TOG蛋白質超家族的演化與摺疊

Abstract

蛋白質摺疊密碼，也是遺傳訊息密碼傳遞的第二部分，為分子生物學的重要研究課題之一。自從Christian Anfinsen提出了蛋白質能自發性折疊的實驗後，科學家們已研究這個問題超過數十年。其中，膜蛋白作為與各種生命現象相關之一類特殊蛋白質、且目前超過一半的藥物分子皆以其作為標靶，其摺疊過程於近十年特別受到注目。在膜蛋白之中，又以G protein-coupled receptors特別受到重視。這類蛋白質在各種感官現象如視覺、嗅覺、味覺都扮演著重要的角色，其中與視覺相關的膜蛋白尤其吸引我們，像是與色彩辨識相關的Cone Opsins及與夜間視覺相關的Rod Opsins。為了解這些蛋白質及其序列-結構-功能之間的關係，我們選擇了TOG(Transporter-Opsin-G protein-coupled receptor)蛋白質超家族來做為我們電腦模擬研究的目標。我們的研究主要分為三個方面：一、蛋白質相似度網路的分析。我們使用一個分群算法來分析蛋白質之間的關係。這個方法可以用來區別各個蛋白質家族間差異以及預測新定序列的功能。並且，藉由此算法所建立的網路結構，我們可以從中研究蛋白質之間可能的演化關係。二、膜蛋白的二級結構螺旋及其上扭角的預測。我們使用了胺基酸的親疏水指數來預測可能屬於穿膜螺旋與半穿膜螺旋的序列片段，以及使用結合分子動力學與摺疊識別演算法的方法來研究穿膜螺旋的結構特徵。三、與Rod Opsins相關的Microbial rhodopsins的折疊研究。我們建立了一個基於胺基酸物理性質的粗粒化模型來了解這些物理交互作用在蛋白質的折疊中扮演怎麼樣的角色，並使用平行回火的蒙地卡羅方法來尋找能量最小的結構。The protein folding code, known as the second half of the genetic code, is one of the most important problems in molecular biology. Since Christian Anfinsen performed experiments concerning spontaneous folding of proteins, scientists have tried to solve this issue over several decades. It is estimated that 20-30% of all genes in most genomes are encoded to membrane proteins (MPs). MPs play key roles in living cells, and are the primary targets of more than 50% of drugs. In human genome, G protein-coupled receptors (GPCRs) constitute the largest family of integral MPs. GPCRs are involved in cell communication processes and mediate senses such as vision, taste, smell, hearing, and touch. In particular, we are interested in understanding the molecular mechanism for color vision in nervous systems. To achieve our goal, we chose the Transporter-Opsin-G protein-coupled receptor (TOG) superfamily to be the target of our computational studies of protein evolution and folding. In this dissertation, we have studied several aspects of the TOG superfamily, listed as followed: (1) The probable evolutionary relations of proteins could be investigated from protein similarity networks. We applied a clustering approach to provide a panoramic view of isofunctional groups of diverse protein superfamilies. This method is useful in predicting the structures and functions of novel protein sequences, and could explore possible evolutionary examples by visualizing the structures of our clustering networks. (2) Recognition of secondary structure helices and investigation of geometric structures of those helices were studied for MPs. We proposed a simplified algorithm to identify both transmembrane (TM) segments and half TM helices by finding the maximum sum of a hydropathic index with a variable window size, and then combined molecular dynamics simulations and a fold identification algorithm to fold the structural features of TM helices. These two methods provide good results in recognizing and folding TM helices. (3) The folding of molecular architecture for microbial rhodopsins was studied. This portion is the most important part of the protein folding code. We constructed a coarse-grained model based on the physical properties of amino acids, and then used replica-exchange Monte Carlo simulations to model their structures. This method has been used to construct protein models that are close to their experimental structures.