D. E. Shaw Research, New York, NY 10036, USA.
J Mol Biol. 2011 Jan 7;405(1):43-8. doi: 10.1016/j.jmb.2010.10.023. Epub 2010 Oct 23.
One of the most important and elusive goals of molecular biology is the formulation of a detailed, atomic-level understanding of the process of protein folding. Fast-folding proteins with low free-energy barriers have proved to be particularly productive objects of investigation in this context, but the design of fast-folding proteins was previously driven largely by experiment. Dramatic advances in the attainable length of molecular dynamics simulations have allowed us to characterize in atomic-level detail the folding mechanism of the fast-folding all-β WW domain FiP35. In the work reported here, we applied the biophysical insights gained from these studies to computationally design an even faster-folding variant of FiP35 containing only naturally occurring amino acids. The increased stability and high folding rate predicted by our simulations were subsequently validated by temperature-jump experiments. The experimentally measured folding time was 4.3 μs at 80 °C-about three times faster than the fastest previously known protein with β-sheet content and in good agreement with our prediction. These results provide a compelling demonstration of the potential utility of very long molecular dynamics simulations in redesigning proteins well beyond their evolved stability and folding speed.
蛋白质折叠过程的原子水平详细理解是分子生物学最重要和最难以捉摸的目标之一。在这种情况下,具有低自由能势垒的快速折叠蛋白质已被证明是特别有成效的研究对象,但快速折叠蛋白质的设计以前主要是通过实验驱动的。分子动力学模拟可达长度的显著进步使我们能够在原子水平上详细描述快速折叠全β WW 结构域 FiP35 的折叠机制。在本研究中,我们将从这些研究中获得的生物物理见解应用于计算设计仅包含天然氨基酸的更快折叠 FiP35 变体。我们的模拟预测的增加稳定性和高折叠速率随后通过温度跃变实验得到验证。在 80°C 时,实验测量的折叠时间为 4.3μs,比以前已知的最快具有β-折叠含量的蛋白质快约三倍,与我们的预测结果吻合良好。这些结果有力地证明了非常长的分子动力学模拟在重新设计蛋白质方面的潜在效用,远远超出了它们的进化稳定性和折叠速度。