高效研究多肽结合结构域结合特性新策略的建立及其应用

【作者】 宋婀莉；

【导师】 高友鹤；

【作者基本信息】 中国协和医科大学 ， 病理生理学， 2007， 博士

【摘要】 ●结构域是蛋白质相互作用研究的一个简单高效的入手点在蛋白质组学规模上研究蛋白质相互作用网络是全面了解蛋白质功能、揭示疾病发病机制的有效手段之一。节点蛋白在相互作用网络中占据关键位置,在生命调控机制中发挥重要功能,可作为研究相互作用网络的一个有效突破点。很多蛋白由单个或多个结构域组成;绝大多数蛋白质相互作用是通过相互作用结构域(如PDZ,SH2,WW,SH3等)与其识别模体的结合完成的。深入研究节点蛋白的相互作用结构域,不仅可以了解详细的相互作用分子机制,为蛋白质相互作用的药物研发提供理论基础,而且可以在全蛋白质组水平上发现其所有的潜在配体蛋白,绘制以节点蛋白的相互作用为骨架的蛋白质相互作用网络图,从而系统地揭示蛋白质在复杂体系中的多样功能。本研究中,我们以最重要的蛋白质相互作用结构域之一的PDZ结构域作为入手点,建立了高效研究结构域结合特性的新策略,全面系统地研究了数个PDZ结构域的结合特性,发现了它们潜在的新配体蛋白。研究结果表明,PDZ结构域的结合特性同时具有较强的规律性和一定的多样性。由于我们研究的PDZ在多种组织和细胞结构中均有各种功能的潜在新配体被发现,提示PDZ蛋白在不同的体系中执行不同的功能具有生物化学基础,可能是一个相当普遍的现象。PDZ蛋白潜在的多功能性超过了我们的预期。我们猜想可能许多具有常见蛋白质相互作用结构域的蛋白也会具有超过我们以往估计的配体数量和功能多样性。这个猜想需要更多结构域的系统研究方法和大规模研究才能证实。●高效研究多肽结合结构域(peptide-binding domain)结合特性新策略的提出近年来常用的高通量研究结构域的方法主要有定向肽库(oriented peptidelibrary)、SPOT合成、噬菌体展示(phage display)和酵母双杂交(yeast two-hybrid)等。但是这些传统方法都是针对某一个特定结构域所进行的大规模筛选,相对费时费力。为了适应于在蛋白质组学规模上研究结构域的结合特性、提高研究通量和效率,我们提出了一种高效地验证性筛选目的结构域结合配体文库的新研究策略。先用目的结构域家族中有代表性的成员筛选随机多肽文库,收集得到的阳性克隆构建成一个配体库,其它成员直接验证筛选配体库研究其识别特性。根据验证筛选得到的阳性和阴性结合序列推知识别规律,再以此识别规律在蛋白质数据库中预测潜在的天然配体,最后在酵母双杂交系统中验证并发现新的配体蛋白。如果直接验证筛选配体库得到的阳性配体数量不足以得出某个结构域的识别特性,该结构域就从头筛选随机多肽文库,所获得的新阳性克隆添加到配体库,作为对配体库的补充,可提高以后通过直接验证筛选来研究此类PDZ的效率。另外在验证预测的天然配体时,构建的克隆也添加到配体库中,这也是对配体库的重要补充,这样直接验证筛选就有可能找到新的配体蛋白。验证筛选体系最大的优点是,与传统筛选相比研究效率得到了极大的提高,而且整个体系是动态发展的,配体库会在体系的应用过程中不断得到补充,扩充后的配体库又会进一步提高研究效率。●高效研究多肽结合结构域结合特性新策略的建立我们首先将验证筛选的研究策略应用于PDZ结构域结合特性的研究。首先,我们收集ZO-1 PDZ1、ZO-1 PDZ3和Erbin PDZ筛选随机多肽文库得到的阳性克隆,构建成一个初始的配体库。通过验证筛选该配体库快速研究了HtrA2 PDZ和LNX1 PDZ2的识别特性。MAGI3 PDZ4等PDZ通过验证筛选没有得到满意的结果,因此从头筛选随机多肽文库研究其结合特性,并将得到的阳性克隆补充到配体库中,作为对配体库的补充和发展。在验证筛选体系建立和初步发展的过程中,我们对上述目的PDZ结构域的结合特性进行了详细的分析。研究结果不仅涵盖了以往大规模研究方法报道的结果,而且还发现了新的PDZ结构域识别特性,即:ZO-1 PDZ3在配体-5位偏好疏水性氨基酸,ZO-1PDZ3、LNX1 PDZ2和MAGI3 PDZ4在配体的0位可选择极性氨基酸,除MAGI3 PDZ4外的五个PDZ在配体-1位都强烈偏好芳香族氨基酸,除ZO-1 PDZ3和MAGI3 PDZ4外的四个PDZ都可识别Ⅰ、Ⅱ、Ⅲ三种类型的PDZ配体,HtrA2 PDZ和LNX1 PDZ2的选择特异性都呈现出一定的前后氨基酸之间的连锁关系。这些新的PDZ结合特性的揭示,为PDZ结构域的重新定义和分类提供了有价值的信息。同时,我们发现了以此六个目的PDZ结构域为节点的40对新的蛋白质相互作用,为深入研究这些PDZ蛋白的生物学功能提供了重要线索。●高效研究多肽结合结构域结合特性新策略的应用LNX蛋白是第一个被发现的含有PDZ结构域的RING类型的泛素连接酶E3,但是其PDZ结构域的生物学功能还不清楚。为了揭示LNX PDZ结构域的功能,我们应用已建立的高效验证筛选体系对人类LNX家族的4个蛋白含有的9个PDZ结构域的结合特性和相互作用蛋白进行了系统的研究。LNX1 PDZ1、LNX1 PDZ2、LNX1 PDZ3、LNX2 PDZ2和LNX4 PDZ通过验证筛选和/或随机多肽文库筛选得到的阳性配体在C末端呈现明显的规律,经分析发现这5个PDZ结构域的结合特性有共性但也有显著的不同。同时,我们发现了LNX家族PDZ结构域之间的三对相互作用,即:LNX1 PDZ1和LNX1 PDZ4、LNX1 PDZ1和LNX2 PDZ4、LNX2 PDZ1和LNX2 PDZ4之间的相互作用。另外,根据总结的LNXPDZ结构域的识别规律,我们在人类蛋白质组水平上发现了新的LNX配体蛋白。这些配体蛋白种类多样且分布广泛,提示LNX可能在多种复杂体系中发挥多样的功能。同时,在PDZ结构域高效验证筛选体系应用的过程中,我们对原PDZ配体库又补充了新的配体,这是对配体库的再次发展,将会使验证筛选体系的高效性在更大规模的PDZ结构域研究中得以体现。●Liddle综合症分子治疗的初步探索Liddle综合症是一种上皮细胞钠离子通道相关的遗传性疾病,是1963年由Liddle等人首先发现的一种常染色体显性、盐敏感型的高血压综合症。Liddle综合症的分子发病机制现已较为明确地认为是上皮钠离子通道(epithelial Na~+ channel,ENaC)的β亚基或γ亚基的胞质侧羧基端的PY模体低频率的点突变或缺失突变,使其与泛素连接酶Nedd4的WW结构域之间的相互作用被打断,导致ENaC不能正常泛素化和降解,因而在质膜上的半寿期延长、活性增加,导致钠离子重吸收的增加,引起Liddle综合症。由于Liddle综合症的发病机制已经研究的比较清楚,而且仅是单基因的突变所致,这为分子治疗提供了可能。我们提出了一种从分子水平上治疗Liddle综合症的设想,即构建一种可识别Liddle病人中ENaC蛋白突变的PY模体的人工泛素连接酶E3,使其结合并降解突变的ENaC蛋白,从而使质膜上ENaC的表达数量和活性恢复到正常水平。而可识别PY突变体的E3,可通过用病人中的PY突变体筛选随机多肽库获得与之结合的随机肽段,然后用其替换PY模体天然配体蛋白Nedd4中的WW结构域,从而得到一个新的人工E3。本论文中,我们以一种Liddle综合症突变体Y620H为诱饵蛋白,通过筛选随机多肽文库获得了一个至少能与两种PY突变体(Y620H和P618L)特异性结合的随机肽段。为下一步的深入研究和探索积累了实验数据。更多还原

【Abstract】 ●Protein interaction domain is a valuable starting point for high-efficiency protein-protein interaction researchCompiling protein-protein interaction network provides many new insights into protein functions and disease mechanisms.Some node proteins in the network usually play key roles in biological processes.A large set of proteins are composed of one or more domains and a substantial proportion of protein-protein interactions are mediated by families of protein interaction domains,such as PDZ,SH2,WW,SH3,etc.Therefore,the protein interaction domain of node proteins values as a simple and high efficient starting point for understanding the mechanisms and networks of protein interactions.In this study,with focus on PDZ domain,one of the most important protein interaction modules,we developed a high-efficiency strategy for binding property characterization of peptide-binding domains,and systematically characterized a number of PDZ domains and discovered a series of novel interactions.We found that the binding properties of the PDZ domains showed strong consensus as well as certain variability. Moreover,many of the identified ligand proteins of the PDZ domains were expressed extensively and had diverse functions.It provides the biochemical basis that PDZ proteins may play diverse roles in multiple systems.The complexity of protein functions exceeds our current understanding and expectation.We propose that proteins composed of common protein interaction domains might have many ligands and diverse functions in different biological systems.●Development of a high-efficiency strategy for binding property characterization of peptide-binding domainsIn recent years,several powerful methods,including oriented peptide library,SPOT synthesis,phage display and yeast two-hybrid,have been successfully employed for characterization of the peptide recognition specificities of individual domain families. However,these methods are all based on large-scale screenings for individual bait domain, which are relatively labor-intensive and time-consuming.For more efficiently characterizing the binding properties of peptide-binding domains,on the basis of the facts that one domain can bind many ligands and one ligand can be recognized by multiple domains,we developed a systematic strategy of high-throughput validation screening of a specialized candidate ligand library using yeast two-hybrid mating array.The library was constructed mainly by collecting the ligand clones isolated from yeast two-hybrid screenings of random peptide libraries with representative members of the selected domain family.By validation screening,consensus sequences were deduced from both positive and negative binding ligands.Novel ligand proteins were identified through database searches with consensus-binding sequences and confirmed in yeast two-hybrid system.The domains that did not interact with enough numbers of ligands from the library were used as bait for de novo screenings of random peptide libraries in the traditional way.This strategy is notably different from other approaches in several features.Firstly,the overall efficiency is dramatically improved by validation screening instead of traditional screening.It can be further improved,as the candidate ligand library expands with the addition of clones from more traditional screenings and clones constructed for confirmations.Secondly,we can achieve high throughput with yeast two-hybrid mating array approach.Multiple bait domains can be tested in parallel and thousands of candidate ligands can be screened on arrays simultaneously.Thirdly,since the library consists of ligands of known sequences,actual positive and negative binding sequences,which are both very important for precisely defining binding properties,can be directly read from arrays without resequencing.●Construction of a high-efficiency strategy for binding property characterization of PDZ domainWe have firstly demonstrated the feasibility of the newly developed strategy by employing it to investigate the PDZ domain interactions.We constructed an initial PDZ ligand library by yeast two-hybrid screenings of random peptide libraries with ZO-1 PDZ1,ZO-1 PDZ3 and Erbin PDZ.Next,we rapidly characterized HtrA2 PDZ and LNX1 PDZ2 by validation screenings of this initial library. MAGI3 PDZ4 and other interested PDZ domains,which could not find enough numbers of ligands from the initial library,were used as bait for de novo screenings of random peptide libraries in the traditional way.The PDZ ligand library was expanded with the addition of new positive clones isolated by new traditional screenings and clones constructed for confirmations.The binding properties of the interested PDZ domains have been successfully defined. Broader binding properties have been identified compared to other methods,including unique novel recognition specificities,such as ZO-1-PDZ3 favors hydrophobic residues at p^-5,ZO-1 PDZ3,LNX1 PDZ2 and MAGI3 PDZ4 select polar or hydrophilic residues at p⁰, five of the six PDZs prefer aromatic residues at p^-1,four of the six PDZs can bind three conventional classes of PDZ ligands,and HtrA2 PDZ and LNX1-PDZ2 exhibit linkage selectivity at the C-termini of ligands.All these findings not only are important for predicting PDZ binding partners but also provide the basis for new definition of PDZ domain and major revision of conventional PDZ classification.In addition,40 novel interactions have been discovered,serving as significant clues for further functional investigation.●Application of the high-efficiency strategy for binding property characterization of PDZ domainLNX is the first described PDZ domain containing member of the RING type E3 ubiquitin ligase.However,the functions of LNX PDZ domains are poorly understood.We systematically characterized the binding properties of human LNX PDZ domains and discovered their potential ligand proteins in human proteome,using our newly developed high-efficiency validation screening method.By validation screenings of the PDZ ligand library and/or yeast two-hybrid screenings of random peptide libraries,the C-terminal binding properties of LNX1 PDZ1,LNX1 PDZ2,LNX1 PDZ3,LNX2 PDZ2 and LNX4 PDZ were characterized.The recognition specificities between these five PDZs showed certain common features and some distinct properties.3 novel PDZ-PDZ interactions in LNX family were identified,which were the interactions between LNX1 PDZ1 and LNX1 PDZ4,LNX1 PDZ1 and LNX2 PDZ4, LNX2 PDZ1 and LNX2 PDZ4.In addition,a few of novel ligand proteins of LNX PDZ domains were discovered.These novel identified interactors are known to be functionally different and expressed extensively.This prompts LNX family proteins may play diverse roles in multiple systems. ●Preliminary investigation on molecular therapy of Liddle syndromeLiddle syndrome,first described by Grant Liddle et al in 1963,is an autosomal dominant form of salt-sensitive hypertension.The pathogenesis of Liddle syndrome has been attributed to missense,deletion or frameshift mutations of PY motifs in cytoplasmic carboxyl termini of beta or gamma subunits of the amiloride-sensitive epithelial sodium channel（ENaC）.Disease mutations lead to absence of ENaC interaction with the WW domain of Nedd4,an ubiquitin ligase.The abnormalities result in increased channel activity and excessive Na⁺ absorption in the kidney.We proposed a novel strategy for molecular therapy of Liddle syndrome,which was use of an artificial ubiquitin ligase to target mutational ENaC.The artificial ubiquitin ligase was constructed by replacement of the WW domain of Nedd4 with a peptide that was able to interact with mutational PY motifs,and this peptide could be obtained from screenings of random peptide libraries.In this study,we firstly chose one of Liddle syndrome mutations,Y620H,to screen random peptide library and obtained one positive peptide.We further demonstrated the positive peptide specifically interacted with at least two Liddle syndrome mutations, Y620H and P618L.更多还原