Convalescent serum samples were collected from 138 (38 ± 11.9 years old) COVID-19 patients (50 males and 102 females) and 39 age-matched controls (37.7 ± 14.3 years old, 18 males and 21 females) between March and December 2020. Furthermore, acute serum samples were collected from a separate group of 14 COVID-19 patients (55.6 ± 6.2 years old; 5 males and 9 females). Clinical characteristics were also collected for these patients (Table 1). The diagnosis of SARS-CoV-2 infection was established based on clinical presentation and was confirmed by qPCR. Convalescent serum samples were collected at various time points post-infection (0–12 months) following standard operating procedures in the hospital for the diagnosis of COVID-19 infection and stored at −80°C. Additionally, 22 (10 male and 12 female) control serum samples collected in 2015–2016 were included into the study. Aliquots (100 μL) of these serum samples were stored at −80°C. All controls were tested for anti-SARS-CoV-2 antibodies and were confirmed sero-negative.

In addition to serum samples, convalescent blood samples from 17 COVID-19 patients (35.9 ± 14.6 years old, 6 male and 11 female) as well as 9 controls (36.7 ± 14.5 years old; 5 male and 4 female) were collected.

The Ethics Committee of the Kazan Federal University approved this study, and signed informed consent was obtained from each patient and control subjects according to the guidelines adopted under this protocol (Protocol 4/09 of the meeting of the Ethics Committee of the KSMA dated September 26, 2019). Sample collection in 2015–2016 was done according to the protocol approved by the Institutional Review Board of the Kazan Federal University and informed consent was obtained from each respective subject according to the guidelines approved under this protocol (Article 20, Federal Law “Protection of Health Right of Citizens of Russian Federation” N323-FZ, 11.21.2011).

Multiple S and N protein peptides for SARS-CoV-2 as well as eCoV NL63, OC43, HKU1 and 229E were synthesized by GenScript (Jiangsu, China). Amino acid (aa) sequences of SARS-CoV-2 S and N protein peptides were selected based on published data (Ahmed et al., 2020; Campbell et al., 2020; Grifoni et al., 2020a; Noorimotlagh et al., 2020). Also, S and N protein peptides for eCoVs were selected using the iedb.org platform (Vita et al., 2019).

SARS-CoV-2 S and N protein peptide aa sequences (purity > 95%) are summarized in Figure 1 and Supplementary Table 1. eCoV S and N proteins peptide aa sequences (purity > 95%) are summarized in Supplementary Table 2.

Alignment of S and N peptide aa sequences from SARS-CoV-2 with each eCoV [NL63, 229E, OC43, and HKU1 (GenBank Accession nos. YP_009724390, AKT07952, AWN62679, APU51916, YP_173238 for S protein and YP_009724397, YP_003771, NP_073556, YP_009555245, YP_173242 for N protein, respectively)] was done using the MegAlign program (Clustal W algorithm), DNASTAR software package Lasergene v. 7.1.0.44 (DNASTAR, Inc., United States) (Tamura et al., 2013). Parameters were adjusted manually.

The SARS-CoV-2-IgG IFA Best diagnostic ELISA kit (Vektor Best Agidel, Ufa, Russia) was used to determine SARS-CoV-2-specific antibody titers as per manufacturer’s instructions. Briefly, control or COVID-19 convalescent serum was diluted 1:100 with PBS and incubated for 60 min at 37°C in a 96-well plate with pre-adsorbed SARS-CoV-2 antigens. Following washes (3×; 0.5% Tween20 in PBS, PBS-T), wells were incubated with anti-human-IgG-HRP conjugated antibodies (1:10,000 in PBS-T, American Qualex Technologies, United States) for 30 min at 37°C. Post-incubation and washes (3×; PBS-T), wells were incubated with 3,3′,5,5′ Tetramethylbenzidine (Chema Medica, Moscow, Russia). The reaction was stopped by adding an equal amount of 10% phosphoric acid (TatKhimProduct, Kazan, Russia). Data were measured using a Tecan 200 microplate reader (Tecan, Switzerland) at OD₄₅₀ with reference OD₆₅₀. OD₄₅₀ values higher than 0.5 were considered positive results, according to the manufacturer’s protocol.

Several peptides were selected for analysis of reactivity with SARS-CoV-2 infected patient as well as control subject sera. Each peptide (1 μg/100 μL) was adsorbed on a 384-well plate at 4°C for 18 h; plates were washed and incubated with serum samples (1:100; 50 μL) at 4°C for 18 h. Following washes (3×; PBS-T), wells were incubated with anti-human-IgG-HRP conjugated antibodies (1:10,000 in PBS-T, American Qualex Technologies, United States) for 30 min at 37°C. Washed (3×; PBS-T), wells were incubated with 3,3′,5,5′ Tetramethylbenzidine (Chema Medica, Moscow, Russia). The reaction was stopped by adding an equal amount of 10% phosphoric acid (TatKhimProduct, Kazan, Russia). Data were captured using a microplate reader Tecan 200 (Tecan, Switzerland) at OD₄₅₀ with reference OD₆₅₀.

Peripheral blood mononuclear cells (PBMCs) were collected from COVID-19 convalescent individuals and used for ELISpot analysis. Leukocytes (10⁵ cells/well) were added to anti–human IFN-γ mAb (Abcam, United Kingdom) pre-coated 96-well plates in RPMI-1640 medium supplemented with 10% FBS (HyClone, South America), 2 mM L-glutamine (PanEco, Moscow, Russia) and 1% mixture of antibiotics penicillin-streptomycin (PanEco, Moscow, Russia). Peptides were added at 1 μg/well and plates incubated for 20 h (37°C, 5% CO₂). Biotinylated mouse anti–human IFN-γ detection antibodies (clone 4S.B3; Abcam) were added for 90 min at room temperature followed by streptavidin alkaline phosphatase (1:1,000 dilution) for 60 min at room temperature. BCIP/NBT substrate solution [5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitro blue tetrazolium (NBT)] (Abcam, United Kingdom) was added for 25 min at room temperature. PBMCs incubated with phytohemagglutinin (PHA; 2.5 μg/ml) served as a positive control, while unstimulated leukocytes as a negative control. The number of spots in negative control wells (range of 0–5 spots) was subtracted from the number of spots in stimulated wells. Peptide stimulations were done in duplicate wells.

Statistical analysis was performed in the R environment (Dessau and Pipper, 2008). Statistically significant differences between comparison groups were accepted as p < 0.05, assessed by the Kruskal-Wallis test with Benjamini-Hochberg adjustment for multiple comparisons. Correlations were analyzed using the R psych package (based on Spearman’s rank correlation coefficient, p-values were adjusted with the Benjamini-Hochberg method).

A total of 152 samples (50 male and 102 female) were collected from COVID-19 acute and convalescent patients with an average age of 38.0 ± 11.9 years old (Table 1). Diagnosis of COVID-19 was established based on epidemiological anamnesis, clinical presentation and confirmed by qPCR. Clinical disease was diagnosed as mild (105 cases), moderate (43 cases) or severe (4 cases). Thirty-nine patients with moderate and 4 severe COVID-19 patients were hospitalized. All mild and four moderate COVID-19 patients did not require hospitalization. Mild COVID-19 was characterized as fever ≤ 38°C, cough, and throat pain. Patients with moderate COVID-19 had fever > 38°C, extent of lung damage as evaluated by computer tomography (CT) at stage 1–2 and oxygen saturation levels < 95%. In severe COVID-19 cases, fever was > 38°C, extent of lung damage by CT evaluated to be at stage 3–4, oxygen saturation levels < 93%, arterial pressure < 90/60 mmHg and some patients required artificial ventilation (Ministry of Health of the Russian Federation, 2020). Lung damage was categorized as < 20%, 20–40% or > 40% in 117, 31 and 4 patients, respectively. Fever was reported in all patients (37.92 ± 0.66°C) with duration of 6.31 ± 4.04 days. Two out of four of the severe COVID-19 patients required artificial ventilation, while none were admitted to the ICU. Control samples were collected from individuals who did not have symptoms of upper respiratory tract infection, did not have contact with COVID-19 patients and had negative anti-SARS-CoV-2 ELISA antibody test results. Control samples collected in 2015–2016 were also tested for anti-SARS-CoV-2 reactivity using ELISA tests and were found to be negative.

Peptide aa sequences were selected based on published data (Coronaviridae Study Group of the International Committee on Taxonomy of Viruses, 2020; Oran and Topol, 2020; WHO, 2020a,b). We selected peptides previously identified on SARS-CoV-2 N and S proteins which are essential for virus replication and binding to the ACE2 receptor (Shah et al., 2020; Supplementary Table 1). The position of these peptides is summarized in Figure 1. Interestingly, many peptides were found in regions identified as containing immunogenic epitopes (Liang et al., 2005; Shrock et al., 2020) suggesting that these peptides could have reactivity with COVID-19 patient cohort serum. There are multiple mutations that have been identified in circulating SARS-CoV-2 viruses (Callaway, 2020; Lauring and Hodcroft, 2021), raising serious concerns about vaccine efficacy and currently used diagnostic tools (Williams and Burgers, 2021). To address the possibility that some of the mutations are located in selected peptides, we have analyzed the location of mutations in currently circulating (release date: 1 July, 2021 to 16 August, 2021) nine delta strains and one reference strain (Supplementary Table 3) sequences for S and N protein of SARS-CoV-2. Mutation in S and N protein was assessed by taking reference sequences YP_009724390.1 and YP_009724397.2, respectively. Mutations in codons coding for sixteen and eight unique amino acids in S and N protein, respectively, were identified (Supplementary Table 4). Most of these mutations are outside of the S and N peptides selected for this study. Among mutations located in selected peptides, there were two S1 (D950N) and S22 (L1063F) located in S protein. Moreover, total of five [N2 (D63G), N18 (G215C), N10 and N5 (T362I), N8 (T362I, D377Y) and N12 (D377Y and R385K)], N protein peptides were found to contain mutations in currently circulating delta strains of SARS-CoV-2. These S and N peptides were found to have mostly a single mutation, which could have limited effect on their immunogenicity.

The reactivity of all 152 COVID-19 acute and convalescent serum samples was tested using B-cell peptides summarized in Figure 1 and Supplementary Table 1. Several S protein peptides (S1, S7, and S18) and the N6 peptide of N protein had significantly higher reactivity with convalescent serum compared to COVID-19 negative controls collected in year 2020 (Figure 2A). However, when compared with control samples collected in 2015, there were more SARS-CoV-2 peptides (S1, S2, S3, S5, S7, S8, S9, S10, S15, S17, S21, N1, N2, N3, N4, N6, N7, N9, N11, and N16) with higher reactivity with COVID-19 serum (Figure 2B). This observed difference between controls from 2015 and 2020 was due to lower reactivity of 2015 controls with SARS-CoV-2 peptides compared to 2020 controls. Similarly, more peptides were found to be reactive with acute COVID-19 serum when compared to 2015 in contrast to when compared to 2020 controls (S1, S7, and S10 peptides vs. S7 peptide) (Figure 2C). This could be explained by the priming of immune responses by endemic coronaviruses and thus the presence of cross-reacting antibodies in 2020 controls. This is further supported by the confirmation of all controls in this study having no previous exposure to SARS-CoV-2. Also, the effect of long-term storage on stability of serum content could contribute to variation in reactivity between 2015 and 2020 controls. Nevertheless, S1, S7, and N6 COVID-19 peptides had consistently high reactivity with patient serum as compared to controls from both 2015 and 2020, suggesting that these peptides could complement current diagnostic methods and could have therapeutic potential.

Studies have demonstrated that serum antibody levels declined in COVID-19 patients 6 months after recovery (Self et al., 2020; Seow et al., 2020). Therefore, we sought to determine reactivity of peptides with serum samples collected at different times after recovery (Figure 3A). Serum samples were separated into three groups: ≤3 months, 4–6 months and ≥ 7 months. Peptides S1, S2, S7, S14, S19, and N6 reacted significantly higher with serum samples collected ≤ 3 months after recovery when compared to 2020 control samples. Most of these peptides remained reactive 4–6 months post-recovery (S1, S2, S7, S14, S18, and N6). Reactivity with half of these peptides was lost later (≥ 7 months), when only S7, S18, and N6 peptides remained highly reactive in COVID-19 convalescent sera when compared to controls. These data demonstrate that reactivity to immunogenic epitopes is most versatile during early convalescence but declines with time after infection.

Next, we analyzed the dynamics of COVID-19 serum reactivity with SARS-CoV-2 peptides (Figure 3B). It appears that reactivity to some peptides (S1, S2 and S14, S19) was reduced few months after recovery. In contrast, reactivity to other peptides (S7, S18, and N6) remained high. Interestingly, reactivity to N6 was increasing with time post-recovery. These data suggest that antibody reactivity declines only to selected epitopes, not to all epitopes initially recognized.

It has been demonstrated that males may have more severe symptoms of COVID-19 and are at a higher risk of fatal outcome when compared to females (Kang and Jung, 2020; Tehrani et al., 2021). Therefore, we sought to determine whether serum reactivity to SARS-CoV-2 peptides differs depending on the sex of patients (Figure 4). We observed an increased serum reactivity with peptides S1, S7, S18, and N6 in male COVID-19 patients when compared with male controls. In female patients, serum reactivity was only higher for peptides S7 and N6 when compared to female controls (Figure 4). This highlights a difference in immune reactivity between male and female COVID-19 patients. A comparison of male and female COVID-19 samples showed an increased reactivity to N6 in female than in male COVID-19 patients.

Multiple studies have demonstrated that COVID-19 severity and fatality rates are increased in older patients (Wong et al., 2004; Li et al., 2020). We aimed to determine whether this age-related severity of COVID-19 could be explained by differences in antibody responses. COVID-19 convalescent samples were divided into younger (<45 years old) and older (≥45) groups (Figure 5) and used to analyze their reactivity to S1, S7, S18, and N6 peptides compared to corresponding controls. Interestingly, reactivity of younger patient serum to S1 was higher compared to that in older COVID-19 patients (Figure 5A). Also, the reactivity to peptides S7 and N6 was higher in both COVID-19 age groups compared to corresponding controls. However, only those in the <45 age group showed higher reactivity to peptides S1, and S18 compared to the same age control (Figure 5A). It appears that younger COVID-19 patients have higher reactivity with more peptides (S1, S7, S18, and N6) compared to older patients (S7 and N6). These data suggest that younger individuals having more COVID-19 reacting epitopes on S and N proteins is a contributing factor to age-related disease pathogenesis.

Next, we analyzed the age dynamics of antibody reactivity with S1, S7, S18, and N6 peptides, which demonstrated age-dependent difference (Figure 5B). This data clearly demonstrates that reactivity with peptides is higher in younger individuals. In contrast, with age, reactivity declines and, when over 50 years old, it becomes at the level of control. These data provide strong indication of more robust antibody response in younger as compared to that in older COVID-19 patients.

We next sought to determine whether there is a correlation between serum reactivity with SARS-CoV-2 peptides as well as the clinical and demographic characteristics of the COVID-19 patients (Figure 6). Severity of COVID-19 positively correlated with increasing age, lung damage, high fever and duration of fever. There was also positive correlation between prolonged fever and lung damage (Figure 6). These data support previous findings and clinical observations in which older patients have more severe course of disease and higher mortality rate (CDC, 2021b; Sagar et al., 2021).

All peptides that demonstrated reactivity with COVID-19 convalescent serum were used for correlation analysis. Only S1 and S18 were identified as having statistically significant correlation with clinical parameters. Analysis of peptide reactivity and correlation with clinical presentation revealed significant negative correlation between severity and reactivity with S1 (Figure 6). We also observed that fever positively correlated with S18. Therefore, we suggest that negative correlation between reactivity with S1 peptide and severity of the disease could reflect its protective role in disease pathogenesis. Interestingly, S1 is located in HR1 region of S2 domain, which is also shown to be a potential target for novel anti-SARS-CoV-2 therapeutics (Liu et al., 2004).

Activation of T-cell immune responses during COVID-19 convalescence was analyzed using peptides containing T-cell epitopes (Figure 1 and Supplementary Table 1). Initially, peptides were screened for the most reactive and those that produced a minimum of two-fold increase compared with controls were selected for further study. Therefore, peptides having T cell epitopes and showing significant reactivity in preliminary screening (S4, S6, S15, N6, N10, and N19) were used in further analysis.

Blood samples were collected from 17 COVID-19 patients and 9 controls. IFN-γ production by SARS-CoV-2 peptide-exposed T-cells was detected using the ELISpot method. We found that when activation of T-cells from all patients was analyzed independently of age, sex and time post-infection, reactivity to S6, N6, and N19 was significantly higher in COVID-19 patients when compared to controls (Figure 7A).

When ELISpot data was analyzed based on sex of the patient, we found no difference between male and female T-cell reactivity to SARS-CoV-2 T-cell peptides (Supplementary Table 5). When ELISpot data was analyzed by separating into two age groups (<45 and ≥45 years old), reactivity to S4 and S15 was higher in older compared to younger patients (Figure 7B).

Studies have demonstrated cross reactivity between SARS-CoV-2 proteins and serum from non-SARS-CoV-2 infected patients (Kuxdorf-Alkirata et al., 2019; Ma Z. et al., 2020; Xie et al., 2020). This cross reactivity could be explained by the identity in aa sequence between SARS-CoV-2 peptides used in this study and the corresponding aa sequence in the same location in eCoV proteins (Supplementary Tables 6, 7). Silvanovitch et al. have demonstrated that immunological cross reactivity can occur when protein sequences are 70% conserved and higher (Graichen, 2021). Therefore, peptides with similarity greater than 70% between eCoVs and SARS-CoV-2 could contribute to immune cross reactivity. In this respect, ≥70% identity between SARS-CoV-2 S protein peptides (S1, S6, S8, S14, and S15) and the corresponding location on HKU1 and OC43 beta-coronaviruses (Table 2) suggests a contribution of these eCoVs to cross reactivity in COVID-19 sera. Interestingly, it appears that the S protein could potentially be the major contributor as none of the N protein peptides had ≥70% identity between SARS-CoV-2 and eCoVs. In contrast to beta-coronaviruses, only S protein peptides of NL63.S1 alpha-coronavirus had a 70% similarity (Supplementary Tables 8, 9), inferring a likely contribution of non-beta-coronaviruses to cross reactivity.

The location of eCoVs S and N immunogenic peptides used in this study differs from that of SARS-CoV-2 peptides. Therefore, we analyzed the identity of aa sequences of S and N peptides of eCoVs to the aa sequence of SARS-CoV-2 peptides in the same location (Supplementary Tables 8, 9). We found that the highest degree of identity was only between the aa sequence of NL63.S1 and the corresponding sequence of S protein of SARS-CoV-2. Other eCoV peptide identity with SARS-CoV-2 was less than 65%.

Next, we sought to determine whether COVID-19 serum displayed cross-reactivity with eCoV peptides. ELISA analysis revealed mostly lack of difference in reactivity of COVID-19 patient serum with eCoV peptides when compared with 2020 control. The only difference was found in reactivity to HKU1-S1 peptide, though it was lower in COVID-19 serum when compared to the control (Supplementary Table 10). We next performed correlation analysis to determine if there was any relationship between eCoV peptide reactivity and severity of COVID-19 (Figure 8). While there was significant correlation between eCoV peptide reactivity and time since infection, there was no correlation with COVID-19 severity. This data suggests that prior exposure to eCoVs does not provide protection against severe forms of COVID-19.