A Computational Study of Gemcitabine Encapsulation in CC3 Porous Organic Cage: Molecular Docking, Molecular Dynamics, Binding Free Energy, and Structure-Activity Relationship Analysis
Gemcitabine, a first-line nucleoside analogue chemotherapy drug, is rapidly inactivated in plasma by cytidine deaminase (CDA), severely limiting its therapeutic efficacy. Porous organic cages (POCs) have emerged as promising drug delivery vehicles due to their discrete, soluble architecture and tunable cavities. Here, we report a comprehensive computational structure-activity relationship (SAR) study of gemcitabine encapsulation across five CC3-based cage variants using molecular docking, molecular dynamics (MD) simulation, and MM-PBSA binding free energy analysis. AutoDock Vina blind docking yielded exclusively endohedral binding (20/20 poses, best ΔG = −5.7 kcal/mol) for pristine CC3. Three independent 200 ns GROMACS MD simulations of pristine CC3 at 300 K reveal stochastic encapsulation-release dynamics consistent with μM-affinity binding (ΔGbind = −13.44 ± 0.41 kcal/mol). A systematic SAR campaign across four additional cage variants establishes critical design rules: N-methyl endo-functionalization (CC3-methyl) eliminates encapsulation by reducing cavity volume; aromatic fluorination (CC3-F12) provides marginal electrostatic improvement but insufficient kinetic stabilization; exo-phenyl functionalization (CC3-exo-phenyl) blocks drug entry entirely; while lowering simulation temperature to 280 K achieves complete, reproducible endohedral encapsulation across all three independent replicas (100% endohedral, mean 2.25 ± 0.25 Å, no exit events in 200 ns each). Probability distribution analysis confirms the transformation from a broad bimodal distribution at 300 K (mean 9.4 Å, 46% endohedral) to a sharp unimodal peak at 280 K (mean 2.25 Å, 100% endohedral). Framework breathing analysis reveals that CC3 is more rigid at 280 K (Rg std = 0.07–0.08 Å) than at 300 K (Rg std = 0.17 Å), consistent with thermally-stabilized encapsulation. These results establish a robust multi-scale computational framework to guide the rational design of porous organic cages for targeted drug delivery applications.