Biocatalysis@TUDelft

Ancestral sequence reconstruction (ASR) is gaining attention as an attractive protein engineering tool for developing highly functional ancestral enzymes. In this study, an ancestral L-tryptophan synthase β-subunit (AncTrpB) with high organic solvent tolerance was identified. The excellent properties of AncTrpB enabled the high-yield synthesis of analogs. This study demonstrates the broader applicability of ASR for developing practical enzymes.

Abstract

Organic solvent-tolerant enzymes expand their applicability in chemical reactions involving poorly water-soluble substrates. The development and discovery of such enzymes remains challenging, even with advanced protein engineering approaches or screening from natural sources. In this study, we explored ancestral sequence reconstruction (ASR) as an alternative approach to identifying organic solvent-tolerant enzymes. Using L-tryptophan synthase β-subunit as a model, an ASR approach on vast sequence data successfully identified AncTrpB1, an ancestral L-tryptophan synthase β-subunit exhibiting high organic solvent tolerance in 50% v/v dimethyl sulfoxide. Furthermore, AncTrpB1 also showed thermostability, high soluble expression levels, comparable catalytic activity to extant TrpBs, and broad substrate scope. X-ray crystallographic analysis of AncTrpB1 suggests the formation of salt bridges at the dimer interface as a plausible factor for its organic solvent tolerance. The excellent properties of AncTrpB1, especially its organic solvent tolerance, enabled the high-yield synthesis (∼1 g/10 mL) of L-tryptophan analogs. Our findings demonstrate the broader applicability of ASR for developing practical enzymes.

Oxidative and reductive reactions are vital processes in cellular metabolism. Like yin and yang in Tai Chi, they are independent and complementary. The non-natural cofactors NCD⁺ and NCDH exhibit substantial potential for the bio-orthogonal regulation of redox pathways. To provide sufficient driving force, it is crucial to maintain the NCD⁺ and NCDH ratio and homeostasis. In the Research Article 10.1002/cbic.202500254, Xueying Wang, Zongbao K. Zhao, and co-workers explain how they reshaped the NADH-binding pocket of NADH oxidase to accommodate NCDH, thereby facilitating the traceless regeneration of the oxidized cofactor NCD⁺ and the selective synthesis of chiral compounds.

Bioreduction of flexible ring N-(3-oxobutyl)-heterocycles mediated by enantiocomplementary recombinant alcohol dehydrogenases [an (S)-selective one from Rhodococcus aetherivorans (RaADH), and an (R)-selective one from Lactobacillus kefir (LkADH)], as whole-cell biocatalysts results in enantiopure (S)- and (R)-alcohols (ee > 99%), which are promising chiral fragments with a high degree of drug-likeness for drug discovery.

In this study, the bioreduction of prochiral N-(3-oxobutyl)heterocycles comprising various (partially) saturated, flexible rings is explored using microbial whole-cell ketoreductases such as wild-type yeast strains including baker's yeast (Saccharomyces cerevisiae) and Escherichia coli cells expressing two enantiocomplementary recombinant alcohol dehydrogenases. Initially, four wild-type yeast strains are screened for ketoreductase activity on a series of nine flexible N-heterocycles with prochiral carbonyl group in the N-(3-oxobutyl) sidechain. The yeast strains resulted in the corresponding (S)-alcohols with a low to moderate conversions. Using recombinant alcohol dehydrogenase whole-cell preparations as biocatalysts ((S)-selective ADH from Rhodococcus aetherivorans (RaADH) and (R)-selective ADH from Lactobacillus kefir (LkADH)) resulted in higher conversions in most cases, while maintaining the full enantiotopic selectivity. Usually, the preparative-scale bioreductions showed comparable or even higher conversions than those observed in the small-scale screening reactions, resulting in virtually enantiopure (S)- and (R)-alcohols (ee > 99%), which are promising chiral fragments with a high degree of drug-likeness. Docking studies confirmed the absolute configuration of the forming (S)- and (R)-alcohols.

Several strategies for biotransformation are compared and a comprehensive analytical characterisation of the novel dihydrodiol is given.

ABSTRACT

Bacterial Rieske non-heme iron oxygenases catalyse the transformation of a wide range of aromatic compounds to vicinal cis-dihydrodiols. Such compounds have been successfully applied in chemoenzymatic synthetic routes for, for example, pharmaceuticals, natural products and polymers. In the case of benzoate, only (1S,2R)-cis-1,2-dihydroxy-2-hydrobenzoate is readily accessible via enzymatic transformation, but not the regioisomeric cis-2,3-dihydroxy-2,3-dihydrobenzoate (2,3-DD) or cis-3,4-dihydroxy-3,4-dihydrobenzoate. While trace amounts of putative cis-2,3-DD have been obtained before by using p-cumate 2,3-dioxygenase (PCDO) or a combination of chlorobenzene dioxygenase and nitrilase, none of these approaches enabled its production and isolation at a greater scale for potential use as a chiral building block in organic synthesis. We here provide a protocol for biotransformation of benzoate yielding (2R,3S)-2,3-dihydroxy-2,3-dihydrobenzoate using the PCDO of Pseudomonas citronellolis strain EB200 with negligible formation of side products. An isolation procedure suitable for production of the 2,3-DD sodium salt monohydrate at high purity (> 95%) at a gram scale, and a comprehensive characterisation of this novel metabolite is given.

Science Advances, Volume 11, Issue 38, September 2025.

Deep learning has accelerated protein design, but most existing methods are restricted to generating protein backbone coordinates and often neglect interactions with other biomolecules. We present RFdiffusion3 (RFD3), a diffusion model that generates protein structures in the context of ligands, nucleic acids and other non-protein constellations of atoms. Because all polymer atoms are modeled explicitly, conditioning the model on complex sets of atom-level constraints for enzyme design and other challenges is both simpler and more effective than previous approaches. RFD3 achieves improved performance compared to prior approaches on a range of in silico benchmarks with one tenth the computational cost. Finally, we demonstrate the broad applicability of RFD3 by designing and experimentally characterizing DNA binding proteins and cysteine hydrolases. The ability to rapidly generate protein structures guided by complex sets of atom-level constraints in the context of arbitrary non-protein atoms should further expand the range of functions attainable through protein design.

We here describe an enzymatic in vitro synthesis route to indigo, the dye that gives denim its renowned color. The proposed 5-step enzymatic pathway is derived from the natural biosynthesis route that starts from anthranilate. In the context of the bio-based economy, this method provides an alternative to the current method of indigo synthesis, which relies on petroleum-based building blocks. First, the enzymes were assessed by stepwise addition to detect intermediate compounds, to ensure cascade function. A two-phase system was designed in response to the difference in melting temperatures of the enzymes. A titer of 183 mg/L indigo was reached through stepwise parameter optimization for each phase, followed by process engineering to enhance the indigo titer. Continuous supplementation of the co-substrate phosphoribosyl pyrophosphate was necessary to maintain indole-3-glycerol phosphate (IGP) formation in the first phase. During the second phase, complete uptake of IGP by ZmBX1 remained challenging, and formation of a significant amount of byproducts was observed. Despite attempts to resolve these issues, the underlying mechanism remained unclear. Although the yield must be significantly improved for an economically viable process, an enzymatic cascade based on bio-based materials remains a potential alternative to facilitate sustainable denim dye synthesis.

Mushrooms have learned twice independently how to make the iconic magic mushroom natural product psilocybin. This article introduces the enzymes of the second pathway, found in a fiber cap mushroom. Curiously, the two pathways do not share any reaction, nor do the enzymes show a close relationship, but both pathways proceed via 4-hydroxytryptamine as a common intermediate.

Abstract

Psilocybin (4-phosphoryloxy-N,N-dimethyltryptamine, 1) is the main indolethyl-amine natural product of psychotropic (so-called “magic”) mushrooms. The majority of 1-producing species belongs to the eponymous genus Psilocybe, for which the biosynthetic events, beginning from l-tryptophan (2), and the involved enzymes have thoroughly been characterized. Some Inocybe (fiber cap) species, among them Inocybe corydalina, produce 1 as well. In product formation assays, we characterized four recombinantly produced biosynthesis enzymes of this species in vitro: IpsD, a pyridoxal-5′-phosphate-dependent l-tryptophan decarboxylase, the kinase IpsK, and two near-identical methyltransferases, IpsM1 and IpsM2. The fifth enzyme, the insoluble monooxygenase IpsH, was analyzed in silico. Surprisingly, none of the reactions intrinsic to the 1 pathway in Psilocybe species takes place in I. corydalina. Contrasting the situation in Psilocybe, the Inocybe pathway is branched and leads to baeocystin (4-phosphoryloxy-N-methyltryptamine, 3) as a second end product. Our results demonstrate that mushrooms recruited distantly or entirely unrelated enzymes to evolve the metabolic capacity for 1 biosynthesis twice independently.

Electrostatic potential surfaces and electric field lines reveal that local fields guide proton delivery through multiple conserved solvent channels in heme peroxidases, enabling efficient access to the active site.

Abstract

The active sites of heme enzymes have evolved to control the formation of highly reactive intermediates in oxidative catalysis. Proton delivery to the heme is essential, yet the mechanisms of proton delivery remain poorly understood. Here, we identify routes and drivers of proton delivery in a heme peroxidase (ascorbate peroxidase) using computational approaches that combine classical, quantum, and hybrid methods with enhanced sampling and local electric field (LEF) analyses. Our results show that networks of active-site water molecules facilitate proton exchange with Arg38, which may act as a transient proton carrier at the γ-heme edge where the substrate binds. The distal His42 residue aids proton transfer into the active site via solvent at the δ-edge. Molecular dynamics simulations of three heme peroxidases identify hydrated channels leading to both γ- and δ-edges, allowing solvent protons to reach the active site. Comparison with eight other heme peroxidases shows that these channels are conserved. LEF analyses reveal a continuous electrostatic funnel drawing protons toward the heme from the γ- and δ-edges, a feature that is broadly conserved across other peroxidases. These results suggest that nature pre-organizes electrostatic funnels and solvent channels to provide multiple well-defined routes for proton delivery in peroxidase catalysis.

The Earths magnetic field plays an important role in the seasonal migrations of many species of animals. A Cryptochrome (CRY)-based radical pair mechanism (RPM) has been suggested to underlie the mechanistic basis of animal magnetosensitivity and navigation. The quantum spin state of a radical pair involving flavin adenine dinucleotide (FAD) bound to CRY in the canonical pocket is sensitive to external magnetic fields that can alter the signalling concentration of activated CRY1-5. However, several experimental observations challenge this model including the finding that the C-terminal fragment of Drosophila CRY (DmCRY), which lacks any canonical FAD binding pocket, and human CRY2, which lacks affinity for FAD, are sufficient to support magnetosensitivity6-9. Here, we use all-atom molecular dynamic (MD) simulations, alongside in vitro and in vivo analyses to reveal that the C-terminus of Drosophila CRY (DmCRY-CT) binds FAD. FAD binding is required for transduction of a magnetic signal within cells, and, in vitro, initiates formation of high molecular weight DmCRY-CT oligomers, including large insoluble aggregates reminiscent of CRY photobodies observed in plants10-14. These results provide a plausible mechanistic basis for several experimental observations that have reported non-canonical magnetosensitivity in animals.

The glutaminase (GLS) isoforms KGA and GAC are expressed in neurons where they catalyze the hydrolysis of glutamine to produce the excitatory neurotransmitter glutamate. Two de novo gain-of-function mutants of GLS, S482C and H461L, were recently identified in patients with developmental delay, epilepsy, and infantile cataract. These patients exhibited high glutamate and low glutamine concentrations in the brain, suggesting that the GLS mutants have abnormal enzymology. Here, we examined the enzymatic properties of these GLS mutants and found that they exhibit a total (S482C) or partial (H461L) loss of glutamate product inhibition, lifting this restriction on glutamate accumulation. The mutant enzymes also no longer require the anionic activator phosphate to stimulate enzymatic activity or induce filament formation. Structural analysis of the S482C GAC mutant shows the mutation shifts the key catalytic residue Y466 into the catalytically competent position and disrupts a key hydrogen bond between it and the glutamate product, explaining how the S482C mutant has enzymatic activity in the absence of phosphate and is insensitive to glutamate product inhibition. These results shed new light on the mechanism of phosphate activation and glutamate product inhibition of GLS and show that loss of these enzymatic properties disrupts glutamate homeostasis in the brain and causes neurological disease.

Equilibration between CO2 and the other forms of dissolved inorganic carbon (DIC) is slow under ambient conditions, bottle-necked by the hydration of dissolved CO2 to form bicarbonate and a proton. This step is often rate-limiting for aqueous CO2 capture processes under ambient temperature and pressure, and is precisely the chemistry catalyzed by the family of carbonic anhydrase enzymes. As a result, incorporation of carbonic anhydrases (CAs) into various CO2 capture schemes may be used to accelerate CO2 hydration, increasing the rates of downstream processes coupled to this DIC equilibrium. While this potential use for CA has been long-studied, extending these basic chemical principles to real applications have been hampered by practical questions of how to economically source carbonic anhydrase enzymes at mass scale. This work explores the use of modified cyanobacterial expression hosts as a potential resolution to that problem. Specifically, engineering the surface display of carbonic anhydrases in fast-growing, marine cyanobacteria would yield a self-sustaining CO2 hydration catalyst that requires only sunlight, CO2 itself, and nutrients freely available in seawater, as inputs for passively generating carbonic anhydrases on the cell envelope during cell growth. These possibilities motivated the development of such constructs; their potential for enhancing CO2 hydration rates are reported here.

Nature Chemistry, Published online: 17 September 2025; doi:10.1038/s41557-025-01935-4

Expanding the complexity of genetically encoded peptides is a long-standing challenge at the intersection of chemistry and biology. Now it has been shown that linear peptides with a reactive N-terminal β- or γ-keto amide can be synthesized ribosomally and elaborated to generate atropisomeric and/or macrocyclic peptides with embedded pharmacophores.

The stereoselective synthesis of optically active γ,γ-dihalo-β-enols remains unexplored. In this study, efficient biocatalysts are identified to enable their synthesis with high conversion and excellent enantiomeric excess. A one-pot, two-step sequential process is developed, combining a light-driven reaction followed by an alcohol dehydrogenase-catalyzed step. The absolute configuration is determined using Mosher ester analysis.

Merging different strategies in one-pot processes is attracting considerable attention due to their straightforward and sustainable potential for synthesizing novel organic compounds. In particular, the exquisite selectivity displayed by enzymes and the possibility of coupling biotransformations with metal-, photo and electrocatalytic processes open new avenues for stereoselective synthesis. Herein, the preparation of chiral (hetero)aryl-3,3-dihalopro-2-en-1-ols is described for the first time. To achieve this, a photochemical and biocatalytic one-pot sequence is developed, employing visible light irradiation and stereoselective alcohol dehydrogenases (ADHs) for the transformation of commercially available alkynes into optically active compounds in an aqueous medium. The one-pot, two-step sequential approach involves a photocatalyst-free reaction between terminal and internal alkynes with polyhalomethanes, leading to gem-dihaloenones, which are subsequently reduced using ADHs. After optimizing the individual steps and identifying suitable conditions for combining both processes, the use of complementary ADHs enables the synthesis of a novel family of optically active allylic alcohols with high stereoselectivity. Their chemical derivatization is further explored, allowing the stereoselective synthesis of a chiral propargylic alcohol from the corresponding γ,γ-dihalo-β-enol.

Phenylalanine ammonia lyases (PALs) were engineered to switch enantioselectivity, enabling hydroamination for the direct synthesis of electron-deficient aromatic d-amino acids (d-AA). Structure-guided engineering provided PAL variants with broad substrate scope, high activity, and excellent d-selectivity, broadening the utility of PALs in asymmetric synthesis.

Abstract

Aromatic d-amino acids (d-AAs) are valuable building blocks in drug discovery and peptide therapeutics, yet their direct and efficient biocatalytic synthesis remains a challenge. Here, we report the rational engineering of phenylalanine ammonia lyase from Planctomyces brasiliensis (PbPAL) to enable asymmetric hydroamination for the enantioselective synthesis of d-aromatic amino acids. By targeting active-site residue L205, we identified variants capable of highly d-enantioselective hydroamination, with L205F enabling the transformation of electron-deficient aryl acrylates with >99% enantiomeric excess (ee). The synthetic utility of this platform was demonstrated by gram-scale synthesis of d-benzoxazole and substituted 2-pyridylalanines. Structural and mutational studies revealed distinct roles for the 4-methylideneimidazole-5-one (MIO) prosthetic group and active-site residues L205, Y64, and K397 in modulating enantioselectivity. These results enabled the identification of PbPAL variants with the opposite selectivity, such as L205V-K397A, which preferentially produce l-amino acids. This work broadens the utility of PALs as programmable biocatalysts for asymmetric synthesis.

With a lens of enzymology, this review compares nanozymes with enzymes using the Michaelis-Menten model and differences in K_M and k_cat along with catalysis in complex biological environment. The implications of these differences in biomedical applications of nanozymes are then described with a statistical analysis of the trend of therapeutic use of nanozymes.

Abstract

Nanozymes are catalytic nanomaterials that transform enzyme substrates into their corresponding products, offering enhanced stability and a cost-effective alternative to traditional enzymes. As nanomaterials, they possess unique physicochemical properties and catalytic mechanisms distinct from those of enzymes. Such differences have profound, yet often neglected, implications in biomedical applications. In the context of enzymology, this review compares nanozymes and enzymes, with a focus on redox reactions. This review begins with the classification of nanozymes based on the types of reactions they catalyze, with the ability to exhibit multiple catalytic activities being a prevalent characteristic. The use of the Michaelis-Menten model for both enzymes and nanozymes is discussed in detail, and the Michaelis constant, maximum reaction rate, and turnover number values are compared. The performance of nanozymes in crowded environments and under extreme conditions is also compared to that of enzymes. We discuss the kinetic factors influencing nanozyme performance, the impact of active site shielding, and the activity under non-physiological conditions. We then compiled recent trends in the biomedical applications of nanozymes, focusing on both the production and scavenging of reactive oxygen species. This review links fundamental enzymology to nanozyme catalysis, providing a key reference for the rational use of nanozymes.

Enzyme engineering with photoswitchable unnatural amino acids (psUAAs) has a high potential for applications such as biocatalysis. Here, we extend the current repertoire of psUAAs to include more versatile properties paving the way to more advanced photocontrol engineering. We report on the synthesis and photochemical behavior, suitable aminoacyl-tRNA synthetases for co-translational incorporation and initial enzymatic studies of these psUAAs.

Abstract

Photoswitchable unnatural amino acids (psUAAs) play a crucial role in the engineering of light-sensitivity in enzymes, which holds significant promise for diverse applications such as biotherapy and biocatalysis. Besides near-quantitative photoconversion, the success and expediency of a psUAA for a certain application is defined by its interaction potential with the enzyme, its thermal stability and its effective wavelength of irradiation. To establish high versatility in the current repertoire, we have designed and synthesized six psUAAs based on azobenzene, arylazopyrazole, arylazothiazole, hemithioindigo and spiropyran photoswitches. The resulting psUAAs exhibit an enhanced interaction potential within an enzyme owing to their capacity for hydrogen bonding, ionic interactions and metal ion coordination. Moreover, we observed diverse photochemical behaviors among the psUAAs, with four of them reversibly switching between the isomers with purely visible light. Notably, we identified orthogonal aminoacyl-tRNA synthetases that facilitate the incorporation of five of the six psUAAs co-translationally and computationally analyzed the synthetase-psUAA interactions. Finally, we evaluated the photochemical behavior of the five psUAAs within an enzymatic model and tested the photocontrol of catalysis confirming their diversity. Ultimately, our findings significantly expanded the repertoire of psUAAs and demonstrated their feasibility for enzyme engineering studies.

A chemoenzymatic strategy is reported for enabling skeletal editing of a target molecule via ring expansion at the level of one or more aliphatic C─H sites. By combining P450-mediated site-selective C─H oxidation with Baeyer–Villiger rearrangement or ketone homologation, a panel of structurally diverse, ring-expanded analogs of a series of natural products was obtained, illustrating the potential value of this approach for the discovery of new bioactive molecules.

Abstract

Methods for introducing subtle modifications at the level of single atoms/bonds (“skeletal editing”) are highly desirable in organic and medicinal chemistry, owing to their potential for fine-tuning the structure and biological activity of organic molecules. Here, we report a chemoenzymatic strategy for enabling the skeletal editing of organic frameworks via ring expansion at the level of one or more aliphatic (methylene) C─H sites, as achieved through the synergistic combination of P450-mediated site-selective oxidation with subsequent Baeyer–Villiger rearrangement or ketone homologation. Combining this approach with engineered P450 catalysts exhibiting divergent regioselectivity enabled the expeditious synthesis of a panel of ring-expanded analogs of various complex natural product substrates. Importantly, the skeletal modification was found to drastically altered the anticancer activity of some of these compounds. By the direct targeting of aliphatic C─H sites with tunable site-selectivity, this strategy provides a powerful tool to rapidly access skeletally edited derivatives of natural products and other bioactive molecules for applications in drug discovery and chemical biology.

We report the construction of hydrogen-bonded organic frameworks that integrate Ru-based photocatalysts with three enzymes of formate/formaldehyde/alcohol dehydrogenase via co-assembly for sustained CO₂-to-methanol conversion through photo-enzymatic cascade catalysis with high activity and recyclability.

Abstract

The integration of photocatalysts and enzymes within confined environments offers a promising approach to developing artificial photosynthetic systems for sustainable CO₂ conversion. However, the efficient coupling of photocatalysts with multiple enzymes to enable photo-enzymatic cascade catalysis remains a significant challenge. Herein, we report the construction of hydrogen-bonded organic frameworks (HOFs) that integrate Ru-based photocatalysts with three-enzyme cascades of formate dehydrogenase (FDH), formaldehyde dehydrogenase (FaldDH), and alcohol dehydrogenase (ADH) via in situ co-assembly in water. The RuHOF exhibits exceptional nicotinamide adenine dinucleotide (NADH) photo-regeneration activity (4.5 mM h⁻¹), while the FDH@RuHOF hybrid converts CO₂ to formic acid with a turnover frequency (TOF) of 681 h⁻¹ (238 µM h⁻¹) over 24 h. By engineering FDH/FaldDH/ADH@RuHOF ternary systems, we achieve sustained CO₂-to-methanol conversion through photo-enzymatic cascade catalysis, delivering 2.2 mM methanol production with an apparent quantum efficiency (AQY) of 5.5% (92 µM h⁻¹) over 24 h with 85% activity retention after five catalytic cycles. This work opens a promising avenue for the development of efficient multi-enzyme cascade artificial photosynthetic systems toward steady and recyclable CO₂ valorization.

Ether lipids play critical roles in membrane dynamics, antioxidant defense, and signaling. They comprise [~]20% of mammalian phospholipids, and disruptions in their metabolism cause severe genetic disorders and are associated with neurodegenerative and metabolic diseases. Ether lipids are synthesized de novo from glycolytic intermediates or salvaged from the diet. While the products of these pathways are known, several key enzymes remain unidentified, including the 1-O-alkylglycerol kinase and the 1-O-alkyl-2-acetyl-sn-glycero-3-phosphate phosphatase. Here, we show that acyl-CoA dehydrogenase member 10 (ACAD10) catalyzes the phosphorylation of 1-O-alkylglycerols and the dephosphorylation of 1-O-alkyl-2-acetyl-sn-glycero-3-phosphate. Worms and mice lacking ACAD10 have reduced ether lipid levels and cannot salvage dietary alkylglycerols. Furthermore, individuals from the Akimel Oodham (Pima) tribe carrying ACAD10 polymorphisms also show decreased plasma ether lipid levels. Collectively, our findings resolve two long-standing gaps in ether lipid biochemistry and reveal a mechanistic link between ether lipid metabolism and a population-associated risk factor for type 2 diabetes.

Avermitilol synthase from Streptomyces avermitilis (SaAS) is a high-fidelity class I terpene cyclase that converts farnesyl diphosphate into a highly-strained, 6-6-3 tricyclic sesquiterpene alcohol. The mechanism of avermitilol formation proceeds through a 10-3 bicyclic intermediate, bicyclogermacrene, which undergoes proton-initiated anti-Markovnikov opening of two separate C=C bonds in a transannulation mechanism that forms the 6-6-3 tricyclic skeleton, with quenching by water to yield avermitilol. Small amounts of a side product, viridifloral, result from Markovnikov opening of one of the reactive C=C bonds. Here, we present enzymological studies of SaAS to establish the substrate scope and metal ion dependence for catalysis, and we present crystal structures of SaAS complexed with a variety of ligands that partially mimic carbocation intermediates in catalysis. Interestingly, these structures show that two water molecules remain trapped in a polar crevice in the active site regardless of the ligand bound. Structure-activity relationships for site-specific mutants yield key insight on the catalytic importance of these trapped water molecules. Specifically, T215 normally hydrogen bonds with water molecule W1, but the T215V substitution breaks this hydrogen bond and causes W1 to shift by 1.3 [A] to form a hydrogen bond with W300. Avermitilol generation is completely blocked in this mutant, but the generation of viridifloral and another side product is enhanced. Thus, the T215V substitution causes water molecule W1 to align for reaction with the tertiary and not the secondary carbon in the reactive C=C bond of bicyclogermacrene. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=108 SRC="FIGDIR/small/675706v1_ufig1.gif" ALT="Figure 1"> View larger version (45K): org.highwire.dtl.DTLVardef@1c02a77org.highwire.dtl.DTLVardef@1240739org.highwire.dtl.DTLVardef@564448org.highwire.dtl.DTLVardef@19d12be_HPS_FORMAT_FIGEXP M_FIG C_FIG