Cytochrome P450 monooxygenases are an extensive and unique class of enzymes, which can regio- and stereo-selectively functionalise hydrocarbons by way of oxidation reactions. These enzymes are naturally occurring but have also been extensively applied in a synthesis context, where they are used as efficient biocatalysts. Recently, a biosynthetic pathway where a cytochrome P450 monooxygenase catalyses a critical step of the pathway was uncovered, leading to the production of a number of products which display high antitumour potency. In this work, we use computational techniques to gain insight into the factors that determine the relative yields of the different products. We use conformational search algorithms to understand the substrate stereochemistry. On a machine learned 3D protein structure, we use molecular docking to obtain a library of favourable poses for substrate-protein interaction. With molecular dynamics, we investigate the most favourable poses for reactivity on a molecular level, allowing us to investigate which protein-substrate interactions favour a given product and thus gain insight into the product selectivity.

中文翻译：

---

探索细胞色素 P450 的选择性以增强新型抗癌药物的合成

细胞色素 P450 单加氧酶是一类广泛而独特的酶，可以通过氧化反应对碳氢化合物进行区域和立体选择性功能化。这些酶是天然存在的，但也广泛应用于合成领域，用作有效的生物催化剂。最近，细胞色素 P450 单加氧酶催化该途径关键步骤的生物合成途径被发现，从而产生了许多具有高抗肿瘤效力的产品。在这项工作中，我们使用计算技术来深入了解决定不同产品相对产量的因素。我们使用构象搜索算法来了解底物立体化学。在机器学习的 3D 蛋白质结构上，我们使用分子对接来获得底物-蛋白质相互作用的有利姿势库。通过分子动力学，我们在分子水平上研究最有利的反应姿势，使我们能够研究哪些蛋白质-底物相互作用有利于给定的产物，从而深入了解产物的选择性。