Rational Design and In Silico Optimization of DNA Aptamers Targeting Ricin B-Chain for Biosensor Applications Using Integrated Molecular Docking, Dynamics, and AI-Guided Approaches

Abstract

Ricin, a highly potent type II ribosome-inactivating protein derived from Ricinus communis, represents an important threat in food safety, environmental monitoring, forensic toxicology, and biosecurity. Rapid and reliable ricin detection is therefore essential. Conventional antibody-based detection systems are effective but may be limited by stability concerns, production complexity, cost, and batch-to-batch variability. Nucleic acid aptamers offer a promising alternative because of their chemical stability, structural versatility, ease of synthesis, and compatibility with biosensor platforms. The present study developed a fully computational, structure-guided workflow for the rational optimization of DNA aptamers targeting the ricin B-chain. Previously reported ricin-binding aptamers, including SSRA1, A3, C1, and C5, were collected and analyzed using multiple sequence alignment, GC-content analysis, and secondary structure prediction. The A3/C1/C5 aptamer family showed a conserved guanine-rich region and higher GC content than SSRA1, suggesting a shared structural motif. Among the candidates, C5 exhibited the most organized predicted secondary structure and was selected for truncation-based optimization. Structure-guided truncation of C5 led to the identification of T1, a 24-nucleotide aptamer that retained the central structural scaffold while reducing sequence length. T1 preserved eleven predicted base-pairing interactions and maintained a compact stem-loop architecture, making it suitable for downstream structural modeling and docking. A supporting evaluation of guaninerich motifs was also performed, but this was interpreted cautiously as a sequence-based structural indication rather than confirmation of G-quadruplex formation. Molecular docking using HADDOCK predicted a favorable interaction between T1 and the ricin B-chain. The best-ranked cluster showed a HADDOCK score of approximately −91.6 ± 3.0 and a Z-score of −2.3. Binding-site analysis indicated that T1 localized near a lectin-proximal region of the ricin B-chain rather than directly occupying the canonical carbohydrate-binding pocket. Residue-level interaction analysis identified Arg24, Arg27, Asn30, Lys40, Lys62, Tyr74, and Tyr78 as important interface residues, suggesting contributions from electrostatic interactions, hydrogen bonding, and surface complementarity. 2 AI-guided candidate comparison was performed using Apta-MCTS-based sequence generation followed by AptaTrans-based scoring. Although several generated variants showed comparable predicted interaction scores, the original T1 aptamer was retained as the final candidate because it combined structural preservation, favorable docking behavior, and reliable predicted interaction potential. Molecular dynamics simulation further supported the stability of the ricin– T1 complex over a 100 ns trajectory, with RMSD, RMSF, radius of gyration, SASA, hydrogenbond, and minimum-distance analyses indicating preservation of the predicted complex. Overall, this study presents a rational in silico framework integrating sequence analysis, secondary structure prediction, truncation, AI-assisted comparison, molecular docking, interaction mapping, and molecular dynamics simulation for ricin-binding aptamer optimization. The optimized T1 aptamer represents a promising computational candidate for future experimental validation and potential development of aptamer-based ricin biosensors. Keywords: Ricin detection; DNA aptamer; ricin B-chain; SELEX; in silico design; aptamer truncation; Apta-MCTS; AptaTrans; HADDOCK; molecular docking; molecular dynamics simulation; guanine-rich motif; biosensor; protein–nucleic acid interaction.