The interactions between metal ions and proteins are ubiquitous in biology. The selective binding of metal ions has a variety of regulatory functions. Therefore, there is a need to understand the mechanism of protein-ion binding. The interactions involving metal ions are complicated in nature, where short-range charge-penetration, charge transfer, polarization, and many-body effects all contribute significantly, and a quantitative description of all these interactions is lacking. In addition, it is unclear how well current polarizable force fields can capture these energy terms and whether these polarization models are good enough to describe the many-body effects. In this work, two energy decomposition methods, absolutely localized molecular orbitals and symmetry-adapted perturbation theory, were utilized to study the interactions between Mg²⁺/Ca²⁺ and model compounds for amino acids. Comparison of individual interaction components revealed that while there are significant charge-penetration and charge-transfer effects in Ca complexes, these effects can be captured by the van der Waals (vdW) term in the AMOEBA force field. The electrostatic interaction in Mg complexes is well described by AMOEBA since the charge penetration is small, but the distance-dependent polarization energy is problematic. Many-body effects were shown to be important for protein-ion binding. In the absence of many-body effects, highly charged binding pockets will be over-stabilized, and the pockets will always favor Mg and thus lose selectivity. Therefore, many-body effects must be incorporated in the force field in order to predict the structure and energetics of metalloproteins. Also, the many-body effects of charge transfer in Ca complexes were found to be non-negligible. The absorption of charge-transfer energy into the additive vdW term was a main source of error for the AMOEBA many-body interaction energies.

中文翻译：

---

通过能量分解分析和AMOEBA力场研究金属离子与蛋白质模型化合物之间的相互作用

金属离子和蛋白质之间的相互作用在生物学中无处不在。金属离子的选择性结合具有多种调节功能。因此，需要了解蛋白质离子结合的机制。涉及金属离子的相互作用在本质上是复杂的，其中短程电荷渗透，电荷转移，极化和多体效应均起重要作用，并且缺乏对所有这些相互作用的定量描述。此外，目前尚不清楚当前的极化力场能否很好地捕获这些能量项，以及这些极化模型是否足以描述多体效应。在这项工作中，利用两种能量分解方法，绝对局部分子轨道和对称适应性扰动理论，研究了镁之间的相互作用。²⁺ /钙²⁺和氨基酸模型化合物。单个相互作用成分的比较表明，尽管Ca络合物具有显着的电荷渗透和电荷转移效应，但这些效应可以通过AMOEBA力场中的范德华（vdW）项来捕获。由于电荷渗透小，因此AMOEBA很好地描述了Mg络合物中的静电相互作用，但与距离相关的极化能量存在问题。已显示多体效应对蛋白质离子结合很重要。在没有多体效应的情况下，带高电荷的结合口袋将过度稳定，并且口袋将始终偏爱镁，因此会失去选择性。因此，为了预测金属蛋白的结构和能量，必须在力场中纳入多体效应。还，发现Ca络合物中电荷转移的多体效应是不可忽略的。电荷转移能吸收到加性vdW项中是AMOEBA多体相互作用能的主要误差来源。