Molecular Docking and Admet of Brucea Javanica Quassinoids on Drug-Resistant Tuberculosis Targets

Imam Prayitno (Universitas Setia Budi)
Rina Herowati (Universitas Setia Budi)
Gunawan Pamudji Widodo (Universitas Setia Budi)

Indonesian Journal of Global Health Research
Vol. 8 No. 5 (2026): Indonesian Journal of Global Health Research

Tuberculosis remains the infectious disease with the highest global burden, and its control is complicated by drug-resistant cases; resistance to the two first-line drugs is largely attributable to the katG S315T and rpoB S531L mutations. Quassinoids of Brucea javanica (L.) Merr. are reported to be antimicrobial, but comparative studies of their interaction with wild-type and mutant target proteins remain limited. To evaluate B. javanica quassinoids as early antitubercular drug candidates through molecular docking and ADMET prediction against KatG and RpoB in both wild-type and mutant forms. An exploratory and predictive in silico study with a comparative design. Twenty-five quassinoids and two controls (isoniazid, rifampicin) were retrieved from PubChem and ChEMBL; target structures were 2CCA (wild-type KatG), 2CCD (KatG-S315T), 5UAC (wild-type RpoB) and 5UAL (RpoB-S531L). Ligands were minimised with the UFF force field and docked with AutoDock Vina in PyRx (exhaustiveness 8, nine poses per combination). The RpoB search space was centred on the co-crystallised rifampicin and validated by redocking, whereas the KatG search space covered the whole enzyme. ADMET was predicted with ADMET-AI against the ATC J04 reference set. On wild-type RpoB all 25 quassinoids bound between −6.7 and −8.6 kcal/mol (mean −7.61 ± 0.52) and 15 matched or exceeded re-prepared rifampicin (−7.5 kcal/mol), although none surpassed the native ligand (−9.2 kcal/mol). S531L weakened every compound (mean ΔE +0.57 kcal/mol), the largest loss occurring in the native ligand itself (+1.20 kcal/mol). On wild-type KatG affinities were weaker and widely dispersed (mean −3.34 ± 3.02 kcal/mol) and none exceeded isoniazid (−6.1 kcal/mol); S315T improved scores (mean ΔE −2.04 kcal/mol) but abolished every contact with the distal heme pocket. Only bruceine D, bruceine H and bruceine I met all four Lipinski criteria, with intestinal absorption ≥ 0.96 and low hERG and cytochrome P450 liability, whereas all sixteen glycosides violated three criteria simultaneously. RpoB is the more promising route for this scaffold; bruceine I is the priority candidate and bruceoside C the glycoside lead. All values are predictive and await experimental confirmation.