Computational Studies of Aqueous Solvation of Proteins

通过该奖项，化学系（CHE）的化学理论、模型和计算方法项目以及分子和细胞生物科学系（MCB）的分子生物物理集群支持乔治城大学的Toshiko Ichiye进行蛋白质水溶液化的计算研究。尽管水是地球上最常见的液体，但其作为液体和溶剂的独特性质的分子基础使其对生命至关重要，目前尚不清楚。虽然人们普遍认为氢键的四面体网络是必不可少的，但水分子中导致氢键的关键特征才刚刚开始被理解。鉴于其独特的液体性质，这并不奇怪，但这使得在计算机模拟中模拟水变得困难。然而，由于使用水溶液中生物大分子的原子表示的计算机模拟已经成为理解许多生物学问题的关键，因此生物大分子及其周围水的准确表示是必不可少的。因此，识别这些特征并确定建模的最佳方法将影响对生物分子的结构和功能的理解，包括蛋白质、核酸、脂质膜及其组装。此外，计算机模拟正被用于理解化学和物理许多领域的科学问题，并且考虑到水作为液体和溶剂的重要性；这些研究将对其他领域产生广泛的影响。此外，实现到一个常用的生物分子模拟包意味着我们的方法将被广泛使用。这项研究旨在发展对水和水溶液的分子理解，并以此为基础，在计算机模拟中为水和蛋白质开发新的原子模型。Ichiye和她的研究小组的方法是通过确定在分子动力学计算机模拟中模拟其独特性质所必需的最小分子特征来理解水，这必须是它的基本分子特征。这种方法被用来理解水是一种纯液体；现在人们正在研究离子水溶液的反常性质。这种模拟水的方法也很独特，因为仅以氧为中心的分子多极被用来表示由于原子核和电子密度而产生的电荷分布，这对于再现水的性质尤其重要。分子多极优于在位点模型中添加代表电子密度的虚拟位点，因为电子不是点粒子。因此，这些模型比4点和5点的水模型更精确，计算速度至少比3点的模型快两倍。这些想法现在将用于开发一个准确而有效的力场，用于离子、共溶质，最终用于蛋白质。分子多极也将被使用：多极将被放置在重原子上，有点像传统的统一原子方法，除了氢和电子密度的静电将通过多极来处理。其目标是开发一种力场，比生物分子最常用的非极化力场更精确，计算速度也快两到三倍。

With this award, the Chemical Theory, Models and Computational Methods program in the Division of Chemistry (CHE) and the Molecular Biophysics Cluster in the Division of Molecular and Cellular Biosciences (MCB) are supporting Toshiko Ichiye of Georgetown University for computational studies of the aqueous solvation of proteins. Although water is the most common liquid on this planet, the molecular bases of its unique properties as a liquid and a solvent that make it critical for life are still unclear. While it is generally accepted that the tetrahedral network of hydrogen bonds is essential, the critical features in a water molecule that lead to hydrogen bonds are just beginning to be understood. Given its unique liquid properties, this is not surprising but it makes modeling water in computer simulations difficult. However, since computer simulations using atomistic representations of biological macromolecules in aqueous solutions have become critical in understanding many biological problems, accurate representations of both the bio-macromolecule and its surrounding water are essential. Therefore, identifying these features and determining the best ways to model them will impact the understanding of structure and function of biological molecules, including proteins, nucleic acids, lipid membranes, and their assemblies. Moreover, computer simulations are being used to understand scientific questions in many areas of chemistry and physics, and given the importance of water as a liquid and solvent; these studies will have a broad impact on other fields. In addition, the implementation into a commonly used biomolecular simulation package means that our methods will be widely available.The study is to develop a molecular understanding of water and aqueous solutions and, based on this, to develop new atomistic models for water and proteins in computer simulations. The approach of Ichiye and her research group is to understand water is by determining the minimal molecular features necessary to model its unique properties in molecular dynamics computer simulations, which must be its essential molecular features. This approach has been used to understand water as a pure liquid; now the anomalous properties of aqueous ionic solutions are being studied. This approach to modeling water is also unique in that molecular multipoles centered only on the oxygen are used to represent the charge distribution due to both the nuclei and the electron density, which is especially important for reproducing the properties of water. Molecular multipoles are superior to adding dummy sites representing electron density in a site model, because electrons are not point particles. Thus, these models are more accurate than four- and five-site water models and at least two times computationally faster than even three-site models. These ideas will now be used to develop an accurate but efficient force field for ions, co-solutes, and eventually proteins. Molecular multipoles will also be used: multipoles will be placed on the heavy atoms somewhat like the traditional united atom approach, except that the electrostatics of both the hydrogens and electron density will be treated via the multipoles. The aim is to develop a force field that is more accurate than the most commonly used non-polarizable force fields for biomolecules and that is also two to three times faster computationally.