Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003, United States.
J Chem Inf Model. 2022 Sep 26;62(18):4523-4536. doi: 10.1021/acs.jcim.2c00974. Epub 2022 Sep 9.
Intrinsically disordered proteins (IDPs) play crucial roles in cellular regulatory networks and are now recognized to often remain highly dynamic even in specific interactions and assemblies. Accurate description of these dynamic interactions is extremely challenging using atomistic simulations because of the prohibitive computational cost. Efficient coarse-grained approaches could offer an effective solution to overcome this bottleneck if they could provide an accurate description of key local and global properties of IDPs in both unbound and bound states. The recently developed hybrid-resolution (HyRes) protein model has been shown to be capable of providing a semiquantitative description of the secondary structure propensities of IDPs. Here, we show that greatly improved description of global structures and transient interactions can be achieved by introducing a solvent-accessible surface area-based implicit solvent term followed by reoptimization of effective interaction strengths. The new model, termed HyRes II, can semiquantitatively reproduce a wide range of local and global structural properties of a set of IDPs of various lengths and complexities. It can also distinguish the level of compaction between folded proteins and IDPs. In particular, applied to the disordered N-terminal transactivation domain (TAD) of tumor suppressor p53, HyRes II is able to recapitulate various nontrivial structural properties compared to experimental results, some of them to a level of accuracy that is almost comparable to results from atomistic explicit solvent simulations. Furthermore, we demonstrate that HyRes II can be used to simulate the dynamic interactions of TAD with the DNA-binding domain of p53, generating structural ensembles that are highly consistent with existing NMR data. We anticipate that HyRes II will provide an efficient and relatively reliable tool toward accurate coarse-grained simulations of dynamic protein interactions.
无定形蛋白质（IDPs）在细胞调控网络中发挥着关键作用，现在人们认识到，即使在特定的相互作用和组装中，它们通常也保持着高度的动态性。由于计算成本过高，使用原子模拟方法准确描述这些动态相互作用极具挑战性。如果高效的粗粒化方法能够在无键和键合状态下提供对 IDP 的关键局部和全局特性的准确描述，它们可能会提供一种有效的解决方案来克服这一瓶颈。最近开发的混合分辨率（HyRes）蛋白质模型已被证明能够对半定量描述 IDP 的二级结构倾向。在这里，我们展示了通过引入基于溶剂可及表面积的隐溶剂项并重新优化有效相互作用强度，可以大大提高对全局结构和瞬态相互作用的描述能力。新模型称为 HyRes II，它可以对半定量地再现一组各种长度和复杂度的 IDP 的广泛局部和全局结构特性。它还可以区分折叠蛋白和 IDP 之间的紧凑程度。特别是，应用于肿瘤抑制因子 p53 的无规则 N 端转录激活结构域（TAD），与实验结果相比，HyRes II 能够再现各种非平凡的结构特性，其中一些达到了几乎与原子显式溶剂模拟结果相当的精度。此外，我们证明 HyRes II 可用于模拟 TAD 与 p53 的 DNA 结合域之间的动态相互作用，生成与现有 NMR 数据高度一致的结构集合。我们预计 HyRes II 将为准确的粗粒化模拟动态蛋白质相互作用提供一种高效且相对可靠的工具。
