Chiricotto Mara, Melchionna Simone, Derreumaux Philippe, Sterpone Fabio
Laboratoire de Biochimie Théorique, IBPC, CNRS UPR9080, University Paris Diderot, Sorbonne Paris Cité, 13 rue Pierre et Marie Curie, 75005 Paris, France.
CNR-ISC, Institute for Complex System, Consiglio Nazionale delle Ricerche, Rome, Italy.
J Chem Phys. 2016 Jul 21;145(3):035102. doi: 10.1063/1.4958323.
Computer simulations based on simplified representations are routinely used to explore the early steps of amyloid aggregation. However, when protein models with implicit solvent are employed, these simulations miss the effect of solvent induced correlations on the aggregation kinetics and lifetimes of metastable states. In this work, we apply the multi-scale Lattice Boltzmann Molecular Dynamics technique (LBMD) to investigate the initial aggregation phases of the amyloid Aβ16-22 peptide. LBMD includes naturally hydrodynamic interactions (HIs) via a kinetic on-lattice representation of the fluid kinetics. The peptides are represented by the flexible OPEP coarse-grained force field. First, we have tuned the essential parameters that control the coupling between the molecular and fluid evolutions in order to reproduce the experimental diffusivity of elementary species. The method is then deployed to investigate the effect of HIs on the aggregation of 100 and 1000 Aβ16-22 peptides. We show that HIs clearly impact the aggregation process and the fluctuations of the oligomer sizes by favouring the fusion and exchange dynamics of oligomers between aggregates. HIs also guide the growth of the leading largest cluster. For the 100 Aβ16-22 peptide system, the simulation of ∼300 ns allowed us to observe the transition from ellipsoidal assemblies to an elongated and slightly twisted aggregate involving almost the totality of the peptides. For the 1000 Aβ16-22 peptides, a system of unprecedented size at quasi-atomistic resolution, we were able to explore a branched disordered fibril-like structure that has never been described by other computer simulations, but has been observed experimentally.
基于简化表示的计算机模拟经常用于探索淀粉样蛋白聚集的早期步骤。然而，当使用具有隐式溶剂的蛋白质模型时，这些模拟会忽略溶剂诱导的相关性对亚稳态聚集动力学和寿命的影响。在这项工作中，我们应用多尺度格子玻尔兹曼分子动力学技术（LBMD）来研究淀粉样蛋白Aβ16 - 22肽的初始聚集阶段。LBMD通过流体动力学的动力学晶格表示自然地包含了流体动力学相互作用（HIs）。肽由灵活的OPEP粗粒化力场表示。首先，我们调整了控制分子和流体演化之间耦合的基本参数，以重现基本物种的实验扩散率。然后将该方法用于研究HIs对100个和1000个Aβ16 - 22肽聚集的影响。我们表明，HIs通过促进聚集体之间寡聚体的融合和交换动力学，明显影响聚集过程和寡聚体大小的波动。HIs还引导最大主导簇的生长。对于100个Aβ16 - 22肽系统，约300 ns的模拟使我们能够观察到从椭圆形组装体到几乎涉及所有肽的细长且略有扭曲的聚集体的转变。对于1000个Aβ16 - 22肽，这是一个准原子分辨率下前所未有的大小的系统，我们能够探索一种分支无序的纤维状结构，这种结构从未被其他计算机模拟描述过，但已在实验中观察到。
