锌蛋白分子力场的开发与应用

Zinc proteins constitute approximately 10% of the total human proteome and play a variety of essential biological roles, and there is an increasing appreciation of their biological and medical importance. However, modeling the Zn-Ligand coordination interaction in zinc-containing proteins is a well-known challenge due to the inherent flexibility feature of zinc coordination structure. The accuracy and transferability of all the current availabe pairwise force fields to describe zinc coordination structure are very poor..This proposal is directed at developing a novel transferable force fields to model zinc proteins and to study the interaction between zinc protein and small molecules. Very rencently a novel Short-Long Effective Function (SLEF) strategy was designed by us to describe charge-charge interactions between divalent zinc cation and all other atoms, then a non-bonded and amber-like pairwise atomic force field Amber99SB-SLEF1 was developed. In this proposal, the short-long effective function will be redesigned, then a much more robust training set and fitting protocal will be set up, and the parameters will be optimized based on Quantum Mechanics/Molecular Mechanics Force Matching (QM/MM FM) method, finally to promote the performance of SLEF force field. Moreover, some selected zinc protein systems will be studied by combining this improved SLEF force field (SLEF2) and other simulation methods such as the Docking, MM-PBSA and Free Energy Perturbation (FEP).

锌蛋白在人类蛋白组中占到了～10%，其各种生物功能的发挥往往需要依赖锌结合位点。然而锌配位结构往往表现出复杂多变的特征，它的配体种类较多，配位模式也多样化。这就使得金属锌蛋白的计算模拟成为当前分子力场方法所面临的重大挑战之一。因此针对锌蛋白的分子力场的开发既有重要的基础研究意义又有广泛的应用前景。. 本项目旨在开发出具有普适性的锌蛋白力场并开展应用研究。最近我们原创性地提出用"短程-长程有效函数（SLEF)"方法来描述锌配位结构。在此基础上，我们将重新设计SLEF函数形式，优化参数拟合方案，精心设计训练集和测试集，结合量子-分子力学组合计算与力匹配方法来拟合得到力场参数，力争开发出在准确性、可移植性方面有优良表现的新一代SLEF力场，并结合Docking、MM-PBSA和FEP方法对一些锌蛋白体系开展蛋白-小分子相互作用方面的研究。

由于锌配位结构内在柔性（可变性，flexibility）高，有多种配位模式，金属锌蛋白的计算模拟一直是分子力学方法所面临的挑战之一。前期我们提出了用“短程-长程有效函数”来描述锌配位结构的方法，获得了第一代SLEF力场(SLEF1)。在此基础上，我们重新设计了全新的SLEF函数形式，随后精心设计了更加完备的训练集和测试集，并结合量子-分子力学组合计算(QM/MM)与力匹配方法(Force Matching)拟合得到了新一代SLEF函数中的4个待拟合参数，从而开发出了在准确性、可移植性方面有优良表现的新一代SLEF力场(简称SLEF2)，并以开源的方式写入了Amber分子动力学程序中，随后应用到一系列锌蛋白体系的测试，我们共测试了16个锌酶体系(含26个锌配位结构：既有单核的锌，也有各种双核\三核模式下的锌配位结构)。模拟结果显示SLEF2表现出了很好的可移植性，能重现晶体结构的各种配位结合模式(配位数和配位键长)，特别是对Zn-Asp， Zn-S, Zn-Zn等都能描述得较为准确，而这些往往在SLEF1或者Amber99SB力场下是描述不好的。近3年，在项目资助下，我们将SLEF2方法还应用到了一些锌蛋白体系蛋白-小分子相互作用机制和抑制剂设计研究中，共计发表SCI论文13篇。