大学化学 >> 2026, Vol. 41 >> Issue (1): 144-158.doi: 10.12461/PKU.DXHX202505028

所属专题: 化学实验数字化设计竞赛获奖作品

专题 上一篇    下一篇

靶向氧化还原系统中谷胱甘肽的抗癌药物的数字化创新设计实验

管堂升, 杨睿智, 耿平越, 林禹宏, 胡水, 陈小娟, 李厚金, 沈勇   

  1. 中山大学化学学院, 广州 510006
  • 收稿日期:2025-05-08 录用日期:2025-07-04 发布日期:2025-12-30
  • 通讯作者: 李厚金, 沈勇 E-mail:ceslhj@mail.sysu.edu.cn;cessy@mail.sysu.edu.cn Houjin Li, Yong Shen
  • 基金资助:
    广东省本科高校教学质量与教学改革工程建设项目(化学实验教研室);中山大学本科教学质量工程项目

Innovative Digital Experiments for Glutathione-Targeted Anticancer Drugs in Redox Systems

Tangsheng Guan, Ruizhi Yang, Pingyue Geng, Yuhong Lin, Shui Hu, Xiaojuan Chen, Houjin Li, Yong Shen   

  1. School of Chemistry, Sun Yat-Sen University, Guangzhou 510006, China
  • Received:2025-05-08 Accepted:2025-07-04 Published:2025-12-30
  • Contact: Houjin Li, Yong Shen E-mail:ceslhj@mail.sysu.edu.cn;cessy@mail.sysu.edu.cn

摘要: 谷胱甘肽参与调节细胞内活性氧水平,消耗癌细胞内的谷胱甘肽,可导致活性氧水平上升,诱导癌细胞失活、凋亡,这成为癌症治疗的潜在靶点。基于α,β-不饱和羰基化合物易与谷胱甘肽发生Michael加成反应而消耗谷胱甘肽,利用ChemDraw、Chem3D、Gaussian、Chemprop等软件创新设计了靶向谷胱甘肽的抗癌药物AI虚拟筛选实验,构建了抗癌药物分子数据库,计算表征了化合物与谷胱甘肽的反应性。AI预测结果与文献记录的实验数据具有高度一致性,说明基于该模型的虚拟筛选方法科学、结果可靠。AI虚拟筛选出潜在的靶向抗癌药物经过化学反应动力学实验、药理实验进一步筛选验证,得到了具有深入研究价值的先导化合物。AI赋能新药研发,降低了筛选成本,提高筛选效率和准确性。本实验可作为化学生物学实验或者化学信息学课程的实践内容。AI技术助力学生更好地理解和掌握知识,让教学内容更具高阶性、创新性和挑战度,有利于调动学习积极性,提升创新思维和创新能力,培养面向未来的卓越人才。

关键词: 谷胱甘肽, 抗癌药物, 人工智能, 虚拟筛选, Michael加成反应

Abstract: Glutathione (GSH) plays a crucial role in regulating intracellular reactive oxygen species (ROS) levels. Depletion of glutathione in cancer cells leads to elevated ROS levels, subsequently inducing cellular inactivation and apoptosis, thereby establishing GSH as a promising therapeutic target for cancer treatment. Capitalizing on the propensity of α,β-unsaturated carbonyl compounds to undergo Michael addition reactions with glutathione, we developed an innovative AI-driven virtual screening protocol using ChemDraw, Chem3D, Gaussian, and Chemprop software. This approach involved constructing a molecular database of anticancer compounds and computationally characterizing their reactivity with glutathione. The AI prediction results demonstrated remarkable consistency with experimentally reported data in the literature, validating the scientific rigor and reliability of our virtual screening methodology. Potential drug candidates identified through AI screening were subsequently evaluated via chemical kinetics and pharmacological assays, yielding lead compounds worthy of further investigation. This AI-enhanced drug discovery approach significantly reduces screening costs while improving both efficiency and accuracy. The experimental framework serves as an effective pedagogical tool for chemical biology or chemoinformatics courses, where AI technology facilitates deeper understanding of core concepts. This innovative teaching method enhances curriculum sophistication, stimulates student engagement, and cultivates critical thinking and innovative capacity, ultimately fostering the development of exceptional talents for future scientific challenges.

Key words: Glutathione, Anticancer drug, Artificial intelligence, Virtual screening, Michael addition reaction