化学学报 ›› 2021, Vol. 79 ›› Issue (4): 472-480.DOI: 10.6023/A20100489 上一篇    下一篇

综述

基于几何约束的蛋白质-配体准确结合自由能计算

付浩浩a, 陈淏川a, 张宏a, 邵学广a,b,*(), 蔡文生a,*()   

  1. a 南开大学化学学院分析科学研究中心 天津市生物传感与分子识别重点实验室 天津 300071
    b 南开大学药物化学生物学国家重点实验室 天津 300071
  • 投稿日期:2020-10-24 发布日期:2020-12-02
  • 通讯作者: 邵学广, 蔡文生
  • 作者简介:

    付浩浩, 本科和博士均毕业于南开大学化学系. 现为南开大学博士后、助理研究员. 研究方向为增强采样算法开发、蛋白质-配体准确结合自由能计算策略研究和复杂体系中高度耦合的运动研究.

    邵学广, 南开大学教授, 博士生导师, 1992年获中国科学技术大学中日联合培养博士学位. 2002 年获教育部第三届高校青年教师奖, 2003 年获国家自然科学基金委杰出青年基金. 主要从事化学计量学及近红外光谱分析方面的研究工作. 建立了小波变换和免疫算法用于复杂信号解析和在线处理的新方法以及一系列用于近红外光谱信号处理和建模的化学计量学方法.

    蔡文生, 南开大学教授, 博士生导师. 1994年获中国科学技术大学博士学位. 主要从事分子模拟与理论化学计算领域的研究工作, 包括优化算法、自由能计算方法、分子模拟及理论化学计算在蛋白质-配体、药物载体、分子机器中的应用研究.

  • 基金资助:
    国家自然科学基金(22073050); 国家自然科学基金(21773125); 南开大学中央高校基本科研业务费专项资金(63191743); 南开大学中央高校基本科研业务费专项资金(63201015); 中国博士后科学基金(bs6619012)

Accurate Estimation of Protein-ligand Binding Free Energies Based on Geometric Restraints

Haohao Fua, Haochuan Chena, Hong Zhanga, Xueguang Shaoa,b,*(), Wensheng Caia,*()   

  1. a Tianjin Key Laboratory of Biosensing and Molecular Recognition, Research Center for Analytical Sciences, College of Chemistry, Nankai University, Tianjin 300071, China
    b State Key Laboratory of Medicinal Chemical Biology, Tianjin 300071, China
  • Received:2020-10-24 Published:2020-12-02
  • Contact: Xueguang Shao, Wensheng Cai
  • About author:
    E-mail: , Tel.: 022-23503430
  • Supported by:
    National Natural Science Foundation of China(22073050); National Natural Science Foundation of China(21773125); Fundamental Research Funds for the Central Universities, Nankai University(63191743); Fundamental Research Funds for the Central Universities, Nankai University(63201015); and the China Post-doctoral Science Foundation(bs6619012)

蛋白质-配体的结合过程伴随着复杂的结构变化, 在分子模拟可及的时间尺度内难以完全捕获, 这使得准确估计蛋白质-配体的结合自由能十分困难. 一种有效的解决途径是采用几何约束减小需要采样的构象空间, 再通过后处理方式扣除约束的影响. 本文综述了三种几何约束策略——漏斗状约束、球形约束和七自由度约束与自由能计算算法结合准确计算结合自由能的原理和进展, 重点概述理论严谨的七自由度约束的最新进展以及与Alchemistry或重要性采样方法的联用策略, 最后, 讨论了如何针对不同体系选择合适的计算策略以及蛋白质-配体准确结合自由能计算在药物设计等领域中的挑战和前景, 并提出了将上述方法进一步运用于研究更复杂的蛋白质-蛋白质问题的可能性.

关键词: 蛋白质-配体, 结合自由能, 分子模拟, 自由能微扰, 重要性采样

Binding free energy is the most crucial physical quantity for describing recognition-association of protein-ligand hybrids. Accurate estimation of protein-ligand binding free energies is of paramount importance in the field of drug design and biological engineering. However, the association process of protein-ligand hybrids is usually coupled with complex conformational changes of molecular objects, which is not amenable to the timescale of classical molecular simulations. This limitation makes it difficult to accurately estimate the protein-ligand binding free energies using classical free-energy calculation strategies. An effective solution is to apply geometric restraints to reduce the configurational space needed to be sampled, so as to boost up the convergence rate of simulations, and then calculate and deduct the contribution of these restraints to the binding free energy by post-processing. In this review, we firstly introduce the recent developments of three geometric restraints, namely, funnel, spherical, and seven-degree-of-freedom restraints, used in accurate binding free-energy calculations, with emphasis on the latest progress of the third one. Specifically, the theoretically rigorous seven-degree-of-freedom restraint describes translational, orientational, rotational, and conformational degrees of freedom by means of a center-of-mass distance, spherical angles, Euler angles and the root-mean-square deviation. Moreover, we demonstrate the theoretical backgrounds and methods of how to achieve accurate protein-ligand binding free-energy estimation by combination of geometric restraints and importance-sampling or alchemical algorithms. In the geometric routes, the degrees of freedom of the relative movement of the protein-ligand complex are addressed in a stepwise fashion by one-dimensional importance-sampling simulations. In the alchemical routes, a special thermodynamics cycle is designed, in which additional simulations are performed to address the contribution of the restraints. A general suggestion for how to choose a suitable strategy for a given molecular assembly based on our experience is provided. Last but not least, we discuss the applications and challenges of using accurate protein-ligand binding free-energy calculation methods in fields such as drug design, and present the possibility of extending these methods for investigating complex protein-protein interaction.

Key words: protein-ligand, binding free energy, molecular dynamics, free-energy perturbation, importance sampling