The self-assembly of the HIV viral capsid is a large-scale macromolecular process that unfolds over extended timescales. Key questions, including the nucleation mechanism, assembly pathways and intermediates, mechanisms of pentamer incorporation, and capsid protein conformational heterogeneity, remain poorly understood. These features are difficult to probe experimentally and are largely inaccessible to all-atom molecular dynamics (MD) simulations due to their size, dynamics, complexity, and computational cost. Coarse-grained (CG) MD offers a practical alternative, but conventional models often lose dynamical fidelity when reducing degrees of freedom. We introduce a moment of inertia-based coarse-grained (MI-CG) method that strikes a balance between efficiency and physical accuracy. MI-CG reduces fine-grained rigid-body models by selecting minimal peripheral sites via convex hulls and adjusting bead masses through constrained minimization to preserve mass and moment of inertia distributions. Applied to HIV capsid assembly, MI-CG achieves substantial computational gains without loss of key structural or dynamical features. These simulations capture nucleation pathways, intermediate dynamics, conformational heterogeneity, and the mechanisms of pentamer incorporation underlying capsid self-assembly. Together, these results provide new insights into the molecular principles of HIV capsid self-assembly while demonstrating MI-CG as a generalizable framework for simulating large biological processes.
Dash et al. (Sun,) studied this question.