Analytical and machine learning approaches identify a sea star steroid with promising activity for COVID-19 therapeutic development

Abstract

The pressing demand for safe and efficient COVID-19 treatments has intensified interest in Natural products, especially those derived from marine organisms. In this study, a bioactive steroidal compound, 5α-cholesta-9(11)-en-3β,20β-diol, was successfully isolated from the starfish Acanthaster planci. Structural elucidation was achieved using HREIMS, FTIR, and advanced 1D/2D NMR spectroscopy, confirming a molecular formula of C₂₇H₄₆O₂ and characteristic functionalities including hydroxyl and double bond moieties. The compound demonstrated notable anti-SARS-CoV-2 activity, attaining 85% viral inhibition at 5 ng/µl with an IC₅₀ of 5.86 µM, as demonstrated by plaque reduction assays. Molecular docking studies demonstrated significant binding affinities toward key viral targets Mpro, NSP10, and RNA-dependent RNA polymerase with free binding energies of -26.85, -27.59, and - 35.08 kcal/mol, respectively. These affinities surpassed those of their respective co-crystallized reference ligands. In-silico ADMET profiling indicated favorable pharmacokinetic properties, including high BBB penetration, moderate intestinal absorption, and non-hepatotoxicity. Toxicity assessments predicted low carcinogenic risk, a high rat MTD, and minimal ocular and dermal irritancy. Additionally, we developed a predictive web application based on machine learning to estimate IC₅₀ values of SARS-CoV-2 inhibitors, streamlining the drug discovery process. The forecasted values nearly matched the experimental outcomes, demonstrating the model's reliability and its potential to reduce time, cost, and risk in early-stage drug development. Moreover, machine learning models, particularly XGBoost, demonstrated excellent performance in predicting pIC₅₀ values (RMSE = 0.1357, MAE = 0.1022), supporting the development of a web-based IC₅₀ prediction application. The bioactivity prediction platform ENHPCG further validated the compound's antiviral potential, estimating an IC₅₀ of 5.95 µM. Overall, these integrated analytical, biological, and computational approaches highlight 5α-cholesta-9(11)-en-3β,20β-diol as a potential SARS-CoV-2 inhibitor and a candidate for further pharmacological development.

Keywords: COVID-19 inhibitor; Drug discovery; Machine learning; Marine natural products; Molecular docking.