Since the first appearance of SARS-CoV-2 in late 2019, its variants have been classified into key groups—such as variants of concern (VOCs), variants of interest (VOIs), variants being monitored (VBMs), and variants of high consequence (VOHCs)—based on predicted virulence, transmissibility, and disease severity (Marino et al., 2020; Jung et al., 2024; Nesamari et al., 2024). Specifically, Alpha, Delta, and Omicron VOCs were highly successful variants that resulted in sequential viral strain replacements over time (Meganck et al., 2024). SARS-CoV-2 infection is initiated by virus interaction with host’s ACE2 cell-surface receptors and proteases, which activates the spike (S) proteins, followed by the fusion of viral and cellular membranes to release the virus genome into the host cell (Pizzato et al., 2022; Steiner et al., 2024). These initial interaction and fusion activities are mediated by SARS-COV-2 spike glycoprotein, which has become a primary target of vaccine development, diagnostic, and therapy (Lucena et al., 2024; Prabhakaran et al., 2024; Yamamoto and Inoue, 2024; Zhang Y. et al., 2024). Specifically, great attention has been directed to immune-based and aptamer-based technologies against SARS-COV-2 spike protein (Chen et al., 2023; Dong et al., 2024).

Aptamers are single-stranded oligonucleotides that possess excellent recognition and binding properties to their targets with remarkable affinity (Brown et al., 2024; Mahmoudian et al., 2024). Compared to antibodies, aptamers have advantageous features such as ease of chemical modification, inexpensive production costs, relatively low batch variability, non-immunogenicity, and in vitro selection (Alkhamis et al., 2024). Informally known as “chemical antibodies,” aptamers have increasingly become antibody alternatives (Maradani et al., 2024). For the detection of intact SARS-CoV-2 VOCs, aptamers that recognize and bind to accessible surface proteins (e.g., S protein) are necessary (Kacherovsky et al., 2021).

The 150 kDa S protein, a glycosylated homotrimeric class I fusion protein, extends from the viral surface, which plays a pivotal role in viral entry and pathogenesis. Its S1 portion is comprised of the receptor-binding domain (RBD) and N-terminal domain (NTD), while the S2 portion contains the fusion peptide (Cao et al., 2021; Yan et al., 2021; Stäb et al., 2024). Aptamers have been identified against S-protein both as potential drug candidates, e.g., lowering pro-inflammatory response (Kim et al., 2024) and binding inhibition and viral neutralization (Gelinas et al., 2023), and as the main detection elements for diagnostic applications, e.g., point of care detection (Shrikrishna et al., 2024) and environmental monitoring (Yu et al., 2024), among others. However, new mutations have emerged and spread, resulting in unique amino acid profiles among SARS-CoV-2 VOCs. Significantly, numerous alterations in S-protein have undergone positive selection, leading to considerable changes in viral traits, including increased transmissibility and enhanced immune evasion, posing challenges to the current SARS-CoV-2 mitigation approaches (Yao et al., 2024; Yue et al., 2024).

In this study, we aimed to identify and generate aptamer sequences targeting the S-protein of SARS-CoV-2 VOCs through systematic evolution of ligands by exponential enrichment (SELEX). We conducted extensive in silico and in vitro characterizations to identify robust and specific aptamer sequences. These sequences are potential candidates for addressing SARS-CoV-2 variants.

In this study, a two-step PCR procedure was conducted per each round of SELEX. The first step (1) preparative PCR was performed to determine the optimum cycle number, while the subsequent step was the amplification PCR (2), where the sub-library was generated. The optimum cycle number was defined as the number of PCR cycles that displayed the highest yield of aptamer fragments while having low amounts of PCR by-products (Supplementary Figure S2). After nine selection cycles, the generated ssDNA pools were processed and sequenced. The most prevalent sequences are as follows: Alpha VOC or Alpha Apt (5′-AGTTAGCCGTCATTCAGACGATTTGTCCGTGTTACGATTGGGGAACGTG GATGACATTCTGGACACTCGCCATGCTCGAT-3′), Delta VOC or Delta Apt (5′-ATCGAGCATGGCGAGTGTCCTATCCCCCATCAACTCAACCCAACCAGTACGACTCCTCCTCGTCTGAATGACGGCTAACT-3′), Omicron VOC or Omicron Apt (5′-CTGGCCTCACTGGATACTCTAAGACTATTGGTCAAGTTTGCCTTGTCAAGGCTATTGGTCAAGGCAAGGCTGGCCAACCCATGGGTGGAGTTTAGCCAG-3′). These sequences were also chosen for further characterization. The most stable secondary structures, i.e., lowest Gibb’s free energy ∆G, are shown in Figure 3A, which also indicate the highest thermodynamic stability. Similar to the secondary structure, Figure 3B shows the tertiary structures of the identified aptamer sequences for the three VOCs. Each of these structures has stems and hairpin loops. Molecular docking models are presented in Figure 3C, showing the highest-ranked models based on the docking scores. Supplementary Table S2 summarizes the top eight docking and confidence scores and ligand root-mean-square deviation (RMSD) value, which is the mean distance between the atoms of superimposed molecules (Parisien et al., 2009; Oliveira et al., 2022). Representative macromolecule (S protein)-ligand (Alpha Apt, Delta Apt, and Omicron Apt) complexes’ binding site sections indicate non-covalent interactions such as hydrogen bond, salt bridge, and hydrophobic interactions (Supplementary Figures S3–S5). Details on atomic-level contacts of aptamers, as well as information on amino acid residues, participating ligand/aptamer atoms, and geometry of the interaction (e.g., distance of interacting atoms), are presented in Supplementary Tables S3–S5.

The dissociation constant (K_d) of VOC aptamers was estimated by absorbance using the direct-enzyme-linked oligonucleotides assay (ELONA) method. This approach allowed immobilization of S-protein on a microplate and utilized biotinylated VOC aptamers and S-HRP for a colorimetric reaction to occur via the oxidation of TMB substrate (3,3′,5,5′-tetramethylbenzidine diamine). By adjusting the aptamer concentrations in relation to the fixed concentration of S-protein followed by quantifying the bound aptamer, the binding affinity could be determined as a dissociation constant (K_d). Figures 4A–C show the binding affinity curves of VOC aptamers. The estimated (K_d) values were 193.53 ± 30.84 nM for Alpha Apt, 111.51 ± 55.41 nM for Delta Apt, and 32 ± 11.84 nM for Omicron Apt (Supplementary Figure S3). The S.D., which was less than 60% of the calculated K_d is typical with this type of affinity estimation (McKeague et al., 2014).

In SELEX, non-specific amplicons or parasite DNA appear to form via primer-primer hybridization and partial binding of primers to the random region of the library, forming products with several base pairs longer than the expected aptamer (Tolle et al., 2014). Unfortunately, due to the reliance of aptamer development on a random library, the amplification of nonspecific sequences appears to be unavoidable once complementary sequences between the primers and library sequences exist. Therefore, selecting a PCR cycle number that both maximizes aptamer amplification while minimizing by-product formation is critical for the efficiency of SELEX. To ensure the high binding of aptamers to the protein target in this study, the concentration of target molecules was reduced with the progression of SELEX to create a more selective SELEX condition. The initial aptamer-to-target ratio used was 50:1 since it has been reported to provide substantial competition between ligands to facilitate the enrichment of high-affinity aptamers (Seo et al., 2010; Liu et al., 2022). The ratio gradually decreased in subsequent rounds of incubation. The poly-T20 with Sp9 bridge allowed denaturing gel electrophoresis-based single-stranded DNA separation (Supplementary Figure S4). Another strategy used in this study for selecting high-binding aptamers was decreasing the incubation time from 1 h in the first round of SELEX to 10 min in the eighth round of SELEX. To stabilize the selection conditions, Mg²⁺ was included in the binding buffer since it has been shown to stabilize the secondary structure of aptamers which results in enhanced binding (Reuss et al., 2014). Additionally, primers were designed to include reasonable annealing temperatures and less than 50% GC content to avoid primer heterodimers and primer self-dimers. The consideration of PCR by-product is especially important since parasite DNA is often the cause of the failure of SELEX (Tolle et al., 2014). By-product formation is generally avoided by lowering the PCR cycle number. This, however, tends to reduce the yield of the desired aptamers. To account for the lower PCR yield, the amplification PCR was conducted to allow larger numbers of PCR reactions. Over time, by-products can still have a large presence in the library. To mitigate this problem, parasite DNA was removed from the library via PAGE gel separation, which allowed the yield of pure desired ssDNA needed for further rounds of SELEX which was completed until round nine.

The NGS data was processed using clustering strategies based on similarities, which identified sequence groups by sequence distribution to aptamer families. Sequence features such as (1) full length, (2) fixed k-mers, and (3) sub-sequences of different lengths can facilitate the clustering of aptamers from sequenced random regions (Sun et al., 2022). In this study, every sequence in the pool was compared and placed in the same cluster based on the similarity threshold. Each representative sequence from clusters was further analyzed, aligned, and truncated. Sequences were filtered even further, which generated the desired aptamer sequences (approx. 80-nt) with the highest frequency. The secondary and tertiary structures of the chosen aptamer sequences are the results of hydrogen bond-based intramolecular base-pairing as determined by structural modeling (Patel et al., 1997; Curtale and Citarella, 2013). For aptamers, the likelihood of its predicted framework relies on the superior flexibility of the deoxyribose-phosphodiester backbone that has six different torsion angles which provide the formation of diversified secondary and tertiary structures (Laing and Schlick, 2011; Navien et al., 2021).

For the molecular docking simulation, the combination of SARS-CoV-2 VOCs aptamers as ligand and S-protein as the receptor was used and evaluated by HDOCK. SARS-CoV-2 VOCs aptamers formed structural motifs (e.g., hairpin loops, stems, and pseudoknots), suggesting that these configurations may have allowed its binding with S-protein via electrostatic interactions, hydrogen bonding, pep stacking or non-covalent interactions between peptides particularly involving aromatic residues or other stacking-prone motifs, hydrophobic interaction, and van der Waals forces or combinations of these various forces (Tan et al., 2016; Navien et al., 2021). The molecular docking scores are estimated by an iterative scoring function of the HDOCK server that allows the ranking of possible binding models based on negative docking score values (Yan et al., 2017). For the ligand RMSD, the values range from 351.8 Å–540.65 Å. The RMSD is a common metric used to assess the quality of protein docking models by quantifying the deviation between the predicted model and a reference structure. It provides an objective and straightforward way to assess the alignment between predicted and reference structures. Lower RMSD values typically suggest that the docked model is more like the reference structure, while higher RMSD values imply greater structural divergence. These values can be highly sensitive to the initial orientation of the docked molecule or the reference structure. Small changes in the pose or the rotation of the ligand (or complex) such as aptamers, may lead to large changes in RMSD, even though the actual binding mode or functional relevance might not have changed significantly. Though the ligand RMSD value is not the most accurate metric for each docking model, it has been used to analyze the structural stability of macromolecules during molecular dynamics simulation analysis (Yan et al., 2017; Hu et al., 2019).

The functionality of aptamers is intrinsically linked to their conformational structure, as their effectiveness relies on the physical compatibility between the oligonucleotide and its target. This structure is determined by the nucleotide sequence and how these nucleotides assemble into a three-dimensional form. Thus, predicting the tertiary structure is essential for pinpointing the residues that dictate the aptamer’s conformation and its interaction with the target. In this study, the presented workflow can successfully simulate molecular docking between VOC aptamers and S-protein, which identifies the physical–chemical interactions. The workflow also functions as a preliminary tool to analyze the potential impact of post-SELEX modifications on aptamer conformation and target interactions. This crucial information will facilitate the identification of residues in various aptamer motifs for potential substitution and will provide a suitable in silico platform to guide subsequent laboratory work. While factors other than affinity, such as toxicity, binding specificity, and deliverability, need to be considered in evaluating aptamers, the VOC aptamers—Alpha Apt, Delta Apt, and Omicron Apt presented in this work should certainly be further explored.

The binding of the VOC aptamers to S-protein was assessed using ELONA, wherein the proteins were coated on polystyrene high bind 96 microplates. Previous studies on the adsorption of SARS-CoV-2 spike protein on polystyrene surfaces, such as the one reported by Sahihi and Faraudo (2022), have shown no significant structural changes in the conformation of S-protein. Sahihi and Faraudo (2022) also found that Val and Phe amino acid residues in the receptor binding domain (RBD) have crucial roles in contact initiation to polystyrene surface, and its primary driving forces for adsorption are π–π and hydrophobic interactions of polystyrene with the residues and glycans. Their investigation on RMSD, solvent accessible surface area (SASA), volume of protein, radius of gyration (R_g), and secondary structure of the protein conformations showed that adhesion to the polystyrene surface did not cause any substantial secondary or tertiary structural modifications in the protein conformations. Moreover, a report by Xie et al. (2020) on the characterization of the interaction of SARS-CoV-2 virions with surfaces using an atomic force microscopy (AFM) which measured the adhesion force and energy of the SARS-CoV-2 S-protein with a series of inanimate surfaces including a large diversity of hydrophobic and hydrophilic materials such as metals and glass, fabrics, and plastics found that polystyrene had the strongest adhesion force among the materials tested.

A report on aptamer development for SARS-CoV-2 employed a viro-SELEX method using active SARS-CoV-2 (pseudotyped) with S-protein as the target and negative selection against UV inactivated SARS-CoV-2 (pseudotyped), SARS-CoV-1 (pseudotyped) with S-protein and H5N1 (pseudotyped) generated a 45-nt aptamer with binding constant (K_d) of 79 ± 28 nM (Peinetti et al., 2021). Similarly, Martínez-Roque et al. (2022) utilized a combination of ideal-filter capillary electrophoresis SELEX (IFCE-SELEX), NGS and slot blot binding assay to isolate and validate ssDNA aptamers (aptamer C7; K_d = 89.41 ± 18 nM and aptamer C9; K_d = 231.9 ± 15 nM) that can bind to SARS-CoV-2 spike glycoprotein. These previous investigations have generated aptamers with relatively higher binding constants than this study, highlighting the potential benefits of adapting the current methods and technologies to isolate and produce highly specific and stable SARS-CoV-2 aptamers. To minimize the degradation of aptamers over time, they can be protected by conjugation to stabilizing molecules (e.g., PEGylation, nanoparticles), which would prevent their interactions with nucleases and extend their half-life. Binding aptamers to nanoparticles (e.g., gold nanoparticles) or other carrier molecules can also protect them from enzymatic degradation while maintaining their functional properties. In addition, structural modifications like circularization (i.e., joining the 5′ and 3′ ends) can protect aptamers from exonucleases since circular aptamers tend to have increased resistance to degradation. These strategies can be considered in the design of stable aptamers for therapeutic, diagnostic, and research applications.

SARS-CoV-2 continues to evolve, and several newer variants have also emerged. A recombinant lineage of Omicron variant—named XBB, appeared in late 2022, and its descendants have evolved successfully (Zhang Q. E. et al., 2024). The members of the XBB lineage were noted for their improved immune evasion and transmissibility. As new mutations in the spike protein emerge, SELEX libraries can be updated to specifically target these changes, helping to recognize the virus even in the presence of mutations that confer immune escape. Moreover, SELEX can be re-executed using viral particles or synthetic spike protein fragments of these newer variants. By using molecular docking in combination with SELEX, reliable models of newly selected aptamers and their interaction with mutated versions of the spike protein or other viral proteins can be created. These approaches can facilitate the designing of more specific and up-to-date aptamer sequences that could recognize viruses even when mutations arise.

While in silico predictions offer a valuable initial step in understanding protein-ligand interactions, they have a few limitations related to the accuracy of structural models, scoring functions, and conformational flexibility. Experimental techniques such as X-ray crystallography, nuclear magnetic resonance (NMR), SPR, isothermal titration calorimetry (ITC), cryo-EM, and mass spectrometry are critical for validating and refining these predictions. A combination of computational and experimental approaches can offer a more comprehensive and reliable understanding of protein-ligand interactions, essential for aptamer discovery and other applications.

This study illustrated that the SELEX procedure was effective in generating in-house aptamers but can also be further optimized with in silico techniques by using such as those applied in this work. The molecular docking simulation has been useful in providing a deeper understanding of SARS-CoV-2 VOCs aptamers and S-protein interactions, which is meaningful for the design and optimization of aptamer applications.

In this report, we presented ssDNA aptamer selection by an in silico approach (SELEX) specifically targeting the receptor-binding domain of the S-protein across SARS-CoV-2 VOCs and elucidated the molecular docking and binding dynamics of the identified aptamer sequences against it. The identified aptamer sequences bind to relevant VOCs via the spike proteins with high affinity. The dissociation constant values (K_d) of the aptamers were in the low nM range, while the docking models displayed another layer of understanding of how SARS-CoV-2 VOCs aptamers and S-protein interact at the molecular level, contributing to future antiviral and diagnostic innovation. Future work will be focused on experimental validation of the identified aptamers’ binding affinity and specificity across a broader range of SARS-CoV-2 variants, including new variants that may arise. Further optimization of aptamer stability, selectivity, and performance in diagnostic platforms will be crucial for their clinical application. Additionally, incorporating aptamers into detection systems or therapeutic interventions, such as biosensors or antiviral agents, represents a promising direction. Expanding this research to other viral targets could also contribute to the broader field of aptamer-based diagnostics and therapeutics, paving the way for versatile tools to combat future viral outbreaks.