Biocatalysis@TUDelft

Recent developments in DNA synthesis and sequencing have allowed the construction of comprehensive gene variant libraries and their functional analysis. Achieving high-replication and thorough mutation characterization remains technically and financially challenging for long genes. Here, we developed an efficient, affordable and scalable library construction approach that relies on low-cost DNA synthesis and standard cloning technologies, which will increase accessibility to systematic mutational studies and help advance the field of protein science.

Multiheme cytochromes c are versatile enzymes which contribute to key steps in diverse biogeochemical cycles of elements and are also key players in microbial electrochemical technologies. We previously showed that these enzymes evolve by grafting and pruning of cytochrome c modules. Here, we extended our analysis and found that grafting can also involve the incorporation of domains of other protein families and of small peptide sequences that affect the heme coordination environment. We show that the addition of hemes to multiheme cytochromes c occurs exclusively by integrating heme-binding peptides and that gain of hemes is equally probable along the protein sequence. By contrast, heme loss occurs by the loss of heme-binding peptides and the accumulation of point mutations. Notably, heme-binding motif loss is disproportionately more prevalent at the position nearer to the N-terminus. This observation contrasts with the general trend observed in proteins, which are usually more conserved at the N-terminus, and likely reflects the way in which multiheme cytochromes c are assembled. Overall, this work completes the picture of how multiheme cytochromes c evolved to achieve the diversity of structures and functions found in nature, and sets them apart from other proteins with respect to the drivers for their evolution. This has contributed to the diversity of roles that they play in various biogeochemical cycles and has implications for engineering artificial variants to enhance biotechnological applications.

A de novo designed single-chain coiled coil protein is unfolded in apo-state but folds up upon addition of divalent transition metal ions Co(II), Ni(II), Cu(II), and Zn(II). The resulting metal-bound complexes exhibit distinct thermal stability and metal stoichiometry, making this protein a promising starting point for the design of fully metal-selective de novo proteins.

Metal-binding selectivity in natural proteins is determined by multiple factors such as the protein's structure, metal concentration within cellular compartments, and the presence of metallochaperones. The in vitro selectivity of proteins for transition metal ions is largely governed by the Irving–Williams series, which states protein-metal complex stability follows the order Co(II) < Ni(II) < Cu(II) > Zn(II). A de novo protein has been designed that folds in the presence of certain transition metal ions into a monomeric α-helical bundle, with the least stable protein-metal complex being formed with Cu(II). Moreover, when increasing the metal concentration of Cu(II) or Zn(II), more metal ions are incorporated into the protein accompanied by a concurrent decrease in the amount of secondary structure. One reason may be that there is a balance between stability conferred by the coordination of the metal ion(s) and stability conferred by hydrophobic packing of the α-helical bundle. Metals may therefore adopt distorted coordination geometries, or binding of multiple ions may cause distortion of the protein backbone, leading to compromised folding of the protein scaffold, or variable thermal stabilities of the metalloprotein complexes. This protein scaffold therefore contributes to the deciphering of design rules for metal selectivity in proteins.

Nature Biotechnology, Published online: 23 May 2025; doi:10.1038/s41587-025-02654-4

Protein structures are predicted by integrating deep learning potentials with iterative threading fragment assembly simulations.

Nature, Published online: 02 June 2025; doi:10.1038/d41586-025-01674-z

Most proteins are left-handed, but scientists have found an ancient molecule that works in both mirror-image forms.

Nature, Published online: 28 May 2025; doi:10.1038/s41586-025-09178-6

Electricity-driven enzymatic dynamic kinetic oxidation

Nitrogen-nitrogen (N-N) bond formation is an inherently challenging chemical process that plays a key role in the global nitrogen cycle. An array of microbial metalloenzyme complexes has evolved to shuttle nitrogen between biologically accessible reduced or oxidized states and its inert form as dinitrogen (N2) gas. More recently, N-N bond formation has been observed in a more specialized context, natural product biosynthesis. Here, we report the discovery of a unique metalloenzyme complex that forms hydrazine functional groups in the biosynthetic pathways of structurally diverse natural products. This heterodimeric system consists of a heme enzyme from a previously unidentified family and a partner ferredoxin. Together, these enzymes effect the unprecedented four-electron reduction of nitrite (NO2-) to form a hydrazine functional group on a substrate amino acid in an oxygen-independent reaction that resembles primary microbial nitrogen metabolism. These enzymes are unexpectedly widespread among bacteria and are present in diverse genomic contexts, including cryptic biosynthetic gene clusters, highlighting the importance of this previously uncharacterized protein family. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=100 SRC="FIGDIR/small/656397v1_ufig1.gif" ALT="Figure 1000"> View larger version (30K): org.highwire.dtl.DTLVardef@1715eeaorg.highwire.dtl.DTLVardef@e8dceorg.highwire.dtl.DTLVardef@1800f03org.highwire.dtl.DTLVardef@599e1d_HPS_FORMAT_FIGEXP M_FIG TOC C_FIG

The inclusion of chemical modifications into oligonucleotides stabilizes their backbones against nuclease mediated degradation and improves their biological activity. Even though cellular, tissue, and organ specific delivery remains a challenging undertaking, conjugation to triantennary N-acetyl galactosamine (GalNAc) ligands permits efficient delivery of nucleic acids to hepatocytes. GalNAc ligands are usually added to the termini of therapeutic oligonucleotides by costly and time-consuming chemical methods that often require specialized equipment and knowledge. Here, we have explored the possibility of enzymatically labelling oligonucleotides with triantennary GalNAc ligands which would be directly amenable to any type of sequence without requiring any synthetic effort. We have modified the gamma-phosphate of ATP with a GalNAc moiety and explored the possibility of transferring the ligand to 5’-termini of oligonucleotides using kinases. In parallel, we have equipped the 3’-position of an LNA nucleoside triphosphate with a GalNAc residue and used this nucleotide analog for labelling of oligonucleotides with template-independent polymerases. While kinases do not seem to tolerate the presence of such a bulky residue on ATP, template-independent polymerases readily incorporate the modified nucleotide at the 3’-end of DNA and RNA oligonucleotides. Overall, this article represents a first step towards the development of a universal, enzymatic GalNAc-labelling method for therapeutic oligonucleotides.

The asymmetric mixed carboligation of aldehydes catalyzed by thiamine diphosphate (ThDP)-dependent enzymes provides a sensitive system for monitoring changes in activity, chemo-, and enantioselectivity. While previous studies have shown that organic cosolvents influence these parameters, we now demonstrate that similar effects occur upon addition of water-miscible ionic liquids (ILs). In this study, six ThDP-dependent enzymes were analyzed in the presence of 14 ILs under comparable conditions to assess their influence on enzymatic carboligation reactions yielding 2-hydroxy ketones. ILs exerted a moderate to strong influence on activity and, more notably, altered enantioselectivity. (R)-selective reactions were generally stable upon IL addition, while (S)-selective reactions frequently showed reduced selectivity or even inversion to the (R)-enantiomer. The most significant change was observed for the ApPDC_E469G variant of pyruvate decarboxylase from Acetobacter pasteurianus, where the enantiomeric excess shifted from 86% (S) to 60% (R) in the presence of 9% (w/v) Ammoeng 102. Control experiments indicated that this shift was primarily due to the Ammoeng cation rather than the anion. To explore the molecular basis of this phenomenon, all-atom molecular dynamics (MD) simulations were performed on wild-type ApPDC and the E469G variant in Ammoeng 101 and Ammoeng 102. The simulations revealed that hydrophobic and hydrophilic regions of the Ammoeng cations interact with the (S)-selective binding pocket, thereby favoring formation of the (R)-product. These results highlight the potential of solvent engineering for modulating enzyme selectivity and demonstrate that MD simulations can capture functionally relevant enzyme-solvent interactions at the atomic level.

α-Angelica lactone is a readily available small molecule from biomass source, which, coupled with cheapness and versatile reactivity, attracted significant interest for applications in asymmetric synthesis. In this minireview, stereoselective organo-, metal-, and enzymatic-based approaches for the construction of new butenolide-containing heterocycles, using α-angelica lactone as a reagent, are illustrated.

Abstract

α-Angelica lactone is a five-membered heterocycle belonging to the rich family of butenolides, a key motif, present in a great variety of bioactive and natural products. Moreover, it is a small compound attainable from the platform molecule levulinic acid, which in turn is readily accessible from lignocellulosic biomass. These important features contributed to arousing interest in α-angelica lactone over the last years. Beyond its upgrading to added value products, such as γ-valerolactone, polymers, and biofuels, α-angelica lactone has been identified as a building block in natural product synthesis and a useful reagent in asymmetric catalysis. In this minireview, we provide an overview of diastereo- and enantioselective organo-, metal-, and enzymatic approaches reported from 2020 up to 2024, where this heterocycle has been used as the starting material. α-Angelica lactone proved to be a versatile scaffold in stereoselective vinylogous 1,4- and 1,2-addition, cycloaddition, isomerization, γ-arylation, and reduction reactions.

We develop a minimal model for an enzyme that allows us to extract a set of golden rules regarding its optimal design. We discover that two key ingredients of momentum conservation and a generically present dissipative coupling give rise to a bifurcation in the dynamical phase portrait of the low-dimensional configuration space, which enables enzymatic activity as an emergent feature of the dynamics, via a fundamentally novel mechanism not accessible to Kramers-like energy-barrier-crossing descriptions in terms of reaction coordinates.

Protein structure prediction and design have traditionally been limited to the 20 canonical amino acids. Expanding this space to include noncanonical amino acids (NCAAs) offers new opportunities for probing novel interactions and engineering proteins with enhanced or entirely new functions. Some NCAAs also offer practical advantages, such as increased proteolytic stability and reduced immunogenicity, as they are rarely encountered by the human immune system. Here, we present RareFold, a deep learning model capable of accurate structure prediction for proteins containing both the 20 canonical amino acids and an additional 29 NCAAs. By treating each amino acid as a distinct token, RareFold learns residue-specific atomic interaction patterns, enabling precise modelling of chemically diverse sequences. This tokenised representation also supports sequence-structure co-optimisation, allowing efficient inverse design. We leverage this capability in EvoBindRare, a design framework for generating linear and cyclic peptide binders that incorporate NCAAs. Applying EvoBindRare, we design binders targeting a ribonuclease and experimentally validate these, obtaining M affinity in both the linear and cyclic cases. RareFold thus enables binder design with an expanded chemical vocabulary, opening the door to next-generation peptide therapeutics with improved stability, specificity, and immune evasion. RareFold is available at: https://github.com/patrickbryant1/RareFold

E. coli BL21 DE3 strain is commonly used to express and purify non-toxic prokaryotic proteins in high yields. Traditionally, IPTG based induction of the host strain at reduced temperatures for extended period (12-16 h) is performed to obtain high yield of functional proteins. It is desirable to explore methods which result in high yield of protein within a short period of time. We report rapid purification of pyruvate decarboxylase (PDC) enzyme from Zymomonas mobilis using E. coli BL21 pLysY/Iq as a host strain. High yield of purified PDC at 0.33 ({+/-}0.02) mg was obtained after two-hour induction by 0.6 mM IPTG at 37{degrees}C. The enzyme yield was comparable to 0.37 ({+/-}0.08) mg obtained in E. coli BL21 DE3 strain (used as control) after 16 h induction by 0.6 mM IPTG at 18{degrees}C. Similar values of the maximum specific activity of the enzyme expressed and purified at 37{degrees}C and 18{degrees}C were obtained at 78.31 ({+/-}1.13) in strain LysY/Iq and 85.73 ({+/-}4.39) {micro}mol/min/mg protein in strain DE3, respectively. In almost all IPTG treatments, the kinetic parameters of the purified enzyme - app Km, app Vmax, Kcat and Kcat/Km -also did not vary remarkably between the two temperature regimes. Based upon the data presented here, we propose that E. coli BL21 LysY/Iq strain has potential to serve as a host for efficient and rapid expression (2 h) of non-toxic proteins. Results of this study will aid in cell free system study which require rapid scale up of the complexity of metabolic pathways by utilizing multiple purified enzymes involved in bioconversion of the substrate of interest.

The pikromycin polyketide synthase (PKS) catalyzes formation of 12-membered macrolactone 10-deoxymethynolide, and 14-membered macrolactone narbonolide. Herein, we show the efficient diversification of novel 14-membered macrolactones from a series of unnatural pentaketides using the PikAIII/PikAIV PKS in vitro system. New macrocycles were further elaborated by the addition of D-desosamine and late-stage C-H hydroxylation. Molecular dynamics (MD) simulations and density functional theory (DFT) calculations were conducted to probe the reactivity and selectivity of this terminal catalytic step on the assembled unnatural macrolides. This approach demonstrates the flexibility and applicability of sequential biocatalytic steps for chemoenzymatic creation of complex antibiotic scaffolds.

The design of artificial selective catalysts that function effectively across structurally diverse substrates remains a long-standing challenge. In this study, we demonstrate that a random heteropolymer, synthesized via radical polymerization, can function as a general catalyst for the selective epoxidation of structurally diverse olefinic quaternary ammonium salts. This three-monomer-based polymer catalyst exhibits exceptional functional-group selectivity and remarkable structural adaptability that cannot be achieved by conventional small-molecule catalysts. Beyond reusability, our approach highlights the often-overlooked advantages of polymer catalysis, paving the way for its application in a broad spectrum of selective chemical transformations.

Predicting enzyme kinetic parameters is crucial for enzyme engineering and mining, but development of accurate tools for this task is still challenging due to the complexity of influencing factors. (LLMs)-based deep learning methods show satisfactory performance in Kcat prediction by encoding substrate and enzyme sequence information, but this further enhancement of the performance can be largely limited by the ignorance of the substrate-binding information. Here, we introduce GraphKcat, a deep learning framework that integrates enzyme-substrate 3D binding conformations for precise kinetic parameter prediction. The method employs Chai-1 to generate enzyme-substrate complex conformations, followed by a developed multi-scale hierarchical graph neural network to systematically characterize active site features from all-atom (AA) level to coarse-grained (CG) level, which are subsequently fused with LLMs embedded substrate, sequence, and environmental related information for prediction. To enable efficient multimodal feature integration, we proposed a multi modal cross-attention fusion (MMCAF) module for feature alignment and update. Experimental evaluations demonstrate that GraphKcat can effectively identifying catalysis-critical residues, learning sequence similarity-independent features through its multi-scale graph architecture, and maintaining robust performance under low sequence similarity conditions. Critically, GraphKcat captures conserved structural and physicochemical patterns related to enzymatic activity, enabling accurate identification of high-activity enzyme variants and functional mutants even for low-homology targets. These capabilities highlight its potential as a transformative tool for enzymatic activity prediction, rational enzyme engineering, and enzyme mining in industrial applications.

By harnessing the synergy between enzymes and photoredox catalysts, cooperative photobiocatalysis has recently emerged as a promising strategy for developing stereoselective radical reactions. While various cofactor-dependent enzymes have been repurposed, the use of cofactor-independent enzymes in such cooperative catalysis without requiring expensive cofactors remains rare. Herein, we report the successful repurposing of class I aldolases, a prominent family of naturally occurring, cofactor-independent enzymes, to catalyze unnatural radical α-alkylation of aldehydes in a highly enantioselective fashion. Through directed evolution of Escherichia coli 2-deoxy-D-ribose-5-phosphate aldolase (EcDERA), we developed an effective radical alkylase bearing five mutations and inverted π–facial selectivity relative to wild-type EcDERA, allowing a range of aldehydes to couple with α-iodoesters, α-iodoketones and α-iodonitriles with excellent enantiocontrol. This study represents the first demonstration of leveraging the nucleophilic enamine intermediate in class I aldolases for radical-mediated stereoselective C–C bond formation. Mechanistic investigations suggested that when irradiated at 440 nm, cooperative catalysis with an exogenous Ir photocatalyst more effectively induces enzymatic enamine radical activity than charge-transfer complex photochemistry. Together, these findings underscore the potential of class I aldolases to enable general and stereoselective new-to-nature radical transformations.