Search for: All records

Various machine learning-assisted directed evolution (MLDE) strategies have been shown to identify high-fitness protein variants more efficiently than typical wet-lab directed evolution approaches. However, limited understanding of the factors influencing MLDE performance across diverse proteins has hindered optimal strategy selection for wet-lab campaigns. To address this, we systematically analyzed multiple MLDE strategies, including active learning and focused training using six distinct zeroshot predictors, across 16 diverse protein fitness landscapes. By quantifying landscape navigability with six attributes, we found that MLDE offers a greater advantage on landscapes which are more challenging for directed evolution, especially when focused training is combined with active learning. Despite varying levels of advantage across landscapes, focused training with zero-shot predictors leveraging distinct evolutionary, structural, and stability knowledge sources consistently outperforms random sampling for both binding interactions and enzyme activities. Our findings provide practical guidelines for selecting MLDE strategies for protein engineering. 

Biocatalysis has the potential to address the need for more sustainable organic synthesis routes. Protein engineering can tune enzymes to perform in cascade reactions and for efficient synthesis of enantiomerically enriched compounds, using both natural and new-to-nature reaction pathways. This review highlights recent achievements in biocatalysis, especially the development of novel enzymatic syntheses to access versatile small molecule intermediates and complex biomolecules. Biocatalytic strategies for the degradation of persistent pollutants and approaches for biomass valorization are also discussed. The transition of chemical synthesis to a greener future will be accelerated by implementing enzymes and engineering them for high performance and new activities. 

Lactones are cyclic esters with extensive applications in materials science, medicinal chemistry, and the food and perfume industries. Nature’s strategy for the synthesis of many lactones found in natural products always relies on a single type of retrosynthetic strategy, a C−O bond disconnection. Here, we describe a set of laboratory-engineered enzymes that use a new-tonature C−C bond-forming strategy to assemble diverse lactone structures. These engineered “carbene transferases” catalyze intramolecular carbene insertions into benzylic or allylic C−H bonds, which allow for the synthesis of lactones with different ring sizes and ring scaffolds from simple starting materials. Starting from a serine-ligated cytochrome P450 variant previously engineered for other carbene-transfer activities, directed evolution generated a variant P411-LAS-5247, which exhibits a high activity for constructing a five-membered ε-lactone, lactam, and cyclic ketone products (up to 5600 total turnovers (TTN) and >99% enantiomeric excess (ee)). Further engineering led to variants P411-LAS-5249 and P411-LAS-5264, which deliver six-membered δ-lactones and seven-membered ε-lactones, respectively, overcoming the thermodynamically unfavorable ring strain associated with these products compared to the γ-lactones. This new carbene-transfer activity was further extended to the synthesis of complex lactone scaffolds based on fused, bridged, and spiro rings. The enzymatic platform developed here complements natural biosynthetic strategies for lactone assembly and expands the structural diversity of lactones accessible through C−H functionalization. 

The selective conversion of mixtures of Z/E alkenes into chiral products is a synthetic challenge. Biocatalytic strategies can transform isomeric alkenes into stereopure compounds, but enzymes typically convert only one alkene isomer, thereby limiting the overall yield. Additional strategies have been used to interconvert alkene isomers, often at the cost of increasing energy consumption and chemical waste. Here, we present engineered hemoproteins derived from a bacterial cytochrome P450 that efficiently catalyze α-carbonyl alkylation of isomeric silyl enol ethers, producing stereopure products. Through screening and directed evolution, we generated P450BM3 variant P411-SCA-5188, which catalyzes stereoconvergent carbene transfer in Escherichia coli with high efficiency and stereoselectivity to various Z/E mixtures of silyl enol ethers. In contrast to established stereospecific transformations that leave one isomer unreacted, P411-SCA-5188 converts both isomers to a stereopure product. This biocatalytic approach simplifies the synthesis of chiral α-branched ketones by eliminating the need for stoichiometric chiral auxiliaries, strongly basic alkali-metal enolates, and harsh conditions, delivering products with high efficiency and excellent chemo- and stereoselectivities. 

In nature and synthetic chemistry, stereoselective [2+1] cyclopropanation is the most prevalent strategy for the synthesis of chiral cyclopropanes, a class of key pharmacophores in pharmaceuticals and bioactive natural products. One of the most extensively studied reactions in the organic chemist’s arsenal, stereoselective [2+1] cyclopropanation, largely relies on the use of stereodefined olefins, which can require elaborate laboratory synthesis or tedious separation to ensure high stereoselectivity. Here we report engineered hemoproteins derived from a bacterial cytochrome P450 that catalyze the synthesis of chiral 1,2,3-polysubstituted cyclopropanes, regardless of the stereopurity of the olefin substrates used. Cytochrome P450BM3 variant P411-INC-5185 exclusively converts (Z)-enol acetates to enantio- and diastereoenriched cyclopropanes and in the model reaction delivers a leftover (E)-enol acetate with 98% stereopurity, using whole Escherichia coli cells. P411-INC-5185 was further engineered with a single mutation to enable the biotransformation of (E)-enol acetates to α-branched ketones with high levels of enantioselectivity while simultaneously catalyzing the cyclopropanation of (Z)-enol acetates with excellent activities and selectivities. We conducted docking studies and molecular dynamics simulations to understand how active-site residues distinguish between the substrate isomers and enable the enzyme to perform these distinct transformations with such high selectivities. Computational studies suggest the observed enantio- and diastereoselectivities are achieved through a stepwise pathway. These biotransformations streamline the synthesis of chiral 1,2,3-polysubstituted cyclopropanes from readily available mixtures of (Z/E)-olefins, adding a new dimension to classical cyclopropanation methods. 

Widespread availability of protein sequence-fitness data would revolutionize both our biochemical understanding of proteins and our ability to engineer them. Unfortunately, even though thousands of protein variants are generated and evaluated for fitness during a typical protein engineering campaign, most are never sequenced, leaving a wealth of potential sequence-fitness information untapped. Primarily, this is because sequencing is unnecessary for many protein engineering strategies; the added cost and effort of sequencing is thus unjustified. It also results from the fact that, even though many lower cost sequencing strategies have been developed, they often require at least some sequencing or computational resources, both of which can be barriers to access. Here, we present every variant sequencing (evSeq), a method and collection of tools/standardized components for sequencing a variable region within every variant gene produced during a protein engineering campaign at a cost of cents per variant. evSeq was designed to democratize low-cost sequencing for protein engineers and, indeed, anyone interested in engineering biological systems. Execution of its wet-lab component is simple, requires no sequencing experience to perform, relies only on resources and services typically available to biology labs, and slots neatly into existing protein engineering workflows. Analysis of evSeq data is likewise made simple by its accompanying software (found at github.com/fhalab/evSeq, documentation at fhalab.github.io/evSeq), which can be run on a personal laptop and was designed to be accessible to users with no computational experience. Low-cost and easy to use, evSeq makes collection of extensive protein variant sequence-fitness data practical.