PDZ结构域结合中间序列结合特性的研究

【作者】 牟一；

【导师】 高友鹤；

【作者基本信息】 北京协和医学院 ， 生理与病理生理学， 2012， 博士

【摘要】 ·第一部分研究了PDZ结构域结合中间序列的结合特性蛋白质-蛋白质相互作用在生命活动中广泛存在,主要由蛋白质结构域来高效介导。大规模的蛋白质相互作用可以通过研究结构域的结合特性来实现。PDZ、SH3、WW等结构域通过一个或多个识别“口袋”来识别和结合配体蛋白的一段保守的短肽序列。然而蛋白质相互作用可能比我们现在认识到的复杂,还有很多的相互作用未被研究清楚。就PDZ结构域而言,它通常结合配体蛋白C末端4-5个氨基酸残基,近来研究发现其也能够结合配体蛋白的中间序列,与自身或其他结构域聚合,或与膜上的脂类结合。这些结合方式使蛋白质的相互作用呈现出多样性与复杂性。首先,为了系统地高通量地研究PDZ结构域结合中间序列的结合特性,我们在酵母双杂交系统中构建了中间序列随机八肽文库。文库的构建策略是将所有序列的C末端设计成相同的序列,从而保证PDZ结构域结合的是中间序列。而后,使用文库筛选了24个PDZ结构域,其中包括已知文献报道的能够结合中间序列的PDZ结构域；实验室已经构建好的PDZ结构域和一些能够结合脂类的PDZ结构域。实验表明有14个结构域具有结合中间序列的特性,其中6个结构域未曾被报道。序列具有偏好性,偏好较强的具有一致序列。与C末端序列的结合特性相比,中间序列的结合特性表现出多样性,只有ZO1-PDZ1结构域的中间序列特性与其C末端特性相似。筛选结果也反映出PDZ结构域对于中间序列和C末端序列的偏好性。另外,HtrA2-PDZ的偏好性较弱,没有一致序列,表现为结合疏水多肽。中间序列的筛选结果说明有更多的PDZ结构域可以结合中间序列。此外,应用中间序列的结合特性可以来寻找已知相互作用的结合位点,也可以利用中间序列或一致序列来搜索数据库发掘潜在的天然相互作用蛋白。·第二部分研究了日本血吸虫蛋白GIPC3的分子特征及其PDZ结构域的配体结合特性血吸虫病仍然是一个全球性的严重的公共卫生问题,流行于77个国家和地区,感染者超过2.3亿。依赖于单一药物吡喹酮大规模长期治疗引起对抗药性问题的关注。PDZ结构域蛋白被认为是下一代药物开发的潜在靶点。在本论文的第二部分主要研究日本血吸虫PDZ结构域蛋白SjGIPC3的分子特征及利用随机多肽文库探索其PDZ结构域的结合特性。SjGIPC3是含有单个PDZ结构域的蛋白。多重序列比对分析显示SjGIPC3与其同源物在GH1与PDZ结构域相对保守,但PDZ结构域中的C末端结合环处并不保守,其中的经典的GLGF模体被SFGL替代。进化分析表明日本血吸虫SjGIPC3与吸虫同源物形成一分支,而与其他物种相距较远。分期转录分析显示,SjGIPC3基因在侵入宿主后转录水平显著升高；在雄性成虫的转录水平为最高。在酵母双杂交系统中,我们初步探索了SjGIPC3-PDZ结合C末端序列的结合特性。用SjGIPC3全长蛋白或其PDZ结构域为诱饵蛋白,筛选了本室利用人基因组构建的新型随机多肽文库,实验表明SjGIPC3PDZ结构域主要结合Ⅰ类和Ⅱ类配体,但以结合Ⅰ类配体为主；用SjGIPC3全长蛋白为诱饵筛选文库得到的C末端序列比用PDZ结构域为诱饵得到的序列的规律性更强,提示SjGIPC3蛋白的N端和C端区域可能对PDZ结构域的配体结合特性产生影响。根据PDZ结构域的结合规律,利用本室开发的Tailfit软件预测并在酵母双杂交系统中验证了4个潜在的SjGIPC3的天然配体,其中之一为NMDA受体。更多还原

【Abstract】 ●In Part I, we studied the non-canonical binding properties of PDZ domains to internal sequences.There are enormous protein-protein interaction that mainly mediated by protein domains in multiple biological processes. Large-scale protein interactions can be investigated by studying the binding properties of protein domain. Protein domains (such as PDZ, SH3, SH2, WW, etc.) can canonically recognize a conservative short peptide sequence through one or more binding "pocket". However, there are probably more ways of protein interaction than we currently realize, some non-canonical protein interactions were still needed to be investigated. Like PDZ domains, they are important protein interaction modules that are considered mainly recognize C-terminal4-5amino acids, they can also non-canonically bind internal sequences, polymerize with themselves or other PDZ domains, or can bind to some lipids. Protein interactions become much more diverse and complicate by discovering those non-canonical interaction modes which expand the research field and provide more powerful evidence.First, to systematically and high-throughput study the binding characteristics of PDZ domains to internal sequences, a nearly random octapeptide library with no stop codon designed in the insert was constructed in the yeast two-hybrid vector. To ensure the PDZ domains really interact with internal sequences, all the C-termini of sequences of the whole library were the same. Then, a total of twenty-four PDZ domains were used as baits to screen the library, including some reported internal sequence-binding PDZ domains, some constructed PDZ domain clones of our lab and some PDZ domains which can bind with membrane lipids. Among the twenty-four PDZ domains, fourteen were found to bind internal sequences, including six that were not previously reported to bind internal sequences. PDZ domains showed strong amino acids preferences; only the consensuse of ZO1-PDZ1was similar to its C-terminal ligands, but the other consensuses were more diverse than those of common C-terminal ligands. In addition, HtrA2-PDZ with no strong consensus recognized extended hydrophobic peptides. The results suggested that more PDZ domains can bind internal sequences. Moreover, using the binding properties of internal sequences can help to find the potential binding site of reported interactions and to search the database to find potential natural binding partners. ●In Part Ⅱ, we presented the molecular characteristics of Schistosoma japonicum PDZ domain-containing protein, SjGIPC3and investigated the ligand binding properties of its PDZ domain.Schistosomiasis remains a serious global health problem, which afflicts more than230million people in77countries. The long-term mass treatments with the only available drug, praziquantel, have caused the concern of the drug resistance. PDZ domain containing proteins are the potential target for next-generation of drug development. In this second part, we presented the molecular characteristics of the Schistosoma japonicum PDZ domain-containing protein, SjGIPC3, and explored its ligand binding specificity by yeast two-hybrid (Y2H) using random peptide libraries.SjGIPC3is a single PDZ domain-containing protein. Multiple sequences alignment revealed that GH1and PDZ domain of SjGIPC3was relatively conserved with its orthologs, however, the carboxylate-binding loop within the PDZ domain was not, and the classic GLGF motif is substituted by SFGL. The phylogenetic analysis revealed that SjGIPC3and other trematode orthologs clustered into a well-defined cluster but was distinguishable from those of other phyla. Transcriptional analysis by qRT-PCR showed that SjGIPC3gene was particularly expressed in the stage within the host, especially in male adult worms.In the system of Y2H, we investigated the C-termini binding properties of SjGIPC3-PDZ. We used SjGIPC3full-length protein or only its PDZ domain as bait, to screen an arbitrary peptide library. We found that the SjGIPC3PDZ domain can both bind class Ⅰ and Ⅱ ligands, but more preferred to bind class I ligand. The binding sequences of SjGIPC3full-length protein have more proterties than those of SjGIPC3-PDZ, indicating that the N-terminal and C-terminal regions flanking the PDZ domain may affect the binding proterties of the PDZ domain. Further, native ligands were predicted based on the C-termini binding properties with Tailfit software, four potential ligands were confirmed by Y2H system, one of them was NMDA receptor.更多还原