The β-subunit of tryptophan synthase is a latent tyrosine synthase

Aromatic amino acids and their derivatives are diverse primary and secondary metabolites with critical roles in protein synthesis, cell structure and integrity, defense and signaling. All de novo aromatic amino acid production relies on a set of ancient and highly conserved chemistries. Here we introduce a new enzymatic transformation for L-tyrosine synthesis by demonstrating that the β-subunit of tryptophan synthase-which natively couples indole and L-serine to form L-tryptophan-can act as a latent 'tyrosine synthase'. A single substitution of a near-universally conserved catalytic residue unlocks activity toward simple phenol analogs and yields exclusive para carbon-carbon bond formation to furnish L-tyrosines. Structural and mechanistic studies show how a new active-site water molecule orients phenols for a nonnative mechanism of alkylation, with additional directed evolution resulting in a net >30,000-fold rate enhancement. This new biocatalyst can be used to efficiently prepare valuable L-tyrosine analogs at gram scales and provides the missing chemistry for a conceptually different pathway to L-tyrosine.

Opportunities and Challenges for Machine Learning-Assisted Enzyme Engineering

Enzymes can be engineered at the level of their amino acid sequences to optimize key properties such as expression, stability, substrate range, and catalytic efficiency-or even to unlock new catalytic activities not found in nature. Because the search space of possible proteins is vast, enzyme engineering usually involves discovering an enzyme starting point that has some level of the desired activity followed by directed evolution to improve its "fitness" for a desired application. Recently, machine learning (ML) has emerged as a powerful tool to complement this empirical process. ML models can contribute to (1) starting point discovery by functional annotation of known protein sequences or generating novel protein sequences with desired functions and (2) navigating protein fitness landscapes for fitness optimization by learning mappings between protein sequences and their associated fitness values. In this Outlook, we explain how ML complements enzyme engineering and discuss its future potential to unlock improved engineering outcomes.

Iron Heme Enzyme-Catalyzed Cyclopropanations with Diazirines as Carbene Precursors: Computational Explorations of Diazirine Activation and Cyclopropanation Mechanism

The mechanism of cyclopropanations with diazirines as air-stable and user-friendly alternatives to commonly employed diazo compounds within iron heme enzyme-catalyzed carbene transfer reactions has been studied by means of density functional theory (DFT) calculations of model systems, quantum mechanics/molecular mechanics (QM/MM) calculations, and molecular dynamics (MD) simulations of the iron carbene and the cyclopropanation transition state in the enzyme active site. The reaction is initiated by a direct diazirine-diazo isomerization occurring in the active site of the enzyme. In contrast, an isomerization mechanism proceeding via the formation of a free carbene intermediate in lieu of a direct, one-step isomerization process was observed for model systems. Subsequent reaction with benzyl acrylate takes place through stepwise C-C bond formation via a diradical intermediate, delivering the cyclopropane product. The origin of the observed diastereo- and enantioselectivity in the enzyme was investigated through MD simulations, which indicate a preferred formation of the cis-cyclopropane by steric control.

Directed evolution of enzymatic silicon-carbon bond cleavage in siloxanes

Volatile methylsiloxanes (VMS) are man-made, nonbiodegradable chemicals produced at a megaton-per-year scale, which leads to concern over their potential for environmental persistence, long-range transport, and bioaccumulation. We used directed evolution to engineer a variant of bacterial cytochrome P450BM3 to break silicon-carbon bonds in linear and cyclic VMS. To accomplish silicon-carbon bond cleavage, the enzyme catalyzes two tandem oxidations of a siloxane methyl group, which is followed by putative [1,2]-Brook rearrangement and hydrolysis. Discovery of this so-called siloxane oxidase opens possibilities for the eventual biodegradation of VMS.

Enzymatic Assembly of Diverse Lactone Structures: An Intramolecular C-H Functionalization Strategy

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-to-nature 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.

Biocatalytic Construction of Chiral Pyrrolidines and Indolines via Intramolecular C(sp3)-H Amination

Nature harnesses exquisite enzymatic cascades to construct N-heterocycles and further uses these building blocks to assemble the molecules of life. Here we report an enzymatic platform to construct important chiral N-heterocyclic products, pyrrolidines and indolines, via abiological intramolecular C(sp3)-H amination of organic azides. Directed evolution of cytochrome P411 (a P450 enzyme with serine as the heme-ligating residue) yielded variant P411-PYS-5149, capable of catalyzing the insertion of alkyl nitrene into C(sp3)-H bonds to build pyrrolidine derivatives with good enantioselectivity and catalytic efficiency. Further evolution of activity on aryl azide substrates yielded variant P411-INS-5151 that catalyzes intramolecular C(sp3)-H amination to afford chiral indolines. In addition, we show that these enzymatic aminations can be coupled with a P411-based carbene transferase or a tryptophan synthase to generate an α-amino lactone or a noncanonical amino acid, respectively, underscoring the power of new-to-nature biocatalysis in complexity-building chemical synthesis.

Directed evolution for Si-C bond cleavage of volatile siloxanes in glass bioreactors

Volatile methylsiloxanes (VMS) are a class of non-biodegradable anthropogenic compounds with propensity for long-range transport and potential for bioaccumulation in the environment. As a proof-of-principle for biological degradation of these compounds, we engineered P450 enzymes to oxidatively cleave Si-C bonds in linear and cyclic VMS. Enzymatic reactions with VMS are challenging to screen with conventional tools, however, due to their volatility, poor aqueous solubility, and tendency to extract polypropylene from standard 96-well deep-well plates. To address these challenges, we developed a new biocatalytic reactor consisting of individual 2-mL glass shells assembled in conventional 96-well plate format. In this chapter, we provide a detailed account of the assembly and use of the 96-well glass shell reactors for screening biocatalytic reactions. Additionally, we discuss the application of GC/MS analysis techniques for VMS oxidase reactions and modified procedures for validating improved variants. This protocol can be adopted broadly for biocatalytic reactions with substrates that are volatile or not suitable for polypropylene plates.

Directed evolution of P411 enzymes for amination of inert C-H bonds

Functionalizing inert C-H bonds selectively is a formidable task due to their strong bond energy and the difficulty of distinguishing chemically similar C-H bonds. While enzymatic oxygenation of C-H bonds is ubiquitous and well established, there is currently no known natural enzymatic process for direct nitrogen insertion. Instead, nature typically relies on pre-oxidized compounds for nitrogen incorporation. Direct biocatalytic C-H amination methods developed in the last few years are only selective for activated C-H bonds that contain specific groups such as benzylic, allylic, or propargylic groups. However, we recently used directed evolution to generate cytochrome P411 enzymes (engineered P450 enzymes with axial ligand mutation from cysteine to serine) that directly aminate inert C-H bonds with high site-, diastereo-, and enantioselectivity. Using these enzymes, we demonstrated the regiodivergent desymmetrization of methylcyclohexane, among other reactions. This chapter provides a comprehensive account of the experimental protocols used to evolve P411s for aminating unactivated C-H bonds. These methods are illustrative and can be adapted for other directed enzyme evolution campaigns.

Enzymatic Nitrogen Incorporation Using Hydroxylamine

Hydroxylamine-derived reagents have enabled versatile nitrene transfer reactions for introducing nitrogen-containing functionalities in small-molecule catalysis, as well as biocatalysis. These reagents, however, result in a poor atom economy and stoichiometric organic waste. Activating hydroxylamine (NH2OH) for nitrene transfer offers a low-cost and sustainable route to amine synthesis, since water is the sole byproduct. Despite its presence in nature, hydroxylamine is not known to be used for enzymatic nitrogen incorporation in biosynthesis. Here, we report an engineered heme enzyme that can utilize hydroxylammonium chloride, an inexpensive commodity chemical, for nitrene transfer. Directed evolution of Pyrobaculum arsenaticum protoglobin generated efficient enzymes for benzylic C-H primary amination and styrene aminohydroxylation. Mechanistic studies supported a stepwise radical pathway involving rate-limiting hydrogen atom transfer. This unprecedented activity is a useful addition to the "nitrene transferase" repertoire and hints at possible future discovery of natural enzymes that use hydroxylamine for amination chemistry.

Reversing the Enantioselectivity of Enzymatic Carbene N-H Insertion Through Mechanism-Guided Protein Engineering

We report a computationally driven approach to access enantiodivergent enzymatic carbene N-H insertions catalyzed by P411 enzymes. Computational modeling was employed to rationally guide engineering efforts to control the accessible conformations of a key lactone-carbene (LAC) intermediate in the enzyme active site by installing a new H-bond anchoring point. This H-bonding interaction controls the relative orientation of the reactive carbene intermediate, orienting it for an enantioselective N-nucleophilic attack by the amine substrate. By combining MD simulations and site-saturation mutagenesis and screening targeted to only two key residues, we were able to reverse the stereoselectivity of previously engineered S-selective P411 enzymes. The resulting variant, L5_FL-B3, accepts a broad scope of amine substrates for N-H insertion with excellent yields (up to >99 %), high efficiency (up to 12 300 TTN), and good enantiocontrol (up to 7 : 93 er).

DeCOIL: Optimization of Degenerate Codon Libraries for Machine Learning-Assisted Protein Engineering

With advances in machine learning (ML)-assisted protein engineering, models based on data, biophysics, and natural evolution are being used to propose informed libraries of protein variants to explore. Synthesizing these libraries for experimental screens is a major bottleneck, as the cost of obtaining large numbers of exact gene sequences is often prohibitive. Degenerate codon (DC) libraries are a cost-effective alternative for generating combinatorial mutagenesis libraries where mutations are targeted to a handful of amino acid sites. However, existing computational methods to optimize DC libraries to include desired protein variants are not well suited to design libraries for ML-assisted protein engineering. To address these drawbacks, we present DEgenerate Codon Optimization for Informed Libraries (DeCOIL), a generalized method that directly optimizes DC libraries to be useful for protein engineering: to sample protein variants that are likely to have both high fitness and high diversity in the sequence search space. Using computational simulations and wet-lab experiments, we demonstrate that DeCOIL is effective across two specific case studies, with the potential to be applied to many other use cases. DeCOIL offers several advantages over existing methods, as it is direct, easy to use, generalizable, and scalable. With accompanying software (https://github.com/jsunn-y/DeCOIL), DeCOIL can be readily implemented to generate desired informed libraries.

Enantio- and Diastereoenriched Enzymatic Synthesis of 1,2,3-Polysubstituted Cyclopropanes from (Z/E)-Trisubstituted Enol Acetates

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.

Enantio- and Diastereoenriched Enzymatic Synthesis of 1,2,3-Polysubstituted Cyclopropanes from (Z/E)-Trisubstituted Enol Acetates

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 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 P450 _BM3 variant IC-G3 exclusively converts ( Z )-enol acetates to enantio- and diastereoenriched cyclopropanes and in our model reaction delivers a leftover ( E )-enol acetate with 98% stereopurity, using whole Escherichia coli cells. IC-G3 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.

MicroED Structure of a Protoglobin Reactive Carbene Intermediate

Microcrystal electron diffraction (MicroED) is an emerging technique that has shown great potential for describing new chemical and biological molecular structures. Several important structures of small molecules, natural products, and peptides have been determined using ab initio methods. However, only a couple of novel protein structures have thus far been derived by MicroED. Taking advantage of recent technological advances, including higher acceleration voltage and using a low-noise detector in counting mode, we have determined the first structure of an Aeropyrum pernix protoglobin (ApePgb) variant by MicroED using an AlphaFold2 model for phasing. The structure revealed that mutations introduced during directed evolution enhance carbene transfer activity by reorienting an α helix of ApePgb into a dynamic loop, making the catalytic active site more readily accessible. After exposing the tiny crystals to the substrate, we also trapped the reactive iron-carbenoid intermediate involved in this engineered ApePgb's new-to-nature activity, a challenging carbene transfer from a diazirine via a putative metallo-carbene. The bound structure discloses how an enlarged active site pocket stabilizes the carbene bound to the heme iron and, presumably, the transition state for the formation of this key intermediate. This work demonstrates that improved MicroED technology and the advancement in protein structure prediction now enable investigation of structures that was previously beyond reach.

Chemodivergent C(sp3)-H and C(sp2)-H Cyanomethylation Using Engineered Carbene Transferases

The ubiquity of C-H bonds presents an attractive opportunity to elaborate and build complexity in organic molecules. Methods for selective functionalization, however, often must differentiate among multiple chemically similar and, in some cases indistinguishable, C-H bonds. An advantage of enzymes is that they can be finely tuned using directed evolution to achieve control over divergent C-H functionalization pathways. Here, we demonstrate engineered enzymes that effect a new-to-nature C-H alkylation with unparalleled selectivity: two complementary carbene C-H transferases derived from a cytochrome P450 from Bacillus megaterium deliver an α-cyanocarbene into the α-amino C(sp3)-H bonds or the ortho-arene C(sp2)-H bonds of N-substituted arenes. These two transformations proceed via different mechanisms, yet only minimal changes to the protein scaffold (nine mutations, less than 2% of the sequence) were needed to adjust the enzyme's control over the site-selectivity of cyanomethylation. The X-ray crystal structure of the selective C(sp3)-H alkylase, P411-PFA, reveals an unprecedented helical disruption which alters the shape and electrostatics in the enzyme active site. Overall, this work demonstrates the advantages of enzymes as C-H functionalization catalysts for divergent molecular derivatization.

Protoglobin-Catalyzed Formation of cis-Trifluoromethyl-Substituted Cyclopropanes by Carbene Transfer

Trifluoromethyl-substituted cyclopropanes (CF3 -CPAs) constitute an important class of compounds for drug discovery. While several methods have been developed for synthesis of trans-CF3 -CPAs, stereoselective production of corresponding cis-diastereomers remains a formidable challenge. We report a biocatalyst for diastereo- and enantio-selective synthesis of cis-CF3 -CPAs with activity on a variety of alkenes. We found that an engineered protoglobin from Aeropyrnum pernix (ApePgb) can catalyze this unusual reaction at preparative scale with low-to-excellent yield (6-55 %) and enantioselectivity (17-99 % ee), depending on the substrate. Computational studies revealed that the steric environment in the active site of the protoglobin forced iron-carbenoid and substrates to adopt a pro-cis near-attack conformation. This work demonstrates the capability of enzyme catalysts to tackle challenging chemistry problems and provides a powerful means to expand the structural diversity of CF3 -CPAs for drug discovery.

New Additions to the Arsenal of Biocatalysts for Noncanonical Amino Acid Synthesis

Noncanonical amino acids (ncAAs) merge the conformational behavior and native interactions of proteinogenic amino acids with nonnative chemical motifs and have proven invaluable in developing modern therapeutics. This blending of native and nonnative characteristics has resulted in essential drugs like nirmatrelvir, which comprises three ncAAs and is used to treat COVID-19. Enzymes are appearing prominently in recent syntheses of ncAAs, where they demonstrate impressive control over the stereocenters and functional groups found therein. Here we review recent efforts to expand the biocatalyst arsenal for synthesizing ncAAs with natural enzymes. We also discuss how new-to-nature enzymes can contribute to this effort by catalyzing reactions inspired by the vast repertoire of chemical catalysis and acting on substrates that would otherwise not be used in synthesizing ncAAs. Abiotic enzyme-catalyzed reactions exploit the selectivity afforded by a macromolecular catalyst to access molecules not available to natural enzymes and perhaps not even chemical catalysis.

Enzymatic Nitrogen Insertion into Unactivated C-H Bonds

Selective functionalization of aliphatic C-H bonds, ubiquitous in molecular structures, could allow ready access to diverse chemical products. While enzymatic oxygenation of C-H bonds is well established, the analogous enzymatic nitrogen functionalization is still unknown; nature is reliant on preoxidized compounds for nitrogen incorporation. Likewise, synthetic methods for selective nitrogen derivatization of unbiased C-H bonds remain elusive. In this work, new-to-nature heme-containing nitrene transferases were used as starting points for the directed evolution of enzymes to selectively aminate and amidate unactivated C(sp³)-H sites. The desymmetrization of methyl- and ethylcyclohexane with divergent site selectivity is offered as demonstration. The evolved enzymes in these lineages are highly promiscuous and show activity toward a wide array of substrates, providing a foundation for further evolution of nitrene transferase function. Computational studies and kinetic isotope effects (KIEs) are consistent with a stepwise radical pathway involving an irreversible, enantiodetermining hydrogen atom transfer (HAT), followed by a lower-barrier diastereoselectivity-determining radical rebound step. In-enzyme molecular dynamics (MD) simulations reveal a predominantly hydrophobic pocket with favorable dispersion interactions with the substrate. By offering a direct path from saturated precursors, these enzymes present a new biochemical logic for accessing nitrogen-containing compounds.

Biocatalytic Carbene Transfer Using Diazirines

Biocatalytic carbene transfer from diazo compounds is a versatile strategy in asymmetric synthesis. However, the limited pool of stable diazo compounds constrains the variety of accessible products. To overcome this restriction, we have engineered variants of Aeropyrum pernix protoglobin (ApePgb) that use diazirines as carbene precursors. While the enhanced stability of diazirines relative to their diazo isomers enables access to a diverse array of carbenes, they have previously resisted catalytic activation. Our engineered ApePgb variants represent the first example of catalysts for selective carbene transfer from these species at room temperature. The structure of an ApePgb variant, determined by microcrystal electron diffraction (MicroED), reveals that evolution has enhanced access to the heme active site to facilitate this new-to-nature catalysis. Using readily prepared aryl diazirines as model substrates, we demonstrate the application of these highly stable carbene precursors in biocatalytic cyclopropanation, N-H insertion, and Si-H insertion reactions.

Biocatalytic One-Carbon Ring Expansion of Aziridines to Azetidines via a Highly Enantioselective [1,2]-Stevens Rearrangement

We report enantioselective one-carbon ring expansion of aziridines to make azetidines as a new-to-nature activity of engineered "carbene transferase" enzymes. A laboratory-evolved variant of cytochrome P450_BM3, P411-AzetS, not only exerts unparalleled stereocontrol (99:1 er) over a [1,2]-Stevens rearrangement but also overrides the inherent reactivity of aziridinium ylides, cheletropic extrusion of olefins, to perform a [1,2]-Stevens rearrangement. By controlling the fate of the highly reactive aziridinium ylide intermediates, these evolvable biocatalysts promote a transformation which cannot currently be performed using other catalyst classes.

evSeq: Cost-Effective Amplicon Sequencing of Every Variant in a Protein Library

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 are 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 access to and experience with 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 the collection of extensive protein variant sequence-fitness data practical.

Combining chemistry and protein engineering for new-to-nature biocatalysis

Biocatalysis, the application of enzymes to solve synthetic problems of human import, has blossomed into a powerful technology for chemical innovation. In the past decade, a threefold partnership, where nature provides blueprints for enzymatic catalysis, chemists introduce innovative activity modes with abiological substrates, and protein engineers develop new tools and algorithms to tune and improve enzymatic function, has unveiled the frontier of new-to-nature enzyme catalysis. In this perspective, we highlight examples of interdisciplinary studies which have helped to expand the scope of biocatalysis, including concepts of enzymatic versatility explored through the lens of biomimicry, to achieve both activities and selectivities that are not currently possible with chemocatalysis. We indicate how modern tools, such as directed evolution, computational protein design and machine learning-based protein engineering methods, have already impacted and will continue to influence enzyme engineering for new abiological transformations. A sustained collaborative effort across disciplines is anticipated to spur further advances in biocatalysis in the coming years.

An Enzymatic Platform for Primary Amination of 1-Aryl-2-alkyl Alkynes

Propargyl amines are versatile synthetic intermediates with numerous applications in the pharmaceutical industry. An attractive strategy for efficient preparation of these compounds is nitrene propargylic C(sp³)-H insertion. However, achieving this reaction with good chemo-, regio-, and enantioselective control has proven to be challenging. Here, we report an enzymatic platform for the enantioselective propargylic amination of alkynes using a hydroxylamine derivative as the nitrene precursor. Cytochrome P450 variant PA-G8 catalyzing this transformation was identified after eight rounds of directed evolution. A variety of 1-aryl-2-alkyl alkynes are accepted by PA-G8, including those bearing heteroaromatic rings. This biocatalytic process is efficient and selective (up to 2610 total turnover number (TTN) and 96% ee) and can be performed on preparative scale.

Informed training set design enables efficient machine learning-assisted directed protein evolution

Directed evolution of proteins often involves a greedy optimization in which the mutation in the highest-fitness variant identified in each round of single-site mutagenesis is fixed. The efficiency of such a single-step greedy walk depends on the order in which beneficial mutations are identified-the process is path dependent. Here, we investigate and optimize a path-independent machine learning-assisted directed evolution (MLDE) protocol that allows in silico screening of full combinatorial libraries. In particular, we evaluate the importance of different protein encoding strategies, training procedures, models, and training set design strategies on MLDE outcome, finding the most important consideration to be the implementation of strategies that reduce inclusion of minimally informative "holes" (protein variants with zero or extremely low fitness) in training data. When applied to an epistatic, hole-filled, four-site combinatorial fitness landscape, our optimized protocol achieved the global fitness maximum up to 81-fold more frequently than single-step greedy optimization. A record of this paper's transparent peer review process is included in the supplemental information.

Biocatalytic, Intermolecular C-H Bond Functionalization for the Synthesis of Enantioenriched Amides

Directed evolution of heme proteins has opened access to new-to-nature enzymatic activity that can be harnessed to tackle synthetic challenges. Among these, reactions resulting from active site iron-nitrenoid intermediates present a powerful strategy to forge C-N bonds with high site- and stereoselectivity. Here we report a biocatalytic, intermolecular benzylic C-H amidation reaction operating at mild and scalable conditions. With hydroxamate esters as nitrene precursors, feedstock aromatic compounds can be converted to chiral amides with excellent enantioselectivity (up to >99 % ee) and high yields (up to 87 %). Kinetic and computational analysis of the enzymatic reaction reveals rate-determining nitrenoid formation followed by stepwise hydrogen atom transfer-mediated C-H functionalization.

Asymmetric Alkylation of Ketones Catalyzed by Engineered TrpB

The β-subunit of tryptophan synthase (TrpB) catalyzes a PLP-mediated β-substitution reaction between indole and serine to form L-Trp. A succession of TrpB protein engineering campaigns to expand the enzyme's nucleophile substrate range has enabled the biocatalytic production of diverse non-canonical amino acids (ncAAs). Here, we show that ketone-derived enolates can serve as nucleophiles in the TrpB reaction to achieve the asymmetric alkylation of ketones, an outstanding challenge in synthetic chemistry. We engineered TrpB by directed evolution to catalyze the asymmetric alkylation of propiophenone and 2-fluoroacetophenone with a high degree of selectivity. In reactions with propiophenone, preference for the opposite product diastereomer emerges over the course of evolution, demonstrating that full control over the stereochemistry at the new chiral center can be achieved. The addition of this new reaction to the TrpB platform is a crucial first step toward the development of efficient methods to synthesize non-canonical prolines and other chirally dense nitrogen heterocycles.

Biocatalytic Transformations of Silicon-the Other Group 14 Element

Significant inroads have been made using biocatalysts to perform new-to-nature reactions with high selectivity and efficiency. Meanwhile, advances in organosilicon chemistry have led to rich sets of reactions holding great synthetic value. Merging biocatalysis and silicon chemistry could yield new methods for the preparation of valuable organosilicon molecules as well as the degradation and valorization of undesired ones. Despite silicon's importance in the biosphere for its role in plant and diatom construction, it is not known to be incorporated into any primary or secondary metabolites. Enzymes have been found that act on silicon-containing molecules, but only a few are known to act directly on silicon centers. Protein engineering and evolution has and could continue to enable enzymes to catalyze useful organosilicon transformations, complementing and expanding upon current synthetic methods. The role of silicon in biology and the enzymes that act on silicon-containing molecules are reviewed to set the stage for a discussion of where biocatalysis and organosilicon chemistry may intersect.

New-to-nature chemistry from old protein machinery: carbene and nitrene transferases

Hemoprotein-catalyzed carbene and nitrene transformations have emerged as powerful tools for constructing complex molecules; they also nicely illustrate how new protein catalysts can emerge, evolve and diversify. These laboratory-invented enzymes exploit the ability of proteins to tame highly reactive carbene and nitrene species and direct their fates with high selectivity. New-to-nature carbene and nitrene transferases catalyze many useful reactions, including some that have no precedent using chemical methods. Here we cover recent advances in this field, including alkyne cyclopropenation, arene cyclopropanation, carbene CH insertion, intramolecular nitrene CH insertion, alkene aminohydroxylation, and primary amination. For such transformations, biocatalysts have exceeded the performance of reported small-molecule catalysts in terms of selectivity and catalyst turnovers. Finally, we offer our thoughts on using these new enzymatic reactions in chemical synthesis, integrating them into biological pathways and chemo-enzymatic cascades, and on their current limitations.

Protein sequence design with deep generative models

Protein engineering seeks to identify protein sequences with optimized properties. When guided by machine learning, protein sequence generation methods can draw on prior knowledge and experimental efforts to improve this process. In this review, we highlight recent applications of machine learning to generate protein sequences, focusing on the emerging field of deep generative methods.

Origin and Control of Chemoselectivity in Cytochrome c Catalyzed Carbene Transfer into Si-H and N-H bonds

A cytochrome c heme protein was recently engineered to catalyze the formation of carbon-silicon bonds via carbene insertion into Si-H bonds, a reaction that was not previously known to be catalyzed by a protein. High chemoselectivity toward C-Si bond formation over competing C-N bond formation was achieved, although this trait was not screened for during directed evolution. Using computational and experimental tools, we now establish that activity and chemoselectivity are modulated by conformational dynamics of a protein loop that covers the substrate access to the iron-carbene active species. Mutagenesis of residues computationally predicted to control the loop conformation altered the protein's chemoselectivity from preferred silylation to preferred amination of a substrate containing both N-H and Si-H functionalities. We demonstrate that information on protein structure and conformational dynamics, combined with knowledge of mechanism, leads to understanding of how non-natural and selective chemical transformations can be introduced into the biological world.

Navigating the Unnatural Reaction Space: Directed Evolution of Heme Proteins for Selective Carbene and Nitrene Transfer

Despite the astonishing diversity of naturally occurring biocatalytic processes, enzymes do not catalyze many of the transformations favored by synthetic chemists. Either nature does not care about the specific products, or if she does, she has adopted a different synthetic strategy. In many cases, the appropriate reagents used by synthetic chemists are not readily accessible to biological systems. Here, we discuss our efforts to expand the catalytic repertoire of enzymes to encompass powerful reactions previously known only in small-molecule catalysis: formation and transfer of reactive carbene and nitrene intermediates leading to a broad range of products, including products with bonds not known in biology. In light of the structural similarity of iron carbene (Fe=C(R¹)(R²)) and iron nitrene (Fe=NR) to the iron oxo (Fe=O) intermediate involved in cytochrome P450-catalyzed oxidation, we have used synthetic carbene and nitrene precursors that biological systems have not encountered and repurposed P450s to catalyze reactions that are not known in the natural world. The resulting protein catalysts are fully genetically encoded and function in intact microbial cells or cell-free lysates, where their performance can be improved and optimized by directed evolution. By leveraging the catalytic promiscuity of P450 enzymes, we evolved a range of carbene and nitrene transferases exhibiting excellent activity toward these new-to-nature reactions. Since our initial report in 2012, a number of other heme proteins including myoglobins, protoglobins, and cytochromes c have also been found and engineered to promote unnatural carbene and nitrene transfer. Due to the altered active-site environments, these heme proteins often displayed complementary activities and selectivities to P450s.

Using wild-type and engineered heme proteins, we and others have described a range of selective carbene transfer reactions, including cyclopropanation, cyclopropenation, Si–H insertion, B–H insertion, and C–H insertion. Similarly, a variety of asymmetric nitrene transfer processes including aziridination, sulfide imidation, C–H amidation, and, most recently, C–H amination have been demonstrated. The scopes of these biocatalytic carbene and nitrene transfer reactions are often complementary to the state-of-the-art processes based on small-molecule transition-metal catalysts, making engineered biocatalysts a valuable addition to the synthetic chemist’s toolbox. Moreover, enabled by the exquisite regio- and stereocontrol imposed by the enzyme catalyst, this biocatalytic platform provides an exciting opportunity to address challenging problems in modern synthetic chemistry and selective catalysis, including ones that have eluded synthetic chemists for decades.

Advances in machine learning for directed evolution

Machine learning (ML) can expedite directed evolution by allowing researchers to move expensive experimental screens in silico. Gathering sequence-function data for training ML models, however, can still be costly. In contrast, raw protein sequence data is widely available. Recent advances in ML approaches use protein sequences to augment limited sequence-function data for directed evolution. We highlight contributions in a growing effort to use sequences to reduce or eliminate the amount of sequence-function data needed for effective in silico screening. We also highlight approaches that use ML models trained on sequences to generate new functional sequence diversity, focusing on strategies that use these generative models to efficiently explore vast regions of protein space.

High Throughput Screening with SAMDI Mass Spectrometry for Directed Evolution

Advances in directed evolution have led to an exploration of new and important chemical transformations; however, many of these efforts still rely on the use of low-throughput chromatography-based screening methods. We present a high-throughput strategy for screening libraries of enzyme variants for improved activity. Unpurified reaction products are immobilized to a self-assembled monolayer and analyzed by mass spectrometry, allowing for direct evaluation of thousands of variants in under an hour. The method was demonstrated with libraries of randomly mutated cytochrome P411 variants to identify improved catalysts for C-H alkylation. The technique may be tailored to evolve enzymatic activity for a variety of transformations where higher throughput is needed.

Scalable continuous evolution for the generation of diverse enzyme variants encompassing promiscuous activities

Enzyme orthologs sharing identical primary functions can have different promiscuous activities. While it is possible to mine this natural diversity to obtain useful biocatalysts, generating comparably rich ortholog diversity is difficult, as it is the product of deep evolutionary processes occurring in a multitude of separate species and populations. Here, we take a first step in recapitulating the depth and scale of natural ortholog evolution on laboratory timescales. Using a continuous directed evolution platform called OrthoRep, we rapidly evolve the Thermotoga maritima tryptophan synthase β-subunit (TmTrpB) through multi-mutation pathways in many independent replicates, selecting only on TmTrpB's primary activity of synthesizing L-tryptophan from indole and L-serine. We find that the resulting sequence-diverse TmTrpB variants span a range of substrate profiles useful in industrial biocatalysis and suggest that the depth and scale of evolution that OrthoRep affords will be generally valuable in enzyme engineering and the evolution of biomolecular functions.

Nature's Machinery, Repurposed: Expanding the Repertoire of Iron-Dependent Oxygenases

Iron is an especially important redox-active cofactor in biology because of its ability to mediate reactions with atmospheric O2. Iron-dependent oxygenases exploit this earth-abundant transition metal for the insertion of oxygen atoms into organic compounds. Throughout the astounding diversity of transformations catalyzed by these enzymes, the protein framework directs reactive intermediates toward the precise formation of products, which, in many cases, necessitates the cleavage of strong C-H bonds. In recent years, members of several iron-dependent oxygenase families have been engineered for new-to-nature transformations that offer advantages over conventional synthetic methods. In this Perspective, we first explore what is known about the reactivity of heme-dependent cytochrome P450 oxygenases and nonheme iron-dependent oxygenases bearing the 2-His-1-carboxylate facial triad by reviewing mechanistic studies with an emphasis on how the protein scaffold maximizes the catalytic potential of the iron-heme and iron cofactors. We then review how these cofactors have been repurposed for abiological transformations by engineering the protein frameworks of these enzymes. Finally, we discuss contemporary challenges associated with engineering these platforms and comment on their roles in biocatalysis moving forward.

Selective Enzymatic Oxidation of Silanes to Silanols

Compared to the biological world's rich chemistry for functionalizing carbon, enzymatic transformations of the heavier homologue silicon are rare. We report that a wild-type cytochrome P450 monooxygenase (P450BM3 from Bacillus megaterium, CYP102A1) has promiscuous activity for oxidation of hydrosilanes to give silanols. Directed evolution was applied to enhance this non-native activity and create a highly efficient catalyst for selective silane oxidation under mild conditions with oxygen as the terminal oxidant. The evolved enzyme leaves C-H bonds present in the silane substrates untouched, and this biotransformation does not lead to disiloxane formation, a common problem in silanol syntheses. Computational studies reveal that catalysis proceeds through hydrogen atom abstraction followed by radical rebound, as observed in the native C-H hydroxylation mechanism of the P450 enzyme. This enzymatic silane oxidation extends nature's impressive catalytic repertoire.

Signal Peptides Generated by Attention-Based Neural Networks

Short (15-30 residue) chains of amino acids at the amino termini of expressed proteins known as signal peptides (SPs) specify secretion in living cells. We trained an attention-based neural network, the Transformer model, on data from all available organisms in Swiss-Prot to generate SP sequences. Experimental testing demonstrates that the model-generated SPs are functional: when appended to enzymes expressed in an industrial Bacillus subtilis strain, the SPs lead to secreted activity that is competitive with industrially used SPs. Additionally, the model-generated SPs are diverse in sequence, sharing as little as 58% sequence identity to the closest known native signal peptide and 73% ± 9% on average.

Diversity-Oriented Enzymatic Synthesis of Cyclopropane Building Blocks

While biocatalysis is increasingly incorporated into drug development pipelines, it is less commonly used in the early stages of drug discovery. By engineering a protein to produce a chiral motif with a derivatizable functional handle, biocatalysts can be used to help generate diverse building blocks for drug discovery. Here we show the engineering of two variants of Rhodothermus marinus nitric oxide dioxygenase (RmaNOD) to catalyze the formation of cis- and tran- diastereomers of a pinacolboronate-substituted cyclopropane which can be readily derivatized to generate diverse stereopure cyclopropane building blocks.

Enzymatic Primary Amination of Benzylic and Allylic C(sp3)-H Bonds

Aliphatic primary amines are prevalent in natural products, pharmaceuticals, and functional materials. While a plethora of processes are reported for their synthesis, methods that directly install a free amine group into C(sp³)-H bonds remain unprecedented. Here, we report a set of new-to-nature enzymes that catalyze the direct primary amination of C(sp³)-H bonds with excellent chemo-, regio-, and enantioselectivity, using a readily available hydroxylamine derivative as the nitrogen source. Directed evolution of genetically encoded cytochrome P411 enzymes (P450s whose Cys axial ligand to the heme iron has been replaced with Ser) generated variants that selectively functionalize benzylic and allylic C-H bonds, affording a broad scope of enantioenriched primary amines. This biocatalytic process is efficient and selective (up to 3930 TTN and 96% ee), and can be performed on preparative scale.

Tailoring Tryptophan Synthase TrpB for Selective Quaternary Carbon Bond Formation

We previously engineered the β-subunit of tryptophan synthase (TrpB), which catalyzes the condensation of l-serine and indole to l-tryptophan, to synthesize a range of noncanonical amino acids from l-serine and indole derivatives or other nucleophiles. Here we employ directed evolution to engineer TrpB to accept 3-substituted oxindoles and form C-C bonds leading to new quaternary stereocenters. Initially, the variants that could use 3-substituted oxindoles preferentially formed N-C bonds on N1 of the substrate. Protecting N1 encouraged evolution toward C-alkylation, which persisted when protection was removed. Six generations of directed evolution resulted in TrpB Pfquat with a 400-fold improvement in activity for alkylation of 3-substituted oxindoles and the ability to selectively form a new, all-carbon quaternary stereocenter at the γ-position of the amino acid products. The enzyme can also alkylate and form all-carbon quaternary stereocenters on structurally similar lactones and ketones, where it exhibits excellent regioselectivity for the tertiary carbon. The configurations of the γ-stereocenters of two of the products were determined via microcrystal electron diffraction (MicroED), and we report the MicroED structure of a small molecule obtained using the Falcon III direct electron detector. Highly thermostable and expressed at >500 mg/L E. coli culture, TrpB Pfquat offers an efficient, sustainable, and selective platform for the construction of diverse noncanonical amino acids bearing all-carbon quaternary stereocenters.