Biocatalysis@TUDelft

The novel ADO from Pseudomonas plecoglossicida (PsADO) contains an extended loop with a disulfide bond that opens the substrate tunnel directly to the active site. Unlike other ADOs with an L-shaped hydrophobic tunnel, this structural adaptation enhances thermostability and catalytic efficiency. PsADO converts alkanes 34 times more efficiently than Prochlorococcus marinus ADO, making it a promising candidate for industrial biofuel production.

Aldehyde deformylating oxygenase (ADO) plays a crucial role in hydrocarbon biosynthesis by converting C_n fatty aldehydes into C_n−1 alkanes, key components of biofuels. However, ADO's low catalytic efficiency and thermostability hinder its industrial application. In this study, we identified a novel ADO from Pseudomonas plecoglossicida (PsADO) using the Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST). PsADO contains a novel loop motif with a disulfide bond that forms a new substrate tunnel, enhancing both thermostability and catalytic efficiency. PsADO exhibited a melting temperature (T _m) of over 61 °C, significantly higher than that of Prochlorococcus marinus ADO (PmADO, T _m = 41 °C), indicating superior stability. PsADO achieved its highest alkane yield at 10% oxygen, with a k _cat of 1.38 min⁻¹, 106 times higher than that of PmADO for tridecane formation. A hybrid reducing system, combining ferredoxin from Synechocystis sp. PCC6803 and ferredoxin–NADP⁺ reductase from Escherichia coli, further enhanced PsADO's activity compared with traditional chemical systems (PMS/NADH). AlphaFold 3 and CaverDock studies revealed that deleting PsADO's extended loop reduced alkane production by up to 9.4-fold, while the N47A variant reduced tridecane formation by 1.25-fold, confirming the importance of these structural features for substrate access and stability. These findings highlight PsADO's potential for biofuel applications, particularly in the production of long-chain alkanes for jet fuel. PsADO's improved stability and efficiency make it a promising candidate for industrial biotechnology and biofuel production, with further optimization potential through genetic and metabolic engineering.

Ferredoxin-dependent flavin thioredoxin reductases (FFTRs) regenerate reduced thioredoxin, sustaining dithiol–disulfide exchange reactions that regulate protein activity in select organisms. In cyanobacterial FFTRs, we describe the formation of a thiolate–flavin charge transfer complex, asynchronous reduction of the two FAD cofactors within the homodimer, and coexistence of two catalytically competent conformations—flavin-reducing (FR) and flavin-oxidizing (FO). These mechanistic insights deepen our understanding of FFTRs, offering a framework applicable to other flavoenzymes to expand our understanding of their redox biochemistry.

Ferredoxin-dependent flavin thioredoxin reductases (FFTRs) catalyze the reduction of the disulfide bond in thioredoxins using electrons transferred from ferredoxin, and therefore play a pivotal role in cellular disulfide relay reactions. FFTRs are essential in cyanobacteria such as Gloeobacter and Prochlorococcus, in which they serve as the sole thioredoxin reduction system, as well as in certain Clostridium species, where they are implicated in processes such as sporulation. Despite the well-established role of ferredoxin in reducing FFTRs, the underlying mechanistic details remain poorly understood. This study examines the catalytic cycle of FFTR from Gloeobacter violaceus, focusing on the role of its redox-active disulfide in electron transfer. We demonstrate that FFTR has a highly negative flavin adenine dinucleotide (FAD) midpoint reduction potential, which explains its preference for ferredoxin over nicotinamide adenine dinucleotide phosphate (NADPH) as an electron source. Spectroscopic detection of a thiolate–flavin charge transfer complex along the enzyme reduction pathway provides the first experimental evidence of a previously elusive FFTR catalytic conformation. Our results further reveal sequential FAD reduction within the enzyme homodimer that strongly suggests monomer asymmetry. Moreover, the impaired flavin reduction observed in an enzyme variant lacking the disulfide highlights the essential role of this redox group in efficient electron transfer. These findings deepen our understanding of FFTR's unique functional adaptations and evolutionary significance. More broadly, they provide a framework for exploring similar electron transfer mechanisms in other flavoproteins with a view to expanding our understanding of their redox biochemistry.

Polyethylene terephthalate (PET) accounts for approximately 6% of global plastic production and is a major contributor to plastic pollution. Enzymatic recycling offers a promising solution, but current PET-degrading enzymes often lack sufficient thermostability and catalytic efficiency. Through in silico design, we engineered a more robust variant, LCC-ICCG-C09, which exhibits enhanced thermal stability and twice the depolymerization efficiency of its predecessor, LCC-ICCG. This mutant shows strong potential for industrial-scale PET recycling applications.

Polyethylene terephthalate (PET) accounts for ≈6% of global plastic production, contributing considerably to the global solid-waste stream and environmental plastic pollution. Since the discovery of PET-depolymerizing enzymes, enzymatic PET recycling has been regarded as a promising method for plastic disposal, particularly in the context of a circular economy strategy. However, because the PET-degrading enzymes developed so far suffer from relatively limited thermostability and low catalytic efficiency, as well as degradation product inhibition, their large-scale industrial applications are still largely hampered. To overcome these limitations, we engineered the current PET-hydrolyzing enzyme gold standard [the ICCG variant of leaf-branch compost cutinase (LCC-ICCG)] using in silico protein design methods to develop a PET-hydrolyzing enzyme that features enhanced thermal stability and PET depolymerization activity. Our mutant, LCC-ICCG-C09, features a 3.5 °C increase in melting temperature relative to the LCC-ICCG enzyme. Under optimal reaction conditions (68 °C), the engineered enzyme hydrolyzes amorphous PET material into terephthalic acid (TPA) with a two-fold higher efficiency compared to LCC-ICCG. Owing to its enhanced properties, LCC-ICCG-C09 may be a promising candidate for future applications in industrial PET recycling processes.

Proceedings of the National Academy of Sciences, Volume 122, Issue 34, August 2025.
SignificanceBerberine bridge enzyme (BBE)-like enzymes, a class of enzymes that mainly catalyze oxidative cyclization or oxidative dehydrogenation, play a crucial role in natural product biosynthesis. However, the mechanism governing the selectivity ...

Proceedings of the National Academy of Sciences, Volume 122, Issue 34, August 2025.
SignificanceChemical routes to thioether-containing cyclic peptides rely on preinstalled electrophiles and often complex syntheses, restricting scaffold diversity and limiting applications. We found that the radicalS-adenosyl-L-methionine enzyme, PapB, ...

New microbial hosts with superior phenotypes, such as fast growth, are attractive for research and biotechnology, but often lack systematic evaluation of functional genetic parts. A reference set of working plasmids would increase reproducibility and encourage use of these hosts. Here, we use the POSSUM toolkit, a collection of 23 origins-of-replication and 6 antibiotic markers, to identify functional genetic parts for strains of Vibrio natriegens. We applied this to the wild-type strain ATCC 14048 and an engineered variant NBx CyClone, evaluating 414 combinations of origins of replication and antibiotic selection conditions. We show that both strains support five replicons (pNG2, pSa, pSC101ts, p15A, and RSF1010) with NBx CyClone supporting an extra replicon RK2. The assay can be performed in under a week and is compatible with multiple DNA delivery methods. This work demonstrates the feasibility of rapidly establishing reference information to accelerate the adoption of new microbial hosts.

The regio- and stereoselective hydroxylation of unactivated C(sp3)-H bonds is an important reaction in organic synthesis. While bacterial alkane monooxygenase AlkB catalyzes the terminal hydroxylation of aliphatic esters with excellent regioselectivity, the molecular principles of substrate recognition and selectivity of this integral membrane enzyme are still poorly understood. In this study, we investigated the substrate scope and engineered the medium-chain alkane monooxygenase from Marinobacter sp. (M_AlkB) for the terminal hydroxylation of linear and branched esters of fatty acids and alcohols. For the first time, we demonstrated the stereoselectivity of AlkB towards prochiral substrates containing terminal gem-dimethyl groups, leading to the corresponding chiral {beta}-methyl primary alcohols in good optical purity (51-79% ee). The hydroxylation products can be further derivatized to chiral diols and lactones. Substitution of the highly conserved active site residue F169 to leucine increased the activity towards short and medium-chain esters up to two-fold. While the wildtype enzyme does not accept long-chain substrates, activity towards n-dodecyl acetate could be unlocked by reducing the size of the tryptophan residue 60 situated in the putative substrate tunnel. Substitution of the peripheral I238 with valine increased activity regardless of the chain length of the substrate. Our results lay the groundwork for the establishment of a whole-cell process for the regio- and stereoselective hydroxylation of linear and branched esters, leading to valuable bifunctionalized products. The insights gained from mutating key residues and the substrate acceptance of AlkB will guide future protein engineering campaigns.

Our study successfully explores strategies to effectively improve the photocontrol efficiency of light-sensitive enzymes, dubbed photoxenases, with photoswitchable unnatural amino acids (UAAs). The engineering of photoxenases is a versatile method for the reversible photocontrol in various applications. To boost the photocontrol of an established allosteric and heterodimeric photoxenase based on imidazole glycerol phosphate synthase, we turned from an ineffective tuning of the UAA photochemistry to a semi-rational enzyme design. Remarkably, mutations at the catalytically important heterodimer interface increased the light-regulation factor (LRF) for the kcat up to ~100 with near-quantitative reversibility. Steady-state kinetic investigations combined with computationally determined correlation-based Shortest-Path-Map analysis and conformational landscapes revealed how photocontrol was altered in the two best hits. The LRF(kcat) correlated with a shift of a conformational equilibrium between an active and inactive population at the targeted active site and a tuned population productivity upon irradiation. While the overall reduced kcat values originated from a rewiring of the allosteric signal transmission, the increased LRF(kcat) resulted from a change in i) the size of the conformational shift, ii) the population productivity, and iii) the conformational heterogeneity. With this, our findings provide initial guidelines to boost photocontrol and underscore the power of photoxenase engineering.

Combining bioorthogonal protecting groups with localized catalysts that can unmask them is a powerful approach to spatially and temporally modulate molecular activity. Enzymes are appealing catalysts in this context because they are genetically targetable, but enzymes are not always available to unmask a protecting group of interest. Here, we report a platform for ultrahigh-throughput enzyme evolution by combining yeast surface display with masked acylating probes, which selectively label yeast cells based on target biocatalytic activity. We introduce the phenylcyclopropyl (pCP) ester protecting group, which has improved bioorthogonality compared to existing ester protecting groups, and use our platform to evolve BS2 esterase for enhanced pCP unmasking. Evolved BS2 mutants are up to 232-fold more active toward the pCP group. Taking advantage of the enhanced bioorthogonality of the pCP group, we applied a pCP probe together with evolved BS2 to perform spatially-resolved RNA tagging with high spatial specificity, including in mammalian cell lines with high endogenous esterase activity. Overall, this work delivers a new bioorthogonal protecting group and engineered enzymes capable of unmasking it, and more broadly, it provides a platform to rapidly engineer enzymes for protecting group removal, opening opportunities in imaging, proximity tagging, induced cell signaling, and therapeutics.

A synergistic photoredox biocatalysis approach was developed to realize new catalytic mechanism of enamine-dependent class I pyruvate aldolase. Both enantiomeric products were obtained in a stereoconvergent fashion through radical alkylation by wild-type and engineered aldolases.

Abstract

Class I aldolases, a unique link among biochemistry, organic chemistry, and computational chemistry are powerful C─C bond-forming enzymes in synthetic chemistry and industry because of their unparalleled selectivity, extensive substrate scope and scalability. However, the types of reactions catalyzed by class I aldolases are restricted and radical reactions have yet to be accomplished. Here, we demonstrate a proof-of-concept study in which a synergistic photoredox biocatalysis strategy can be applied to realize new catalytic functions of enamine-dependent aldolases. This new reactivity enables asymmetric alkylation of a prochiral radical under exclusive stereocontrol, a challenging task for amine catalysts. Both enantiomeric products were obtained in a stereoconvergent fashion from wild-type and engineered aldolases. This synergistic photoredox biocatalysis strategy has resulted in a new-to-nature enzymatic reaction and led to an asymmetric transformation that is not feasible for organocatalysis. We envision that this discovery will motivate the development of enzymatic enamine and iminium catalysis for valuable asymmetric radical transformations, complementing the prevailing organocatalysts.

A rare thermostable Fatty Acid Hydratase ortholog from Marinitoga Piezophila, a thermo-piezophilic organism, displays novel properties. While the enzyme has excellent thermostability (retaining full activity after incubation at 70°C), quite interestingly, it shows the highest activity at 20 °C. Moreover the enzyme has a broad substrate scope and unique regioselectivity.

Hydroxy fatty acids (HFAs) are valuable derivatives of fatty acids (FAs) with interesting bioactivities. Moreover, they are used in materials industry as additives, starting materials and surfactants. HFAs can be produced from FAs either by hydroxylation or by hydration reaction, if FA is unsaturated, using chemical or enzymatic methods. FA hydratases (FAHs) are promising biocatalysts for HFA synthesis thanks to their non-redox nature, high efficiency and excellent selectivity. Although FAHs are relatively more stable compared to other enzymes like monooxygenases, their tolerance to high temperature and organic solvents is limited. In this study, we characterized a rare thermostable FAH ortholog through database gene mining. This enzyme from Marinitoga Piezophila, a thermo-piezophilic organism, displayed novel properties, including broad substrate scope, broad pH range, unique regioselectivity and excellent thermostability (retaining full activity after 30 min incubation at 70 °C); however, quite interestingly, its temperature optimum was at 20 °C. Although kinetic parameters indicate a less efficient enzyme compared to some other FAHs, the enzyme can reach over 90% conversion within 24 h at a 100 mL scale reaction containing 1.75 mM substrate. Furthermore, mutagenesis of key active-site residues indicated a possibly different reaction mechanism compared to earlier proposed mechanisms.

Integrating new metal-catalysed transformations into enzymes is a key objective in biocatalysis. This study introduces photoinduced ligand-to-metal charge transfer (LMCT) as a new strategy for enabling abiotic cross-coupling reactions in metalloenzymes. By tailoring the primary coordination sphere to establish a 2-histidine metal binding site and replacing the iron center with nickel, the ethylene-forming enzyme from Pseudomonas savastanoi (PsEFE) was activated for nickel-catalysed C(sp2)‒S cross-coupling between aryl bromides and thiols. Directed evolution of PsEFE produced highly active variants capable of generating over 50 thioether products in up to 98% yield and 530 total turnover numbers. Mechanistic investigations suggest that this photoenzymatic reaction involves a Ni(II)/Ni(I)/Ni(III) catalytic cycle with generation of a reactive Ni(I) species and thiyl radical via photoinduced LMCT. We anticipate that these findings will inspire further exploration of integrating abiotic cross-coupling transformations into enzymatic catalysis.

Chloramines are an important class of compounds containing a covalent nitrogen-chlorine bond. Despite the growing interest in their applications in the small molecule and polymer industries, selective and sustainable catalyst systems for their synthesis have remained elusive. We recently discovered that the vanadium-dependent chloroperoxidase from Curvularia inaequalis (CiVCPO) is an effective biocatalyst for selective chlorination of a broad range of structurally diverse amines to give the corresponding chloramines and chlorimines in moderate to high yield and with excellent chemoselectivity. The catalyst system is readily scalable and applied to chemoenzymatic nitrile and amide synthesis. Finally, halide divergent reactivity is demonstrated in chloride-selective chlorimine formation and bromide-selective aldehyde formation using the same biocatalyst.

Accurately modeling the complex electronic structure of metalloenzyme active sites remains a significant challenge for density functional theory (DFT). The MME55 test set [J. Chem. Theory Comput. 2023, 19, 8365] is a recent benchmark providing high-level DLPNO-CCSD(T)/CBS reference values for large, practically relevant metalloenzyme models with up to 116 atoms. So far, the best-performing rung-4 hybrids were MPW1B95-D3(BJ) with a mean absolute error (MAE) of 2.67 kcal/mol and ωB97M-V (2.78 kcal/mol). This work focuses on range-separated local hybrids (RSLHs), a recent class of rung-4 functionals that use both a real-space dependent and inter-electronic distance dependent admixture of exact (Hartree-Fock-like) exchange. Strong-correlation corrected RSLHs (scRSLHs) based on real-space non-dynamical correlation models are also assessed. The best-performing functional is the scRSLH ωLH23tdB-D4 with an MAE of 2.46 kcal/mol. Overall, the results reflect a systematic improvement in accuracy along the functional development series from LH to RSLH to scRSLH. However, our analysis also indicates that the real-space non-dynamical correlation model is not the primary driver of this improvement, which is instead dominated by other features of the functional design. Additionally, we studied (sc)(RS)LHs for the ENZYMES22 set of enzyme reactions comprising metal-free active site models as well as the ECR20 benchmark of net reaction energies of enzyme catalyzed reactions. MAEs of 1.45 kcal/mol and 0.63 kcal/mol confirm ωLH23tdB-D4 as one of the top-performing functionals for these sets. Due to inaccuracies in the original CCSD(T)/aug-cc-pVDZ + [SCS-MP2/aug-cc-pVTZ - SCS-MP2/aug-cc-pVDZ] references values for the ECR20 set, we devise new canonical CCSD(T)/CBS reference values using a aug-cc-pVXZ (X = D, T, Q) based basis set extrapolation. Because scRSLHs have only very recently emerged, this benchmark study is accompanied by critical assessments of their use as a practical tool. We analyze the grid dependence of their semi-numerical implementation in Turbomole and evaluate the computational cost and parallel scaling on a shared-memory OpenMP architecture using up to 192 physical CPU cores.

The ability to engineer enzymes for desired reactions is a cornerstone of modern biotechnology, yet identifying suitable starting proteins remains a critical bottleneck. Dual-encoder contrastive learning models have emerged as a promising approach for enzyme discovery, learning to match chemical reactions with catalyzing enzymes through a shared embedding space. However, their practical performance beyond computational benchmarks remains unproven. Here, we close this gap with Horizyn-1, a deep learning framework for direct reaction-to-enzyme recommendation validated through comprehensive experimental testing. Leveraging a computationally efficient combination of reaction fingerprints and protein language models, we trained Horizyn-1 on 8.9 million reaction-enzyme pairs to achieve state-of-the-art performance, recovering an enzyme with correct activity within the top 100 hits for over 75% of test reactions. We experimentally validate Horizyn-1 across three enzyme discovery scenarios: identifying enzymes for orphan reactions, predicting enzyme promiscuity for both characterized and uncharacterized enzymes, and discovering enzymes for non-natural biochemical reactions including lysine-driven transaminations that enable efficient synthesis of non-canonical amino acids. On underrepresented reaction classes, we find that fine-tuning with fewer than 10 additional reactions can dramatically improve performance. Furthermore, a logarithmic scaling of model performance with training dataset size suggests continued improvement with larger and more diverse reaction datasets. Horizyn-1 addresses the critical bottleneck of sourcing initial enzymes for optimization campaigns, enabling efficient and scalable in silico screening for enzymes with desired activities and promising to accelerate future efforts in biocatalysis and metabolic engineering.

Abstract

Mirror-image poly(ethylene terephthalate) (PET) plastic-degrading enzymes have emerged as promising biocatalytic platforms due to their exceptional enzymatic stability and low immunogenicity. Currently, the sole reported mirror-image plastic-degrading enzyme, the D-form of 231-residue PET hydrolase ICCG (engineered leaf-branch compost cutinase variant), suffers from thermophilic activity requirements, which limits its practical applications. Here, the first total chemical synthesis of a mirror-image 271-residue D-Fast-PETase was presented by using an enzyme-cleavable solubilizing tag strategy. Comparative kinetic analysis revealed that D-Fast-PETase showed a remarkable increase (∼20-fold within 24 hours) in PET degradation efficiency compared to D-ICCG at temperatures of 37 °C, making it a promising candidate for prolonged PET decomposition in open environments and holding potential in addressing microplastic-related health issues within the biomedical field. This work not only expands the chemical biology toolbox for mirror-image enzyme synthesis but also establishes D-Fast-PETase as a candidate in combating the dual crises of global plastic pollution and microplastic-associated health risks.

A novel catalytic mode of aldoxime dehydratases for the abiotic radical ring-opening reaction of cyclic ketoximes is reported. Aldoxime dehydratase from Nocardioides simplex was found to efficiently generate iminyl radicals from challenging “non-activated” cycloketone oximes and to promote radical ring-opening reactions to produce γ- and ε-sulfinylated nitriles under mild conditions.

Abstract

The iminyl radical is a distinctive N-centered radical which serves as a versatile synthon in preparation of nitrogen-containing compounds. In principle, iminyl radicals can be directly generated by single electron reduction of oximes through elimination of OH group. However, due to the low reactivity of the oxime N─OH bonds, direct conversion of the oximes does not proceed efficiently, thereby enforcing chemical activation of the oxime OH group which results in the formation of stoichiometric by-products. To overcome this problem, we are developing a new biocatalytic system using aldoxime dehydratases. Through a series of enzyme screenings, we identified an aldoxime dehydratase from N. simplex (NsOxd) which is capable of catalyzing iminyl radical-mediated ring-opening reactions. Notably, NsOxd efficiently converts the “non-activated” 2-phenylcyclobutanone oxime within 10 min under ambient conditions and quantitatively produces the corresponding γ-sulfinylated nitrile in >95% yield. This enzyme activity is even faster than that of previously-reported chemo-catalysts. Furthermore, evaluation of the scope of potential substrates indicates that NsOxd has a versatile N─O bond cleaving activity which efficiently generates iminyl radicals from various “non-activated” oximes. These findings highlight the utility of aldoxime dehydratases for managing the reactivity of “non-activated” oximes and for achieving challenging iminyl radical-mediated catalytic reactions.