蛋白質結構反映演化約束之研究

Abstract

蛋白質序列的演化會受到其功能和生物物理條件所約束，此約束也反映在序列個別位置上氨基酸保留度的差異。根據過去文獻的記載，蛋白質結構區域摺疊密度（Local Packing Density, LPD）和個別氨基酸相對暴露面積（Relative Solvent Accessibility, RSA）是用來探討序列演化上，受到其結構約束的兩個最主要的結構特徵。然而，探討這兩個結構特徵，在序列演化上相對影響性的研究相當匱乏。此外，也難以直接比較過去相關的文獻，因其使用了不同的資料來源、探討不一樣的結構特徵以及採用的序列保留度計算方式亦不盡相同。為了了解及量化序列演化如何受到其結構約束的影響，我們採用了兩種LPD的計算模型（Weighted Contact Number, WCN和Contact Number, CN）、四種氨基酸暴露面積（Solvent Accessibility, SA）的計算模型以及五種不同的序列變異度的計算方法。我們發現包含了系統樹資訊的序列保留度計算方法與結構特徵有較高的相關性。更重要的是，我們亦發現相較於RSA而言，LPD對於蛋白質序列演化上的影響更為重要。本研究的結果可促使對序列演化的機制有更進一步的了解，並對未來建立較好的模型提供有利的資訊。

The evolutionary divergence of protein sequences is subject to purifying selection against amino acid substitutions imposed by functional and biophysical requirements, resulting in emergent site-dependent patterns of sequence conservation. Two site-specific structural properties have emerged as the best candidate structural constraints: local packing density (LPD) and relative solvent accessibility (RSA). Studies addressing the relative contributions of LPD and RSA to evolutionary constraints are scarce and comparison between different studies is difficult because they use different datasets, structural properties, and measures of sequence variability. To advance our understanding of structural constraints on sequence divergence, we compared two LPD measures, the Weighted Contact Number (WCN) and the Contact Number (CN), and four measures of solvent accessibility (SA), with five measures of sequence variability for a diverse dataset of enzymes. We found that the best sequence variability measures take into account the phylogenetic tree topology. In contrast with recent studies, we found that both LPD measures (WCN and CN) significantly outperform all of the SA measures as structural constraints on sequence divergence at site level. This should have far reaching implications for our understanding of the mechanisms of sequence evolution and the development of better models.