All-atom molecular dynamics (MD) simulation is a useful method for studying the relationship between the structure and the function of a protein. It has become a popular method owing to the advancements of hardware, like GPU, and software, such as Gromacs and NAMD. However, all-atom MD simulation has limitations in its applicable time and space scales. While isotropic spherical approximations of protein structures have been applied to overcome the limitations, they ignore the proteins’ shapes and anisotropy in the intermolecular interactions. In this study, we developed a coarse-grained rigid-body MD simulation software called CGRig for studying long-time and large-scale biological phenomena, considering protein shape and interaction anisotropy. In our software, each protein is coarse-grained at an amino-acid level and assumed as a rigid body. Its translational and rotational dynamics are described by the overdamped Langevin equation in which a shape-dependent resistance tensor was applied to a protein molecule. In an isolated ubiquitin system, calculated translational and rotational diffusion coefficients closely agreed with the experimental values, which confirmed the validity of our implementation. In multimolecular systems, short-range and long-range intermolecular interactions are calculated approximately using the force-matching method and the Debye-Hückel electrostatics, respectively. We next evaluated the validity of our intermolecular interaction model on a barnase-barstar complex system and confirmed that the complex structure was stably maintained and the correct association rate was obtained. Finally, we measured the simulation speed using a system comprised of multiple tubulin molecules, anticipating its future applications. We achieved a simulation speed of over 60 μs/day on a 128-molecule system. Thus, CGRig will provide structural insights for cellular-scale biological phenomena, such as microtubule formation and virus capsid formation.
Teshirogi et al. (Sun,) studied this question.