Coarse-grained force field parameters for natural product-based small molecules: from structural mapping to pharmaceutical applications

Abstract

Context

Natural product-derived bioactive small molecules represent a major source of lead compounds for drug discovery and pharmaceutical innovation. The performance of drug delivery systems is fundamentally governed by intermolecular noncovalent interactions, yet their multiscale characterization remains computationally challenging. All-atom molecular dynamics (AA-MD) simulations provide atomic-level mechanistic resolution but are limited by their high computational cost, whereas coarse-grained molecular dynamics (CG-MD) simulations offer improved computational efficiency but lack standardized and transferable parameters for structurally diverse natural product-derived small molecules. In this study, we developed a Martini 3-based CG modeling framework for natural product-derived small molecules. Coarse-grained mapping rules were established for four representative scaffolds, including phenylpropanoids, quinones, flavonoids, and terpenoids, and a CG force-field parameter set was developed for 24 representative bioactive natural products. The models and parameters were systematically validated using two complementary criteria based on structural descriptors and energetic metrics. The resulting CG models were further applied to investigate β-cyclodextrin inclusion mechanisms and predict transdermal penetration enhancement effects, which successfully reproduced the experimentally observed macroscopic trends. These results demonstrate the reliability, transferability, and practical utility of the developed parameter set. Overall, this work established a standardized CG simulation strategy for natural product-derived small molecules and provided efficient computational guidance for the rational design and optimization of drug delivery systems for natural product-derived drugs.

Methods

All-atom simulations were performed using the GROMOS 54a7 force field. Coarse-grained models were subsequently constructed based on the Martini 3.0 force field framework, and bonded parameters were obtained through iterative optimization using the Swarm-CG software. Structural rationality was validated by comparing the radius of gyration (Rg), solvent-accessible surface area (SASA), and distributions of bond lengths, bond angles, and dihedral angles between CG and AA models. Thermodynamic properties were validated by calculating the solvation free energy (ΔGₛₒₗ) using the free energy perturbation method. Finally, drug delivery system models were established to explore β-cyclodextrin inclusion mechanisms and predict the penetration-enhancing effects of transdermal penetration enhancers, thereby evaluating the applicability of the developed models. The reliability of the constructed systems, workflow, and model parameters was assessed by comparing the simulated encapsulation and penetration-enhancement efficiencies with the corresponding experimental values.