以蛋白質結構為基礎之合成可行性先導藥物最適化電腦輔助藥物設計

Structure-Based Lead Optimization with Synthetic Accessibility in Computer-Aided Drug Design

林芳宇(Fang-Yu Lin)

指導教授 ： 曾宇鳳

國立臺灣大學/電機資訊學院/生醫電子與資訊學研究所/碩士(2011年)

https://doi.org/10.6342/NTU.2011.10951

摘要

先導藥物結構最適化是藥物開發中重要的一環，然而在抑制劑設計上，化學合成可行性亦是設計的一項考量。因此為設計出結構適化的抑制劑，並考慮合成可行性，本研究將針對先導藥物結構最適化的設計方法探討，根據其設計的演算法分為兩種，一為跳躍示式分子子結構置換，另一為考慮合成可行性的藥物設計。
在論文第一部分，將詳述LeadOp系統，此系統以蛋白質結構為基礎，針對已知的抑制劑結構，利用跳躍示式分子子結構置換方法結構優化已知的抑制劑結構，進而找出更有效的抑制劑。LeadOp考慮已知抑制劑分子的子結構，藉由計算Group efficiency (GE)的計分，評估子結構與蛋白質活性區的作用關係，並從資料庫中選出子結構較佳者進行替代；同時，我們考慮到分子性質並作為篩選條件，排除不穩定的子結構，並選出具有增加分子之間作用性質的子結構做置換。置換完的分子經過結構最佳化以及分子與蛋白質結合能量的計算，篩選出較佳的分子作為抑制劑。在此論文第二部分中，將詳述LeadOp+R 系統, 此系統同時評估抑制劑分子與蛋白質活性區的作用關係以及先導藥物設計上之分子可合成性，進而達到先導藥物結構最適化的設計。從預先擷取的化學反應合成上的規則，並以蛋白質結構為基礎，根據化學反應規則進行蛋白質活性區塊中之抑制劑分子合成，同時每一化學反應步驟的產物皆以GE排序計分，分數較高者將被視為新的反應物進行下一步合成直到系統執行結束，最後合成的產物將是結構最適化後的分子。
本論文中，分別以三組藥物及相對之蛋白質系統，依序BRaf、5-LOX、Tie-2，進行演算法上結構最適化的評估以驗證LeadOp、LeadOp+R的效能，進而設計出已知的藥物，其抑制活性與預測的結果有相同的趨勢。

關鍵字

片段設計 ； 結構跳躍式 ； 先導藥物最適化 ； 以蛋白質結構為基礎設計 ； 電腦輔助分子設計 ； 電腦輔助合成設計

並列摘要

In this thesis, we describe two systems of structure based de novo optimization process, called “LeadOp” (short for Lead Optimization), and “LeadOp+R” (short for Lead Optimization with chemical Reaction).
In the first system, LeadOp, we described a structure based de novo optimization process, “LeadOp”, by decomposing a structure into fragments of different parts either by chemical rules or user-defined, evaluate each fragment at each part in a pre-docked fragment database that ranked fragments according to specific fragment-receptor binding interactions, replace fragments with less contribution to binding, and finally reassemble fragments from each part to form a ligand. The fundamental idea was to replace “bad” fragments of an inhibitor and replace with “good” fragments while leaving the rest of the inhibitor in the original core to improve the activity for lead optimization. The fragments were selected from a collection of 27,417 conformers by exhaustive docking at the target binding sites from synthesizable docked molecular building blocks and fragments from decomposing all known inhibitors from DrugBank database and related inhibitors. However, even with the fragment based design from common building blocks, it is still a challenge for synthesis. In the second system, “LeadOp+R” was developed based on 198 classical chemical reactions to consider the synthetic accessibility while optimizing leads. LeadOp+R first allows user to identify a preserved space defined by the volume occupied by a fragment of the query molecule to be preserved. Then LeadOp+R searches for building blocks with the same preserved space as initial reactants and grows molecules towards the preferred receptor-ligand interactions according to reaction rules from reaction database in LeadOp+R. Multiple conformers of each intermediate product were considered and evaluated at each step. The conformer with the best group efficiency score would be selected as the initial conformer of the next building block until the program finished optimization for all selected receptor-ligand interactions.
The two systems were examed with three biomolecular sysmtes, including mutant B-Raf kinase, Tie-2 kinase, and human 5-LOX inhibitor design. The “LeadOp” methodology was able to optimize the query molecules and systematically developed improved analogs for each of our example systems. The “LeadOp+R” methodology optimized the query molecule and systematically developed improved analogs along with their proposed synthetic routes. The suggested synthetic route (proposed from our synthesis algorithm) was the same as the published synthetic route devised by a synthetic/organic chemist.

並列關鍵字

Fragment-based ； Scaffold-Hopping ； Lead Optimization ； Structure-based Drug Design ； Computer-aided Molecular Design ； Computer-assisted Synthesis

參考文獻

REFERENCE

1. Brown, N.; Lewis, R. A., Exploiting QSAR Methods in Lead Optimization. Curr. Opin. Drug Discov. Dev. 2006, 9, 419-424.

2. Zhang, Q.; Muegge, I., Scaffold Hopping through Virtual Screening Using 2D and 3D Similarity Descriptors: Ranking, Voting, and Consensus Scoring. J. Med. Chem. 2006, 49, 1536-1548.

3. Senger, S., Using Tversky Similarity Searches for Core Hopping: Finding the Needles in the Haystack. J. Chem. Inf. Model. 2009, 49, 1514-1524.

4. Bohm, H.-J.; Flohr, A.; Stahl, M., Scaffold Hopping. Drug Discov. Today 2004, 1, 217-224.