Protonation states of titratable amino acids play a key role in many biomolecular processes. Knowledge of protonatable residue charges at a given pH is essential for a correct understanding of protein catalysis, inter- and intramolecular interactions, substrate binding, and protein dynamics for instance. However, acquiring experimental values for individual amino acid protonation states of complex systems is not straightforward; therefore, several in silico approaches have been developed to tackle this issue. In this work, we assess the accuracy of our previously developed constant pH MD approach by comparing our theoretically obtained pK_a values for titratable residues with experimental values from an equivalent NMR study. We selected a set of four pentapeptides, of adequately small size to ensure comprehensive sampling, but concurrently, due to their charge composition, posing a challenge for protonation state calculation. The comparison of the pK_a values shows good agreement of the experimental and the theoretical approach with a largest difference of 0.25 pK_a units. Further, the corresponding titration curves are in fair agreement, although the shift of the Hill coefficient from a value of 1 was not always reproduced in simulations. The phase space overlap in Cartesian space between trajectories generated in constant pH and standard MD simulations is fair and suggests that our constant pH MD approach reasonably well preserves the dynamics of the system, allowing dynamic protonation MD simulations without introducing structural artifacts.

中文翻译：

---

探究肽段的显式溶剂恒定pH分子动力学模拟的准确性。

可滴定氨基酸的质子化状态在许多生物分子过程中起着关键作用。例如，对于正确理解蛋白质催化，分子间和分子间的相互作用，底物结合和蛋白质动力学，了解在给定pH值下质子化残基电荷至关重要。但是，获取复杂系统中各个氨基酸质子化状态的实验值并不容易。因此，已经开发了多种计算机方法来解决此问题。在这项工作中，我们通过比较理论上获得的p K _a来评估先前开发的恒定pH MD方法的准确性。可滴定残留物的平均值与同等NMR研究的实验值。我们选择了一组四种五肽，大小足够小以确保全面采样，但是同时，由于它们的电荷组成，对质子化态的计算提出了挑战。p K _a值的比较表明实验方法和理论方法吻合良好，最大差异为0.25 p K _a单位。此外，尽管在模拟中并不总是再现希尔系数从1值的偏移，但相应的滴定曲线也相当吻合。在恒定pH和标准MD模拟中生成的轨迹之间的笛卡尔空间中的相空间重叠是合理的，这表明我们的恒定pH MD方法合理地很好地保留了系统动力学，从而允许在不引入结构伪影的情况下进行动态质子化MD模拟。