Amyloid formation by proteins and peptides is the hallmark of many diseases. Growing evidence suggests that oligomeric species arising during aggregation are toxic, but the molecular mechanism of aggregation and oligomer generation remains unclear. Recent discoveries that amyloid fibrils can convert soluble proteins into oligomeric nuclei to facilitate aggregation highlight the role played by fibrils in protein aggregation. We review here computational studies conducted to elucidate the molecular mechanism of two fibril-dependent processes during protein aggregation, namely, secondary nucleation and fibril elongation. Secondary nucleation occurs on the lateral surface of a fibril to generate nuclei while fibril elongation, through addition of proteins to the ends of fibrils increases the lateral surface of the fibril. We summarize the molecular insights into each process unraveled by computational methods at levels ranging from coarse-grained to atomic and discuss the connection between these insights and experimental observations. The computational challenges faced by these studies and their solutions are also discussed. Finally, we propose possible computational studies that could shed light on the mechanistic aspects of secondary nucleation and fibril elongation that have been unaddressed.

由蛋白质和肽形成的淀粉样蛋白是许多疾病的标志。越来越多的证据表明，聚集过程中产生的低聚物种具有毒性，但聚集和低聚物生成的分子机制仍不清楚。淀粉样蛋白原纤维可以将可溶性蛋白转化为寡聚核以促进聚集的最新发现突出了原纤维在蛋白聚集中所起的作用。我们在这里回顾进行的计算研究，以阐明蛋白质聚集过程中两个原纤维依赖过程的分子机制，即次级成核和原纤维伸长。二次成核发生在原纤维的侧表面上以产生核，而原纤维的伸长通过向原纤维的末端添加蛋白质而增加了原纤维的侧表面。我们总结了从计算方法到分解过程的分子洞察力，涵盖了从粗糙到原子的各个层次，并讨论了这些洞察力与实验观察之间的联系。还讨论了这些研究及其解决方案所面临的计算挑战。最后，我们提出可能的计算研究，以阐明尚未解决的二次成核和原纤维伸长的机械方面。