Search

The combinatorial composition of proteins has triggered the application of machine learning in enzyme engineering. By predicting how protein sequence encodes function, researchers aim to leverage machine learning models to select a reduced number of optimized sequences for laboratory measurement with the aim to lower costs and shorten timelines of enzyme engineering campaigns. In this review, we will highlight successful algorithm-aided protein engineering examples, including work carried out within the NCCR Catalysis. In this context, we will discuss the underlying computational methods developed to improve enzyme properties such as enantioselectivity, regioselectivity, activity, and stability. Considering the rapid maturing of computational techniques, we expect that their continued application in enzyme engineering campaigns will be key to deliver additional powerful biocatalysts for sustainable chemical synthesis.

Enzymes are potent catalysts with high specificity and selectivity. To leverage nature's synthetic potential for industrial applications, various protein engineering techniques have emerged which allow to tailor the catalytic, biophysical, and molecular recognition properties of enzymes. However, the many possible ways a protein can be altered forces researchers to carefully balance between the exhaustiveness of an enzyme screening campaign and the required resources. Consequently, the optimal engineering strategy is often defined on a case-by-case basis. Strikingly, while predicting mutations that lead to an improved target function is challenging, here we show that the prediction and exclusion of deleterious mutations is a much more straightforward task as analyzed for an engineered carbonic acid anhydrase, a transaminase, a squalene-hopene cyclase and a Kemp eliminase. Combining such a pre-selection of allowed residues with advanced gene synthesis methods opens a path toward an efficient and generalizable library construction approach for protein engineering. To give researchers easy access to this methodology, we provide the website LibGENiE containing the bioinformatic tools for the library design workflow.

Squalene–hopene cyclases (SHCs) have great potential for the industrial synthesis of enantiopure cyclic terpenoids. A limitation of SHC catalysis has been the enzymes’ strict (S)-enantioselectivity at the stereocenter formed after the first cyclization step. To gain enantio-complementary access to valuable monocyclic terpenoids, an SHC-wild-type library including 18 novel homologs was set up. A previously not described SHC (AciSHC) was found to synthesize small amounts of monocyclic (R)-γ-dihydroionone from (E/Z)-geranylacetone. Using enzyme and process optimization, the conversion to the desired product was increased to 79 %. Notably, analyzed AciSHC variants could finely differentiate between the geometric geranylacetone isomers: While the (Z)-isomer yielded the desired monocyclic (R)-γ-dihydroionone (>99 % ee), the (E)-isomer was converted to the (S,S)-bicyclic ether (>95 % ee). Applying the knowledge gained from the observed stereodivergent and enantioselective transformations to an additional SHC-substrate pair, access to the complementary (S)-γ-dihydroionone (>99.9 % ee) could be obtained., BT/Biocatalysis