高通量酵母双杂交平台的建立及其在蛋白质相互作用组学当中的应用

【作者】 王玮；

【导师】 霍克克；

【作者基本信息】 复旦大学 ， 遗传学， 2007， 博士

【摘要】 蛋白质通过相互作用所形成的具有一定拓扑结构的互相关联的网络是生物体进行复杂生命活动的基础。通过传统的手段进行逐个的蛋白相互作用研究来逐步构建蛋白质相互作用图谱是一件无法完成的任务,但是研究者又如何摆脱“盲人摸象”的尴尬,而从一个更加宏观的角度审视全蛋白组范围内的蛋白质相互作用的情况呢?高速发展的自动化技术为生物学研究者提供了全新的理念和手段。大多数的重复性试验操作可以通过程序控制的机器人高效率自动化完成,这样既把试验人员从繁重的实验室劳动中解放出来,使他们有精力进行数据的整合和分析。人类蛋白质组学计划(Human Proteomics Project,HPP)是继人类基因组学计划之后,为了更加深入的解析人类生命活动的内在规律而启动的又一项全世界范围内的大型合作研究计划。研究人员已经不再满足于DNA水平的知识,将目光汇聚于蛋白质组的系统研究上来。基因组中绝大部分基因及功能仍然处于未知状态,需要从蛋白质水平予以解释。基因组决定生命体的基本形式,而蛋白质组决定生命的多样性、复杂性及其功能。蛋白质组是指“一种基因组所表达的全套蛋白质的集合”。其研究可以实现与基因组的对接与确认,直接揭示人类重大疾患发生与发展的病理机制。人类蛋白质相互作用图谱(HumanProtein-Protein Interaction Map,HPPIM)是蛋白质组学计划当中重要的一个组成部分。蛋白质相互作用组学在这个背景之下方兴未艾。通过自动化高通量的技术平台,研究者已经可以利用多种手段构建人类蛋白质相互作用的连锁图,其中高通量酵母双杂交就是一个重要的方法。传统酵母双杂交费时费力,无法完成组学计划的任务。本文在已有试验技术的基础上,逐步摸索并初步建立了大规模酵母双杂交的技术平台,并已应用到了人类蛋白质相互作用连锁图的构建过程中。在此基础上,本文获得了大量的人类蛋白质相互作用数据,并且初步构建了人类蛋白质相互作用图。后续的研究可以分为两个方面:一对于其中的重要而且尚未报道的相互作用进行深入的研究,揭示其生理生化意义;二利用积累的数据进行生物信息学分析和预测,对试验结果进行分析和评估。本论文以高通量酵母双杂交技术平台的初步建立过程入手,转入对筛选得到的具体蛋白质相互作用进行深入研究。首先,本文筛选得到了人类蛋白质MRKβ与MSK1这对相互作用。MRKβ是MAPK途径当MAKPKKK家族成员,它具有一个N端催化结构域,参与蛋白质的磷酸化修饰。MSK1是MAPK途径当中关键的激酶,直接参与调控核内重要转录因子如CREB、ATF1以及NFkBp65等的活性,并且是多种癌症和炎症的潜在药物靶点。最近已经成为研究的热点。该蛋白包括两个激酶结构域,通过一个调控序列连接起来。N端的结构域(N-terminal Kinase,NTK)属于AGC激酶家族,参与对其底物的磷酸化。C端的结构域(C-terminalKinase,CTK)属于CamK家族,目前认为它对N端有激活作用。通过生物信息分析本文发现,MRKβ的N端激酶区域与p38、ERK的磷酸化结构域具有高度的保守性。有报道推断调控MSK1的除了p38和ERK1/2之外,目前为止似乎没有其他蛋白参与这个过程的信号调控,而这与MSK1功能非常相似RSK则不同。RSK除了受到ERK的磷酸化调控之外还受到PDK1的磷酸化修饰,而PDK1与ERK,p38和MRKβ在磷酸化的domain上具有高度的保守性。本文通过体外结合实验、免疫共沉淀实验和亚细胞免疫荧光共定位实验证实了MRKβ与MSK1之间的特异性相互作用,并且通过体外实验发现MRKβ能够将MSK1磷酸化。体外和体内试验表明MRKβ可以通过磷酸化激活MSK1,并且进一步激活下游的转录因子CREB。这表明MRKβ参与了MSK1的磷酸化过程,从而调控信号在细胞中的传导,影响下游基因的转录。这对于经典的MAPK途径中MSK1的激活途径可能是一个新的发现和补充。本文进一步利用质谱技术鉴定了MRKβ在MSK1上的磷酸化位点,其次,本文还筛选得到了人类蛋白质MRKβ与14—3—3 zeta(YWHAZ)的相互作用。利用上述方法本文也确定了二者之间的特异性相互作用。这对相互作用对于MRKβ磷酸化MSK1的意义值得进一步的深入研究。另外,本文也验证了人类蛋白质CCNH与CtBP2之间的相互作用并且分析了这对相互作用的未来研究意义。生物信息学正处于急速发展上升的阶段,已经成为生命科学研究人员必不可少的工具和重要的方法,也成为了蛋白质组学当中不可或缺的重要组成部分,大量的试验数据分析成为了生物研究者理解认识蛋白质分类和功能的有力手段。结构域是蛋白质的基本组成部分,它具有三维的定义。首先,它应该是一段具有序列保守性的蛋白质序列(motif);其次,它具有保守的结构特征;再次,它具有保守的功能。对结构域的研究和认识是理解蛋白质功能的主要方面。本文针对链霉菌噬菌体整合酶phiC31的结构域进行了预测,并结合该整合酶的识别序列特征,预测了将该系统应用于人类基因治疗的风险。本文也利用动态规划以及基于核方法的机器学习方法预测蛋白质结构域,并且以网络服务器的形式免费提供预测服务。为了更好的研究丝氨酸整合酶家族,本文也建立了针对这个蛋白家族的数据库,为生物试验和计算分析提供了良好的平台。更多还原

【Abstract】 Functional proteins are connected in a dynamic and highly organized network in living cells. And it remains a problem how to parse the complex network and find more meaningful interactions that are related with human diseases. The results are valuable for drug discovery in pharmaceutical industry. Classic methods, such as the yeast two hybrid, are laborious and time-consuming, and it is a mission impossible for researchers finally build up a reliable and integrated protein interaction network even in human cells. Those methods do not meet the needs of proteomics any more. Thus, we have developed the high through-put yeast two hybrid system aiming to build up protein interaction networks in various species including human. It can liberate scientists from repeated operations and help them in exploring the more deep mechanism inside living cells. On setting up this platform, we have built up a interaction network of human proteins and further we focused on human protein MSK1 and we screened its preys and one of them is MRKβ. MRKβis a member of MAPKKK kinases. It has an N terminal catalytic domain which is responsible for the phosphorylation of its substrates. MSK1 is also an important kinase in the MAPK pathway and its activation depends on the phosphorylation of serines and threonines by ERK1/2 or p38. Compared with its relative RSK protein families, MSK1 should also be activated by some certain kinase besides ERK1/2 or p38. However, no such kinases have been found up to now. Our results show that MRKβcould interact with and then activate MSK1 through phosphorylation. Thus, downstream transcription factor CREB can be activated by the active MSK1 at Ser 133. This results in the gene expression in the nucleus responding to the extra signal. We have also use mass spectrum method to detect the phosphorylation sites in MSK1 by MRKβ. MRKβcan possibly phosphorylate MSK1 at Ser 436, Ser522 and Thr523. More research is needed to find out the in-depth mechanism and its physiological meaning.We have also identified MRKβcan also interact with YWHAZ which is a member of 14-3-3 protein family. GST pull-down and Co-IP resuts show that they interact with each other. The co-localization of these two proteins may imply an interaction-induced trans-localization of MRKβ. More details are waiting to be discovered.We also found human CtBP2 interacts with CCNH in Hela cell nucleus.The high through-put yeast two hybrid platform has been applied into the research of interactomics successfully and at the same time we have also found problems in it. Modification and optimization are needed.Bioinformatics is now in its blossom. Computational methods are now applied in many fields of biological research. Protein informatics is one of them. Identification of novel domains and the verification of their functions are focus of the third part in this thesis. We have developed a free of charge web server KemaDom to identify domains in proteins and we hope this could be of help to experimental scientists in labs. On the basis, we have identified a C₄ zinc ribbon domain in the C terminus of the phiC31 integrase which is successfully applied in the fundamental research of human gene therapy. We have also used motif finding methods to identify the recognition motif of the phiC31 integrase and thus we calculate the distribution of the motif in the human genome which help us in a conclusion that this system is fairly safe while some hidden risks are still inside the genome. SerRD is a database we have developed to collect information about the serine integrase and we hope this free database could help researchers in the study of this protein family.Protein interactomics is a crucial component of the proteomics. By constructing different protein-protein interaction map, researchers could discover those meaningful protein interactions. High through-put yeast two hybrid system inherits from the pervious normal yeast two hybrid system and it has already been applied in the large-scale identification of protein interactions as a main technique platform. It offers a great number of reliable data for researchers to take a comprehensive analysis in the protein function and evolution from the view of system biology. At the same time, the platform also gives reasonable clues to researchers in the medical research and drug targets discovery. Opportunities for both computational scientists and lab researchers are now on the table. On the one hand, we produce high through-put data by using automated equipments and then make detailed verifications in molecular methods; on the other hand, we use computational methods on the accumulated data to find those statistical results which can not be easily be discovered by conventional lab methods. This is a new era in the life science, and we need to be more ready to think computationally and do corporately.更多还原