Twenty COVID-19 patients were recruited at Xiyuan and Renmin Hospitals in Shiyan City, Hubei province, China. Ten healthy donors were recruited at Renmin Hospital in Shiyan City. The protocol for this study has received approval from the Clinical Ethics Committee of Hubei University of Medicine (No. 2020-TH-017). COVID-19 patients and healthy donors were all free from tumors, other infections, and diseases. All individuals had given their informed consent to participate in this study. The median ages of COVID-19 patients and healthy subjects were 42.8 and 39.3, respectively. The sex ratio for males and females is 12/8 in COVID-19 patients and 6/4 in healthy donors.

Total RNA was extracted separately from the peripheral blood of COVID-19 patients and healthy donors. One microgram of total RNA was then converted into cDNA using a reverse transcription system. Primer sequences complementary to upstream V regions and downstream C regions were used to amplify the CDR3 regions. The primer sequences were listed in Table 1.

The purified PCR products were ligated into pGEM-T easy vector (Invitrogen, Carlsbad, CA, United States) and sequenced by using T7 primer (Sangon Biotech Inc., Shanghai, China). The CDR3γ region was considered to contain conserved “CALW” at its N-terminus and conserved “KVFG” at its C-terminus. While CDR3δ region was considered to contain conserved “CA” at its N-terminus and conserved “FGXG” at its C-terminus.

The SARS-CoV-2 specific CDR3 sequences were separately inserted into full-length γ9 and δ2 chains to substitute their original CDR3 sequences based on our previous report (Xi et al., 2011b). The obtained γ9 and δ2 chains were then inserted into pREP7 and pREP9 vectors (Figure 1A), respectively. Full-length pREP7-γ9 and pREP9-δ2 chains were co-transfected into J.RT3-T3.5 cells. After 48 h, the transfected cells were cultured in a selection medium with hygromycin and neomycin for 4 weeks. The expression of transfected γδTCR in the cells was then evaluated by flow cytometry.

The transfected cells expressing potential SARS-CoV-2 specific γδTCR were used as probe cells to perform subtractive screening in a 12-peptide phage-display library. Four rounds of screening with conditions such as increased Tween concentration, increased action time with control cells as well as decreased action time with SARS-CoV-2 specific γδTCR transfected cells were conducted in order to enrich epitope peptides that could specifically bind with SARS-CoV-2 specific γδTCR transfected cells.

Sangon Biotech Inc. synthesized peptides with a purity of more than 95% as determined by high-performance liquid chromatography analysis. Half of the synthesized peptides were linked with FITC at their N-terminals.

Cells were incubated with FITC-conjugated peptide or control peptide for 30 min at 4°C. The cells were then analyzed by flow cytometry on a MoFlo XDP flow cytometer (Beckman Coulter, Fullerton, CA, United States).

γδT cells were isolated from healthy donors’ peripheral blood mononuclear cells (PBMCs) using an anti-TCR γ/δ MicroBead Kit from Miltenyi company (130-050-701) according to the manufacturer’s instructions.

The transfected cells and sorted γδT cells were separately incubated with 10 ng/mL NSP8 protein (Sino Biological Inc., Beijing, China) or control protein for 30 min at room temperature. After extensive washing with RPMI-1640 culture medium, the transfected cells and natural γδT cells were then plated into 24-well plates at 1 × 10⁶ cells per well. The supernatants were harvested after 24 h and the level of IL-2 was detected by using Human IL-2 ELISA Kit (BD Biosciences, San Jose, CA, United States) according to the manufacture’s instructions.

The homologous analysis and sequence alignment were performed by using the Basic Local Alignment Search Tool (BLAST) to identify the matched proteins. After the screening, the obtained epitope peptide candidates were analyzed on the Heliquest website.² The sequence alignment between peptide candidates and the downloaded SARS-CoV-2 ORF1ab sequence was performed by using DNAMAN8 software.

Statistical comparisons between the experiment group and control group were performed by using the Student’s t-test. All data were analyzed either by SPSS 19.0 software or by GraphPad 8.0 software. P < 0.05 was considered statistically significant.

The specificity in antigen recognition by TCR is primarily determined by the sequences within CDR3 region that are highly diverse. We isolated peripheral γδT cells from the patients infected by SARS-CoV-2 viruses and analyzed both Vγ9 CDR3 and Vδ2 TCR regions in comparison to the sequences derived from healthy donors. There was no significant variation in Vγ9 CDR3 region identified between the infected and control groups (Table 2). However, in Vδ2 region, we found a CDR3 sequence specifically present in most of the infected cases (Table 3).

We amplified the sequences encoding γδTCRs derived from either infected patients or healthy individuals. The γ9 sequence (CALWEVISELGKKIKVFG) was identical between the two groups, whereas the δ2 sequences were different, with CACDPLLGDASYTDKLIFGKG from COVID-19 patients and CACDRLGDTGTDKLIFGKG from healthy individuals. We then established the expressions of full length γ9 and δ2 chains via introducing the designated vectors into J.RT3-T3.5 cells by electroporation (see more details in section “Materials and Methods”). After 4 weeks of selection with hygromycin and neomycin, the SARS-CoV-2-specific γδTCRs lines established were verified by both PCR (Figure 1B) and flow cytometry (Figure 1C).

Next, we used this cell model to screen potential epitopes recognized by this γδTCR based on a 12-mer random peptide phage-display library (E8110S, New England Biolabs, Hitchin, United Kingdom) (Xi et al., 2011b). There were 20 positive clones obtained as indicated by the phage-ELISA. Sequences derived from these clones were sequenced and gave rise to the dominant epitope candidates (SP1 to SP8) (Table 4).

We used the SARS-CoV-2-specific γδTCR cell model to verify the binding between the epitopes identified and γδTCR. IL-2 secretion was monitored by ELISA in the cells upon the stimulation with individual epitope peptides. Among three representative epitopes SP1, SP2, and SP3 highlighted in Figure 2A (The predicted spiral structures were obtained using bioinformatics analysis tools proved by Heliquest website.), SP1 and SP2 triggered significant IL-2 production in the cells (Figure 2B) (P < 0.05). FACS analysis using FITC-conjugated peptides also confirmed that SP1 and SP2 exhibited strong affinity toward the SARS-CoV-2-specific γδTCR cells (Figure 2C).

A BLAST search was performed to identify SARS-CoV-2 proteins that contain SP1 and SP2 epitopes (Table 5). The top hits were located in ORF1ab region that encodes non-structural polyproteins involved in virus assembly, transcription, and replication. Further analysis using DNAMAN8 software revealed NSP8, among the ORF1ab region-derived polypeptides, as the origin of these γδTCR-specific epitopes (Figure 3A). NSP8 stimulates the production of IL-2 in the SARS-CoV-2-specific γδTCR cells, which were evident at both transcriptional (Real-time PCR, Figure 3B) and translational (ELISA, Figure 3C) levels. Furthermore, INF-γ production has been linked to γδT cell activation (Xi et al., 2011b,2013, 2021). Our findings also demonstrated that NSP8 could activate peripheral γδT cells isolated from healthy donors and increase INF-γ secretion in these cells, implying that NSP8 could bind to natural γδT cells (Figure 3D).