thus clearly separated from the smaller S6-P_ΔTM ( Figure 2C ), confirming that these proteins have two distinct conformations (49, 50). However, these findings also showed that it was difficult to determine the exact molecular weight of these two glycoproteins by SEC and simply assign it to the molecular weights expected from SDS-PAGE (i.e., considering an approximately 180 kDa molecular weight of a monomer). In our system, we determined higher molecular weights for both the trimeric S2-P_ΔTM/GCN4-IZ antigen and the monomeric S6-P_ΔTM antigen that differ from the calculated values, i.e., 670 kDa instead of 540 kDa for the trimer and 370 kDa instead of 180 kDa for the monomer. This was not unexpected, since their migration behavior during SEC could be affected among others by their glycosylation state possibly leading to an overestimation of the molecular weight up to 100% (54), their conformation, and from the molecular weight standard proteins (49, 50). Furthermore, the level of glycosylation might vary depending on the antigen itself, the cells used for production, and the level and/or time of expression.

An interesting observation was that, in contrast to full-length S- and S-2P proteins, the recombinant trimeric S2-P_ΔTM/GCN4-IZ protein was sensitive to anionic SDS-PAGE conditions and quantitatively disassembled into the typical monomeric 180 kDa protein band in both reducing and non-reducing SDS-PAGE ( Figure 2D ). This suggested that the natural TM domain, but not artificial trimerization motives, mediated a better stabilization of the high molecular S-trimers and their resistance to anionic SDS-PAGE conditions.

To determine the antigenicity of cell-associated vs. secreted S proteins, we immunized young (2–3 months) C57BL/6J mice twice (at day 0 and 22), either with equal amounts of pCI/S, pCI/S-2P, or pCI/S6-P_ΔTM DNA and the recombinant, cell-associated S and S-2P- or secreted S6-P_ΔTM proteins in saponin-based Quil-A adjuvant (55) ( Figure 3 ). S-RBD-specific antibody titers were measured with the commercial Elecsys Anti-SARS-CoV-2 S immunoassay at day 14 after the second vaccination ( Figure 3A ). Immunization of B6 mice with pCI/S, pCI/S2-P, or pCI/S6-P_ΔTM DNA induced very similar S-RBD-specific antibody titers ( Figure 3A ). This showed that neither the stabilization of the prefusion conformation in the pCI/S-2P vs. the pCI/S constructs (47, 48) nor the expected trimeric (pCI/S-2P and pCI/S) vs. monomeric (pCI/S6-P_ΔTM) antigen structure had a striking effect on the priming of S-RBD-specific antibody responses in young mice through DNA-based immunization ( Figure 3A ). Furthermore, equal amounts of recombinant S and S2-P proteins isolated from cell lysates of transfected HEK-293T cells and secreted monomeric S6-P_ΔTM induced comparable S-RBD-specific antibody titers in young mice ( Figure 3A ). The differences in the antibody titers induced through the different DNA- and protein-based vaccines did not reach statistical significance, although recombinant proteins showed a pronounced trend to induce higher antibody titers than vector DNAs at the dosages used here ( Figure 3A ). Notably, the amount of recombinant S6-P_ΔTM protein and its formulation with adjuvant was crucial for the magnitude of antibody titers ( Supplementary Figure S2 ).

We next used a vesicular stomatitis virus (VSV)-based SARS-CoV-2 pseudovirus carrying the S protein to determine the neutralizing activity of the elicited antibodies (40, 41). Irrespective of the antigen conformation, the S-specific antibodies developed in different DNA- (pCI/S, pCI/S2-P, pCI/S6-P_ΔTM) or protein-immunized (S, S2-P, S6-P_ΔTM) young mice efficiently suppressed S-mediated cell entry of pseudovirus particles ( Figure 3B ). This showed that both trimeric and monomeric S proteins elicited substantial pseudovirus-neutralizing antibodies in young mice, confirming previous reports using recombinant S1 or S-RBD subunit vaccines (28, 32, 36–39).

To further curtail the antigenicity of the NH₂-terminal domain of the S protein and to determine possible interactions with cellular chaperones, we generated vectors expressing secreted forms of NH₂-terminal S fragments, namely, pCI/S-300, pCI/S-450, and pCI/S-596, of which only the S-596 antigen covers the entire S-RBD (position S_319–541) ( Figure 4A ). All three protein fragments were secreted in transiently transfected HEK-293T cells ( Figure 4B ), and we could gain approximately 100 µg of S-300, 20 µg of S-450, and 80–120 µg of S-596 proteins from the cell culture supernatants of 5×10⁸ cells. Only the cell-associated forms of S-300, S-450, and S-596 proteins bound Grp78, but only S-450 and S-596 proteins bound detectable levels of Hsp73 ( Figure 4C ). This was expected, as Hsp73 generally showed a higher binding affinity to larger proteins (53). Furthermore, immunization of mice with pCI/S-596, but not pCI/S-300 or pCI/S-450, induced S-RBD-specific antibodies ( Figure 4D ) and a pronounced pseudovirus neutralization activity ( Figure 4E ). Notably, also the S-300 protein failed to induce detectable levels of S-RBD-specific antibodies and also a neutralization activity ( Figures 4D, E ). However, both the pCI/S-300 and recombinant S-300 induced S-specific serum IgG antibodies ( Figure 4F ). We thus concluded that antibodies directed against the NH₂-terminal 300 residues of the S protein did not exert a neutralizing activity detectable in our sensitive pseudovirus neutralization assay, supporting the concept that the entire S-RBD was crucial for the induction of neutralizing antibodies.

To determine S-specific immune responses in aging mice, we initially tested the recombinant S6-P_ΔTM protein vaccine in young (2–3 months), middle-aged (16 months), and old (23–24 months) mice. Both S-RBD- and S-specific serum antibody titers, and pseudovirus neutralization activities, were similar in young and middle-aged mice, but significantly decreased in old mice ( Supplementary Figure S3 ). Therefore, we used 23–24 months old mice to evaluate the age-dependent efficacy of S-specific vaccines.

To determine the antigenicity of the monomeric (S6-P_ΔTM) versus the trimeric S protein (S2-P), we immunized young (2–3 months) and old (23–24 months) mice with pCI/S2-P or pCI/S6-P_ΔTM DNAs. Interestingly, the pCI/S2-P vaccine efficiently induced very similar S-RBD-specific antibody titers and pseudovirus neutralization activity in both young and old mice ( Figures 5A, B ). In clear contrast, both the pCI/S6-P_ΔTM and the recombinant S6-P_ΔTM protein vaccines induced significantly lower S-RBD-specific antibody titers in old than in young mice, along with an overall weakened or absent pseudovirus neutralization activity in the old ( Figures 5C, D ). This highlighted an inefficient priming of neutralizing antibodies in old mice by monomeric S6-P_ΔTM antigen.

To determine whether heterologous prime/boost injections of the S6-P_ΔTM antigen could enhance its immunogenicity in old mice, we injected groups of young (2–3 months) and old (23–24 months) mice with pCI/S6-P_ΔTM DNA followed by a booster injection with recombinant S6-P_ΔTM protein (DxP). In young mice, the heterologous DxP vaccination regimen induced very similar S-RBD-specific neutralizing antibody titers as compared to homologous protein-based vaccination ( Figures 5C, D ). Most interestingly, very high S-RBD-specific antibody titers were induced in old mice by heterologous DxP vaccination, which even reached similar levels as in young mice and also conferred a very similar pseudovirus neutralization activity ( Figures 5C, D ). This suggested that heterologous DxP vaccination of old mice primarily affects the quantity but not the quality of the induced antibodies. In contrast, the reversed S6-P_ΔTM protein-prime/pCI/S6-P_ΔTM DNA-boost (PxD) vaccination could not rescue the impaired antibody response in old mice and also the S-RBD-specific antibody response in young mice was low ( Figures 5E, F ).

We next asked whether the high S-RBD-specific serum antibody titers developed in young and old mice upon S6-P_ΔTM-specific heterologous DxP vaccination could also cross-protect target cells from pseudoviruses expressing S proteins from different variants of concern (VOCs), namely, Alpha/B.1.1.7, Beta/B.1.351, Delta/B.1.617.2, or Omicron/BA.1 (40, 41, 56). Indeed, S-specific antibodies developed in young and old mice showed a broad cross-neutralization of the different pseudovirus variants, with a weakened recognition and neutralization of Omicron/BA.1 ( Supplementary Figure S4 ) (44).

Our findings thus showed that particularly the antigenicity of monomeric S6-P_ΔTM antigen was significantly improved in old mice by the heterologous DxP vaccination regimen.