Ouyang Yanhua, Zhao Likun, Zhang Zhuqing
College of Life Science, University of Chinese Academy of Sciences, Beijing, 100049, China.
Phys Chem Chem Phys. 2018 Mar 28;20(13):8676-8684. doi: 10.1039/c8cp00067k.
Intrinsically disordered regions (IDRs) or proteins (IDPs), which play crucial biological functions in essential biological processes of life, do not have well-defined secondary or tertiary structures when isolated in solution. The highly dynamic properties and conformational heterogeneity of IDPs make them challenging to study with traditional experimental techniques. As a powerful complementary tool for experiments, all-atom molecular dynamics simulation can obtain detailed conformational information on IDPs, but the limitation of force field accuracy is a challenge for reproducing IDP conformers. Here, we compared five empirical all-atom force fields AMBER03, AMBER99SB-ILDN, CHARMM27, OPLS-AA/L and CHARMM36m in modeling the conformational ensembles of wild-type peptide TAD2(41-62) from the human p53 tumor suppressor. Our results show that for the model peptide, the newest force field CHARMM36m produces more expanded coil ensemble followed by AMBER99SB-ILDN; CHARMM27 displays a predominant propensity for a helical structure; whereas OPLS-AA/L exhibits a apparent preference for a β-sheet structure and yields the most compact conformation. In the comparison of the simulated dimensions with theoretical prediction and the back-calculated chemical shifts with experimental measurements, AMBER99SB-ILDN gives a more consistent agreement than the other force fields. In addition, the region from residues 47 to 55, which commonly forms an amphipathic α-helix upon binding target proteins according to experimental observation, could form a helical structure with a different probability population in our simulations with different force fields. This implies that the binding process might be conducted by, or partly by "conformation selection" for this peptide. This work indicates that force field development for modeling general IDPs accurately has a long way to go, and more detailed experimental data of IDPs are also in demand.
内在无序区域(IDRs)或蛋白质(IDPs)在生命的基本生物学过程中发挥着关键的生物学功能,当它们在溶液中单独存在时,没有明确的二级或三级结构。IDPs的高度动态特性和构象异质性使得用传统实验技术研究它们具有挑战性。作为实验的有力补充工具,全原子分子动力学模拟可以获得IDPs的详细构象信息,但力场精度的限制是再现IDP构象的一个挑战。在这里,我们比较了五个经验性全原子力场AMBER03、AMBER99SB - ILDN、CHARMM27、OPLS - AA/L和CHARMM36m在模拟人类p53肿瘤抑制因子野生型肽TAD2(41 - 62)的构象集合方面的表现。我们的结果表明,对于模型肽,最新的力场CHARMM36m产生的卷曲构象集合更为扩展,其次是AMBER99SB - ILDN;CHARMM27表现出形成螺旋结构的主要倾向;而OPLS - AA/L表现出对β - 折叠结构的明显偏好,并产生最紧凑的构象。在将模拟尺寸与理论预测以及将反算化学位移与实验测量进行比较时,AMBER99SB - ILDN比其他力场给出了更一致的结果。此外,根据实验观察,通常在结合靶蛋白时形成两亲性α - 螺旋的47至55位残基区域,在我们使用不同力场的模拟中可能以不同的概率群体形成螺旋结构。这意味着该肽的结合过程可能是通过“构象选择”进行的,或者部分是通过“构象选择”进行的。这项工作表明,准确模拟一般IDPs的力场开发还有很长的路要走,同时也需要更多IDPs的详细实验数据。