區域-區域交互作用地圖探索蛋白質交互作用介面

Abstract

蛋白質交互作用(PPIs)的機制對許多生物途徑以及設計藥物上都扮演著重要的角色。目前在PPIs的分析上已經有研究利用單一殘基或一對交互作用殘基的保守性來推論蛋白質交互作用介面上的熱點(hotspots)。然而這些表面包含著許多交互作用的殘基，因此在解釋交互作用介面上的功能性以及在結構上扮演的角色時，藉由觀察一群具有密集交互作用以及在多個物種上具有保守性的殘基，會比只分析單一交互作用的殘基還要符合生物上的意義。而在實驗室之前的研究所提出的SiMMap概念，是利用大量的小分子化合物和蛋白質的交互作用，來觀察蛋白質對於官能基的偏好以及在結合介面上的物化特性。因此我們將應用此概念以及跨多物種的同源蛋白交互作用家族(PPI family)來分析蛋白質交互作用介面。 在本論文中提出了一個新的概念稱為區域-區域交互作用地圖(site-site interaction maps)，用來分析蛋白質介面上的數個微環境，此外，為了分析介面上的微環境，透過此概念可以將交互作用介面區分為數個具有保守性且較密集交互作用之區域，我們稱為錨點(anchor)，而這些錨點內的殘基多為具有高度保守性且具有生物功能。除此之外，這些錨點可被分為兩種型態，分別為：核心錨點(core anchor)表示該錨點的物化特性在PPI families間具有保守性，以及專一性錨點(specific anchor)表示該錨點在特定PPI family內會形成特定的結合環境。而為了驗證我們所定義的錨點具有生物功能性，因此我們統計了29個蛋白質複合體包含332個丙胺酸點突變(alanine mutations)的資料，可以發現在錨點內的殘基被突變後結合能的改變(ΔΔG)都會比錨點外的殘基還要大。 在此我們利用forkhead-associated domain (FHA domain)的蛋白質家族來展現site-site interaction maps的概念，根據這些結果我們推論出在許多含有FHA domain的蛋白質中，為什麼其可以與不同的下游蛋白質結合且執行不同的功能。接著利用非小細胞肺癌病患EGFR對下游蛋白質磷酸化的蛋白質體資料，包含wild type和L858R突變兩種型態，發現EGFR兩種突變型態對於下游蛋白質的磷酸化程度具有差異，因此藉由探討兩種形態上在蛋白質交互作介面的錨點，根據突變影響到結構部分以及專一性錨點解釋兩種EGFR型態對下游蛋白質的磷酸化為什麼具有差異，並且利用已知結合在Mdm2家族上的抑制劑，來驗證我們所定義的錨點其物化特性是否與蛋白質-抑制劑間的交互作用力相似，實驗結果顯示抑制劑會結合到的殘基多會被我們辨識在錨點內，因此我們相信可以利用此新概念來分析蛋白質複合體的交互作用機制，以及做為未來設計多標靶或是專一性藥物的參考。

Studying protein-protein interactions (PPIs) mechanisms is essential to understand cellular processes and drug discovery. Many studies of PPIs have been applied to infer the hotspots in a given protein-protein interface by considered the conservation of a single residue or pair of residues. However, these approaches often were residue-based methods and lacked of the site-site interactions for PPI interfaces. To elucidate functional and structural roles of a PPI interface, identifying a group of contact residue pairs (site-site interactions), which are usually compact region and conserved interaction across multiple species, is very important for understanding the mechanisms of PPI interfaces. In our previous studies, we have proposed a concept, SiMMap, which uses immense screening compounds to describe the relationship between the moiety preferences and physicochemical properties of the binding site. Here, we use the concept of SiMMap and the PPI family to characterize the PPI interfaces. In this thesis, we propose a new concept, called site-site interaction map to analyze the binding environments of a PPI interface. To identify binding environments of a PPI interface, the site-site map can be used to divide a PPI interface into several conserve and packing regions (called anchors), which residues are highly conserved and play a key role in certain biological functions. In addition, the anchors can be distinguished into two types, core and specific anchor. The core anchor often possesses conserved physicochemical properties between PPI families, and the specific anchor often forms a selective environment for a target PPI interface. To verify the identified anchors, we collected 29 complexes with 332 mutated residues in the database of alanine scanning mutagenesis. Our experimental results showed that the binding free energies (ΔΔG) of anchor residues are statistically significant higher than the non-anchor residues in the PPI interfaces. Here, we use the site-site interaction map to analyze FHA domain families. The results may explain why the various FHA families can recognize the different substrates and trigger different biological functions. Furthermore, we applied our method to identify the site-site interaction map of wild type (WT) and L858R EGFR, which are common mutated types of EGFR in non-small lung cancer. According to the structures and the specific anchor may explain the differently interacting mechanisms between WT and L858R EGFR. We also applied the concept to Mdm2 proteins complexes. According to the known inhibitors of the Mdm2 family, we found that PPI anchors shared similar physicochemical properties with protein-ligand interactions. In addition, the consensus interacting residues of protein-ligand are usually located on the PPI anchors. These results showed that our concept can reflect the biological functions and binding mechanisms of the Mdm2 family. We believe that site-site interaction maps can be used to explore mechanisms of PPI and design PPI inhibitors.