Data-Driven Protein Engineering for Improving Catalytic Activity and Selectivity.

Protein engineering is essential for altering the substrate scope, catalytic activity and selectivity of enzymes for applications in biocatalysis. However, traditional approaches, such as directed evolution and rational design, encounter the challenge in dealing with the experimental screening process of a large protein mutation space. Machine learning methods allow the approximation of protein fitness landscapes and the identification of catalytic patterns using limited experimental data, thus providing a new avenue to guide protein engineering campaigns. In this concept article, we review machine learning models that have been developed to assess enzyme-substrate-catalysis performance relationships aiming to improve enzymes through data-driven protein engineering. Furthermore, we prospect the future development of this field to provide additional strategies and tools for achieving desired activities and selectivities.

A Universal, Continuous Assay for SAM-dependent Methyltransferases.

Enzyme-catalyzed late-stage functionalization (LSF), such as methylation of drug molecules and lead structures, enables direct access to more potent active pharmaceutical ingredients (API). S-adenosyl-l-methionine-dependent methyltransferases (MTs) can play a key role in the development of new APIs, as they catalyze the chemo- and regioselective methylation of O-, N-, S- and C-atoms, being superior to traditional chemical routes. To identify suitable MTs, we developed a continuous fluorescence-based, high-throughput assay for SAM-dependent methyltransferases, which facilitates screening using E. coli cell lysates. This assay involves two enzymatic steps for the conversion of S-adenosyl-l-homocysteine into H2 S to result in a selective fluorescence readout via reduction of an azidocoumarin sulfide probe. Investigation of two O-MTs and an N-MT confirmed that this assay is suitable for the determination of methyltransferase activity in E. coli cell lysates.

Structure- and Data-Driven Protein Engineering of Transaminases for Improving Activity and Stereoselectivity.

Amine transaminases (ATAs) are powerful biocatalysts for the stereoselective synthesis of chiral amines. Machine learning provides a promising approach for protein engineering, but activity prediction models for ATAs remain elusive due to the difficulty of obtaining high-quality training data. Thus, we first created variants of the ATA from Ruegeria sp. (3FCR) with improved catalytic activity (up to 2000-fold) as well as reversed stereoselectivity by a structure-dependent rational design and collected a high-quality dataset in this process. Subsequently, we designed a modified one-hot code to describe steric and electronic effects of substrates and residues within ATAs. Finally, we built a gradient boosting regression tree predictor for catalytic activity and stereoselectivity, and applied this for the data-driven design of optimized variants which then showed improved activity (up to 3-fold compared to the best variants previously identified). We also demonstrated that the model can predict the catalytic activity for ATA variants of another origin by retraining with a small set of additional data.

Creation of (R)-Amine Transaminase Activity within an α-Amino Acid Transaminase Scaffold.

The enzymatic transamination of ketones into (R)-amines represents an important route for accessing a range of pharmaceuticals or building blocks. Although many publications have dealt with enzyme discovery, protein engineering, and the application of (R)-selective amine transaminases [(R)-ATA] in biocatalysis, little is known about the actual in vivo role and how these enzymes have evolved from the ubiquitous α-amino acid transaminases (α-AATs). Here, we show the successful introduction of an (R)-transaminase activity in an α-amino acid aminotransferase with one to six amino acid substitutions in the enzyme's active site. Bioinformatic analysis combined with computational redesign of the d-amino acid aminotransferase (DATA) led to the identification of a sextuple variant having a specific activity of 326 milliunits mg-1 in the conversion of (R)-phenylethylamine and pyruvate to acetophenone and d-alanine. This value is similar to those of natural (R)-ATAs, which typically are in the range of 250 milliunits mg-1. These results demonstrate that (R)-ATAs can evolve from α-AAT as shown here for the DATA scaffold.