Computational and experimental antibody affinity and diagnostic accuracy quantification of SARS-CoV-2 SD2 major disulfide loop analog

Synthetic oligopeptides provide a rapid and cost-efficient approach to developing antibodies and diagnostics for emerging viral variants. This study computationally and experimentally characterized a synthetic peptide analog of the SARS-CoV-2 spike subdomain 2 major disulfide loop (SD2MDL), designated S621 (CPVAIHADQLTPTWRVYSTC). Binding affinity was computationally estimated using the Heuristic Affinity Prediction Tool for Immune Complexes (HAPTIC), while experimental validation was performed using enzyme-linked immunosorbent assay (ELISA) with rabbit-derived antipeptide antibodies. Clinical diagnostic accuracy testing was done using plasma samples from RT-PCR–confirmed COVID-19 patients and pre-COVID-19 controls. S621 demonstrated nanomolar binding affinity (K d app = 1.14 nM) and high avidity (3.67 nM), closely matching HAPTIC predictions (3.54 nM). Diagnostic evaluation yielded a sensitivity of 75.14% and specificity of 53.18%, corresponding to an overall accuracy of 71.79%. These findings demonstrate that a single synthetic peptide derived from a conserved spike subdomain can function as a high-affinity surrogate for full-length antigens, supporting its potential application in rapid peptide-based immunodiagnostics.

The global experience with COVID-19 has emphasized urgency in quickly developing and testing diagnostics and therapeutics (1). Since its identification in late 2019, SARS-CoV-2 has evolved into multiple variants, each with altered antigenic properties, thus affecting the performance of existing vaccines and complicating serological surveillance (2). Vaccines based on the spike protein have demonstrated substantial protection, yet the production of full-length recombinant proteins requires extensive infrastructure, purification steps, and temperature-controlled storage (3). These limit the speed at which new biotechnological products can be generated in response to viral evolution.

In response to this, synthetic oligopeptides serve as alternatives, combining specificity with chemical simplicity (4, 5). By reproducing defined antigenic regions, peptides can provide targets for antibody detection or vaccine design without the logistical constraints of full-length proteins (6). Chemical synthesis allows rapid modification of sequences and the incorporation of stabilizing features (7). Multivalent constructs, achieved through polymerization or crosslinking, can simulate epitope clustering, enhancing apparent antibody binding through avidity (8). While affinity is a general term referring to binding strength, avidity is a subset of affinity that represents multivalent binding strength (9–11). Despite these advantages, peptides have lower intrinsic affinity compared with native proteins (12). Linear sequences may fail to capture conformational epitopes, and the choice of sequence, chemical modifications, and presentation format all influence assay performance.

Research on SARS-CoV-2 has focused primarily on peptides derived from spike or nucleocapsid proteins (13–15). Conserved loops in subdomain 2 of the spike protein have been previously identified in structural studies as both accessible to antibodies and relatively conserved across variants (16). Computational methods can prioritize sequences likely to engage antibodies effectively, yet most studies do not combine in silico predictions with experimental validation and testing in clinical samples. The lack of an integrated workflow limits the translation of peptide candidates into practical diagnostic tools and hinders assessment of their performance in diverse patient populations.

This study aimed to determine whether synthetic peptides derived from structurally conserved regions of SARS-CoV-2 spike protein can serve as functional surrogates for full-length antigens in diagnostic applications. The focus is on the major disulfide loop of subdomain 2, a region predicted to be immunogenic and surface-exposed. This study aimed to computationally and experimentally determine the affinity of the SARS-CoV-2 spike SD2 major disulfide loop (SD2MDL) toward cognate antibodies. Specifically, The study aimed to develop immunogenic, conserved synthetic peptide analogs of SD2MDL, experimentally estimate the affinity of these analogs using antipeptide antibodies, and demonstrate the binding of said analogs to cognate antibodies in clinical samples.

Design and in silico evaluation of SD2MDL peptide analogs

The peptide analogs S621 and S623 were designed by B-cell epitope prediction (Table 1). S621 is the primary SD2MDL analog under study, while S623 (a shortened version of S621) was the immunizing peptide used for generation of antipeptide antibodies. S623 also served as a comparator for depolymerization. Affinity was computationally estimated using the Heuristic Affinity Prediction Tool for Immune Complexes (HAPTIC) (17). In brief, paratope-epitope binding affinity was expressed in terms of the standard free-energy change of binding Δ G b °:

where R is the gas constant, T the absolute temperature, C the standard concentration (set by convention to 1M), and K d the dissociation constant. Paratope-footprint radius was set at 10.0Å, which corresponds to one complete turn (3.2 amino acid residues) of a polyproline II helix, the underlying assumed bound state of the linear epitope in HAPTIC (18). The temperature of immunization at 312.35K, which is the normal core body temperature for rabbits, the host for our generated antipeptide antibodies (19). The temperature of immunoassay was set at 298.15K, ambient room temperature.

Structural and functional analyses were conducted by retrieving three-dimensional structures from the Protein Data Bank (PDB; https://www.rcsb.org/) and UniProt (https://www.uniprot.org/, RRID: SCR_002380), visualized using PyMOL (https://www.pymol.org/, RRID: SCR_000305). Sequence homology was assessed using BLASTp to ensure specificity to SARS-CoV-2 epitopes. Sequence-based mutation effects were analyzed using resources available through the ExPASy bioinformatics portal. Peptides were modified by adding cysteine residues at the N- and C-termini to facilitate polymerization through disulfide bond formation.

Peptide synthesis, polymerization, and depolymerization

The antigenic peptides were synthesized by Genscript (Singapore), a commercial provider. Solid-phase peptide synthesis was employed, and the peptides were analyzed for purity and identity using high-performance liquid chromatography (HPLC). The peptides were reconstituted in dimethyl sulfoxide (DMSO), then subjected to polymerization under mild oxidizing conditions. This involved initiation of polymerization by adding an equal amount of sterile water, to achieve a final concentration of 50% DMSO in water as the polymerization solvent. The mixture was then incubated at room temperature for 24 hours to promote disulfide bond formation. The resulting solution served as the stock solution which was diluted in coating buffer prior to ELISA.

S621 was subjected to depolymerization to assess the contribution of the polymeric structure to avidity. N-acetylcysteine (Nac) was used as a reducing agent to break disulfide linkages. Aliquots of the polymeric peptide were incubated with various concentrations of Nac at 37 °C for 30 minutes. To match the composition of the test solutions with the negative control, DMSO was added to a final concentration of 20%. The negative control was prepared by heating Nac in the presence of DMSO to degrade it. The complete degradation of Nac was confirmed using Ellman assay and by the evolution of the characteristic odor of dimethylsulfide. Peptide polymerization and depolymerization was confirmed with SDS-PAGE (Supplementary Figure 1). The apparent dissociation constant (K d app) was calculated for both the polymeric and depolymerized forms of the peptide. The “polymeric” form of the peptide was the peptide subjected to 50% DMSO oxidation, while the “depolymerized” form of the peptide was the polymeric peptide subjected to Nac depolymerization. The prepolymerized of the peptide (i.e., not subjected to DMSO oxidation) was not used in our experiments.

Generation and purification of antipeptide antibody

New Zealand White rabbits were immunized with keyhole limpet hemocyanin (KLH)-conjugated S623 peptide emulsified in Freund’s adjuvant (Day 0, 14, 28). Sera were collected on Day 42 and affinity-purified on M1-Sepharose 4B columns.

S621 ELISA and binding analysis

To assess peptide-antibody interactions, indirect ELISA was performed as previously described in the detailed protocol available on protocols.io (20). Peptides were diluted in coating buffer to achieve a final concentration of 10 nM peptide and used to coat high-binding 96-well microtiter plates. The plates were incubated at 4 °C overnight, followed by washing with phosphate-buffered saline containing 0.05% Tween-20 (PBST). Blocking was performed with 2% skimmed milk in PBST at 37 °C for 30 minutes. Serial dilutions of antipeptide antibody were then applied, followed by incubation with 1:2000 protein A peroxidase conjugate (0.5μg/mL). Chromogenic development was achieved using tetramethylbenzidine and hydrogen peroxide, and the reaction was halted with 0.1 M sulfuric acid. Absorbance was recorded at 450 nm, and K d app values were calculated from the slope of linearized plots.

To experimentally estimate K d, ELISA absorbance values were then plotted against antibody concentration and linearized according to the equation of Orosz & Ovádi (2002) (21):

where i is the relative saturation, [Ab]tot the total antibody concentration, and K d app the apparent dissociation constant (calculated from the slope of the resulting linear plot). The K d app for polymeric S621 was taken as the avidity, while that for depolymerized S621 was taken as the affinity. Calculations are shown in Supplementary Methods 1.

Assessment of antibody recognition in clinical plasma

Clinical samples

Clinical samples were prospectively collected from patients who satisfied the following eligibility criteria: (1) adults (≥18 years), (2) with RT-PCR-confirmed COVID-19, (3) admitted at Philippine General Hospital, the national referral hospital, between October 2020 and February 2021, prior to the national vaccine rollout. All enrolled participants with successfully collected serum specimens were included in the analysis, and further sample selection was performed. Participants discharged or absconded before specimen collection were not included because samples were unavailable for testing. Potential participants were identified by the attending physician using convenience sampling. Blood was drawn on days 1, 7, and 14 of hospitalization, coinciding with routine clinical sampling. All sera were heat-inactivated at 56 °C for 30 minutes and stored at −80 °C until use. Disease severity was classified based on Department of Health guidelines: asymptomatic (RT-PCR positive), mild (with symptoms but no pneumonia), moderate (with pneumonia), severe (requiring oxygen supplementation), or critical (requiring mechanical ventilation).

Controls

Negative control sera were obtained from healthy blood donors prior to December 2019. Samples were treated identically to patient sera, including heat inactivation and storage conditions, to ensure assay consistency.

Diagnostic accuracy determination

The result of the S621 ELISA was compared against the index test, RT-PCR as performed by non-blinded laboratory personnel, following standard operating procedures (cycle threshold: 40 or less). Clinical data was available to medical technologists and pathologists for interpretation. Diagnostic performance was evaluated using statistical software (MedCalc version 23.0.8, RRID: SCR_015044). The cutoff for ELISA positivity was determined via Youden index (J = Sensitivity + Specificity − 1), a widely used ROC-based criterion that maximizes the combined sensitivity and specificity. Calculations are further described in Supplementary Methods 2. Accuracy measures were calculated for both prespecified and exploratory thresholds. No indeterminate results were reported, and all samples were included in the analysis. Sample size estimates were determined using PASS software (version 20.0.7, RRID: SCR_019099) based on an expected prevalence of 42.3% and a binary outcome assumption (22, 23). Figures were created using BioRender (https://BioRender.com, RRID: SCR_018361).

Ethics

The study was conducted following the principles of the Declaration of Helsinki and the standards for reporting diagnostic accuracy (STARD) guidelines (24). Ethics approval was obtained from the University of the Philippines Manila Research Ethics Board (Code 2022-0492-01). The full protocol is registered under HERDIN (https://www.herdin.ph/) (PHRR230601-005772). All authors had access to study data and approved the final manuscript. Laboratory personnel performing ELISA were blinded to patient disease status throughout testing.

Conservation and immunogenic features of SD2MDL analogs

The peptide analog S621 was selected to represent a conformationally flexible segment of the SD2 major disulfide loop (SD2MDL) of the SARS-CoV-2 spike protein (Figure 1A). Sequence alignment and structural mapping using PDB entries 6XR8 and 7CWL revealed that this segment is highly conserved across multiple variants, including Alpha, Beta, Gamma, Delta, Epsilon, and Omicron (Figure 1B). Notably, several residues within S621 were unresolved in the crystal structures, indicative of conformational disorder and surface exposure, features that are consistent with B-cell epitope accessibility. Computational B-cell epitope prediction using HAPTIC highlighted residues AIHADQLTPTWRVY as the most immunodominant within this segment. These analyses indicate that the selected peptide region preserves key structural and antigenic features across major viral lineages and is likely to retain antibody recognition potential despite variant-specific amino acid changes. Structural visualization highlighted the surface orientation of the loop and potential accessibility to circulating antibodies, with a notable absence of steric hindrance from neighboring S1 subdomains.

High-affinity interactions between peptide analogs and antibodies

The affinity of S621 for its cognate antibodies was examined both computationally and experimentally. HAPTIC-based in silico modeling predicted an apparent dissociation constant (K d app) of 3.54 nM, reflecting strong interaction potential. Experimental evaluation using an in-house ELISA confirmed high-affinity binding, with a K d app of 1.14 nM for depolymerized interactions and an overall avidity of 3.67 nM for polymeric forms (Figure 2). The modest 3.54% deviation between predicted and measured values supports the accuracy of the computational affinity estimates. Observed binding curves demonstrated a typical sigmoidal pattern consistent with saturable, high-affinity interactions, with negligible background signal observed in negative control wells containing irrelevant peptides or pre-pandemic plasma samples. Comparative analysis between depolymerized and polymeric S621 revealed a measurable avidity gain, suggesting that the multivalent presentation of epitopes enhances effective binding. The EC 50 value of polymeric S621 was 1.982 (95%CI: 1.826 – 2.153) nM, which was significantly lower than that of the depolymerized form at 5.053 (95%CI: 4.391 – 5.943) nM, (p < 0.0001, unpaired t -test), indicating a higher avidity. These findings provide quantitative evidence that polymeric S621 maintains structural integrity while promoting enhanced antibody recognition.

Recognition of SD2MDL analogs by antibodies in clinical samples

The diagnostic performance of S621 was evaluated using plasma samples collected from 1,443 individuals with RT-PCR–confirmed SARS-CoV-2 infection prior to vaccination rollout (Supplementary Table 1, Figure 3A). The ELISA incorporating S621, with a cutoff value set by Youden J to an OD450nm of 0.04, achieved a sensitivity of 75.14% (95% CI: 72.62–77.54%) and a specificity of 53.18% (95% CI: 46.36–59.92%) (Figures 3B, C, Table 2). Overall diagnostic accuracy was 71.79%, with positive and negative likelihood ratios of 1.60 and 0.47, respectively. Signal intensity was consistently higher in confirmed positive samples, whereas pre-pandemic negative controls exhibited low optical density readings, indicating minimal non-specific binding. Despite relatively high sensitivity, the comparatively low specificity suggests potential cross-reactivity with antibodies generated in response to related epitopes, warranting further evaluation with variant-specific panels. Subgroup analyses by age, sex, and disease severity showed no statistically significant differences in ELISA signal, supporting broad applicability across diverse patient populations. Nevertheless, given the considerable number of false negatives, we have reported them in Supplementary Table 2.

The peptide analog S621 corresponds to a highly conserved and surface-exposed segment of the SD2 major disulfide loop of the SARS-CoV-2 spike protein. This region is preserved across major variants, with minimal substitutions affecting residues implicated in B-cell epitope recognition (Figure 1B), as demonstrated with sequence alignment and structural mapping. These findings are consistent with previous structural analyses suggesting that SD2 loops maintain a stable surface orientation across SARS-CoV-2 variants while exhibiting conformational flexibility (25). The unresolved residues in PDB entries 6XR8 and 7CWL correspond to regions likely to experience dynamic disorder, a characteristic that enhances antibody accessibility. This observation aligns with reports that conformationally disordered, flexible loops often serve as immunodominant epitopes due to their increased solvent exposure and adaptability in binding to diverse paratopes (26, 27). The computational identification of residues AIHADQLTPTWRVY as a dominant B-cell epitope further validates the selection of S621 as a candidate diagnostic peptide. Collectively, these results emphasize the importance of targeting conserved, flexible regions when designing peptide-based diagnostics, as such epitopes retain cross-reactivity potential despite antigenic drift.

Experimental evaluation of S621 confirmed the high-affinity interactions predicted computationally. The apparent dissociation constant (K d app) measured by ELISA closely matched the computationally derived estimate, demonstrating the reliability of in silico modeling using HAPTIC for pre-screening peptide candidates (Figure 2). The modest deviation between predicted and measured K d app values may be attributable to differences between idealized computational models and the physicochemical environment in vitro, including multivalent interactions and potential peptide folding. Note that solvent effects are already taken into consideration in the calculation of HAPTIC. Polymeric S621 exhibited enhanced avidity relative to the depolymerized form, indicating that multivalent presentation significantly improves effective binding. This finding is consistent with previous studies reporting that polymerization or clustering of epitopes can amplify apparent antibody affinity via avidity effects, a strategy successfully employed in both vaccine design and serological assays (28, 29). While previous studies have demonstrated the feasibility of peptide serodiagnostics using linear epitopes, our findings add that multivalent presentation may further enhance antigenicity (30). Further, the choice to use depolymerized peptide to estimate affinity (as opposed to prepolymerized peptide) may represent an improvement over prior efforts in avidity gain estimation, given that cysteine-linked peptides are known to undergo air oxidation, potentially leading to unwanted aggregation and antigenicity change (31, 32). Overall, the results suggest that computational pre-screening combined with polymerization strategies can accelerate the development of high-affinity peptide analogs suitable for diagnostic applications.

Application of S621 in clinical ELISA demonstrated strong sensitivity but limited specificity. The assay correctly identified 75.14% of COVID-19 samples (i.e., sensitivity) (Figure 3, Table 2). The relatively low specificity (53.18%) is likely attributable to cross-reactivity with antibodies targeting structurally similar epitopes. This limitation highlights the trade-off between maximizing sensitivity and maintaining high specificity in peptide-based serological assays (33, 34). These findings provide valuable context for the deployment of peptide-based diagnostics in surveillance and epidemiological studies, where sensitivity may be prioritized over specificity in early detection contexts. As for the specific explanations for the false positives in pre-COVID-19 controls, it may have been due to shared or partially homologous epitopes between SARS-CoV-2 and other viruses or pathogens. This has been previously reported for HIV as well as malaria, both of which were not excluded in our study (35, 36). Further, matrix effects could not be discounted, as heterophile/polyreactive antibodies as well as rheumatoid factor are known to interact with immunoassays (37, 38).

The present findings highlight the feasibility of designing synthetic peptide analogs that recapitulate key immunogenic features of viral proteins. By targeting conserved, flexible loops, peptide-based diagnostics may retain effectiveness against emerging variants, addressing a critical limitation of conventional protein-based systems. Moreover, the demonstration that polymerization enhances avidity provides a methodological framework for optimizing peptide presentation in diagnostic assays. This approach may be extended to other viral antigens or conserved immunogenic regions in emerging pathogens, offering a scalable and modular alternative to recombinant protein-based diagnostics.

Several limitations must be acknowledged. Antigenicity was assessed using rabbit antipeptide antibodies rather than human clinical sera. Although antibodies from SARS-CoV-2-infected individuals may better represent natural immune response, clinical sera are inherently heterogeneous and contain polyclonal antibodies directed against multiple epitopes, with affinity varying depending on infection stage and disease severity (39, 40). Rabbit antipeptide antibodies were therefore used to provide a more controlled estimate of S621 avidity (41, 42). Another limitation was the use of Youden index for cutoff selection, which is known to introduce optimism bias (43). Thus the reported clinical performance here may have been overstated, and independent validation for specific settings is needed (44). Also, the relatively low specificity observed for S621 necessitates further optimization, potentially through inclusion of multiple peptide epitopes. The study was performed on pre-vaccine sera, and performance may differ in post-vaccine populations or in samples with complex immunological histories. Quantitative comparisons of binding affinity relied on ELISA-derived K d app measurements, which may not capture all kinetic parameters of peptide-antibody interactions. Future studies could incorporate surface plasmon resonance or biolayer interferometry for enhanced kinetic resolution Future studies should evaluate the performance of S621 in combination with complementary epitopes, assess longitudinal stability of antibody detection, and explore alternative multimerization strategies to further enhance specificity and avidity. Additionally, integration of peptide-based assays into high-throughput platforms may facilitate large-scale epidemiological surveillance and vaccine efficacy monitoring. This is especially because peptide-based approaches have already been shown to discriminate antibody responses, including those elicited by mRNA vaccination (15, 45). Such efforts could identify antigenic hotspots in whole Spike protein (46, 47). Future work ought to incorporate comparator peptides to more rigorously determine whether the avidity gain depends on the selected peptide sequence. Comparison against established commercial immunoassays would also further benchmark the performance of the developed diagnostic.

S621, a synthetic analog of the SARS-CoV-2 spike SD2 major disulfide loop, has nanomolar affinity toward cognate antibodies, shown both computationally and experimentally. Its strong binding and broad antibody recognition support its usefulness as surrogate for the full-length spike protein in immunodiagnostics. Work is underway to evaluate the binding of anti-S621 to intact SARS-CoV-2 virions. This oligopeptide-based strategy may be adapted for rapid development of locally produced immunodiagnostics, prophylactics, and therapeutic antibodies against endemic or emerging pathogens in the Philippines.