Department of Biophysics, Graduate School of Science, Kyoto University, Kitashirakawa-Oiwakecho, Sakyo-ku, Kyoto 606-8502, Japan; Bioinformatics Institute, Agency for Science, Technology and Research (A*STAR), 30 Biopolis Str., #07-01 Matrix, 138671, Singapore.
Delft University of Technology, Department of Biotechnology, Van der Maasweg 9, Delft 2629HZ, the Netherlands.
Biochim Biophys Acta Gen Subj. 2021 Jan;1865(1):129766. doi: 10.1016/j.bbagen.2020.129766. Epub 2020 Oct 15.
Prediction of ligand binding and design of new function in enzymes is a time-consuming and expensive process. Crystallography gives the impression that proteins adopt a fixed shape, yet enzymes are functionally dynamic. Molecular dynamics offers the possibility of probing protein movement while predicting ligand binding. Accordingly, we choose the bacterial FF ATP synthase ε subunit to unravel why ATP affinity by ε subunits from Bacillus subtilis and Bacillus PS3 differs ~500-fold, despite sharing identical sequences at the ATP-binding site.
We first used the Bacillus PS3 ε subunit structure to model the B. subtilis ε subunit structure and used this to explore the utility of molecular dynamics (MD) simulations to predict the influence of residues outside the ATP binding site. To verify the MD predictions, point mutants were made and ATP binding studies were employed.
MD simulations predicted that E102 in the B. subtilis ε subunit, outside of the ATP binding site, influences ATP binding affinity. Engineering E102 to alanine or arginine revealed a ~10 or ~54 fold increase in ATP binding, respectively, confirming the MD prediction that E102 drastically influences ATP binding affinity.
These findings reveal how MD can predict how changes in the "second shell" residues around substrate binding sites influence affinity in simple protein structures. Our results reveal why seemingly identical ε subunits in different ATP synthases have radically different ATP binding affinities.
This study may lead to greater utility of molecular dynamics as a tool for protein design and exploration of protein design and function.
预测配体结合并设计新的酶功能是一个耗时且昂贵的过程。晶体学给人一种蛋白质采用固定形状的印象，但实际上酶是具有功能动态性的。分子动力学提供了探测蛋白质运动并预测配体结合的可能性。因此，我们选择细菌 FF ATP 合酶 ε 亚基来揭示为什么来自枯草芽孢杆菌和 PS3 芽孢杆菌的 ε 亚基的 ATP 亲和力差异约 500 倍，尽管它们在 ATP 结合位点具有相同的序列。
我们首先使用 PS3 芽孢杆菌 ε 亚基结构来模拟枯草芽孢杆菌 ε 亚基结构，并利用该结构来探索分子动力学（MD）模拟预测 ATP 结合位点以外残基影响的实用性。为了验证 MD 预测，我们进行了点突变并进行了 ATP 结合研究。
MD 模拟预测，枯草芽孢杆菌 ε 亚基中位于 ATP 结合位点之外的 E102 影响 ATP 结合亲和力。将 E102 工程化为丙氨酸或精氨酸分别导致 ATP 结合增加约 10 或 54 倍，证实了 MD 预测，即 E102 极大地影响 ATP 结合亲和力。
这些发现揭示了 MD 如何预测底物结合位点周围的“第二壳层”残基的变化如何影响简单蛋白质结构中的亲和力。我们的结果揭示了为什么在不同的 ATP 合酶中看似相同的 ε 亚基具有截然不同的 ATP 结合亲和力。
这项研究可能会导致分子动力学作为蛋白质设计工具和探索蛋白质设计和功能的更大用途。
