【作者】 郝格非；

【导师】 杨光富； 湛昌国；

【作者基本信息】 华中师范大学 ， 农药学， 2011， 博士

【摘要】 农业在我国的国民经济中占有很大的比重,而农药是保障农业生产的重要“武器”,然而越来越普遍的抗性问题使农药陷入一种十分不利的境地。如何对付已经出现的抗性问题,又如何进行反抗性农药分子设计将是一项非常艰巨的任务,农药分子合理设计是解决以上问题的有效方法之一。然而合理设计是一项系统工程,其中包括多个环节,比如类农药性研究、农药分了作用机制研究、农药抗性预测研究和新型农药分子设计等。如何针对以上的诸多环节,采用有效的方法,并最终发现潜在的新型农药分子先导化合物是合理设计的关键和目标。因为提高新型先导化合物的发现效率,一方面可以极大地缩短农药研发的周期；另一方面可以免受现有抗性机制的困扰。随着结构生物学及计算化学等学科的不断发展,计算模拟的方法在农药分子合理设计中具有无可比拟的优势,并与实验科学形成很好的互补,在新农药的创制过程中扮演着越来越重要的角色。本论文采用计算模拟的方法围绕“农药合理设计的分子基础”这一关键的科学问题,系统地研究农药合理设计过程中的各个主要环节,其中包括类农药性研究、代表性农药分子的作用机制研究、农药抗性预测方法研究以及基于碎片的农药分子设计研究。提出了类农药性的规则；闸明了农药中多个重要靶标的分子作用机制；发展和完善了一些与农药分子合理设计相关的计算方法；并成功发现了一些全新结构的高活性分子。首先,以商品化农药分子为基础开展类农药性研究。对788个商品化农药分子的物理化学性质进行系统性地分析,重点考察了分子量(MW)、脂水分配系数(CLogP)、氢键供体数目(HBA)、氢键给体数目(HBD)、可旋转键数目(ROB)以及芳香键数目(ARB)的分布情况,本文首次将光稳定性引入到类农药性研究中；找出区分不同种类农药分子的关键物理化学参数,并考察了不同时期上市农药品种的物理化学性质随着年代的变化情况。在此基础上提出了类农药性规则,即：MW<435 Da,CLogP<6,HBA<6,HBD<2,ROB<9,ARB<17。其次,以农药上最新最重要的发现之一(生长素及赤霉素受体的发现)为基础,综合运用计算模拟的方法系统研究了生长素和赤霉素与各自受体相互作用的分子机制。在生长素受体TIR1蛋白中,辅因子InsP6扮演了“构象稳定剂”的角色,生长素及其类似物不仅发挥了“分子胶水”的作用,同时还可以诱导Phe82侧链发生构象变化以适应底物Aux/IAA蛋白的结合,并且Phe351的侧链构象变化对TIR1蛋白识别底物也起到了非常重要的作用。此外,在赤霉素受体GID1蛋白中,我们发现了赤霉素进出受体的新通道,并从理论上对赤霉素介导GID1氮端α螺旋变构的调节机制进行了驳斥,证明了赤霉素在通过新通道时并不介导GID1氮端α螺旋变构,而是通过稳定GID1与DELLA蛋白之间的氢键来起到调节作用的。这是关于赤霉素与受体相互作用的一种全新机制。这些研究为将来开展新型人工植物激素的合理设计奠定良好基础,同时对其它农药的作用机制研究具有借鉴意义。再次,针对农药分子抗性发展新型的抗性预测方法。首先通过多种分子模拟方法对近来出现的一例奇特抗性—水麻Gly210缺失导致的PPO酶除草剂的广谱抗性进行抗性机制研究,发现Gly210与Ser424之间重要氢键的丢火可以引起活性腔的局部结构发生微妙的变化,从而导致PPO酶除草剂与Arg128之间的氢键变弱是产生抗性的原因。此外,针对所用到的预测方法的弊端,发展了一种全新的抗性预测方法—计算突变扫描(Computational Mutation Scanning,CMS),实现了在野生型复合物动力学轨迹基础上对氨基酸侧链的任意变换从而提高了预测的效率；相对于计算丙氨酸扫描(Computational Alanine Scanning, CAS)而言,由于引入了快速熵变计算,所以大大地提高了结合自由能计算的精度；在对模板体系的测试中,该方法的预测准确率达到了80%左右,是一种快速而有效的抗性预测计算工具。并且我们利用该方法对一些PPO酶商品化除草剂进行了抗性预测,并发现了一些潜在的抗性突变体。这些研究为今后其它农药分子的抗性预测提供了理论基础。最后,本文针对两个农药体系(以细胞色素bcl复合物为靶标的杀菌剂和以PPO酶为靶标的除草剂)自身的特点发展了新的基于碎片的农药分子设计策略。首先将商品化农药分子按照碎片筛选规则进行切割,并建立适于农药分子设计的碎片库；然后在细胞色素bcl复合物体系中固定现有抑制剂的药效团结构,采用碎片生长方法来优化其与Phe128及Phe274之间的疏水相互作用,成功发现了多个不同结构类型的高活性抑制剂(活性最高达到几十pM级别)；此外我们通过减少抑制剂与抗性位点之间的相互作用、同时引入与靶标内不可变位点(PPO酶辅因子FAD)的相互作用、并筛选与底物作用模式类似的碎片等反抗性策略,成功发现了新型作用模式的PPO酶高活性抑制剂(活性达到几十nM级别)。这些研究策略也可以为其它体系的新型农药分子设计提供理论方法上的借鉴。更多还原

【Abstract】 Agriculture accounts for a great proportion in the national economy of our country and the pesticide is an important "weapon" to ensure its production. However, resistance is increasingly popular, which may bring the pesticide to disadvantage. It is an arduous task how to antagonize resistance and how to do anti-resistance pesticide design. Rational design of pesticide molecule is a way to solve the problem. However, rational design is a systematic work, which includes many components, for example, pesticide likeness, molecular mechanism of pesticide, pesticide resistance prediction, novel pesticide design, and so on. The key of rational design is how to select effective methods based on the above components to finally discover potentially novel lead compound. Improving the lead compound discovering rate can reduce the period of pesticide discovering and render free of the resistance mechanism. With the development of structural biology and computational chemistry, computational simulation, complementary to other experimental methods, plays more and more important role in the process of new pesticide discovery and development.In this thesis, computational simulation method was used to explain "the molecular basis of pesticide rational design". We systematically studied several important components in the process of pesticide rational design, including pesticide likeness, the molecular mechanism of representative pesticide, the methodology of pesticide resistance prediction, and fragment based pesticide design. The rule of pesticide likeness was raised, the molecular mechanisms of several important targets in pesticide were understood, some new computational methods for pesticide rational design were developed, and some high bioactivity molecules were discovered.First of all, the pesticide likeness study was based on the marketed pesticides. We systematically analyzed the distribution of physiochemical properties of 788 marketed pesticides:molecular weight (MW), CLogP, hydrogen bond acceptor (HBA), hydrogen bond donor (HBD), rotational bond (ROB), and aromatic bond (ARB). For the first time, the index of photostability (aromatic bond (ARB)) was introduced to explain pesticide likeness. We found the key physiochemical properties to differentiate different kinds of pesticide and analyzed the variability of physiochemical properties along with the market time. Based on the above, the rule for pesticide likeness was that "molecular weight≤435 Da, ClogP≤6, H-bond acceptor≤6, H-bond donor≤2, the number of rotatable bond≤9, and the number of aromatic bond≤17".Secondly, we explored the molecular interaction mechanism by following one of the most important discoveries in the area of pesticide—the receptors of auxin and gibberellin. In the auxin receptor (TIR1 protein), the co-factor InsP6 acted as a’conformational stabilizer’; auxin and its analogs not only played a role of "molecular glue", but also induced the sidechain of Phe82 to undergo a conformational change in order to accommodate the subsequent binding of the substrate Aux/IAA. The change of the sidechain of Phe351 was also important for the recognition of Aux/IAA. Besides, in the gibberellin receptor (GED1 protein), we discovered a new channel for the entering and leaving of gibberellin, which theoretically refuted the gibberellin induced allosteric transformation mechanism. We proved that the regulation of gibberellin was not by inducing the allosteric transformation of N-terminal a-helix, but by stabilizing the hydrogen bonding interaction between GID1 and DELLA protein. This is a new mechanism of the interaction between gibberellin and its receptor. The research in this chapter provides a good foundation for the rational design of novel phytohormone and can be useful for other pesticide mechanism studies.Thirdly, as for the pesticide resistance, novel prediction method was developed. At first, multiple molecular modeling methods were used to uncover the mechanism of a peculiar mutation (Gly210 deletion in A. tuberculatus PPO) which is able to induce broad-spectrum resistance. This was due to the weakness of the hydrogen bonds between PPO herbicides and Arg128, which was caused by a subtle change of the local structure of the active site because of the loss of a key hydrogen bond between Gly210 and Ser424. Moreover, a new resistance prediction method—Computational Mutation Scanning (CMS) was developed to offset disadvantages of the current method, in which the computational mutation was realized on a wild-type complex MD trajectory to improve the prediction rate. Compared with the traditional method—Computational Alanine Scanning (CAS), the calculation accuracy of the binding free energy was improved by introducing a rapid entropy calculation. In the testing system, the prediction accuracy rate of CMS was about 80%, so it could be used as a quick and effective computational tool for resistance prediction. It is worth pointing out that we found some resistance mutations in PPO system based on the CMS results for some marketed PPO herbicides. All of these can supply the theoretical basis for the resistance prediction in other pesticide system.Finally, we developed novel fragment based pesticide design strategies based on the characteristics of two pesticide systems (i.e., fungicide targeted to cytochrome bcl complex and herbicide targeted to PPO enzyme). In the first place, pesticide fragment library was constructed by decomposing the structures of marketed pesticides according to the fragment rules. Then, in the system of cytochrome bcl complex, common pharmacophore structure of the current inhibitors was fixed and fragments were grafted on it to optimize the hydrophobic interaction with Phe128 and Phe274. Based on this strategy, we successfully discovered several inhibitors with potent bioactivity (the highest bioactivity was in double pM level). At the same time, we also successfully discovered new PPO inhibitors (the bioactivity was in double nM level) with novel binding mode and potential anti-resistance ability by using several anti-resistance strategies, such as reducing the interaction between the inhibitor and resistance position, introducing the interaction with the unchanged position of the target (the cofactor FAD in PPO), and focusing to screen fragments that have similar binding mode with the substrate. These strategies can provide theoretical references for the novel pesticide design in other systems.更多还原