蛋白質結構、結構聚集和序列保留性之間的關係

Abstract

蛋白質通常依照胺基酸序列折疊成獨特、穩定而有序的結構，使得所得到的蛋白質可以執行其特定的功能。因此，對於執行功能和維持結構穩定性重要的氨基酸在序列和結構方面通常是高度保留的。近來有研究探討了序列的保留性賦予蛋白質在結構上的限制，並且發現氨基酸序列保留度（sequence conservation）與溶劑可接觸性（solvent accessibility）以及與根據胺基酸主鏈C原子計算加權接觸數(weighted contact number)所得的局部結構聚集密度（local packing density）間的高度相關性。源自序列比對的位點特異性取代率計算而來的序列保留度和由三維結構得到的局部結構聚集密度之間的關係表明，蛋白質序列和結構的演化約束是相互關聯的。在本論文中，我們首先從相似的蛋白質結構中計算結構聚集密度曲線之間的關係，由SCOP資料庫所定義的有著遠端同源（remote homolog）關係的超家族蛋白質結構域（Superfamily domains）間，相似的結構具有相似的加權接觸數分佈，這顯示局部結構聚集密度分佈可以反映出蛋白質的結構限制。接著，樣本資料集中兩個蛋白質的位點特異性替代率譜（site-specific substitution rate profiles）與彼此的結構，結果亦顯示，相似的序列保留度可能與相似的結構相關。蛋白質序列和結構的進化約束不僅在同一個蛋白質中相互關聯，而且可能在不同的蛋白質中也有著相互關聯的現象。 另一方面，受演化限制的蛋白質序列和結構導致功能結合區域在序列和結構中趨向保留，而且涉及交互作用的氨基酸通常處於鄰近的三維空間中。在本論文中，我們除了分析以上蛋白質整體結構層級，接著探討局部結構的影響。以金屬離子結合位的蛋白質結構為例，我們分析了金屬離子結合位的蛋白質結構，局部聚集密度分佈和氨基酸構成,包含了鈣離子（ Ca2+）, 銅離子（ Cu2+）, 亞銅離子（ Cu+）, 鐵離子（Fe3+）, 亞鐵離子（Fe2+）, 鎂離子（Mg2+）, 錳離子（Mn2+）, 鋅離子（Zn2+）, 鎘離子（Cd2+）, 鎳離子（Ni2+）, 汞離子（Hg2+）和鈷離子（Co2+）。金屬離子結合位的局部聚集密度分析顯示，金屬離子結合位在蛋白質中往往位於較擁擠的環境。我們的結果表明，金屬離子結合氨基酸的局部聚集密度分佈與非金屬離子結合氨基酸的分佈是明顯不同的。這12種金屬離子結合位具有明顯不同的結合模式，顯示出金屬離子結合氨基酸中結構和序列的保留性。根據保留度高的立體結構和氨基酸模式，我們進一步開發了一種通過利用片段變換方法（fragment transformation method）預測金屬離子結合氨基酸和建立金屬離子對接模型，並且構建了MIB伺服器（Metal Ion Binding sites prediction and docking server， http://bioinfo.cmu.edu.tw/MIB/）不僅有十二個金屬離子結合位的預測，而且提供金屬離子對接（metal ions docking）的模型。

During evolution, substitutions at individual residues within amino acid sequences arise under the constraints of structure folding, protein function, and the protein–protein interactions. Amino acid sequences generally fold into unique, stable, and well-ordered conformations so that the resulting proteins can carry out their specific functions. As such, residues that are important for function and structural stabilization are generally highly conserved in terms of both sequence and structure. Recent studies have discussed the structural constraints imparted by site-specific substitutions, and amino acid sequence conservation was found to correlate with solvent accessibility and the local packing density such as weighted contact number. The relationship between sequence conservation, site-specific substitution rates derived from multiple sequence alignment, and the weighted contact number, local packing density derived from three-dimensional structure, revealed that the evolution constraint of protein sequence and structure properties were correlated. In this study, we assessed the relationship between the packing density profiles drawn out from similar protein domain structures in SCOP superfamily with the relationship of remote homologs, and the results showed that similar structures had similar weighted contact number profiles, and demonstrated that the packing density profiles could reflect the structural constraints. Then, we compared the site-specific substitution rate profiles of two proteins and their structures, and the results showed that similar conservation profiles could be linked to similar structures. The evolution constraints of protein sequence and structure were not only related to each other in a protein, but also were interrelated in different proteins. The protein sequence and structure restricted by evolution constraints lead to that the binding regions tend to be conserved in sequence and structure and the interacting residues involved are usually in close three-dimensional space. In this study, we analyzed the protein structure, local packing density and residue constitute of metal ion-binding sites, including Ca2+, Cu2+, Cu+, Fe3+, Fe2+, Mg2+, Mn2+, Zn2+, Cd2+, Ni2+, Hg2+, and Co2+. The analysis of local packing density of metal ion-binding sites revealed that the metal ion-binding sites tend to be more crowded in proteins. Our result showed that the distributions of local packing density of metal ion-binding residues were distinct from those of non-metal ion-binding residues. The results showed that there were distinct different binding patterns of these twelve kinds of metal ion-binding sites, indicating the conservation of structure and sequence in metal ion-binding residues. According to the conserved patterns of structure and residue, we further developed a method to predict the metal ion-binding residues and build the model of metal ions docking by using fragment transformation method, and built up the MIB server (Metal Ion Binding sites prediction and docking server, http://bioinfo.cmu.edu.tw/MIB/) for not only twelve metal ion–binding sites prediction but also metal ions docking.