定向进化在蛋白质工程中的应用研究进展

Recent development of directed evolution in protein engineering

定向进化旨在通过基因多样化和突变体库筛选的迭代循环,加速实现在胞内或胞外进行的自然进化过程。近年来,因其强大的功能而被广泛应用于酶工程当中。本文概述了近十年助力定向进化发展的最新技术,包括胞外和胞内高效构建基因突变体库的方法、高通量筛选突变体库的方法、连续定向进化策略、自动化生物合成平台助力定向进化的策略、计算机技术辅助定向进化的应用实例。为了阐述定向进化在酶工程中的应用价值,本文着重讨论了利用定向进化技术对酶进行改造的代表性案例,其中包括改善酶在有机溶剂中的耐受性、提高酶的热稳定性、增强天然酶对非天然底物的催化能力、提高酶催化化学反应的选择性(包括区域选择性、立体选择性和对映选择性)以及拓展酶催化的反应类型。最后,本文对定向进化在未来可能遇到的挑战及应用前景进行了归纳总结。

Directed evolution aims to accelerate the natural evolution process in vitro or in vivo through iterative cycles of genetic diversification and screening or selection.It has been one of the most solid and widely used tools in protein engineering.This review outlines the representative methods developed in the past 10 years that increase the throughput of directed evolution,including in vitro and in vivo gene diversification methods,high-throughput selection and screening methods,continuous evolution strategies,automation-assisted evolution strategies,and AI-assisted protein engineering.To illustrate the significant applications of directed evolution in protein engineering,this review subsequently discusses some remarkable cases to show how directed evolution was used to improve various properties of enzymes,such as the tolerance to elevated temperature or organic solvent,the activities on non-native substrates,and chemo-,regio-,stereo-,and enantio-selectivities.In addition,directed evolution has also been widely used to expand the biocatalytic repertories by engineering enzymes with abiotic activities.In addition to the native enzymes,directed evolution has also been used to engineer de novo designed enzymes and artificial metalloenzymes with activities comparable to or exceeding the ones of the native enzymes.Finally,this review has pointed out that further improving the efficiency and effectiveness of directed evolution remains challenging.Some advanced continuous evolution and high throughput screening strategies have been succesfully demonstrated in improving the throughput of directed evolutions extensively,but they have been limited to engineering certain protein targets.To resolve those issues,continuously improved computational modeling tools and machine learning strategies can assist us to create a smaller but more accurate library to enhance the probabilities of discovering variants with improved properties.Additionally,laboratorial automation platforms coupled with advanced screening and selection techniques also have great potential to extensively explore the protein fitness landscape by evolving multiple targets continuously in a high throughput manner.