Biocatalysis@TUDelft

Flavin-dependent halogenases provide a biocatalytic approach for site-selective halogenation of aromatic compounds, but their use in late-stage functionalization of peptides has remained limited. Here, we show that the tryptophan (Trp) 7-halogenase RebH and an engineered variant (4V) originally optimized for larger small-molecule scaffolds can brominate peptidyl-Trp residues across a broad range of sequence and positional contexts. Through extensive analysis of diverse substrate sequences, we define features that enable RebH activity and reveal 4Vs expanded sequence tolerance. We applied 4V for enzymatic bromination of diverse bioactive peptide scaffolds, including an antimicrobial peptide, a cell-penetrating peptide, and a G protein-coupled receptor agonist, without the need for sequence modification. These brominated peptides served as substrates for Suzuki-Miyaura coupling, enabling installation of functional groups that conferred new functional properties or tuned the biological activity of these peptides. Our results expand the substrate landscape of FDHs and establish bromination-enabled cross-coupling as a general approach for late-stage diversification of bioactive peptides.

Styrene monooxygenases with different enantioselectivities allow olefins to adopt distinct binding orientations within the catalytic cavity, leading the O-atom of C4a-hydroperoxyflavin (FAD_OOH) to attack either the Re-face or the Si-face of the substrate. This difference in attack direction ultimately determines the resulting enantioselectivity.

Numerous styrene monooxygenases (SMOs) have been identified and extensively applied in the asymmetric epoxidation of alkenes to prepare enantiopure epoxides, which are important intermediates in pharmaceuticals, agrochemicals, and fine chemicals. A current challenge is development of engineered SMOs capable of catalyzing a broad range of substrates with complementary enantioselectivity, thereby enabling the production of both (R)- and (S)-epoxides. Achieving this requires a deep understanding of the molecular basis underlying enantiocontrolling in SMO-catalyzed epoxidation. In this concept article, recent progress in elucidating the mechanisms of enantiocontrol by SMOs is summarized, with particular emphasis on the structural and mechanistic features that differentiate (R)- and (S)-selective SMOs. These insights not only expand fundamental knowledge of SMO catalysis but also provide a critical foundation for the rational design and protein engineering of highly efficient and versatile SMOs.

A biochemical characterization of Staphylococcus aureus serine proteases SplA and SplB allelic variants (recombinantly produced in E. coli and purified) reveals that allelic variation adds an extra layer of complexity to the functioning of virulence factors in S. aureus colonization and infection. This highlights the importance of considering allelic variation in the study of host-pathogen interactions.

Staphylococcus aureus is an opportunistic pathogen that is persistently colonizing nearly 30% of the human population and can cause life-threatening infections. S. aureus secretes a variety of virulence factors, such as a set of extracellular serine protease-like proteins (Spls). Spls are expressed by most clinical isolates of S. aureus, but their pathophysiological substrates and role during infection are largely unknown. Pathogens use allelic variation of virulence factors to allow an adaption to different host cells and their defense mechanisms. The differences of these variants are marginally characterized so far. Here, we performed a biochemical characterization of selected allelic variants of the S. aureus SplA and SplB. Our data suggests different variants show differences in their stability, enzymatic activity, and substrate specificity. For the recently identified Spl target proteins RickULP and SseL, different cleavage patterns were observed, upon treatment with different Spl allelic variants. One SplB variant strongly differed in its substrate recognition at the P3/P4 position, more closely resembling SplE than SplB wildtype in its substrate selectivity. Our data provide valuable insights into the evolution of bacterial virulence factors and highlight the importance of including allelic variation of virulence factors to fully understand their role in host–pathogen interaction.

Applied and Environmental Microbiology, Volume 92, Issue 1, January 2026.

The Michaelis constant (Km) is a key parameter in enzymology, yet its experimental measurement is often low-throughput and costly. While machine learning (ML) models offer a promising alternative, they have predominantly relied on separate feature representations for the enzyme and substrate, lacking explicit information about their interaction interface. In this work, we introduce a novel approach that explicitly incorporates enzyme-substrate interface information by encoding the enzyme’s active site as a feature. Using a simple multilayer perceptron (MLP) with a gated layer, we demonstrate that this explicit active site information enables our model, Active Site for Km (AS4Km), to achieve state-of-the-art performance on the independent HXKm test set, rivaling more complex architectures. Ablation studies confirm that active site features significantly enhance generalization to unseen data. Furthermore, our analysis highlights a critical limitation in current enzymology databases: predictive performance is heavily reliant on substrate identity due to low substrate diversity and a bias towards active enzyme-substrate complexes. Our results show that AS4Km, a data-driven approach combined with explicit interaction interface features, displays competitive performance in the prediction of Km values for enzyme-substrate complexes, and may be able to assist in the identification of novel substrates for known enzymes.

Direct electrochemistry (DET), whereby a redox enzyme exchanges electrons with an electrode without the intervention of any mediator, has emerged at the turn of the XXIth century as a powerful tool to investigate redox enzymes. In this review we focus on the electrochemistry of metalloenzymes that are involved in the recycling of biological solar fuels. We summarize the strategies that made it possible to wire them to electrodes, from their mere physisorption onto friendly electrodes to more sophisticated covalent attachment. We then describe a number of qualitative and quantitative electrochemical investigations of their mechanisms. These studies have disclosed the functional diversity within each family of enzymes considered here, suggesting that structural features that are remote from the active site play a major role in determining catalytic properties such as catalytic directionality, reversibility, and resistance to oxygen. DET electrochemistry contributed to elucidating some of these outer sphere effects, which are crucial in molecular catalysis, and also to discovering and characterizing new enzymes, which have the properties and robustness required in solar fuel devices.

Vitamin B12 is a structurally unique tetrapyrrole cofactor that catalyzes a wide range of radical and polar reactions. Its unique structure, diverse axial ligands, and ability to access distinct redox states underpin the reactivity of three major classes of cobalamin-dependent enzymes: adenosylcobalamin-dependent isomerases, methylcobalamin-dependent methyltransferases, and reductive dehalogenases. These enzymes leverage precise scaffold-controlled interactions to direct radical rearrangements, methyl group transfers, and reductive bond cleavages with remarkable selectivity and efficiency. Despite these capabilities, the native reactivity of B12 enzymes remain relatively underutilized for biocatalysis, and expansion of their activity toward non-native reactions has only recently emerged as a promising frontier. Mechanistic studies and advances in protein engineering and synthetic biology are beginning to establish B12 enzymes as versatile platforms for selective C-C bond formation, C-H functionalization, alkylation, and other transformations. This perspective summarizes key structural and mechanistic foundations of B12 enzyme catalysis, highlights progress in native biocatalysis, and discusses emerging strategies to exploit B12-dependent enzymes as a platform for non-native biocatalysis.

Natural Diels–Alderases catalyse [4+2] cycloadditions by preorganizing substrates into reactive conformations. However, the role of other catalytic factors, such as electrostatic effects, remain elusive. Here, we combine conceptual Density Functional Theory (CDFT) descriptors and electric field analysis to unravel the electrostatic basis of activity in the Diels-Alderase AbyU. Previously, four different enzyme-substrate poses were identified, of which two showed catalytically favorable free energy barriers based on quantum mechanical/molecular mechanical (QM/MM) reaction simulations. Here, we show that atom-condensed Fukui functions can predict the reactivity from reactant conformations alone, focusing on the diene carbons involved in bond formation. The importance of the enzyme-diene interaction is supported by electric field analysis, which shows how reactivity of enzyme-substrate poses correlates with alignment of the enzyme field along the diene moiety. Our findings establish a basis for predicting and engineering Diels–Alderase activity based on electrostatic and electronic reactivity features.

In the past decades, usage of enzymes to catalyse halogenation reactions has emerged as a greener approach compared to usual ones using toxic reagents and yielding to a bad atom economy. Vanadium dependent haloperoxidases (VDHals) are the most commonly used enzymes for such transformations in organic chemistry due to their robustness. Among them, haloperoxidase from Curvularia inaequalis (CiVCPO) is the most cited in the literature but only barely characterized aside from its substrate spectrum and kinetic parameters. In the present study, we evaluated the melting temperature, thermostability, thermoactivity and solvent stability of CiVCPO in order to have a better overview of its potential in organic synthesis. We have also performed the first immobilisation study of this enzyme on 3 types of supports: 3 EziG™ glass beads coated with different polymers, 5 Relizyme™ polymethacrylate beads with different functional groups, the commercial Amberlite IRA900 and Dowex 50WX8, and 3 metal-organic frameworks from the UiO-66(Zr) family. As a result, we could retain 50% of total immobilized activity on 2 Relizyme supports (ethylamine (EA403) & iminodiacetic (IDA403), with 55% activity kept over 5 recycling steps. The unconventional UiO-66(Zr) family also proved to be an interesting material for this enzyme. Finally, we show that free enzyme and supported enzyme are suitable for bromination and chlorination of a range of phenolic compounds with excellent yields with up to 65% conversion in 24 h in the tested conditions.

In this study, we reveal a long-standing mechanistic conundrum of the functions of two very similar isozymes, CYP1A1 and CYP1A2, that possess nearly identical active-site architectures yet display a perplexing difference in their reactivity toward the flavonoid α- naphthoflavone (ANF). CYP1A1 efficiently catalyses epoxidation, whereas CYP1A2 shows negligible activity. This contrast is especially striking because the crystallographic orientation of ANF is the same in both enzymes, and the reactive carbon lies far from the heme-oxo centre. To resolve this long-standing puzzle, we combined extensive molecular dynamics simulations with hybrid QM/MM calculations. Our results reveal that the divergent reactivity arises not from differences in substrate binding, but from distinct water architecture within the catalytic pocket. CYP1A1 forms an open, well-organised aqueduct connecting the heme to the reactive carbon centre, facilitating epoxidation, whereas CYP1A2 lacks such an organised channel. This contrast is attributed to the different synchronized movement of the F and I helices, resulting in altered side-chain packing between key residues controlling the solvent gate. Site-directed mutations confirm the reopening of the closed water gate in CYP1A2 and reestablish water occupancy. Hybrid QM/MM calculations further reveal that ANF epoxidation proceeds through a sequential water-mediated relay culminating in an asynchronous proton-coupled electron transfer (PCET) step that yields the experimentally observed 5,6 oxide. These findings establish that subtle second-shell variations reshape water topology and thereby control catalytic competence in two deceptively similar P450 isozymes, providing a unified mechanistic explanation for their divergent reactivity.

Despite the remarkable versatility of enzymes for catalyzing complex organic transformations under mild aqueous conditions, no biocatalytic system has been reported for intermolecular C–H fluorination. A radical photoenzymatic approach is introduced that leverages a designed de novo protein scaffold with an unnatural amino acid to achieve precise benzylic monofluorination by a hydrogen atom transfer mechanism.

Abstract

Organofluorine compounds are vital in pharmaceuticals, and enzymes, nature's most efficient catalysts, offer tremendous potential for precise fluorination. However, no enzymatic strategies for intermolecular C–H fluorination have been realized—until now. We present the first radical photoenzymatic system enabling intermolecular C–H fluorination using an unnatural amino acid within a robust de novo protein scaffold. This system achieves chemoselective benzylic monofluorination in aqueous solutions with Selectfluor, driven by hydrogen atom transfer from the photoexcited amino acid. It successfully fluorinated various aromatic compounds and enabled biosynthesis of fluorinated polyketides and chiral fluorinated alcohols. These results establish radical photoenzymatic systems as a powerful new approach for efficient, selective biocatalytic fluorination, with direct relevance to pharmaceuticals.

Nature Chemical Biology, Published online: 18 December 2025; doi:10.1038/s41589-025-02110-z

Machine learning-based tools have revolutionized how scientists study protein structure. Here, Nature Chemical Biology speaks to Cecilia Clementi, Bruno Correia and Peilong Lu about progress in developing computational tools for predicting protein structure and properties, how these programs can be used for protein design, and the developments they would like to see in the field.

Nature Biotechnology, Published online: 18 December 2025; doi:10.1038/s41587-025-02937-w

Base editors are made more precise through deaminase and gRNA engineering.