Biocatalytic control of site-selectivity and chain length-selectivity in radical amino acid halogenases

Discovery and substrate specificity engineering of nucleotide halogenases

C2'-halogenation has been recognized as an essential modification to enhance the drug-like properties of nucleotide analogs. The direct C2'-halogenation of the nucleotide 2'-deoxyadenosine-5'-monophosphate (dAMP) has recently been achieved using the Fe(II)/α-ketoglutarate-dependent nucleotide halogenase AdaV. However, the limited substrate scope of this enzyme hampers its broader applications. In this study, we report two halogenases capable of halogenating 2'-deoxyguanosine monophosphate (dGMP), thereby expanding the family of nucleotide halogenases. Computational studies reveal that nucleotide specificity is regulated by the binding pose of the phosphate group. Based on these findings, we successfully engineered the substrate specificity of these halogenases by mutating second-sphere residues. This work expands the toolbox of nucleotide halogenases and provides insights into the regulation mechanism of nucleotide specificity.

Machine learning guided rational design of a non-heme iron-based lysine dioxygenase improves its total turnover number

Highly selective C-H functionalization remains an ongoing challenge in organic synthetic methodologies. Biocatalysts are robust tools for achieving these difficult chemical transformations. Biocatalyst engineering has often required directed evolution or structure-based rational design campaigns to improve their activities. In recent years, machine learning has been integrated into these workflows to improve the discovery of beneficial enzyme variants. In this work, we combine a structure-based machine-learning algorithm with classical molecular dynamics simulations to down select mutations for rational design of a non-heme iron-dependent lysine dioxygenase, LDO. This approach consistently resulted in functional LDO mutants and circumvents the need for extensive study of mutational activity before-hand. Our rationally designed single mutants purified with up to 2-fold higher yields than WT and displayed higher total turnover numbers (TTN). Combining five such single mutations into a pentamutant variant, LPNYI LDO, leads to a 40% improvement in the TTN (218±3) as compared to WT LDO (TTN = 160±2). Overall, this work offers a low-barrier approach for those seeking to synergize machine learning algorithms with pre-existing protein engineering strategies.

Controllable multi-halogenation of a non-native substrate by SyrB2 iron halogenase

Geminal, multi-halogenated functional groups are widespread in natural products and pharmaceuticals, yet no synthetic methodologies exist that enable selective multi-halogenation of unactivated C-H bonds. Biocatalysts are powerful tools for late-stage C-H functionalization, as they operate with high degrees of regio-, chemo-, and stereoselectivity. 2-oxoglutarate (2OG)-dependent non-heme iron halogenases chlorinate and brominate aliphatic C-H bonds offering a solution for achieving these challenging transformations. Here, we describe the ability of a non-heme iron halogenase, SyrB2, to controllably halogenate non-native substrate alpha-aminobutyric acid (Aba) to yield mono-chlorinated, di-chlorinated, and tri-chlorinated products. These chemoselective outcomes are achieved by controlling the loading of 2OG cofactor and SyrB2 biocatalyst. By using a ferredoxin-based biological reductant for electron transfer to the catalytic center of SyrB2, we demonstrate order-of-magnitude enhancement in the yield of tri-chlorinated product that were previously inaccessible using any single halogenase enzyme. We also apply these strategies to broaden SyrB2's reactivity scope to include multi-bromination and demonstrate chemoenzymatic conversion of the ethyl side chain in Aba to an ethylyne functional group. We show how steric hindrance induced by the successive addition of halogen atoms on Aba's C4 carbon dictates the degree of multi-halogenation by hampering C3-C4 bond rotation within SyrB2's catalytic pocket. Overall, our work showcases the synthetic potential of iron halogenases to facilitate multi-C-H functionalization chemistry.

Engineering Hydroxylase Activity, Selectivity, and Stability for a Scalable Concise Synthesis of a Key Intermediate to Belzutifan

Biocatalytic oxidations are an emerging technology for selective C-H bond activation. While promising for a range of selective oxidations, practical use of enzymes catalyzing aerobic hydroxylation is presently limited by their substrate scope and stability under industrially relevant conditions. Here, we report the engineering and practical application of a non-heme iron and α-ketoglutarate-dependent dioxygenase for the direct stereo- and regio-selective hydroxylation of a non-native fluoroindanone en route to the oncology treatment belzutifan, replacing a five-step chemical synthesis with a direct enantioselective hydroxylation. Mechanistic studies indicated that formation of the desired product was limited by enzyme stability and product overoxidation, with these properties subsequently improved by directed evolution, yielding a biocatalyst capable of >15,000 total turnovers. Highlighting the industrial utility of this biocatalyst, the high-yielding, green, and efficient oxidation was demonstrated at kilogram scale for the synthesis of belzutifan.

Synergistic Binding of the Halide and Cationic Prime Substrate of l-Lysine 4-Chlorinase, BesD, in Both Ferrous and Ferryl States

An aliphatic halogenase requires four substrates: 2-oxoglutarate (2OG), halide (Cl- or Br-), the halogenation target ("prime substrate"), and dioxygen. In well-studied cases, the three nongaseous substrates must bind to activate the enzyme's Fe(II) cofactor for efficient capture of O2. Halide, 2OG, and (lastly) O2 all coordinate directly to the cofactor to initiate its conversion to a cis-halo-oxo-iron(IV) (haloferryl) complex, which abstracts hydrogen (H•) from the non-coordinating prime substrate to enable radicaloid carbon-halogen coupling. We dissected the kinetic pathway and thermodynamic linkage in binding of the first three substrates of the l-lysine 4-chlorinase, BesD. After addition of 2OG, subsequent coordination of the halide to the cofactor and binding of cationic l-Lys near the cofactor are associated with strong heterotropic cooperativity. Progression to the haloferryl intermediate upon the addition of O2 does not trap the substrates in the active site and, in fact, markedly diminishes cooperativity between halide and l-Lys. The surprising lability of the BesD•[Fe(IV)=O]•Cl•succinate•l-Lys complex engenders pathways for decay of the haloferryl intermediate that do not result in l-Lys chlorination, especially at low chloride concentrations; one identified pathway involves oxidation of glycerol. The mechanistic data imply (i) that BesD may have evolved from a hydroxylase ancestor either relatively recently or under weak selective pressure for efficient chlorination and (ii) that acquisition of its activity may have involved the emergence of linkage between l-Lys binding and chloride coordination following the loss of the anionic protein-carboxylate iron ligand present in extant hydroxylases.

Regioselective Radical Alkylation of Arenes Using Evolved Photoenzymes

Substituted arenes are ubiquitous in molecules with medicinal functions, making their synthesis a critical consideration when designing synthetic routes. Regioselective C-H functionalization reactions are attractive for preparing alkylated arenes; however, the selectivity of existing methods is modest and primarily governed by the substrate's electronic properties. Here, we demonstrate a biocatalyst-controlled method for the regioselective alkylation of electron-rich and electron-deficient heteroarenes. Starting from an unselective "ene"-reductase (ERED) (GluER-T36A), we evolved a variant that selectively alkylates the C4 position of indole, an elusive position using prior technologies. Mechanistic studies across the evolutionary series indicate that changes to the protein active site alter the electronic character of the charge transfer (CT) complex responsible for radical formation. This resulted in a variant with a significant degree of ground-state CT in the CT complex. Mechanistic studies on a C2-selective ERED suggest that the evolution of GluER-T36A helps disfavor a competing mechanistic pathway. Additional protein engineering campaigns were carried out for a C8-selective quinoline alkylation. This study highlights the opportunity to use enzymes for regioselective radical reactions, where small molecule catalysts struggle to alter selectivity.

Synergistic Binding of the Halide and Cationic Prime Substrate of the l-Lysine 4-Chlorinase, BesD, in Both Ferrous and Ferryl States

An aliphatic halogenase requires four substrates: 2-oxoglutarate (2OG), halide (Cl - or Br - ), the halogenation target ("prime substrate"), and dioxygen. In well-studied cases, the three non-gaseous substrates must bind to activate the enzyme's Fe(II) cofactor for efficient capture of O 2 . Halide, 2OG, and (lastly) O 2 all coordinate directly to the cofactor to initiate its conversion to a cis -halo-oxo-iron(IV) (haloferryl) complex, which abstracts hydrogen (H•) from the non-coordinating prime substrate to enable radicaloid carbon-halogen coupling. We dissected the kinetic pathway and thermodynamic linkage in binding of the first three substrates of the l -lysine 4-chlorinase, BesD. After 2OG adds, subsequent coordination of the halide to the cofactor and binding of cationic l -Lys near the cofactor are associated with strong heterotropic cooperativity. Progression to the haloferryl intermediate upon addition of O 2 does not trap the substrates in the active site and, in fact, markedly diminishes cooperativity between halide and l -Lys. The surprising lability of the BesD•[Fe(IV)=O]•Cl•succinate• l -Lys complex engenders pathways for decay of the haloferryl intermediate that do not result in l -Lys chlorination, especially at low chloride concentrations; one identified pathway involves oxidation of glycerol. The mechanistic data imply that (i) BesD may have evolved from a hydroxylase ancestor either relatively recently or under weak selective pressure for efficient chlorination and (ii) that acquisition of its activity may have involved the emergence of linkage between l -Lys binding and chloride coordination following loss of the anionic protein-carboxylate iron ligand present in extant hydroxylases.