A Simplified Potential for Protein Folding Simulations

描述（由申请人提供）： 生物体的功能很大程度上取决于其每种组成蛋白在生理条件下采用独特的结构（所谓的天然结构）；这种结构是由氨基酸序列及其环境决定的。根据安芬森的热力学假设，蛋白质的天然构象是其自由能超曲面上的全局能量最小值。因此，如果精确的势能函数可用，则可以将自然结构寻求为该超曲面中的全局最小值。蛋白质自由能超表面的形状非常复杂且难以描述。出于效率原因，必须使用多肽链的简化模型，其中每个氨基酸残基由一个或几个相互作用位点代表，而不是全原子分辨率模型。虽然我们之前的重点是全局优化方法，但现在重点是基于物理的联合残基 UNRES 势能函数。主要目标是仅基于势能的全局优化，预测链长最多为 200 个氨基酸残基且均方根偏差在 4-6 埃以内的所有主要结构类别（α、β、αβ 和 α/β）的蛋白质结构。这将通过改进各个能量成分的函数形式和参数，并通过优化总能量函数以反映部分展开结构的能量层次来确保势的折叠特性来实现。了解物理相互作用在蛋白质天然结构形成中的作用将使我们不仅能够仅根据氨基酸序列的知识来预测蛋白质的最终结构，而且可用于研究蛋白质折叠或错误折叠过程。预测蛋白质三维结构或预测其折叠途径的能力可以极大地有助于针对癌症、阿尔茨海默病或朊病毒疾病的合理药物设计。

DESCRIPTION (provided by applicant): The functioning of living organisms is largely dependent on the fact that each of its constituent proteins adopts a unique structure (the so-called native structure) under physiological conditions; this structure is determined by amino-acid sequence and its environment. According to Anfinsen's thermodynamic hypothesis, the native conformation of a protein is a global-energy minimum on its free-energy hypersurface. The native structure can therefore be sought as the global minimum in this hypersurface, if an accurate potential energy function is available. The shape of the free-energy hypersurface of proteins is very complex and difficult to describe. For efficiency reasons, simplified models of polypeptide chains, in which each amino-acid residue is represented by one or a few interaction sites rather than all-atom resolution models, must be used. While our previous focus was on global optimization methods, it is now focused on our physics-based united-residue UNRES potential-energy function. The main goal is to predict the structure of proteins of all major structural classes (alpha, beta, alpha+beta and alpha/beta) with chain lengths of up to 200 amino-acid residues within 4-6 Angstrom root mean square deviation based solely on global optimization of the potential energy. This will be accomplished by improving the functional forms and parameters of individual energy components and assuring the folding property of the potential by optimizing the total energy function to reflect the energetic hierarchy of partially unfolded structures. Understanding of the role of physical interactions in the formation of the native structur e of the protein will enable us not only to predict the final structure of the protein based only on knowledge of the amino-acid sequence but can be used to study protein folding or misfolding processes. The ability to predict three-dimensional structures of proteins or to predict their folding pathways can greatly contribute to rational drug design against cancer, Alzheimer or prion deseases.