Si-Han Liu , Ligai Bai , Xu-Dong Wang , Qi-Qiang Wang , De-Xian Wang , Uwe T. Bornscheuer , Yu-Fei Ao
{"title":"以机器学习为指导的蛋白质工程改善转氨酶在中性pH值条件下的催化活性","authors":"Si-Han Liu , Ligai Bai , Xu-Dong Wang , Qi-Qiang Wang , De-Xian Wang , Uwe T. Bornscheuer , Yu-Fei Ao","doi":"10.1039/d5qo00423c","DOIUrl":null,"url":null,"abstract":"<div><div>Biocatalysis provides an eco-friendly and efficient method for the synthesis of fine chemicals, pharmaceuticals, and biofuels. However, the catalytic performance of enzymes is greatly reduced when they react under non-optimal pH conditions. Despite efforts in protein engineering to improve enzymatic pH dependence, predicting and optimizing catalytic activity under various pH conditions remain challenging. Recent advancements in machine learning (ML) provide a promising solution by enabling data-driven predictions of enzyme properties, thus reducing the need for labor-intensive experiments. This study aims to enhance enzyme activity and regulate pH dependence through ML-guided protein engineering. Using a transaminase from <em>Ruegeria</em> sp. TM1040 as the mutation template, we first collected high-quality experimental data from a series of variants by measuring their activity under a series of pH conditions. Based on these data, we developed an ML model to predict catalytic activity under different pH conditions. This work provides a powerful ML tool to guide the rational design of variants, which showed improved activity (up to 3.7-fold compared to the starting variant) at pH 7.5, and offers a data-driven protein engineering strategy to achieve co-optimization of enzyme activity and pH dependence, demonstrating the potential of ML to accelerate enzyme engineering for industrial applications.</div></div>","PeriodicalId":94379,"journal":{"name":"Organic chemistry frontiers : an international journal of organic chemistry","volume":"12 17","pages":"Pages 4788-4793"},"PeriodicalIF":0.0000,"publicationDate":"2025-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Machine learning-guided protein engineering to improve the catalytic activity of transaminases under neutral pH conditions†\",\"authors\":\"Si-Han Liu , Ligai Bai , Xu-Dong Wang , Qi-Qiang Wang , De-Xian Wang , Uwe T. Bornscheuer , Yu-Fei Ao\",\"doi\":\"10.1039/d5qo00423c\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Biocatalysis provides an eco-friendly and efficient method for the synthesis of fine chemicals, pharmaceuticals, and biofuels. However, the catalytic performance of enzymes is greatly reduced when they react under non-optimal pH conditions. Despite efforts in protein engineering to improve enzymatic pH dependence, predicting and optimizing catalytic activity under various pH conditions remain challenging. Recent advancements in machine learning (ML) provide a promising solution by enabling data-driven predictions of enzyme properties, thus reducing the need for labor-intensive experiments. This study aims to enhance enzyme activity and regulate pH dependence through ML-guided protein engineering. Using a transaminase from <em>Ruegeria</em> sp. TM1040 as the mutation template, we first collected high-quality experimental data from a series of variants by measuring their activity under a series of pH conditions. Based on these data, we developed an ML model to predict catalytic activity under different pH conditions. This work provides a powerful ML tool to guide the rational design of variants, which showed improved activity (up to 3.7-fold compared to the starting variant) at pH 7.5, and offers a data-driven protein engineering strategy to achieve co-optimization of enzyme activity and pH dependence, demonstrating the potential of ML to accelerate enzyme engineering for industrial applications.</div></div>\",\"PeriodicalId\":94379,\"journal\":{\"name\":\"Organic chemistry frontiers : an international journal of organic chemistry\",\"volume\":\"12 17\",\"pages\":\"Pages 4788-4793\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2025-04-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Organic chemistry frontiers : an international journal of organic chemistry\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/org/science/article/pii/S2052412925002955\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Organic chemistry frontiers : an international journal of organic chemistry","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/org/science/article/pii/S2052412925002955","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Machine learning-guided protein engineering to improve the catalytic activity of transaminases under neutral pH conditions†
Biocatalysis provides an eco-friendly and efficient method for the synthesis of fine chemicals, pharmaceuticals, and biofuels. However, the catalytic performance of enzymes is greatly reduced when they react under non-optimal pH conditions. Despite efforts in protein engineering to improve enzymatic pH dependence, predicting and optimizing catalytic activity under various pH conditions remain challenging. Recent advancements in machine learning (ML) provide a promising solution by enabling data-driven predictions of enzyme properties, thus reducing the need for labor-intensive experiments. This study aims to enhance enzyme activity and regulate pH dependence through ML-guided protein engineering. Using a transaminase from Ruegeria sp. TM1040 as the mutation template, we first collected high-quality experimental data from a series of variants by measuring their activity under a series of pH conditions. Based on these data, we developed an ML model to predict catalytic activity under different pH conditions. This work provides a powerful ML tool to guide the rational design of variants, which showed improved activity (up to 3.7-fold compared to the starting variant) at pH 7.5, and offers a data-driven protein engineering strategy to achieve co-optimization of enzyme activity and pH dependence, demonstrating the potential of ML to accelerate enzyme engineering for industrial applications.