肝醣合成激酶-3β抑制劑設計/搜尋之嵌合計算研究

Abstract

GSK-3β為一蛋白質激酶，在生物體中的作用與許多疾病諸如癌症、阿滋海默症以及糖尿病等息息相關。在本研究中，我們利用分子嵌合計算來研究GSK-3β抑制劑。 對於GSK-3β抑制劑的研究，Staurosporine（STU）是已被發現具有良好抑制效果化合物。在我們研究中，首先檢視兩STU相似結構化合物Bis-3與Bis-8，我們的計算結果顯示Bis-8評分值優於Bis-3，與實驗相符，凡得瓦作用力為主要貢獻者。在結合模式上，那些與實驗結構類似之構型--單五環與GSK-3β鉸鏈片段位置有氫鍵作用--是具有較高親和力之構型。我們分析∕討論此兩化合物及相關化合物結合模式的原理。此外，活性位置中結晶水，對於嵌合結果的影響，亦予以分析∕討論。 此外，為了搜尋其他結構類之GSK-3β抑制劑，我們以rule of five規則，篩選美國國家癌症研究所化學資料庫中24,535個分子。進行高速篩選前，我們以25個已知實驗IC50分子作測試，結果顯示所選之評分函數所得之值與實驗數值具有良好的相關性，顯示此評分函數之適用性。我們討論數種具有較佳親合力化合物與GSK-3β之間的作用及其結合模式構型。這些計算結果將有助於實驗學家設計與搜尋GSK-3β抑制劑。GSK-3β is a kinase protein that plays significant role in a number of biochemical reaction networks associated with diseases, including cancers, Alzheimer disease, and diabetics. In the present study, we used docking computation to aid in design and discovery of GSK-3β inhibitors. In the development of GSK-3β inhibitors, the natural compound, Staurosporine（STU）, has drawn much attention due to its strong inhibition effect. In the first part of this study, we used docking computation to investigate binding modes and affinities of several STU-like ligands and related ligands binding with GSK-3β. The computed results of two examined STU-like ligands, Bis-3 and Bis-8, showed that Bis-8 binds stronger than Bis-3, in accord with experimental results. Van der Waals interaction is the major contributor to the larger binding affinity of Bis-8. The binding modes with the classical hydrogen bondings of STU-like ligands with the hinge loop of GSK-3β were found to be among the binding modes of higher affinity. The binding principles of these two ligands with other examined ligands were examined. The effect of crystal water molecules in the docking calculation was examined and discussed. To further aid in developing GSK-3β inhibitors, we carried out high-throughput virtual screening for 24,535 compounds in the National Cancer Institute（NCI, USA）database which were selected based on the rules of five of drug compounds. Docking for 25 ligands with available experimental IC50 values were examined first. The computed scores are in good correlation with the IC50 values, showing feasibility of the chosen scoring function for docking for GSK3β. Binding affinities and modes of several top-ranked ligands were analyzed and discussed. These computed results should be useful for experimentalists in design/discovery of GSK-3β inhibitors.