蛋白质结构识别：从天然折叠到诱导折叠

发布时间：2018-08-17 16:24

【摘要】：蛋白质通常被认为可以从信使RNA翻译而来的氨基酸无规则卷曲链自主折叠成特定的三维结构。折叠而成的蛋白称为其天然状态(或天然结构),三维结构形成的物理过程称为蛋白质天然折叠。一个正确的天然折叠构象通常是蛋白质执行功能必需的。但是,越来越多的研究证明,很多蛋白质整体或部分缺乏有序或稳定的三维结构,却依然在生物界发挥着重要的生物功能。这些蛋白或片段被称为固有无序蛋白或固有无序区域(IDPs or IDRs),与之对应的是天然状态下有稳定结构的有序蛋白。虽然IDPs or IDRs在天然状态下不折叠,但是它们常常与结合配体后经历无序到有序转变,这个过程称为折叠与结合偶联,或诱导折叠。本论文主体主要包括四个部分:(1)首先,设计了一个整合打分函数SVR_CAF从大量计算机预测模型中识别蛋白质天然结构。SVR_CAF的有效性在几个数据集上得到验证,与其他打分函数相比,它表现出相当的识别天然结构的能力。(2)然后,从蛋白质家族结构域数据库Pfam序列中收集了一组预测为完全无序的、但是在蛋白质结构数据库PDB中有结构的结构域。我们对这些结构域的所有种子序列和PDB结构序列进行无序预测,结果显示同一结构域序列的预测无序都表现出多样性。进一步的研究我们将所有结构按照其可能的形成机制分为二硫键、离子结合、配体结合。对于同时含有单体和二聚体(多聚体)的同一蛋白,和同时结合多个配体的蛋白,我们比较了其结构并观察到了复合物形成和结合多个配体的结构变化。(3)其后,还特别关注了配体结合的结构域,并把它们命名为无序结合结构域。我们识别到无序结合结构域三种相互作用:DNA结合、无序配体结合和有序配体结合,并对每种方式进行了详细讨论。(4)最后,我们对一组无冗余的PDB文件进行了基于溶剂可及表面积和接触面积的无序分析,识别到更多可能包含IDPs or IDRs的蛋白复合物。我们举例说明了固有无序在调控蛋白质-蛋白质相互作用中的重要作用,并分析了含有三个不同蛋白的复合物配体间相互作用的多样性。总体而言,我们从大量计算模型中识别了有序蛋白天然折叠结构,从Pfam和PDB数据库中识别了无序蛋白或无序区域的诱导折叠结构。这两部分研究工作经历了“有序到无序”的过渡,是对序列-结构-功能传统范式内容的重要补充。
[Abstract]:Proteins are generally thought to be able to translate from messenger RNA amino acids that are unruly curled chains that fold autonomously into specific three-dimensional structures. The folded protein is called its natural state (or natural structure), and the physical process of three-dimensional structure formation is called protein natural folding. A correct natural folding conformation is usually necessary for the protein to perform its functions. However, more and more studies have proved that many proteins lack ordered or stable three-dimensional structure, but still play an important biological function in the biological world. These proteins or fragments are called intrinsically disordered proteins or inherently disordered (IDPs or IDRs), corresponding to ordered proteins with stable structures in the natural state. Although IDPs or IDRs does not fold in a natural state, they often undergo disordered to ordered transitions after binding with ligands. This process is called folding and binding coupling, or inducing folding. The main contents of this thesis are as follows: (1) at first, an integrated scoring function, SVR_CAF, is designed to identify the natural structure of protein from a large number of computer prediction models. The validity of SVRCAF is verified on several data sets, compared with other scoring functions. It shows considerable ability to recognize natural structures. (2) then, a set of domains which are predicted to be completely disordered but have structures in protein structure database PDB are collected from Pfam sequences of protein family domain database. All seed sequences and PDB structural sequences of these domains are predicted disorderly. The results show that the prediction disorder of the same domain sequence shows diversity. We further study that all the structures can be divided into disulfide bond, ion binding and ligand binding according to their possible formation mechanism. For the same protein containing both monomer and dimer (polymer) and the protein that binds multiple ligands at the same time, we compared its structure and observed structural changes in the formation and binding of complexes to multiple ligands. (3) subsequently, Special attention is paid to ligand binding domains, which are named disorder binding domains. We identify three kinds of interactions of disordered binding domains: DNA binding, disordered ligand binding and ordered ligand binding, and discuss each way in detail. (4) finally, We analyzed a set of non-redundant PDB files based on solvent-accessible surface area and contact area and identified more protein complexes that may contain IDPs or IDRs. We illustrate the important role of inherent disorder in regulating protein-protein interactions and analyze the diversity of complex ligands containing three different proteins. In general, we identify the natural folding structures of ordered proteins from a large number of computational models, and the induced folding structures of disordered proteins or disordered regions from Pfam and PDB databases. These two parts have experienced the transition from order to disorder, which is an important supplement to the traditional paradigm of sequence-structure-function.
【学位授予单位】：苏州大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：Q51

【相似文献】