key: cord-0935553-zes18u3q
authors: Felbinger, Nathaniel; Trudil, David; Loomis, Lawrence; Ascione, Richard; Siragusa, Gregory; Haba, Seiji; Rastogi, Shruti; Mucci, Aidan; Claycomb, Mark; Snowberger, Sebastian; Luke, Brian; Francesconi, Stephen; Tsang, Shirley
title: Epitope Mapping of SARS-CoV-2 Spike Protein Reveals Distinct Antibody Binding Activity of Vaccinated and Infected Individuals
date: 2022-05-19
journal: bioRxiv
DOI: 10.1101/2022.04.13.487697
sha: 01da6a9f520d6b99b50913333d57ba488dbcf9f2
doc_id: 935553
cord_uid: zes18u3q

Previous studies have attempted to characterize the antibody response of individuals to the SARS-CoV-2 virus on a linear peptide level by utilizing peptide microarrays. These studies have helped to identify epitopes that have potential to be used for diagnostic tests to identify infected individuals, however, the immunological responses of individuals who have received the currently available Moderna mRNA-1273 or Pfizer BNT162b2 mRNA vaccines have not been characterized. We aimed to identify linear peptides of the SARS-CoV-2 spike protein that elicited high IgG or IgA binding activity and to compare the immunoreactivity of infected individuals to those who received both doses of either vaccines by utilizing peptide microarrays. Our results revealed peptide epitopes of significant IgG binding among recently infected individuals. Some of these peptides are located near functional domains implicated in the high infectivity of SARS-CoV-2. Vaccinated individuals lacked these distinct markers despite overall binding activity being similar.

Emerging variants of SARS-CoV-2 continue to make up an increased proportion of previous infection or the general protective status of individuals. Viral load detected using PCR microarray approach to study individuals who received the two currently available mRNA vaccines, Moderna mRNA-1273 and Pfizer BNT162b2. It is to be expected that a peptide 78 microarray is capable of differentiating the immune response of vaccinated individuals by 79 analyzing the antibody binding to the targeted SARS-CoV-2 spike protein peptides.

Herein we report a study of IgG and IgA antibody reactivity of individuals vaccinated 81 with the Moderna or Pfizer vaccines against linear peptides. As the Moderna and Pfizer vaccines 82 contain the sequences for the SARS-CoV-2 spike protein, our study has focused mainly on the 83 spike protein region. Data herein identifies epitopes that have the potential to serve as markers to 84 assess SARS-CoV-2 protection in addition to gaining a general characterization of IgG and IgA 85 antibody response against SARS-CoV-2 in vaccinated individuals. An assessment of the 86 potential impact of mutations found in current prevalent variants within these identified epitopes 87 was also performed. These epitopes may be useful in the study of immune response development Table S1 . Demographic information for vaccinated and negative subjects can be found in 105 Supplemental Table S2 . Available symptom information for all infected samples is available in 106 Supplemental Table S3 . All purchased infected sera was inactivated with a 4.0% Triton X-100 107 treatment prior to their arrival at our facility. All samples were immediately aliquoted and stored 108 at -80° C upon arrival to our facility. sequences found in SARS-CoV-2 variants and a set of influenza and polio peptides which were utilized as positive controls.

Prior to performing the procedure, solutions of PBST wash buffer (Phosphate-buffered 118 saline with 0.05% Tween20, at 7.4 pH), and pH 7.4 1mM Tris dipping buffer were prepared. All 119 solutions were filtered using a 0.44-micron vacuum filter kit and pH was adjusted to 7.4 with the 120 addition of 3M HCl if necessary. Microarray slides were initially incubated in a solution of 121 0.05% PBST (7.4 pH) and 10% blocking buffer (Rockland Immunochemicals, Inc., Pottstown, 122 PA) for 15 minutes at 23°C. Microarray slides were aspirated and then blocked with blocking 123 buffer for 30 minutes at 23°C. After blocking, slides were again aspirated and probed with a 124 mixture of fluorescent secondary antibodies that would be used for peptide binding detection 

After performing three PBST washes of one minute each, slides were treated with sera 131 diluted 1:500 in a 10% blocking buffer/PBST solution and incubated overnight at 4°C.

Following this overnight incubation, slides were washed three times with PBST for one minute 133 each and dipped into the previously prepared 1mM Tris Solution. Detection of IgG was 134 performed via treatment with Rabbit Anti-Human IgG DyLight ™ 800 for 45 minutes at 23°C.

After performing three PBST washes of one minute each, slides were then treated a second time 

The tiff image files, acquired after probing slides with secondary antibody solution, were 147 analyzed for any slide artifacts prior to data analysis (16). Artifact sequences from the prescreen 148 of secondary antibody treatment were excluded from further analysis. Any microarray slides that 149 did not display sufficient reactivity with polio and HA positive control peptide spots were 150 considered invalid.

For data analysis, we used the background-subtracted median intensity values acquired potential epitopes, we first averaged these median fluorescent intensity values of each peptide the average median intensity value of the infected group was at least 1.5-fold greater and had pvalues of < 0.05 from an unpaired t-test comparison to the average log2-normalized median 157 intensity value of the negative controls. Among these selected peptides, we compared the 158 average log2-normalized median intensity value of each peptide to their neighboring peptides.

Peptides that had significantly different fluorescent intensity values were screened out, and the 160 remaining peptides were recognized as potential epitopes (5, 17) (Table S4) .

To identify reactive epitopes, we looked at peptides in which the average log2-normalized 179 intensity value of either of the two infected sample groups had a fold-change value of at least 1.5 180 and yielded adjusted p-values of < 0.05 from an unpaired t-test when compared to the mean log2-181 normalized intensity value from the negative control group (15, 17) . Using these criteria, nine 182 peptides with significantly higher IgG antibody binding were identified in the spike protein 183 region ( Figure 1a and Table 1 ) among infected individuals.

Using the same approach of epitope identification with our vaccinated groups, no 185 peptides were found to meet the criteria for significant IgG binding activity. Two peptides were 186 found to have p-values less than 0.05, but these peptides did not meet our fold change criteria as 187 described in our methods (Figure 1b) . Of the nine peptides identified as significant in the 188 infected sample group, five were also found to have significantly greater IgG binding than the 189 vaccine group. Peptides with no significant differences in binding activity between the recently 190 infected group and vaccinated group were peptides S_1141 and S_1247-S_1251. The SARS-appear to be distributed throughout the spike protein ( Figure 2 ). Identified reactive epitopes that 196 meet our criteria include two consecutive peptides in the RBD (S_0343 and S_0345), one located 197 in the putative fusion peptide domain (S_0671), one found in proximity to the identified Heptad

Repeat 2 sequence (S_1141), and three located within a cysteine-rich sequence (S_1247, 199 S_1249, and S_1251) neighboring what is thought to be a transmembrane domain (18) located in 200 the C terminal of the S2 subunit. Among these peptides, those that were found to have relatively 201 higher intensity among the vaccinated group (S_1159 and S_1247) were both found in the S2 202 region. Seven frequent mutations in common variants were identified within at least one peptide 203 sequence (Table 3 ) and were found to be well distributed throughout the spike protein. The same approach used to identify IgG epitopes was utilized to identify IgA epitopes.

However, no significant IgA epitopes were identified in either the infected group or the 229 vaccinated group. The IgA titer information of the samples provided by RayBiotech was limited, 230 but in-house ELISA testing of the RBD antigen revealed low levels of IgA binding throughout 231 both infected and vaccinated groups (Table S4 ). The PEPperCHIP® microarrays contained the entire proteome of SARS-CoV-2, allowing for identification of significant IgA binding to 233 peptides outside of the spike protein. As the scope of this report is solely focused on the spike 234 protein these epitopes have been excluded.

To assess the potential of incorporating the discovered peptide sequences into viable 237 therapeutics or diagnostics, we investigated the conservation of sequences among other closely 238 related coronaviruses and the proximity of the sequences to mutations found in current dominant 239 variants. Previous studies have established that antibodies that bind to certain linear epitopes 240 found in the SARS-CoV-2 proteome are found to be reactive to conserved peptide stretches in 241 other coronaviruses (CoVs) (6). Aligning the epitopes to closely related CoVs helped to assess 242 the potential for nonspecific antibody binding by antibodies present due to exposure to other individuals. Both groups showed overall significantly higher IgG binding activity to epitopes 311 throughout the spike protein relative to the naïve samples. Despite similar levels of IgG binding activity between the two groups, distinct epitope binding patterns were only found in the infected response that is comparable to natural infection and easily detectable with most methods (32, 316 33). We believe multiple factors were contributing to these unexpected results. Many of the 317 acutely infected samples were collected two to three weeks after infection, when antibody 318 production during infection is thought to peak (34). Corresponding symptom information for 319 some infected patients was not available, so the number of infected patients with more severe 320 symptoms is unknown. It has been found that severe symptoms correspond to higher antibody 321 production (35) and more distinct antibody binding patterns (10, 11). If a large portion of these 322 infected patients had severe symptoms, then their antibody binding would be expected to be 323 higher. These variables limit comparisons between the infected and vaccinated group. From our 324 findings, it would appear that the infected sample groups yielded more specific epitopes with 325 significant IgG binding despite overall binding activity between the two groups being similar.

Although our results indicate that the vaccinated individuals did not have SARS-CoV-2 S 327 protein epitopes that showed consistent binding, there were still some peptides that demonstrated 328 the ability to potentially discriminate between vaccinated and naive individuals (S_1159 and 329 S_1247-S_1251). These peptides did not meet the fold change criteria to be identified as 330 significant in this study, so their potential sensitivity when incorporated into other assays is 331 unknown. These peptides are well conserved among the hCoVs which means this enhanced 332 binding may be non-specific. It is very possible that potential distinct epitopes with significant 333 binding would appear in a larger sample pool of vaccinated individuals. Furthermore, it has been 334 shown that previously infected individuals who receive a full vaccine treatment have extremely 335 high antibody titers (36). The study of epitope binding among these individuals after their 336 vaccination could lead to markers of vaccination to be more easily identified due to their higher 337 antibody titers. Though no epitope markers were identified in the vaccinated group, this may be 338 due to the relatively smaller sample size and low serum dilutions tested. An expanded study with 339 a larger sample pool has high potential to yield epitopes with more distinct binding among 340 vaccinated individuals.

The lack of significant IgA response within the spike region was somewhat surprising as 342 previous studies had successfully identified epitopes within the spike protein utilizing peptide 343 microarrays (11). Our ELISA data revealed relatively low IgA levels among some of the infected 344 samples. A potential source of these lower IgA levels may be tied to a portion of the serum 345 samples coming from asymptomatic and mildly symptomatic patients. IgA epitopes identified in 346 previous peptide microarray studies (11) found that IgA epitopes were predominately identified 347 among their severe symptom patient group. Since the quantity of overall IgA in serum is 348 typically found to be significantly less than that of IgG, it is possible that levels of IgA present in 349 some screened infected samples with milder symptoms were not high enough to discriminate 350 from the naive samples. It has been observed that the amount of IgA antibodies reactive to the difficult due to the limitations previously described without purification and concentration of IgA from samples. However, the potential for studying IgA epitopes of SARS-CoV-2 may still be viable with the use of saliva, which has a high content of IgA. Previous work using the peptide 357 microarray to study the epitopes of other viral targets have shown the IgG antibody profiles of 358 blood and saliva were similar (38). Salivary IgA has shown potential as a reliable biomarker for 

The authors declare that the research was conducted in the absence of any commercial or financial 547

relationships that could be construed as a potential conflict of interest. 548 Displayed are the log2 normalized individual fluorescent intensity values of infected (n=30), vaccinated (n=17), and control (n=10) sample groups. The peptide ID along with the results of the unpaired t-test comparing the infected and control groups can be found above the data. The results of the t-test comparing the fluorescent intensities of the infected and negative control groups values are shown (* < .05, ** < .01, *** < .001, **** < .0001). The protein sequence of the SARS-CoV-2 spike protein was annotated to study the proximity of identified potential epitopes to functional regions. Blue bars depict the locations of epitopes with significant IgG binding. Orange bars cover regions that are associated with predicted protein function. Protein sequence is colored according to the Zappo amino acid color scheme (with brightness of some colors increased for clarity).

Depicted above are mean log2 normalized fluorescent intensities of the vaccinated and control groups that were found to be significantly different (p-value < .05) based on the results of an unpaired t-test. The ID and results of the t-test are displayed above the plots. The vaccinated group was not found to have a fold change < 1.5 and, therefore, these peptides were not recognized as epitopes of highest interest based on our criteria. To determine conservation of the potential epitopes, the Wuhan SARS-CoV-2 sequence was aligned with the Omicron variant sequence, the other six human infected CoVs, and the closely related Pangolin and bat-infecting coronaviruses. Peptides IDs are displayed above their corresponding alignments and the color scale for the alignment can be found above. Listed are the peptides that met our criteria to be classified as potential epitopes among infected individuals. The leftmost column contains peptide IDs, which were assigned based on the position of the first amino acid in the sequence. Amino acid sequences and their location in the spike protein sequence are displayed in the second and third columns, respectively. Fold change and p-values acquired are displayed in the fourth and fifth columns respectively Table 2 .

Presented are peptides whose mean fluorescent intensities were at least three standard deviations higher than the overall average intensities in either the Moderna or Pfizer vaccinated groups. The vaccine group in which the peptide met these criteria are displayed in the first column (no peptides were found to be unique to the individuals who received the Pfizer vaccine). Peptide sequences are displayed in the second column. Bolded sequences are among or share significant amino acid identity with one of the 10 potential epitopes identified in the infected sample group. Amino acid position and domain/subregion of each sequence is displayed in the third and fourth columns. Table 3 .

The seven recurring mutations of prevalent SARS-CoV-2 variants found within one of the 10 potential epitopes were identified to study their potential impact on antibody binding. Mutations are displayed in the leftmost column. Variants associated with each mutation are displayed in the middle column. Epitopes that contain the mutation are shown on the right.

CoV-2 variants, spike mutations and immune escape

Effects of SARS-CoV-2 variants on vaccine 388 efficacy and response strategies

Epitope Mapping on Microarrays Highlights Strong Immune-Response to N Protein Region

CoV-2 reveals significant differences between non-structural/accessory and structural proteins

Immunoreactive peptide maps of 440 SARS-CoV-2

Two linear epitopes on 443 the SARS-CoV-2 spike protein that elicit neutralising antibodies in COVID-19 patients

Positional artifacts in 447 microarrays: experimental verification and construction of COP, an automated detection tool

Proteome-wide epitope mapping 451 identifies a resource of antibodies for SARS-CoV-2 detection and neutralization. Sig Transduct 452 Target Ther

Predicted 3D Models of the SARS-CoV-2 Spike Protein Membrane Proximal 455

External Region and Transmembrane Domain