Next-generation computational enzyme design for the creation of efficient artificial biocatalysts

酶是已知的最有效的催化剂。它们可以将化学反应加速10亿倍以上，表现出无与伦比的选择性，并且完全可生物降解。如果我们能够从零开始创造出能够高效催化任何化学反应的人工酶，它将为目前无法使用天然酶实现的高价值生物技术打开大门。其中包括精细化学品和药品的生物催化合成、代谢紊乱的治疗以及受污染的水和土壤的生物修复。在这个拟议的研究项目中，我们将开发下一代酶设计方法，并使用它们来创造高效的人工生物催化剂，用于合成有价值的化学品。这些方法将基于我们在上一个资助周期中开发的多状态蛋白质设计算法，该算法允许对蛋白质进行建模，而不是像传统方法那样将其作为单一结构，而是作为更现实地代表这些分子可以在溶液中采样的结构集合。这一点很重要，因为酶必须“移动”才能以极高的精度和效率催化化学反应。迄今为止，所有已发表的计算酶设计的例子都集中在单一结构状态的设计上，而没有考虑到酶在多种状态之间交换的能力，因此产生的人工酶的催化效率比天然酶差几个数量级。我们提出的多态酶设计算法的开发和实验验证将在以下研究主题中进行。1)我们将开发方法来设计完整的酶催化循环，而不是像以前那样只考虑单一状态，并利用它们来创造高效的生物催化剂，用于工业上感兴趣的多步反应。2)我们将开发方法来预测远离活性位点的突变，这将对催化产生有利的影响，这是创造高活性酶的基本要求，但仍然难以捉摸。3)我们将开发多态设计方法来控制酶的构象平衡，并最终创造更复杂反应所需的大构象变化。从长远来看，我们将结合这些方法，为有价值的生物技术应用创造“按需”生物催化剂。拟议的研究计划将为不同群体的学员提供世界一流的计算和实验技术研究培训，帮助他们获得制药、生物技术和生物制品行业急需的技能。由于酶可以应用于广泛的商业和社会相关问题，我们提出的研究将产生新的技术，以解决21世纪加拿大在工业，环境和医学方面的重要挑战。

Enzymes are the most efficient catalysts known. They can accelerate chemical reactions by more than a billion times, display unmatched selectivity, and are completely biodegradable. If we could create, from scratch, artificial enzymes that can catalyze any chemical reaction with high efficiency, it would open the door to highly valuable biotechnologies that are currently inaccessible using natural enzymes. These include biocatalytic syntheses of fine chemicals and pharmaceuticals, treatment of metabolic disorders, and bioremediation of contaminated water and soil. In this proposed research program, we will develop next-generation enzyme design methods and use them to create efficient artificial biocatalysts for the synthesis of valuable chemicals. These methods will be based on the multistate protein design algorithms that we developed in the previous funding cycle, which allow proteins to be modeled, not as a single structure as is traditionally done, but as structural ensembles that more realistically represent the range of conformations that these molecules can sample in solution. This is important because enzymes must "move" in order to catalyze chemical reactions with extreme precision and efficiency. To date, all published examples of computational enzyme design have focussed on the design of a single structural state, not taking into account the ability of enzymes to exchange between multiple states, and thereby yielded artificial enzymes with catalytic efficiencies several orders of magnitude worse than those of their natural counterparts. The development and experimental validation of our proposed multistate enzyme design algorithms will be undertaken in the following research themes. 1) We will develop methods to design complete enzymatic catalytic cycles, instead of considering only a single state as was done previously, and use them to create efficient biocatalysts for multistep reactions of industrial interest. 2) We will develop methods to predict mutations far from the active site that will have a favorable effect on catalysis, an essential requirement for the creation of highly active enzymes that nevertheless remains elusive. 3) We will develop multistate design methods to control enzyme conformational equilibria and eventually create large conformational changes required for more complex reactions. In the long term, we will combine these methods to create biocatalysts `on-demand' for valuable biotechnological applications. The proposed research program will provide world-class research training, in both computational and experimental techniques, to a diverse group of trainees, helping them acquire skills highly sought after by the pharmaceutical, biotech, and bioproducts industries. As enzymes can be applied to a wide range of commercially and societally relevant problems, our proposed research will yield new technologies required to tackle important challenges of 21st century Canada in industry, the environment and medicine.