Rv1899c, a previously identified HDAC1–ZBTB25-interacting protein of Mycobacterium tuberculosis, plays a crucial role in bacterial adaptation and immune modulation. Recombinant M. smegmatis-expressing Rv1899c (MS_ Rv1899c) showed enhanced survival under acidic and oxidative stress compared to vector controls, along with improved early intracellular growth in THP1-derived macrophages. This was accompanied by reduced reactive oxygen species (ROS), diminished cytokines associated with inflammation and downregulation of autophagy proteins ATG5, Beclin, and LC3, which ultimately skewed the immune response, suppressing the pro-inflammatory M1 macrophage population. Targeting Rv1899c with 3-aminobenzamide (3-AB) impaired intracellular bacterial survival and restored IL-12B expression, while its combination with the HDAC inhibitor C1994 significantly enhanced bacterial clearance. Structural modelling confirmed the high stereochemical quality of the Rv1899c macrodomain, and computational studies identified 3-AB as the strongest ligand (−5.75 kcal/mol), stabilized through hydrogen bonding and hydrophobic interactions with key residues. Molecular dynamics simulations conducted for 200 ns demonstrated stable protein–ligand interactions with consistent parameters, while MM/GBSA analysis indicated favourable binding energy (ΔG_bind = −6.6 kcal/mol), largely influenced by van der Waals and electrostatic forces. Together, these findings highlight Rv1899c as a mediator of stress resistance and immune evasion and propose it as a potential therapeutic target against M. tuberculosis.

SUMOylation is a critical post-translational modification pathway implicated in various ailments, such as diabetes, neurodegenerative disorders, infectious diseases, and cancer. Dysregulation of SUMOylation affects different hallmarks of cancer, including metastasis, angiogenesis, invasion, and proliferation, making it an important therapeutic target. While anacardic acid (AA) has been shown to inhibit the essential SUMO E1 enzyme (IC50 = 2.1 μM), its clinical potential is limited by high lipophilicity and poor bioavailability, stemming primarily from its C15 alkyl chain. This study is therefore aimed at identifying AA derivatives with improved druggability. Using Mcule, 129 AA derivatives were generated and subjected to molecular docking against SUMO E1, which yielded 61 potential inhibitors based on binding energy. Subsequent virtual screening with SwissADME evaluated their pharmacokinetic properties (solubility, logP, and toxicity), narrowing the candidates to 24 hits. Further analysis identified 2-hydroxy-6-methylbenzoic acid as the top lead compound, demonstrating a superior pharmacokinetic profile. Molecular dynamics simulation studies confirmed its stable binding to SUMO E1 through favorable hydrogen bonding. Previous studies from our laboratory have shown AA to regulate the EGF pathway by inhibiting MMPs directly as well as through its modulators, BSG and TIMP. Based on these observations, GPS SUMO analysis demonstrated potential SUMOylation sites in both BSG and TIMP, which have the potential to be modulated by the lead compound, resulting in regulation of the EGF pathway, thereby controlling cancer progression. These observations strongly support the development of 2-hydroxy-6-methylbenzoic acid and its analogues as potent SUMO E1 inhibitors with enhanced pharmacokinetic profiles for cancer therapy.

Solubility of a chemical compound is highly dependent on both its molecular structure and the experimental conditions at which the solubility is measured, which makes it a challenging task to accurately predict solubility across aqueous and organic media. In this study, an interpretable machine learning framework was developed for predicting the solubility of drug-like compounds using large-scale curated data sets, including AqSolDB, AqSolDBc, BigSolDB, and BigSolDB 2.0. CatBoost models were trained and rigorously validated using repeated 5-fold cross-validation, employing scaffold-based splitting for aqueous solubility and cold solute–solvent pair splitting for organic solubility. Feature selection and hyperparameter tuning were systematically applied, with hyperparameter optimization emerging as the primary contributor to performance improvement. The optimized models demonstrated strong and statistically significant performance improvement over baseline configurations. Similarity-based domain-of-applicability analysis on the external data set revealed that prediction error increases with structural dissimilarity from the training set. SHAP analysis provided mechanistic insights, showing that aqueous solubility is primarily governed by polarity and hydrogen-bonding descriptors, whereas organic solubility is influenced by solvent characteristics, temperature, and molecular topology. The consistency of results across data sets demonstrates the robustness and transferability of the learned structure–property relationships and the applicability of the proposed approach for real-world solubility prediction tasks.