Rapid and stable mobilization of CD8+ T cells by SARS-CoV-2 mRNA vaccine

SARS-CoV-2 spike mRNA vaccines1–3 mediate protection from severe disease as early as ten days after prime vaccination3, when neutralizing antibodies are hardly detectable4–6. Vaccine-induced CD8+ T cells may therefore be the main mediators of protection at this early stage7,8. The details of their induction, comparison to natural infection, and association with other arms of vaccine-induced immunity remain, however, incompletely understood. Here we show on a single-epitope level that a stable and fully functional CD8+ T cell response is vigorously mobilized one week after prime vaccination with bnt162b2, when circulating CD4+ T cells and neutralizing antibodies are still weakly detectable. Boost vaccination induced a robust expansion that generated highly differentiated effector CD8+ T cells; however, neither the functional capacity nor the memory precursor T cell pool was affected. Compared with natural infection, vaccine-induced early memory T cells exhibited similar functional capacities but a different subset distribution. Our results indicate that CD8+ T cells are important effector cells, are expanded in the early protection window after prime vaccination, precede maturation of other effector arms of vaccine-induced immunity and are stably maintained after boost vaccination.

SARS-CoV-2 spike mRNA vaccines 1-3 mediate protection from severe disease as early as ten days after prime vaccination 3 , when neutralizing antibodies are hardly detectable [4][5][6] . Vaccine-induced CD8 + T cells may therefore be the main mediators of protection at this early stage 7,8 . The details of their induction, comparison to natural infection, and association with other arms of vaccine-induced immunity remain, however, incompletely understood. Here we show on a single-epitope level that a stable and fully functional CD8 + T cell response is vigorously mobilized one week after prime vaccination with bnt162b2, when circulating CD4 + T cells and neutralizing antibodies are still weakly detectable. Boost vaccination induced a robust expansion that generated highly differentiated effector CD8 + T cells; however, neither the functional capacity nor the memory precursor T cell pool was affected. Compared with natural infection, vaccine-induced early memory T cells exhibited similar functional capacities but a different subset distribution. Our results indicate that CD8 + T cells are important effector cells, are expanded in the early protection window after prime vaccination, precede maturation of other effector arms of vaccine-induced immunity and are stably maintained after boost vaccination.
The current SARS-CoV-2 vaccination campaign provides the unique opportunity to gain important insights into human CD8 + T cell biology in the context of prime or boost mRNA vaccination. Initial data revealed that all arms of adaptive immunity such as neutralizing antibodies, virus-specific CD4 + T cells with T helper 1 (T H 1) polarization and IFNγ-producing CD8 + T cells emerge after prime or boost vaccination 4,5,9 . The onset of mRNA vaccine-mediated protection has been observed as early as 10-12 days after the first dose 3 . During this early phase, T cells and spike-specific antibodies are detectable 7,8 , whereas neutralizing antibodies first appear after boost [4][5][6]10,11 . These observations point towards a key role of vaccine-induced T cells in early protection after prime vaccination. Previous studies focused on the analysis of the overall vaccine-elicited spike-reactive T cell response 4,5,7,8,12 ; however, by this approach, the strength, dynamics and functional capacity are underestimated or even blurred in contrast to analyses performed at the single epitope level 5 . Here, we conducted continuous longitudinal analyses starting at baseline of prime vaccination until 3-4 months after boost on a single epitope level, to track the trajectories of bnt162b2 vaccine-elicited spike-specific CD8 + T cell responses in comparison to spike-specific CD4 + T cells, B cells, antibodies and their neutralizing activity.

Vaccine-elicited CD8 + T cells
We longitudinally collected peripheral blood mononuclear cells (PBMCs) and sera in 3-4-day intervals from 32 healthcare workers (Supplementary Table 1) that had not been previously infected with SARS-CoV-2, starting before prime until day 80-120 after boost (Extended Data Fig. 1a) and analysed the induction of spike-specific CD8 + T cells that target A*01/S 865 , A*02/S 269 and A*03/S 378 epitopes in 4-5 individuals each (Extended Data Fig. 1b). All three epitopes are not highly conserved between SARS-CoV-2 and SARS-CoV-1, MERS or Nature | Vol 597 | 9 September 2021 | 269 common cold coronaviruses (Extended Data Fig. 1c). Thus, the detected spike-specific CD8 + T cells indeed reflect a response to vaccination. The epitopes are not affected by the sequence variations present in the variants of concern (VOC) alpha, beta, gamma and delta (Extended Data Fig. 1c). The tested A*01-, A*02-and A*03-restricted CD8 + T cells that are part of a broader spike-specific CD8 + T cell response, however, proved to be dominant when analysing responses that span the whole S protein (Extended Data Fig. 1d). Ex vivo frequencies of A*01/S 865 -, A*02/ S 269 -and A*03/S 378 -specific CD8 + T cells were rather low after vaccination (Extended Data Fig. 2a). To increase the detection rate and to allow subsequent comprehensive profiling, we performed pMHCI-tetramer enrichment (Extended Data Fig. 2b). We detected a rapid and substantial induction of spike-specific CD8 + T cells that were present in 9 out of 13 tested donors already at days 6-8 and peaked in most donors 9-12 days post prime (dpp) (Fig. 1a). The strong CD8 + T cell activation was also reflected by high expression of CD38 and Ki-67 as early as days 6-8 in most cells (Fig. 1b, c and Extended Data Fig. 2c). Boost vaccination led to a further increase of CD8 + T cell frequencies that peaked 5-6 days post boost (dpb) with a subsequent slow contraction phase that reached nearly pre-boost frequencies at about 80-120 dpb (Fig. 1a). Post-boost and post-prime expansion were accompanied by effector T (T eff ) cell differentiation (high expression of Ki-67, CD38, granzyme B, PD-1, CD39, T-BET and TOX) ( Fig. 1b and Extended Data Fig. 2c-e). However, t-distributed stochastic neighbour embedding (t-SNE) analysis revealed that CD8 + T eff cells are qualitatively different at the peak expansion after boost (obtained at 5-6 dpb) compared with prime (obtained at 9-12 dpp) with a more consolidated cytotoxic effector cell phenotype (increased expression of T-BET, TOX and CD39) post boost (Extended Data Fig. 3a). This consolidated post-boost T eff cell response is further supported by diffusion map analysis ( Fig. 1c and Extended Data Fig. 3b). Specifically, diffusion map embedding revealed a continuous relationship of the longitudinally collected spike-specific CD8 + T cells after prime (depicted in reddish colours)/boost (depicted in grey colours) indicating a directed trajectory of the T eff cell response. Along the trajectory, CD8 + T cells exhibited the highest expression of PD-1, TOX, T-BET and CD38 after boost indicating profound activation and progressing differentiation ( Fig. 1c and Extended Data Fig. 3b). Of note, a single vaccine dose also induced boost expansion and strong activation but lower TOX expression (Extended Data Fig. 4a-c) of spike-specific CD8 + T cells in individuals who recovered from mild to moderate infection approximately 12 months before vaccination (Supplementary Table 1).
We also assessed the induction of spike-specific memory precursor CD8 + T cells that are characterized by CD127, BCL-2 and TCF-1 expression and are relevant for maintaining the CD8 + T cell response 13,14 . Roughly 20-30% of spike-specific CD8 + T cells expressed CD127 after prime followed by a transient reduction and subsequent strong increase after boost ( Fig. 1d and Extended Data Fig. 4d). Expression dynamics of TCF-1 ( Fig. 1e and Extended Data Fig. 4e) and BCL-2 (Extended Data Fig. 4f) were similar to CD127. However, the overall frequency of CD127 + (Fig. 1d) and TCF-1 + (Fig. 1e) spike-specific CD8 + T cells remained constant indicating a stable memory precursor pool induced already early after prime vaccination. Together, bnt162b2 vaccination vigorously induces a lasting spike-specific CD8 + T cell response rapidly after prime vaccination.

CD8 + T cell function after vaccination
After two weeks of peptide-specific in vitro expansion (Extended Data Fig. 5a, b), we detected higher frequencies of spike-specific CD8 + T cells after boost compared to prime vaccination (Extended Data Fig. 5c, d). However, the expansion index, a measure taking the input number of virus-specific CD8 + T cells into account was comparable for spike-specific CD8 + T cells after prime and boost vaccination, but differed between the A*01/S 865 -A*02/S 269 -and A*03/S 378 -specific CD8 + T cell      Nature | Vol 597 | 9 September 2021 | 271 responses (Fig. 2a). Thus, the increased frequencies of spike-specific CD8 + T cells after peptide-specific expansion most probably result from the increased ex vivo frequencies after boost. We also assessed spike-specific production of IFNγ and TNF (Extended Data Fig. 5e, f) and degranulation as indicated by CD107a expression (Extended Data Fig. 5g) in relation to the frequency of spike-specific CD8 + T cells after expansion as a measure of the effector function per cell. We observed reasonable effector capacity of circulating spike-specific CD8+ T cells obtained as early as 6-8 dpp ( Fig. 2b-d). Similar to the expansion capacity, cytokine production and degranulation capacity remained nearly stable after boost compared to prime ( Fig. 2b-e). Hence, functionally competent spike-specific CD8 + T cells that target different epitopes are substantially induced early after prime, and subsequent boost vaccination does not further increase their functional capacities in vitro.

CD4 + T cells, B cells and antibodies
Next, we longitudinally assessed circulating spike-specific CD4 + T cells that target DRB1*15:01/S 236 (Extended Data Fig. 6a) after prime and boost vaccination in eight individuals (Supplementary Table 1). The selected DRB1*15:01/S 236 epitope is unique for SARS-CoV-2 in comparison to SARS-CoV-1, MERS or common cold coronaviruses and conserved in circulating SARS-CoV-2 variants (B.1, alpha, gamma and delta) except for VOC beta (Extended Data Fig. 6c). The frequencies of DRB1*15:01/ S 236 -specific CD4 + T cells were lower than CD8 + T cell responses but detectable after pMHCII tetramer-based enrichment (Extended Data Fig. 6b). At baseline and in historic control samples (banked before August 2019), spike-specific CD4 + T cells were detectable with a primarily naive phenotype (Extended Data Fig. 6d, e), which reflects the presence of antigen-unexperienced precursors. After vaccination, the proportion of naive spike-specific CD4 + T cells decreased, which suggests vaccine-induced activation (Extended Data Fig. 6e). However, compared with CD8 + T cells, we observed a lower mobilization of circulating spike-specific CD4 + T cells indicated by a limited increase of frequencies (Fig. 3a) and a smaller percentage of activated ICOS + CD38 ++ or Ki-67 + subsets ( Fig. 3b and Extended Data Fig. 6f). Most activated DRB1*15:01/S 236 -specific CD4 + T cells exhibited a T H 1 cell phenotype (Fig. 3c). In line with this observation, vaccine-induced spike-specific CD4 + T cells displayed a T H 1 cell rather than a follicular helper T (T FH ) cell phenotype (Extended Data Fig. 6g). We then assessed the kinetics of the vaccine-induced humoral response. The distribution of peripheral B cell subpopulations was stable throughout prime or boost vaccination, with the exception of a progressively slight increase in antibody-secreting cells (ASC) (Extended Data Fig. 7a, b). An increase in the frequency of CD95 + B cells was observed shortly after boost, which indicates ongoing B cell activation via CD40-mediated T cell help and/or B cell receptor activation within secondary lymphoid organs 15 (Extended Data Fig. 7b). In line with the appearance of activated B cells in the periphery, we observed a progressive maturation of the serum antibody response with S1-specific IgM present after prime whereas S1-specific IgG reasonably detectable after boost ( Fig. 3d and Extended Data Fig. 7c), coinciding with a high neutralization capacity in SARS-CoV-2 plaque reduction assays. More precisely, SARS-CoV-2 B.1 and VOC alpha were similarly well neutralized by post-boost sera, whereas the cross-neutralization activity against VOC beta was reduced approximately by a factor of 5 ( Fig. 3e and Extended Data Fig. 7d). Neutralization capacity of post-boost sera was clearly increased compared with time point-matched mild infection (Extended Data Fig. 7e). In line with the progressive maturation of the antibody response, S1-and receptor-binding domain (RBD)-specific B cells (Extended Data Fig. 7f) largely remained below the ex vivo detection limit until the first week post boost (Fig. 3f). The   Article delayed appearance of circulating S1-specific B cells was confirmed by polyclonal restimulation in vitro (Fig. 3g), which showed a limited presence of class-switched B cells that could produce S1-specific IgG before boost. S1-specific B cells were largely unswitched after prime (Extended Data Fig. 7g, h), also reflected by S1-specific IgM production upon polyclonal restimulation in vitro (Extended Data Fig. 7c), and acquired a memory phenotype after boost vaccination (Extended Data Fig. 7g, h). In addition, after boost vaccination, S1-specific B cells showed increased transferrin receptor (CD71) and CD95 expression (Extended Data Fig. 7g, h), which indicates their germinal centre origin 16 . Hence, bnt162b2 vaccination efficiently elicits a protective humoral immune response, composed of ASC and antigen-specific memory B cells that are mobilized to the periphery after boost.

Early memory CD8 + T cells
We compared vaccine-elicited spike-specific early memory CD8 + T cells  Fig. 8c). Possible reasons for this include differences in their MHCI binding and presentation characteristics (Extended Data Fig. 1c). In addition, we also observed differences in T cell memory subset distribution ( Fig. 4b and Extended Data Fig. 9a-c) of spike-specific early memory CD8 + T cells with higher fractions of more early differentiated subsets, for example, early differentiated (T ED ) and central memory (T CM ) T cells for A*01/S 865 -and A*02/S 269 -specific CD8 + T cells and transitional memory cells for A*03/S 378 -specific CD8 + T cells after natural infection (80-120 dps). By contrast, higher frequencies of effector memory 1 T cells (T EM1 ) were detectable after vaccination (80-120 dpb) (Fig. 4b). Spike-specific effector memory 2 and 3 T cells (T EM2 and T EM3 ) and terminally differentiated effector memory T cells that expressed CD45RA (T EMRA ) were hardly detectable in the circulation (Extended Data Fig. 9c). Of note, the memory subset distribution of A*03/S 378 -specific CD8 + T cells differed from A*01/S 865 -and A*02/ S 269 -specific CD8 + T cells with only a minor fraction of T ED and T CM cells targeting A*03/S 378 reflecting an overall further differentiation towards effector memory subsets ( Fig. 4b and Extended Data Fig. 9b). t-SNE analysis of concatenated expression data further supports qualitative differences of spike-specific CD8 + T cells obtained from the early memory phase (80-120 dpb/dps) after vaccination compared to natural infection being less pronounced for A*03/S 378 -specific CD8 + T cells (Fig. 4c). For A*01/S 865 -specific CD8 + T cells we also observed higher expression of TCF-1 and BCL-2 after natural infection (Extended Data Fig. 10a, b). Both t-SNE analysis and manual gating demonstrated a higher and prolonged CD38 expression on spike-specific CD8 + T cells after natural infection ( Fig. 4c and Extended Data Fig. 10c). However, vaccine-and natural infection-associated expansion capacity and cytokine production of spike-specific CD8 + T cells were similar ( Fig. 4d and Extended Data Fig. 10d, e). Hence, compared with natural infection, vaccine-associated spike-specific early memory CD8 + T cell populations exhibit similar functional capacities but a different subset distribution.

Discussion
In summary, a robust, stable and fully functional spike-specific CD8 + T cell response is elicited already after prime vaccination at a time point when neutralizing antibodies were hardly detectable and coincides with the protective effect observed for mRNA vaccines that starts at 10-12 dpp 2,3 . In contrast to CD8 + T cells, peak mobilization of neutralizing antibodies and antigen-specific B cells to the periphery was first detectable after boost. This is in line with previous reports 4,7,11,12 and most probably represents maturation of the response in secondary lymphoid organs 17 with subsequent release to the circulation. After boost, highly cross-neutralizing antibodies are present in the sera, clearly adding a major protective effector mechanism on top of the early-mobilized spike-specific CD8 + T cell response. The humoral and CD8 + T cell response are potentially coordinated by early elicited spike-reactive CD4 + T cells 8 that underwent a limited boost expansion after second dose mRNA vaccination supporting their coordinating role. Fully functional vaccine-elicited early memory CD8 + T cells patrol the periphery for SARS-CoV-2 at least within the first months. The functional capacity of spike-specific early memory CD8 + T cells is similar after vaccination and natural infection up to 3-4 months after boost or symptom onset. Compared with natural infection, however, the early memory pool of spike-specific CD8 + T cells after vaccination exhibits a different memory T cell subset distribution that may affect long-term maintenance characteristics 18 . This difference may be caused by differential duration and location of antigen contact and different inflammatory responses after vaccination versus infection 19,20 , as indicated by a lower CD38 expression on early memory spike-specific CD8 + T cells after vaccination compared with natural infection 4,21 . Follow-up studies including larger cohorts of vaccinees and SARS-CoV-2 convalescent individuals are clearly required to assess longevity of CD8 + T cell immunity. Our study was limited to circulating spike-specific adaptive immunity, and did not address local immunity at the viral entry site, the respiratory tract. However, our data provide insights into the protective mechanisms that underlie bnt162b2 vaccination with implications for the development of vaccination strategies against emerging pathogens and cancer.

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-021-03841-4. Spike proteins of human coronaviruses were aligned according to their homology (amino acid level). Analysed spike SARS-CoV-2 epitopes were then mapped to the corresponding protein alignment. Correspondingly mutation analyses were performed with the spike protein of VOC alpha, beta, gamma and delta.

In vitro expansion and intracellular IFNγ staining with overlapping peptides
A total of 182 overlapping peptides that spanned the SARS-CoV-2 spike sequence (Gene Bank Accession code MN908947.3) were synthesized as 18-mers overlapping by 11 amino acids with a free amine NH 2 terminus and a free acid COOH terminus with standard Fmoc chemistry and a purity of >70% (Genaxxon Bioscience). In vitro expansion with OLPs was performed as follows: 20% of the PBMCs were stimulated with a pool of all 181 SARS-CoV-2 spike OLPs (10 μg ml −1 ) for 1 h at 37 °C, washed and co-cultured with the remaining PBMCs in RPMI medium supplemented 20 U ml −1 with recombinant IL-2. On day 10, intracellular IFNγ staining was performed with pooled OLPs (45 pools with 4 OLP each). Therefore, cells were re-stimulated with OLP pools (50 μM), DMSO as negative control or PMA and ionomycin as positive control in the presence of brefeldin A an IL-2. After 5 h of incubation at 37 °C, cells were stained for surface markers (CD8 + , CD4 + ; Viaprobe) and intracellular markers (IFNγ). Subsequently, on day 12 the single overlapping peptides of positive pools were tested by intracellular cytokine staining. Viral amino acid sequences of positive individual OLPs were analysed for pre-described minimal epitopes or the best HLA-matched predicted candidate using the Immune Epitope Database website (using two prediction algorithms ANN 4.0 and NetMHCpan EL 4.1 23 for 8-mer, 9-mer and 10-mer peptides with half-maximal inhibitory concentration (IC 50 ) of <500 nM).

Peptides and tetramers for T cell analysis
Peptides were synthesized with an unmodified N terminus and an amidated C terminus with standard Fmoc chemistry and a purity of >70% (Genaxxon Bioscience). Peptide was loaded on HLA class I easYmers (immunAware) according to manufacturer's instructions (A*01/ S 865 LTDEMIAQY, A*02/S 269 YLQPRTFLL and A*03/S 378 KCYGVSPTK). SARS-CoV-2 peptide-loaded HLA class I tetramers were produced by conjugation of biotinylated peptide-loaded HLA class I easYmers with phycoerythrin (PE)-conjugated streptavidin (Agilent) according to the manufacturer's instructions. A SARS-CoV-2-specific HLA class II custom tetramer (DRB1*15:01/S 236 TRFQTLLALHRSYLT) was obtained from (MBL).

Magnetic bead-based enrichment of spike-specific CD8 + T cells
Spike-specific CD8 + T cells were enriched as previously described 25 . In brief, 1 × 10 7 -2 × 10 7 PBMCs (with an average of 15.7% CD8 + T cells) were labelled with PE-coupled peptide-loaded HLA class I tetramers for 30 min. Enrichment was then performed using anti-PE beads with MACS technology (Miltenyi Biotec) according to the manufacturer's instructions. Subsequently, enriched spike-specific CD8 + T cells were analysed by multiparametric flow cytometry and frequencies of spike-specific CD8 + T cells were calculated as described before 25 . Only enriched samples with ≥5 spike-specific CD8 T cells were included in further analyses, resulting in a detection limit of 5 × 10 −6 .

Magnetic bead-based enrichment of spike-specific CD4 + T cells
Enrichment of spike-specific CD4 + T cells was adapted from the method described previously 25 . In brief, 1.5 × 10 7 -2 × 10 7 PBMCs of DRB1*15:01-positive donors were labelled with PE-coupled peptide-loaded MHC class II tetramers for 40 min. Then, 5 μl was taken from 1,000 μl pre-enriched sample (1:200) and used for subsequent flow cytometric staining. Subsequent enrichment was performed with anti-PE beads using MACS technology (Miltenyi Biotec) according to the manufacturer's protocol. Enriched spike-specific CD4 + T cells and the pre-enriched sample were used for flow cytometric staining. The complete pre-enriched and enriched samples were recorded. Only enriched samples with ≥5 spike-specific CD4 + T cells were included in further analyses. The frequency of spike-specific CD4 + T cells was calculated as follows: Absolute number of spike-specific CD4 + T cells (enriched sample) divided by the absolute number of CD4 + T cells (pre-enriched sample) × 200. The detection limit as a frequency was calculated as follows: 5 spike-specific CD4 + T cells (enriched sample) divided by the mean number of CD4 + T cells (pre-enriched sample) throughout all tested donors × 200.

Dimensional reduction of multiparametric flow cytometry data
Dimensionality reduction of multiparametric flow cytometry data was done with R version 4.0.2 using the Bioconductor (release (3.11)) CATA-LYST package23. The analyses were performed on gated virus-specific CD8+ T cells including the markers CD69, CD45RA, BCL-2, PD1, CD25, Ki-67, TCF-1, EOMES, CCR7, T-BET, TOX and CD38. Downsampling of cells to 100 or 200 cells (t-SNE or diffusion maps) was performed before dimensionality reduction to facilitate the visualization of different samples. Marker intensities were transformed by arcsinh (inverse hyperbolic sine) with a cofactor of 150. Dimensionality reduction on the transformed data was achieved by t-SNE and diffusion map visualization.

S1-and RBD-tetramerization for B cell analysis
A biotinylated form of recombinant S1 and RBD proteins (BioLegend) were tetramerized by addition of PE-conjugated or BV421-conjugated streptavidin (BioLegend) and used for B cell tetramer staining assays. In brief, streptavidin-PE or streptavidin-BV21 was added in an amount that equals one-fifth of the monomer substrate amount. The streptavidin was added in five equal portions to the monomer and incubated each time at 4 °C for 20 min on a shaker. The tetramers were filled up to 100 μl with 0.1% BSA in PBS and stored at 4 °C.

In vitro PBMCs activation and ELISA
PBMCs of vaccinated individuals and patients with a history of SARS-CoV-2 infection were plated at 0.5 × 10 6 cells ml −1 and polyclonally stimulated for 9 days with thiol-modified CpG (0.25 μM, TCGTCGTTTT-GTCGTTTTGTCGTT) and hIL-2 (100 ng/ml, Immunotools). At day 9, the supernatants of the in vitro culture were cleared from debris by centrifugation and used to determine the presence of SARS-CoV-2 spike-specific IgG antibodies (Anti-SARS-CoV-2-QuantiVac-ELISA (IgG), Euroimmun) according to the manufacturer's instructions. To detect S1 specific IgM, supernatant of the in vitro culture and serum of vaccinated individuals was incubated on a S1 pre-coated plate (Anti-SARS-COV-2, Euroimmun). Bound IgM was detected with alkaline phosphatase-conjugated anti-human IgM ( Jackson ImmunoResearch), and developed with p-nitrophenyl phosphate (Sigma-Aldrich) in DEA buffer.

Neutralization assay
Samples of vaccinated and convalescent individuals were tested in a plaque reduction neutralization assay. In brief, VeroE6 cells were seeded in 12-well plates at a density of 2.8 × 10 5 cells per well 24 h before infection. Serum samples were diluted at ratios of 1:16, 1:32, 1:64, 1:128, 1:256, 1:512 and 1:1,024 in 50 μl PBS total volume. For each sample, one negative control was included (PBS without serum). Diluted sera and negative controls were subsequently mixed with 90 plaque-forming units (PFU) of authentic SARS-CoV-2 (either B.1, alpha or beta variant) in 50 μl PBS (1,600 PFU ml −1 ) resulting in final sera dilution ratios of 1:32, 1:64, 1:128, 1:256, 1:512, 1:1,024 and 1:2,048. After incubation at room temperature for 1 h, 400 μl PBS was added to each sample and the mixture was subsequently used to infect VeroE6 cells. After 1.5 h of incubation at room temperature, inoculum was removed and the cells were overlaid with 0.6% Oxoid-agar in DMEM, 20 mM HEPES (pH 7.4), 0.1% NaHCO 3 , 1% BSA and 0.01% DEAE-Dextran. Cells were fixed 72 h after infection using 4% formaldehyde for 30 min and stained with 1% crystal violet upon removal of the agar overlay. PFU were counted manually. Plaques counted for serum-treated wells were compared to the average number of plaques in the untreated negative controls, which were set to 100%. The PRNT 50 value was calculated using a linear regression model in GraphPad Prism 9 (GraphPad Prism Software).

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.

Code availability
R code to reproduce the analyses of multiparametric flow-cytometry data are available at https://github.com/sagar161286/SARSCoV2_spe-cific_CD8_Tcells.

Reporting Summary
Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend,

Software and code
Policy information about availability of computer code Data collection All software used to perform data collection are described in the methods section of the manuscript or the supportive information. Data Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A list of figures that have associated raw data -A description of any restrictions on data availability Raw data in this study are provided in the Source data. Additional supporting data are available from the corresponding authors upon reasonable request. All requests for raw and analyzed data and materials will be reviewed by the corresponding authors to verify if the request is subject to any intellectual property or confidentiality obligations. Patient-related data not included in the paper were generated as part of clinical examination and may be subject to patient confidentiality. Any data that can be shared will be released via a Material Transfer Agreement. Reference viral sequences 229E (NC_002645), HKU1 (NC_006577), NL63 (NC_005831), OC43 (NC_006213), MERS (NC_019843), SARS-CoV-1 (NC_004718) and SARS-CoV-2 (MN908947.3) were downloaded from the NCBI database (https://www.ncbi.nlm.nih.gov/).

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design
All studies must disclose on these points even when the disclosure is negative.

Sample size
Patients were recruited and patient material was banked at the University Hospital Freiburg; inclusion criteria were: (1) 32 health care workers that received a prime and boost vaccination with the mRNA vaccine bnt162b2/Comirnaty, (2) 59 acutely infected and convalescent individuals following a mild course of SARS-CoV-2 infection, SARS-CoV-2 infection was confirmed by positive PCR testing from oropharyngeal swab and/or SARS-CoV-2 spike IgG positive antibody testing, (3) 2 convalescent health care workers following a mild course of SARS-CoV-2 infection that received a single dose of bnt162b2/Comirnaty and (4) 8 age and sex-matched historic controls. No sample size calculations were performed. 32 vaccinatedhealth care workers gave informed consent and were available to donate blood samples. Therefore, similar numbers of COVID-19 convalescents were selected.
Data exclusions For flow cytometrical analysis, cell populations containing less than 5 cells were excluded. This data exclusion strategy has been applied and validated previously by our group to gain reproducible results in studies investigating virus-specific CD8+ T cells in human viral infections.
April 2020 controls 37,6 years. The gender ratio of vaccinated donors was m/f: 19/13, donors with a history of natural SARS-CoV-2 infection was m/f: 31/28, donors with a history of natural SARS-CoV-2 vaccination and a single vaccination was m/f: 1/1, of historic controls m/f: 5/3.

Recruitment
Vaccinated donors as well as SARS-CoV-2-infected and SARS-CoV-2-convalescent patients were recruited at the University Hospital Freiburg (in-and outpatient section); self-selection bias or other biases can be excluded since several people were included in the recruitment. Samples were banked and retrospectively selected according to the following inclusion criteria: HLA-A*01:01, -A*02:01, -A*03:01, DRB1*15:01. Banked samples from sex-, age-and HLA-matched historic controls were retrospectively selected.

Ethics oversight
Written informed consent was obtained from all participants and the study was conducted according to federal guidelines, local ethics committee regulations (Albert-Ludwigs-Universität, Freiburg, Germany; vote #: 322/20, #21-1135 and 315/20) and the Declaration of Helsinki (1975).
Note that full information on the approval of the study protocol must also be provided in the manuscript.

Flow Cytometry
Plots Confirm that: The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
All plots are contour plots with outliers or pseudocolor plots.
A numerical value for number of cells or percentage (with statistics) is provided.

Methodology
Sample preparation Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.