Department of Biochemistry & Molecular Biology, University of Iowa, Iowa City, Iowa 52242, United States.
J Chem Theory Comput. 2023 Aug 8;19(15):5099-5111. doi: 10.1021/acs.jctc.3c00476. Epub 2023 Jul 6.
The Brownian dynamics (BD) simulation technique is widely used to model the diffusive and conformational dynamics of complex systems comprising biological macromolecules. For the diffusive properties of macromolecules to be described correctly by BD simulations, it is necessary to include hydrodynamic interactions (HIs). When modeled at the Rotne-Prager-Yamakawa (RPY) level of theory, for example, the translational and rotational diffusion coefficients of isolated macromolecules can be accurately reproduced; when HIs are neglected, however, diffusion coefficients can be underestimated by an order of magnitude or more. The principal drawback to the inclusion of HIs in BD simulations is their computational expense, and several previous studies have sought to accelerate their modeling by developing fast approximations for the calculation of the correlated random displacements. Here, we explore the use of an alternative way to accelerate the calculation of HIs, i.e., by replacing the full RPY tensor with an orientationally averaged (OA) version which retains the distance dependence of the HIs but averages out their orientational dependence. We seek here to determine whether such an approximation can be justified in application to the modeling of typical proteins and RNAs. We show that the use of an OA-RPY tensor allows translational diffusion of macromolecules to be modeled with very high accuracy at the cost of rotational diffusion being underestimated by ∼25%. We show that this finding is independent of the type of macromolecule simulated and the level of structural resolution employed in the models. We also show, however, that these results are critically dependent on the inclusion of a non-zero term that describes the divergence of the diffusion tensor: when this term is omitted from simulations that use the OA-RPY model, unfolded macromolecules undergo rapid collapse. Our results indicate that the orientationally averaged RPY tensor is likely to be a useful, fast, approximate way of including HIs in BD simulations of intermediate-scale systems.
布朗动力学（BD）模拟技术被广泛用于模拟包含生物大分子的复杂系统的扩散和构象动力学。为了使 BD 模拟正确描述大分子的扩散性质，有必要包括流体动力相互作用（HIs）。例如，当在 Rotne-Prager-Yamakawa（RPY）理论水平上建模时，可以准确地再现孤立大分子的平移和旋转扩散系数；然而，当忽略 HIs 时，扩散系数可能会低估一个数量级或更多。在 BD 模拟中包含 HIs 的主要缺点是其计算费用，并且之前的几项研究试图通过开发用于计算相关随机位移的快速逼近来加速其建模。在这里，我们探索了使用另一种方法来加速 HIs 的计算，即通过用具有保留 HIs 的距离依赖性但平均其方向依赖性的各向同性平均（OA）版本替换完整的 RPY 张量。我们在这里寻求确定这种近似在典型蛋白质和 RNA 的建模中的应用是否合理。我们表明，使用 OA-RPY 张量可以以非常高的精度模拟大分子的平移扩散，而旋转扩散的估计值低估了约 25％。我们表明，这一发现独立于模拟的大分子类型和模型中使用的结构分辨率水平。然而，我们还表明，这些结果严重依赖于包括描述扩散张量散度的非零项：当从使用 OA-RPY 模型的模拟中省略该项时，未折叠的大分子会迅速坍塌。我们的结果表明，各向同性平均 RPY 张量可能是在 BD 模拟中等规模系统中包含 HIs 的一种有用，快速，近似方法。
